nA La .

,
iv u- Malt
es harm ibn

RACING — naienats &—_ ® aoe & GROUND EFFECTS ¢ THE Driver's _

TESTING & RACE SET-UP © RACING COMPUTERS BRAKING SYSTEMS ¢ TIRE TECHNOLOGY @ WIND TUNNELS
BY PAUL VAN VALKENBURGH
 LIVERPOOL
JOHN MOORES UNIVERSITY
AVRIL ROBARTS LRC
TIT {=D ALA STREET
Bi as| os 4
‘ sek ems :
LIVEF 15
a oF}
oto
* ee final ©

wrk ALES O97 AD

t ® She AS&

a}

NS Livernosi Jom Moores University

4&<.. LEARNING & 0:50 SERVICES

icons 63.9. 208 Wy

 RACE CAR ENGINEERING AND MECHANICS
Paul,We}
Valkenburgh
All photographs and drawings i h LTA py
pyerwiseAse

rove es =
7 pERPOOL
Published by the author
ORA
‘i Yo QSOPRES UNIVERSITY
Box 3611 ~ el befor LRC
Seal Beach, Ca 90740 ii a
VERPOOLL2 2ER
Books are to be returned on or before TEL. 015
the last date below.

I ov store of
knowl ut giving
them ; © proper
person
Firs e d with at
Chevr = mn I listed
in my 3 ‘redit to:
Don > Racing,
who k
Martr wununuc, wuu siaicu Pall UL 1d Yast swieuUUDY UL pracucal puLrouual vaperiences
with me as I worked on his autobiography, The Unfair Advantage.
Carroll Smith, who wrote the most informative race car preparation articles ever pub-
lished, in a series he did for Sports Car Graphic while I was technical editor there.
And finally, all those other acquaintences I admire, whose technical comments have
added to or changed my thinking about race cars: Don Gates and Mike Pocobello of An-
tares Engineering; Bill Milliken and Doug Rowland of Calspan (Cornell Aero Labs); Jim
Hall of Chaparral Cars; Karl Kainhofer, Earl MacMullin, John Woodard of Penske Racing;
Jim Travers and Frank Coon of Traco; Bernie Pershing of Aerospace Corporation, Bob
Liebeck of McDonnell-Douglas; Bill Bettes of Cal Tech; Pete Weismann of Traction
Products; Don Devendorf of Electramotive; Wayne Hartman the bodybuilder; Lee Dykstra;
a Cicale; and Mary, my wife, who really changed my thinking about race cars.

Mii ll |
095 4103
INTRODUCTION...
In a way, I’m using this book to say good- with a lot of fascinating unanswered ques- should ever have to compete with an unfair
bye to all those people I’ve raced with or tions and potential for improvements. disadvantage — whether on the track, or in
who have read my automotive journalism. At any rate, I have tried to cover all the real life. Even if I were still driving, I would
By the time this book is published, I'll be areas of race car engineering, theory, be willing to tell a competitor everything I
finishing my Master’s Thesis on “A development, and mechanics as completely know, and then let the best man win. But
Systematic Integration of the Behavioral as possible and in a balanced manner. That I suppose it is fortunate that there wasn’t
Sciences,’ and moving on to other social is something that (to my knowledge) hasn’t enough room in this book to reveal some
challenges. But as I get out of the racing existed before. Naturally, in trying to cram of my more esoteric accumulated know-
game, I want to leave something to show for the greatest amount of factual information ledge, because someday I might have to go
my efforts. I would like it to matter that I into the least amount of space, I must ex- back to making a living as a race car
was ever involved at all. So with this book, pect arguements about my point of view and engineer or driver. (September 1976)
I’m leaving most of my racing knowledge complaints that I was not thorough enough.
behind. I don’t expect to be using it where But you can’t please all the people all the
I’m going, so I’m putting it down in print time. For those who need more basic in- PS.
before I forget it, in the hope that other formation, or more advanced information, Well, I did go back. That is, I occasion-
racers can put it to good use. or verification of some of my more ally take on an interesting race car engin-
It is possible that many of the things I unbelievable comments, I have provided eering job just to keep abreast of technology.
write here were previously known only by frequent references and an extensive And after ten years my book distributors
a very few professional racers. However, bibliography in the back. asked for an update to the book. At first I
that doesn’t mean I’m so smart that I know If anyone is the least bit curious about said the first edition was still valid — it had
everything there is about racing. It merely why I’m moving away from engineering, no errors and there were few advancements
means that I found the time and the initiative racing, and writing about race cars, when in race car engineering. But then I realized
to do the job before other qualified people thousands of excitement-seeking young men I had said nothing about front wheel drive
I know could have. Don Cox and Mark are anxious to get into the business, I don’t race cars, and too little about computers,
Donohue and I talked about a combined mind explaining. I have been exposed to wind tunnels, and aerospace materials, not
effort for years, but they were so involved some of the greatest, wisest, and most to mention a prediction about the future,
with reality that they couldn’t possibly spare successful people in racing. I’ve seen the based on my research into future
the time. With the exception of Carroll top, and discovered that for me for one technologies. And now, after all these years,
Smith, most other people haven’t been reason or another it didn’t seem worth the I’m willing to say more, even though there
fortunate to have the engineering education, effort to get there — and to stay there. I was isn’t much money in it; it’s simply a labor
or the driving experience, or the mechanical fortunate to know the taste of winning — of love.
experience, or the writing experience to do building and driving my own car — in my Many of the old photographs are carried
the job. I can’t boast, though, because I own limited sphere, and felt that these over from the first edition, which may make
recognize just how little is really known by successes were hollow. It seemed all too the book look dated, but in fact proves that
anybody in this business. I did generate a easy for me to go through life racing my the basic principles still stand. Incidentally,
lot of original knowledge about race cars heart out, winning more than my fair share, most of the new material can be identified
while a Research Engineer for Chevrolet, and still ending up with nothing more than by a different type face in the Contents, if
but so did all the others. But the most a bunch of trophies and some fading fame you want to avoid re-reading the original.
humbling knowledge is that the more you to show for it. I feel that I have relatively Racers are busy people. I don’t want to waste
learn about anything, the more you realize well satisfied my drive for competitive suc- your time any more than I want to waste
how little anyone knows for sure. If anyone cess, and I’m anxious to redirect my efforts mine. So I have tried to avoid entertaining
tries to give the idea that he knows all the into other areas which are more humanistic digressions and most of the material you can
answers, you can be sure that he really and cooperative than racing. find somewhere else.
doesn’t even know the questions very well. It is perhaps a broad statement of my Good luck. I hope this is adequate for
I leave this book — and the profession — philosophy to say that I don’t believe anyone another decade. (January 1986) ©
(Updated 1992)
(Computer information and some photos
updated 2000.)
 1 HOW TO WIN — PEOPLE; KNOWLEDGE; PREPARATION.......ceccccececsccccceces 4
ZETURESTAND AV EIEL BIS iorerslerelcjeveleie
lovecle!siaic\clesalelerstcrercievslere terete Srayelatelahessieisicte
stovcle sislelerelare we TL
THEORY: temperature, slip, camber, pressures, loading
TIRE DESIGN: sizes, treads, wheels
DEVELOPMENT; PREPARATION; DRIVING
3 SUSPENSION GEOMETRY AND ALIGNMENT ..........cccccccccscccccscscsccece 19
THEORY: force and deflection, torques and rotations, compliance
HARDWARE: types of suspensions, miscellaneous components
MAINTENANCE AND SERVICE: balance, alignment
DEVELOPMENT
4 SPRINGS, ANTI-ROLL BARS, SHOCK ABSORBERS ...........ccceccceccececceces 33
THEORY: springs, bars, shocks
HARDWARE: springs, bars, shocks
DEVELOPMENT: springs, bars, shocks
PO MESICA INE SOletetetereterctecsiotercretsisferousiereie.s
clete sic sicyeteiawersaveiece cicceeloie creole ielere sisieiaicre overs wielsienerttorersie 43
THEORY: requirements, heat absorption, forces, balance, fade
HARDWARE: discs, calipers, master cylinder, fluid, pads, valves
DEVELOPMENT: proportioning, cooling
_ ASSEMBLY AND RACING; CARBON BRAKES; ANTI-LOCK SYSTEMS
6 AERODYNAMICS ....... TC oO BEB OR AG GO CIA GAIOG CIS RACE ERI BIC OHS IICEICISIG CIID LOL 3p)
THEORY: drag, downforce, tunnel cars, force balance, stability
HARDWARE: spoilers, wings, ducts
DEVELOPMENT: measurement, balance, pressures
RACING: drafting, stability interferences
PEA ANINDDILIINGe yateleyslaieicleielelsisleraiere + 0le\oie)cfuialols/siere!sielnisia/efeislstefele!
s\eielsiclelers sisieleisiciele ssfateieie oie 70
STABILITY; CONTROL; OVERSTEER/UNDERSTEER; HARDWARE; DEVELOPMENT
8 ENGINE SUPPORT SYSTEMS ................cccccccccccccccccccncccccccseconccs 78
FUEL; OIL; WATER; IGNITION; EXHAUST
GEARING AND DIFFERENTIALS 3) occ scctteee ccevece se scessscssctacecccsceese
RATIO THEORY; TRANSMISSION HARDWARE; TRANSVERSE TRANSMISSIONS;
DEVELOPMENT; RACING AND MAINTENANCE;
DIFFERENTIAL THEORY AND HARDWARE; DRIVELINE; AUTOMATICS
EO IERAMEZAND BODY racials «2s sacs dajseats He vateiem Dn wiiita steals celcsisne sebinemans 92
TRADITIONAL MATERIALS
AEROSPACE COMPOSITES: fibers, resins, pre-pregs, how-to
TYPES OF FRAMES; STRENGTH
11 SAFETY ...........24- mieloneie levoierelelstoleieic efolole atelelersioisoieishelarsteiersielotere eieicicialclcis/avevaiev='ets 101
ACCIDENT AVOIDANCE: fasteners, inspection
DRIVER PROTECTION: deceleration, retention, collapse, fire
Le THESDRIVER Dre ces a ee ee eee ;, Soe 108
FITTING AND COMFORT: LEARNING: EMOTIONAL PROGRAMMING
ES CTESTING ei Rees ORI art oe MS OE aR ES. lhe x, 114
TOOLS: facilities, electronic/pressure instruments, wind tunnels
PROCEDURES: acceleration, brakes, aerodynamics, handling, tires
TRACK TESTING; STATIC TESTING
TQ THE BITS cerca cc ciccicis vice ccc sccc ccs cssveisisics cies ssieececesecicecesseucescs 133
RECORDS; TIMING AND SCORING: PIT STOPS
15 COMPUTERS IN RACING ........... iaelolsletelatel sieletaysietelererelcteloselelsusisieis/slslslele!s)¢
ieeice/ 139
PERFORMANCE SIMULATIONS: suspensions, dynamics, race courses
DESIGN: theory, requirements, who does what, how-to
DATA RECORDING: pros/cons, who does what, hardware, how-to
ENGINE CONTROLS: production equipment, tricks, tools
TIMING AND SCORING: official systems, input, fail-safety
16 FRONT WHEEL DRIVE RACING..... etstorolalalohaveieteteisiareeteloretetelelcielalevorelsioievelo\e)siela.e
stele 156
DRIVING STYLE; DIFFERENTIAL SELECTION; WEIGHT BALANCE;
BRAKE BALANCE; LIFT/DRAG; DOWNFORCE DISTRIBUTION;
SUSPENSION GEOMETRY, TORSIONAL STIFFNESS; TIRE TRANSIENTS
17 RACE CAR OF THE FUTURE .............. ADO RODOONONOBOCOOU 0b SOUGOOOUOUOE 161
AERO: moveable surfaces, pressure/suction, toroidial flow
FRAMES: return to the tube?
WHEEL CONTROL: rising rate, electro-mechanical, computers
SUDDEN IMPACTS: driver cells, fire control
ENGINES: plastics, electronics, turbos
TRANSMISSIONS: automatics, infinitely-variable
TIRES: temperature control
WHEELS/BRAKES: plastics, filament winding, carbon
ELECTRONICS: computers, feedbacks, expert systems
CONTROL: suspension, brakes, steering
BIBLIOGRAPHY ........ aistatetas aieistetersiets cisieveisisiateisters!s ehalelatels/siatetsieinislcla/sisle eisiaisielaiae se Ll
1
HOW TO WIN...
PEOPLE, KNOWLEDGE, PREPARATION
his book will show you how to win any it. He just fixes his fixes the next time valuable. Top men can make more money
automobile race. And it’s very easy to around. But a racing mechanic has to be than engineers or executives, and the work
do. That is, it’s very easy to show someone more professional, because, like a doctor, is a lot more exciting — if also demanding
how to win a race — but it’s usually an in- he may not have a second chance. in terms of hours worked and miles traveled.
credible challenge in time and money to Sometimes he may be able to recheck his The best investment in a race team can be
actually do the job. But the better anyone work before the race starts and correct any one or two of the best mechanics available.
understands and uses everything in this problems if there‘s time. But as the old say- Otherwise, all the money spent on the most
book, the more likely he is to win — assum- ing goes, “When the flag drops, the bullshit exotic equipment, not to mention the other
ing his car is at least equal to the others in stops.” Racing is very unlike school, where expenses of racing, is totally wasted.
basic design factors such as power-to-weight 60 or 80 percent correct may be a passing Unlimited dollars won’t win a race without
ratio, body size and shape, and tire size. Un- grade. There are thousands of components some people with a lot of knowledge and
fortunately, all cars are not created equal, in a race car, and 99.99 percent right is not experience. And if the best man available
and race sanctioning groups can’t always close enough. If the mechanic misses one starts making serious mistakes — he’s out.
make classifications equitable between dif- thing, the results are likely to be serious. He can lose races, he can lose money, and
ferent makes or designs. Until they can Not only is the race certainly lost — and he can lose the driver’s life.
create an accurate handicapping system, any in professional racing the mechanic may be If a mechanic or engineer lacks ex-
car may be at a basic disadvantage, which losing out on a share of the winnings — but perience in racing, a lot of book learning
the driver and mechanic will have to over- the most serious aspect is that a man’s life can make up most of the difference. There
come by even greater efforts. If this chapter is at stake, and possibly the lives of inno- may not be enough space here to go into
seems to be more about the philosophy of cent bystanders. Sometimes, in longer races, great detail about all the aspects of race cars,
racing than the engineering and mechanics, it may be possible to correct an error during but at least there is a mention of almost
it’s because a tremendous personal commit- a pitstop and still have a chance of winning. every area and potential problem that should
ment is of primary importance in winning In that case, the mechanic is more like an be considered, and references for further
races. emergency medic, in that slow, methodical, study in depth. The book also presents a
No book is big enough to include and precise work is not as important as his systematic overall approach to the entire
everything a driver or mechanic needs to ability to instantaneously diagnose and cure package, treating the race car, driver, and
know about race cars. Therefore, it is the problem. team as a system.
assumed that the reader already has a basic Good race car mechanics are hard to find But if anyone wants to keep up with race
knowledge of automobile service and or train, and they are therefore very car technology and the state of the art, it
mechanics. If not, it can be picked up from y
an ordinary auto mechanics text, preferably
one written for the Junior College student.
There are thousands of good potential race
car mechanics who already have a lot of ex-
perience in garages and dealerships, but
who don’t know all the necessary specific
details of racing. Qualified aircraft
mechanics perhaps have even better train-
ing. Similarly, there are thousands of
graduate engineers who could be a great
asset to a racing team, but who don’t know
what all the problems are, or the factors in-
volved. So this book was written for either
the experienced non-racing mechanic, or the
non-racing engineer. If you know how to
engineer a car, or how to rebuild one com-
pletely, this book will show how to apply
; Halbert Racing 0
your knowledge specifically to racing.
amoune TEN Thows aug
The first distinction between a race car
mechanic and an ordinary repair mechanic CAME,
is perfection. The average garage mechanic
can make a few mistakes and get away with _ as
4
will be necessary to do a lot of reading. nals. The Society of Automotive Engineers aerospace, or race car engineer. Experi-
First, it’s mandatory to have a copy of the also puts out hundreds of technical papers enced race car mechanics are also good con-
particular sanctioning body’s regulations — each year. Most of the information will be sultants, although they have a terrible
and to know them by heart. Even if it never directed toward passenger car applications, tendency to say, “It won’t work — because
becomes necessary to recite a certain rule but the fanatic race car engineer will always it hasn’t been done before.’ Or, perhaps,
to a technical inspector or official, everyone be able to see any conceivable connection because they didn’t think of it first.
on the team must know exactly what they with his own needs. And finally, there are Besides specific information, here are a
can and cannot do — to the car, or in the a few other books which cover limited areas few basic philosophical generalizations
pits, or on the track. In addition, automotive of automotive — if not race car — engin- worth mentioning. There is such an infinite
enthusiast periodicals, whether monthly eering, most of which are listed in the number of tasks to be performed in winning
slick magazines or weekly newsletters, often bibliography. races that there has to be some established
have the latest technical advances or scoops As far as the information in this book series of priorities. It isn’t possible to build
on a competitor’s new device. At the very goes, most of it is assembled relatively and race a car successfully if everyone
least, they are always a good sorce of in- sequentially. The first half, for example, simply concentrates on what is fun, or
formation on the availability of new products builds slowly up to a total comprehension familiar, or simply easy to do. In the basic
through the advertisements. Very little in- of handling. Therefore, the chapter on assembly of a race car, there will be three
formation is given in this book on brand Handling won’t make much sense if you categories of jobs to do: Must, Important,
names or sources of racing products. Both haven’t read the previous chapters. However, and Also. These will range from the safety
change so rapidly that it’s impossible to keep a mechanic with a specific problem or area items necessary to get through technical in-
up in a book which may take a year to of interest can concentrate on individual spection, to final appearance detailing. It
publish. Any good racer should have chapters, or specific areas within a chapter, won't do a bit of good to have a polished
enough initiative and imagination to find such as theory, hardware, development, or and pinstriped body if the engine can’t be
whatever he needs through the ads or by maintenance. installed in time. This assembly priority list
simply asking around. Another good source NOTE: This is an important disclaimer is also a good reference for scheduling
of information is through the manufacturers to keep in mind, regarding all general critically important components, which
themselves. Any speed equipment or statements and especially mathematical must be bought or built before many other
automobile manufacturer who appreciates equations in this book. In the interest of assembly jobs can follow in sequence.
good racing exposure will probably have a saving space, a lot of the theories and equa- There is no reason to ever go to a race
lot of up-to-date technical data available on tions have been simplified a great deal. It until the car is ready to be raced. That ought
their products. It can also be helpful to read is possible for a person who doesn’t have to be obvious. But so many people are so
race reports to get an idea of the causes of a good engineering background to get into anxious to be racing that they will take their
vehicle retirements. A person with a great trouble by misinterpretation or extension unfinished, unprepared car to the track
deal of excess time could compile a tremen- into unrelated areas. There are innumerable anyhow, hoping to be able to get it ready
dous check list of potential failures — and pitfalls for the person who doesn’t do some at the last minute. They literally push
their frequency — by tabulating the pub- research behind a theory or equation, or themselves into total frustration and exhaus-
lished race DNF’s (Did Not Finish) from who doesn’t follow up with some sort of test tion by packing, unpacking, searching for
the past few years. verification, or who bases an entire vehicle parts and tools, correcting mistakes, making
For the engineers who want to keep up on a single unproven theoretical premise. It‘s continual adjustments by intuition, and go-
with the latest theoretical advances and more happened to the best of us. It could save a ing without food or sleep. Then they wonder
esoteric technological information, there are lot of time, money, and frustration to get what‘s wrong with the car or the driver that
a number of valuable engineering trade jour- further advice from a qualified mechanical, kept them from winning. So the first
preparation priority is to have the car
finished before ever leaving for the track.
The oldest truism in racing is — you‘ve
got to be able to finish before you can win.
In any given race it's quite likely that half
of the cars which start will not even finish
the race at all. Therefore, any car that is
prepared well enough to stay together for
the length of a race has already beaten at
least half the field before the race even
starts. Those priorities ought to be obvious
— first, the car has to be safe and durable,
and second, it can be made faster. So if
there is ever any thought about what ought
to be done next — and that should be a con-
stant consideration — then the most rational
sequence is: (1) have the car assembled on
time, (2) make it safe and durable, and (3)
try to make it faster. If the car is ready to
race as it is rolled off the trailer, except for
adding fuel and checking tire pressures, then
it is probably already in the lead at that
point.
 A few things ought to be mentioned about among themselves, but that makes it doubly dissolvable carburetor restrictor plate. These
radical innovation, and a related subject, in- difficult for a newcomer to become com- are prosecutable violations. On the other
tentional cheating. Simple observation and petitive. Sometimes you can’t simply follow hand, creative rule-bending includes all
a lot of practical experience shows time after the book. changes that could have happened by acci-
time that the wildly innovative car will prac- Creative rule-bending, however, is dent, or by the accumulation of factory
tically never win a race — until it is also something else entirely. This is where you tolerances. For example, a fuel tank might
well developed. Almost invariably, it is the study the rule book over and over, reading have expanded due to accidental over-
best developed, old reliable design that between the lines, to stretch interpretations pressurization, or extra airflow might get
finishes first. There can be thousands of test and to find what it doesn’t say you can’t do. past a restrictor plate when loose bolts allow
miles on a faster, more exotic race car, and In the first case, you may come up with an air leak in the manifold. This is
yet it hardly stands a chance of finishing it’s some technological advantage that is within debatable, and you might not be penalized
first races. If innovation and experimenta- the Jetter of the rule, if not perhaps within — the first time.
tion are felt to be more important than win- the spirit of the rule. For example, when a While it may seem unfair for some teams
ning early races, that’s fine, but once you’re Showroom Stock rule said that stock camber to come up with such technological “unfair
behind, it will be extremely difficult if not was required, I had the tires shaved advantages,” it is impossible to avoid them.
impossible to catch up again in a particular (allowable) with a camber cut into the tread If perfect laws could be written, we wouldn’t
championship series. surface itself — which gave the same end need courts and lawyers to interpret them.
As for intentional cheating — don’t result. In the second case, you look for an The only way to guarantee no tricks in
bother. It isn’t worth the time. There will idea that has not been prohibited because pitstops is to specify the stop time. And the
never be enough time to do all the perfectly no one has thought of it yet. A good exam- only way to eliminate questionable
legal assembly and development work ple of this is in gravity-feed fuel rigs, where technological advancements in race cars is
described in this book. It may be easy to fueling speed is a direct function of the to demand that they turn identical lap times.
increase engine displacement or to use ex- height of the fuel level. Since the overall Auto racing seems to be moving toward
otic fuels or to increase fuel capacity, but height of the rig was legislated, I picked up better equality in cars, if not proven and
it’s even easier to recognize and uncover about ten seconds per pitstop by having an guaranteed absolute equality between cars,
such blatant cheating. The more sophis- auxiliary transfer pump to keep the fuel level so that the contest will be more between
ticated techniques take a great deal of time topped up in it while the car was being fuel- drivers. Until then, it is the intent of this
and money to make them stand up under ed (until the trick was noticed, and a new book to make the situation at least
scrutiny. Racers are usually suspicious rule was written). Rule books almost en- academically more equal. There is nothing
enough so that a winning car will be care- courage innovation, although once the in- here that most professional and other con-
fully scrutinized, officially or otherwise. novation is noticed there will probably be sistently winning teams shouldn’t know
And once a cheater is exposed, no matter an attempt to supress it. already. But this may balance the scale a
how light the penalty, he will never again A perhaps subtle but important distinc- little by making most of the necessary
be able to celebrate an unquestioned victory. tion should therefore be made between information available to everyone. If every
The only exception is under those sanction- cheating and creative rule-bending. racer were to follow everything in this book,
ing bodies which may have unwritten rules Cheating is where blatant modifications at least more cars ought to finish more races
of their own, regardless of the official rules. could only have been made intentionally, — and hopefully more competitively.
At least they still remain relatively equal such as a hidden extra fuel tank or a
 2
TIRES AND WHEELS
acing tires seem simple enough. and road surface. The ratio of these amount of friction work being done
You buy the biggest ones you can fit numbers, or force/weight, is referred to as generates a given amount of heat, which
on your car, or the biggest ones the rules the coefficient of friction (Cf), and is one raises the temperature of the tire until it
will allow, or whatever is being used on a of the most important considerations in the balances with the amount of heat it can
faster car. You put them on your car, inflate discussion of tires. This coefficient has dissipate in airflow. That is the temperature
them and race. Since you can’t design your many important relationships to ordinary at which the tire compounder tries to design
own or modify what you get, what else mat- driving, but in this book, it represents the the tire to have its highest Cf. Various fac-
ters? However, to understand practically maximum coefficient of friction for a tire, tors which affect tire temperature are:
everything else about race car engineering, since race cars ideally operate at that max- weight of the car, cornering capability,
you have to know how a tire works. And, imum at all times. The greatest Cf that can average speed, airflow to the tire, road
when you understand tires, it becomes ap- be obtained from production passenger car temperature, air temperature, heat transmit-
parent that there are a large number of fac- tires is about 0.8 to 1.0 under ideal condi- ted from brakes or engine, tire pressure, tire
tors that can be controlled to optimize their tions. A race car tire, however, is generally tread thickness, and, of course, tire con-
performance. The racer who knows how to capable of a Cf between 1.2 and 1.5, depen- struction. The last two are primarily. the
get the most out of his tires has a definite ding on tire construction and the vehicle. responsibility of the tire engineer, and ex-
advantage over anyone else who is using the Now we will see how a number of factors plain some common characterisitics of race
identical tire. affect this all-important number. tires. Tread thickness is always at a
minimum, to prevent the unreinforced rub-
THEORY TEMPERATURE ber mass from storing excess heat to the
There is more to tire knowledge than the The most significant consideration with point where it degenerates and separates
“black art” technology the racing tire com- respect to racing tires is the temperature of from the cord body. Likewise, the sidewall
panies would lead us to believe. It’s true that the rubber under operating conditions. As has only enough rubber to keep air from
the design of internal cord bodies is based Fig. la shows, there is an optimum leaking past the cords. Heat is also
largely on experience, and the formulation temperature at which the rubber operates, generated by flex of the carcass as it con-
of rubber compounds from hundreds of and there is a significant loss in traction on tacts the ground, which is one reason for
basic elements is more experimental than either side of this peak. The temperature stiff cord bodies, as in belted or radial tires.
scientific, but a great deal of research has peak is determined by the rubber com-. The rest of the factors are a matter of match-
gone into the theory and operation of tires. pound, and is commonly believed to vary ing the proper tire to the proper conditions,
To get down to the very basics, it‘s impor- between perhaps 120 degrees for a rain tire or modifying the conditions, and are
tant to know how rubber acts in contact with on a lightweight car, to over 250 degrees for covered in other sections.
the ground. Some of the earliest and most a NASCAR stocker on high banks. A given At least this has been the accepted ‘“‘con-
complete research and theorizing in this area
was by Kummer and Meyer, although they 1.5
DRY_TIRE
ES
are probably outdated now by work in the RACE TIRE
hotter topic of rolling drag. Also, most of 1.0 RAIN 7, 1.0
this theorizing is probably irrelevant to most -
oOo
Cle tama
Re
tire engineers, much less race car
0.5
mechanics. To over-simplify, they explain
how rubber generates friction by a combina- la
tion of mechanical gripping of irregularities, 0
100 200 300 0 10 20 30 40
and molecular adhesion between the rub- temperature in degrees percent slip
ber and these microscopic contact areas.
They further explain how both conditions
can be optimized through modification of
many other factors, most of which, unfor-
tunately, are up to tire builders.
These theories explain how a tire is able
to produce “something for nothing,’ or is
able to produce a cornering, accelerating,
0
or braking force greater than the weight 0 10 20 30 40 —1 —2 0 Eero er 1000 2000
resting on it. In other words, a tire carry- slip angle in degrees camber angle in degrees load in pounds
ing a load of 1000 pounds could have max-
imum friction force of well over 1200
pounds, depending on the tire construction Figure 1. Coefficient of Friction as a Function of Tire Variables
 =e

CHART NO. PZ-260

Figure 2. An illustration of how tire surface temperatures can reach a peak (of an amazing 370 degrees) over two seconds
after the car reaches steady-state lateral acceleration in a corner.

ventional wisdom’ for a number of decades, become dull again.” knowledge.

with respect to relatively static race tire Clue 5: In verifying some computer Let’s look at what really happens, using
temperatures. But now let us consider a new simulations of vehicle dynamics, I learned typical but disguised data to protect my
theory, looking at the dynamics of tire that real vehicles usually responded more sources (Figure 2). It is common knowledge
temperature, using clues that have ac- slowly than the math models said they that a tire temperature of 250-300 degrees
cumulated over 20 years. should — and experimental “fudge factors” is about the upper limit of performance —
Clue 1: While skid-pad testing race tires were sometimes added to make up the as measured immediately after the car
in the mid-60’s, I learned that there was an difference. comes to a stop! But the first thing you
optimum operating temperature (as Clue 6: Some tire data obtained from an notice about the temperature traces, is that
measured in the pits), above and below independent testing lab had unexplainable there are peaks of up to 370 degrees in the
which the traction coefficient dropped off peculiarities — until I noticed that test turns.
noticeably. Who knows when that was first temperatures apparently had not even been All right , let’s say that the tire actually
discovered? (How old is the practice of recorded, much less controlled. starts to “go away” at something over 370
dragstrip “burnouts” to heat the tires?) Clue 7: A few years back I was running degrees, and the 250-300 reading in the pits
Clue 2: In 1968 a NASCAR stocker was some “lane change” tests (a hard left swerve is just representative. That is, perhaps 300
instrumented to record tire temperatures immediately followed by a hard right in the pits translates to 370 on the track.
continually on the track, to study loading swerve), and noticed that invariably the se- However, look at the continual variation in
on the banks at Daytona. I noticed that the cond swerve produced a higher g-level. temperature, depending where the car is on
temperatures reached unbelievably high Clue 8: At the same time, it became ap- the track and what it is doing. If the pits
peaks — and they lagged well behind the parent that the above “‘cold”’ test produced are right after a series of low-speed corners,
cornering force peaks. But it wasn’t impor- slower response figures than a continual temperatures will probably read a lot greater
tant to that particular study and was ignored. “sine-wave’’ steering oscillation test. than after high-speed corners. And that’s not
Clue 3: All the years I spent doing brake What did all this mean? One day the clues to mention the effect on reading temperature
tests, there was a noticeable “rise time”’ in just sort of clicked into place. Calling upon “profiles” across the tread, which even out
recorded braking g’s — which I assumed to some high-level contacts, I managed to get rapidly due to conductive heat transfer.
be due to the time required for weight some continual data recordings of Another interesting point to notice is the
transfer, or the chassis settling out. temperatures on Formula One tires, as the rate at which the temperature rises and falls
Clue 4: Working on Donohue’s biography car accelerated, braked, and cornered left (lagging behind the turn, or desired loca-
in 1974, I quoted him as saying about In- and right. It was amazing data! I soon came tion of peak performance). For example, it
dianapolis, “‘As the car settles into the turn, to the conclusion that perhaps no one real- typically takes the tire over 4 seconds to rise
the lateral forces rise gently to an ex- ly knows the true effect of real-world tire from 150 to 370 degrees after entering a
treme. . .When the tires start working in a temperatures on race car performance. And turn. Of course we don’t really know where
corner, the rubber heats up and becomes if they did, I suspect they could make some the optimum temperature is, but wherever
shiny. As they cool off in a straightaway, they big improvements in lap times with the it is, the tire doesn’t spend much time there.

8
If it is actually best at 350-370 degrees, then
how much performance is wasted before it
gets there?
Secondly, consider the rate at which the
tire cools, or loses its optimum performance
temperature. You can watch the dropoff in
temperature on a pyrometer even as the car
sits in the pits. But at high temperatures and
high airflow rates, tires seem to cool much
faster — about 50 degrees per second, or
roughly as fast as they heat up in a corner.
There could be a great difference in trac-
tion in the next corner, depending on how
soon it comes up. Also, of course, it makes ope PRC IB /

a difference whether the next turn is in the J

same or opposite direction, since the inside
tire tends to run cooler by 100-150 degrees
at the peak.
Given the awareness that (1) tire perfor-
mance varies considerably with
temperature, and (2) “‘tire temperature” is A visual indication of tire temperature. A pyrometer isn’t needed to tell which
not a constant, but can vary rapidly between part of the inside front tire is doing the most work. (Geoffery Hewitt)
150 and 370 degrees on the track, what are
the practical consequences?
In production cars on the street, it might reasonable to assume that we do not im- angle, or the direction the tire is traveling
be considered academic, because the tires mediately go from hard braking to hard cor- compared to the direction it is pointed.
are usually relatively “cold.” That is, a sud- nering simply because the tires are not Longitudinal slip (Fig. 1b) is obvious in its
den emergency maneuver is seldom capable of it. They require some few extremes: in acceleration the limit is
preceeded by a warming period, and the seconds to get up to the required working wheelspin, and in braking the limit is
emergency may not last long enough to heat temperature, and we simply must wait for lockup. The extreme of cornering slip angle
the tires. Yet that doesn’t mean the effect them. is 90 degrees, at which point the tire is
can’t be felt, if someone is looking for it. Also, what does it mean to the more sliding sideways. What is less obvious is that
For example, most brake testers know from esoteric question of transient the tire has its maximum Cf at some very
experience that a car will stop harder oversteer/understeer? If you are accelerating small degree of slip—say 5 percent of rota-
(higher g’s) as it slows. But while this has hard down a straightaway, then the rear tires tion or 5 to 10 degrees slip angle, and that
been assumed a speed effect, it is probably are working hard, and the fronts are the Cf drops slightly beyond that point,
in fact a tire heating effect. And although coasting along in a free air stream. If you- when the tire goes into a slide. (Fig 1c) Both
I normally use the “blend technique” of brake before the next corner, then the fronts of these factors have been explained in
high-g cornering (gradually easing off the will heat up more and be able to “pull” the detail, by Harvey and Ressler, although tire
brakes and easing into steer angle), the test car in better. But if the next turn is taken companies seldom report them and the only
technique of “‘step steer” has shown me flat out, then the fronts will be relatively meaning they have to the racer is in the feel
something else. After a hard, fast, fixed cool and understeer more at first, with the of the tire, or the warning it gives at the limit
steering wheel input of say 90 to 120 understeer diminishing over a period of of breakaway. A race driver automatically
degrees, I would swear that I usually feel seconds. tries to maintain his car at these limits by
a two-stage response. It feels like there is Again, this may all be considered steering angle and throttle control. No mat-
a noticeable step change in g-level or yaw academic. Race drivers are incredibly adap- ter what slip angle the peak Cf is at,
rate. I had previously attributed this to roll table. The best ones have intuitively been however, whether 5 degrees for a race tire
steer coming in late — except that it is doing their job better than we can com- or 10 degrees for a production tire, the peak
always a rise in effect, and roll steer could prehend or analyse, in spite of all these should be rounded and not drop sharply, or
occasionally cause a decrease in g’s or yaw dynamic complications. They unconsciously the driver will be unable to get close to the
rate. feel these limits experimentally and adapt limit without losing control. The reason for
Assuming all this is true, perhaps it to them, and will continue to do so whether small slip figures in race tires is that the slip
should be accounted for in tests of vehicle we get the theory right or not. At least now accounts for wasted forces in the form of
response, either subjective tests, or the there is one plausible explanation for the drag, and because this wasted energy con-
quantitative measures which have become eight clues. tributes to heat buildup. Passenger car tires
popular. Of course it could also be signifi- are designed to accept these losses for soft-
cant in the evaluation and/or tailoring of SLIP ness in ride and handling. The concept of
tires, as a function of ambient conditions A second consideration is the relationship slip angle also affects oversteer/
and tire compounds. of the Cf to slippage in the rubber contact understeer in the consideration of an entire
Of more immediate interest, what is the area. Whether a tire is accelerating, brak- vehicle, but as long as the front and rear
application to race car dynamics? To start ing, or cornering, it has some small degree tires have about the same characteristics
with, this should cause us to re-evaluate our of slippage. In a longitudinal sense it is (which is why you should never mix bias-
explanation for the “blend technique’’ of referred to in percent slippage, and in cor- ply tires with radials) it is primarily
race car driving. Is it cause or effect? It’s nering it is represented in terms of slip theoretical and mathematical. Likewise,

9
 BRAKING
Cf more than the inner tire gains it. Typical
15
figures are as follows:
POSSIBLE FORCE PATHS
IN TRANSITION FROM Cornering force with no weight transfer:
BRAKING TO CORNERING
outside tire 500 Ibs X 1.30 = 650
inside tire 500 Ibs X 1.30 = 650
= 1300 lbs
Force with typical weight transfer:
outside tire 750 Ibs X 1.27. = 952
inside tire 250 Ibs X 1.31 328
RIGHT LEFT 1280 Ibs
CORNERING CORNERING

These figures don’t seem to show a great

deal of loss between the two conditions, but
even the loss of twenty pounds’ force in cor-
nering power can affect the handling when
it is at one end of the car. Also, these figures
are for lightweight race cars on very large
tires. As the loading increases, as with
15 NASCAR sedans on 9.5-inch-wide treads,
ACCELERATING the effect increases. This factor
Figure 3. Ellipse Representing Peak Tire Coefficient in Any Direction demonstrates one reason why the race car
builder tries for the lowest center of gravity
to minimize lateral weight transfer in
longitudinal slip figures are most valuable some bump absorption, and yet it must be comering.
in the design of anti-skid brake systems and high enough to minimize heat-generating
drivelines. tire deformations. In addition, the genera- OTHER FACTORS
tion of heat during a race will raise
CAMBER pressures by 10 to 20 psi and cause Finally, it is important to realize that the
pressures to vary from one side of the car Cf varies a great deal with respect to the
The Cf is also affected a great deal by the surface the tire is on. This is true not only
camber angle or inclination of a tire. Ob- to the other depending on which tires are
working hardest at any one time. Regardless in such an obvious condition as a wet or oily
viously, as a wide tire is tilted, one edge is track, where the Cf can drop to 0.3 or less,
raised off the pavement and traction is lost. of what pressures are recommended as ideal
by engineers from their experiments on flat but in the variation from a dirty to a well-
A rubber tire may be expected to flex used section of the same dry track. Even
somewhat to keep an even pressure foot- test surfaces, it always comes down to deter-
mining the actual pressures to fit a particular an invisible layer of dust or rubber particles
print, but this flex also allows the tire car- can have a noticeable effect. Measured coef-
cass to roll under when subjected to extreme car and track surface. The most general
statement that can be made is that in the ficients can vary by 10 percent around the
cornering forces. The curve in Fig. 1d in- same track with the same tire on the same
dicates how camber reduces the Cf of a range from 20 to 40 psi, the Cf goes up
slightly with increasing pressures, except as day. The driver of a race car automatically
typical race tire and also shows that in fact, discovers these variations and takes them in-
a slight inward inclination of perhaps 1/2 it can be greatly decreased by track
roughness. It’s worth mentioning that to consideration, even if it’s not always
degree actually increases the Cf. It is not possible to find and drive on the areas with
clear whether such test data is accurate to enough pressure is also required to keep the
tire bead seated on the wheel rim, as cen- the highest Cf. The point is that a specific
this degree of precision, however, because Cf can never be assigned to a tire, except
it becomes difficult to measure angles under trifugal forces try to pull it off.
under impossibly limited conditions. The
these forces and deflections. It is quite possi- numbers are always relative, but so impor-
ble for the wheel to be deflected half a LOADING
tant to the ultimate performance of a race
degree, and the tire to be deflected over one Another property of the tire Cf is that it car that they must always be considered.
degree with respect to the wheel. Since tire varies with the total load on the tire, as seen There are other factors in tire construc-
companies don’t provide the information in Fig. le. As unit pressures rise anywhere tion or use that ordinarily would be of no
anyhow, it’s necessary to determine the ac- in the rubber contact patch, the rubber is interest, except that they should be mention-
tual optimum camber angle for each tire on less able to resist frictional shearing forces. ed. As the tire is essentially a flexible
each vehicle. In other words, a tire that has a Cf of 14 device, it has a spring rate and some inter-
on a 2000-pound car, might have a Cf of nal damping. These rates are so high com-
PRESSURES 1.2 on a 3000-pound car. Where this factor pared to the suspension spring rates that they
Tire pressure is another factor that can af- gets more complicated, however, is in the are generally ignored. However, they can
fect the Cf, although it is difficult to deter- lateral weight transfer under cornering, become harmonic with the chassis under
mine with respect to other factors, and prac- when the outside tires are more heavily certain conditions, and create a destructive
tically impossible to generalize upon. The loaded and the inside tires less so. At first chatter or wheel hop. A tire also has a self-
proper pressure is one at which there is an glance, it would appear that there would be aligning torque, which tries to keep it travel-
even rubber pressure in contact with the no loss in total cornering capability, because ing in a straight line, and obviously, the
road, all across the tire patch. Pressure must the outer tire gains as much weight as the wider the tire, the greater the tendency. If
also be at a minimum, to get the maximum inner one loses. In fact, however, the change the steering kingpin were located in the ex-
rubber area on the ground and to provide in Cf is not so linear, and the outer tire loses act center of the tire, it would be very dif-

10
ficult to steer it. The tire would have to be from any number of articles or tire com-
twisted in its own contact patch. Offsetting pany handouts. However, the choice be-
this pivot point allows the tire to tend to roll tween bias and radial construction is usu-
around one edge. Another property is ally decided by the tire companies and
pneumatic trail, which means that the forces rules makers. For low-level amateur rac-
on the tire move the contact patch away from ing you may get relatively inexpensive
the geometric center, upsetting static racing bias or street radial tires, while
geometry. It should also be noted that tires high-buck pro teams running on
can be made (intentionally or unintentional- superspeedways will get cord angles that
ly) asymmetrical in cornering capabilities. are closer to, but not exactly, radial in
This might be an advantage on an oval track, construction. The lower slip angle with
or a definite disadvantage if it happens ac- the radials will give more cornering power
cidentally on a car intended to corner in with less drag, and therefore be faster and
both directions. Finally, most of these fac- more durable and run cooler, but will
tors are related to steady-state conditions, require more driving finesse at the limit.
whether in high-speed straightaway driving
or low-speed cornering. In practice, SIZES
everything is always changing, or transient,
Tire size is the most apparent criterion in
and can’t be isolated. When the suspension
allows the tire to oscillate vertically and/or selection. It’s obvious that with all other fac-
tors such as tire compound and pressures
move laterally over bumps, the Cf always
being equal, the more rubber on the ground,
loses.
The most ignored concept in tire the greater the traction. But another advan-
dynamics (at least before the first edition of tage is that the greater contact area permits
this book came out) is the friction-circle the use of softer or stickier compounds. Not
theory. This theory is related to the obvious only is there more rubber to wear away for Tire width illustrated by the stacked
fact that a tire has relatively equal traction the same tire mileage in a race, but there height of four Formula One tires.
capabilities in any direction—accelerating, is a larger surface area to dissipate the
braking, or cornering, as shown in Fig. 3. greater heat generated. There appears to be
In reality, it’s more of an ellipse, with more no basic physical limit to the overall width that a wider tire will not always be faster,
traction in the longitudinal plane for most of a tire tread. The greatest difficulty in because of other over-riding effects such as
tires with tread grooves, but a circle is a making them wider has been in construc- compound or geometry changes. Width
close enough approximation. (Actually, a tion and the machines that are required. Any alone is not an adequate criterion.
real driver/vehicle capability limit would be flexible, inflated structure tends to take a A secondary consideration is tire
a complex truncated pear-shaped curve, as spherical shape, while it’s necessary to keep diameter. Here it’s not quite so easy to make
discovered from computer simulations, but a tire tread area flat across its surface. a clear-cut decision. There are conflicting
that is too esoteric to explain here, and is Because thick rubber shoulders are not rigid considerations. In the first place, there can
only of theoretical interest anyhow.) In any enough, and they store too much heat, the . be a choice of wheel diameters, which have
transient operation of a race car, say from tire engineers have been slowed by the recently ranged from 16 inches down to 10
braking to cornering, there are two ways to technology in designing flat cord bodies. inches on various radical race cars. The
get from one limit of traction to the other. Tire sizes and size designations change so current selection for most professional
The driver may either suddenly get off the fast that any specific description would soon formula racers is between 16 and 13 inches,
brakes and turn the wheel, or ease off the become obsolete. The only sure way to and it’s not uncommon to see different sizes
brakes while he gradually applies more and make a selection is with a tape measure. The front and rear.Then the question of section
more steer angle. As can be seen from the first interest is in tread width, and specifi- height, or overall tire diameter, is added to
force paths, the first option allows the car cally, that part of the tread in contact with that.
to go through a short period where no work the ground. Currently, maximum tire or Since the tendency in front seems to be
is being done by the tires. This is wasted rimwidths are limited by the rules in al- to try and reduce diameter from the typical
time, no matter how short, because the car most every racing series. While there maybe 26-inch tire on a 15-inch wheel, we should
cannot instantaneously change its motion too wide for a small-displacement race car, consider both the advantages and disadvan-
from straight-line braking to a curved path. it has never been conclusively demonstrated tages of this reduction. The most obvious
It has been mathematically and experimen- that there is such a thing as too much tire. advantage is that it reduces the rotational in-
tally proven (by Mark Donohue) that the The greatest current problem is in sanction- ertia of the assembly, which is equivalent
second option—trying to keep the tire and ing groups which limit the rim width, to weight in pounds when accelerating or
car on the perimeter of the friction circle forcing tire engineers to design for radical braking. Reducing diameter also lowers the
(or envelope) at all times — is a measurably overhangs. When SCCA sedan rim widths center of gravity of each tire and wheel,
faster way around a race track. Not only were at 8 inches, some treads were over 10 although this has a small effect on the car
that, but it reduces the stresses on the car inches wide. As long as there are no restric- as a whole. And considering there is a re-
and tires which are created by sudden tions, however, select the tire first, and then quired minimum ground clearance, it
movements, and it is almost the definition follow the manufacturer’s recommendations doesn’t lower the chassis any further.
of smoothness in driving technique. as to proper rim size. Each tire is designed Another advantage can be gained in suspen-
to have the tread flat with the beads a cer- sion geometry, in making it possible to get
TIRE DESIGN tain distance apart, and if this distance varies the roll center closer to the ground without
The internal construction of different by a few inches, evenness of the rubber con- other negating effects. And finally, on open-
types of tires ought to be familiar already tact area may suffer. Finally, it is possible wheeled cars, the smaller-diameter tire ob-

11
 tolerance for error. But it’s still possible to the least cut out is the best bet, all else be-
get a tire that has a 0.l-inch diameter dif- ing equal. Even the slickest of tires,
ference or out-of-roundness. It’s worth however, needs some sort of wear indicator,
checking with a dial gauge and tape such as holes sparingly molded into the
measure, especially if the car has a high- tread surface. All race tire treads are as thin
speed vibration, or pulls to one side under as possible anyhow, to reduce heat and in-
acceleration or braking. In oval-track rac- ertia, and ideally, should be no thicker than
ing, an intentional left-right diameter dif- required to last one race in a really com-
ference, or “stagger;’ is often used. petitive class. If the tire is only half-worn
However, this is unique to cars that corner at the end of a race, perhaps a softer com-
only one direction, and is best dealt with pound would be better, or perhaps even start-
by trial and experience rather than engineer- ing the race on half-worn tires. Of course,
ing analysis. where street-stock tires are required, tread
Clearances are another very important shaving may be mandatory to avoid ex-
consideration, particularly on production cessive heat buildup and subsequent blister-
sedans and sports cars. It’s hard to visualize ing or chunking.
the extent to which a wheel and tire nor- By the time all else is decided, there may
mally travel, and the amount which the tire not be much in tire compounds to choose
will deflect on a rim. The only real way to from unless it is an important enough race
find whether a tire will fit in a limited for tire companies to be participating. The
wheelwell is to mount the tire on the cor- only way to determine compound on a cold
Suspension components cut through rect wheel and try them on the car. It’s tire sitting in the pits is to read the manufac-
three cord layers on this inner sidewall. usually easiest to remove the suspension turer’s identification number stamped in the
springs and manually move the suspension sidewall, and then ask their engineer. At
and steering through all extremes of travel. that, about all he can say is whether it is
Then, additional clearance must be provid- a gumball, or very sticky qualifying tire, an
viously has less air drag. ed to allow for wheel and bearing deflec- intermediate, or a very hard compound. If
The disadvantages are almost as great. As tion (which can amount to one degree off some testing has been done recently, he
diameter goes down, the contact patch the axle centerline), and lateral tire deflec- might even know about how many laps it
grows shorter—in theory, at least—and the tion on the wheel. The worst examples of will last, and whether it’s likely to overheat
shorter and wider it gets, the more unstable tire shift are probably the cantilevered and blister the tread off. For the amateur
it is for the driver to control. More impor- sidewall tires on narrow rims, which can racer who buys tires at the warehouse,
tant, as section height is reduced, there is move as much as 1.5 inches laterally. A especially older or discontinued lines, it’s
less distance for the sidewalls to deflect, and quicker but potentially much more expen- important to remember that they aren't
the tire becomes more rigid and less able sive method is to try the tires on the track. designed for storage or long life. Advances
to absorb road roughness. On a very smooth With a good ear and good luck, a driver can
track this is no problem, but it can cause hear a tire rubbing before it gets cut or
severe chatter on other surfaces. On a heavy worn. With bad luck he can destroy the tire
car, a smaller-diameter tire will also have and possibly the car. When there are doubts
a tendency to run hotter. Not only does it as to whether the tire is occasionally
have less working rubber in its cir- touching somewhere, the inner sidewall can
cumference, and rotate faster, but there is be marked with chalk and then examined
less surface area to radiate heat to the air. to see if it gets rubbed off. Clay can also
And finally, on heavy, enclosed cars, a be used, by building it up on suspicious
small-diameter wheel simply doesn’t have metal surfaces and checking how far the
room for an adequate brake assembly— rotating tire wears into it. Chassis construc-
whether disc or drum—much less enough tors often use a template of the tire’s radial
cooling airflow for it. The most common cross-section mounted on the axle to check
application of 13-inch wheel is on the front clearances.
of formula cars, which have little weight on
the front, and therefore less braking TREADS
requirements. Once the tire size is determined, there
It may be too obvious to mention, but it’s may not be much choice in tread design, ex-
worth keeping in mind that tire diameters cept between different manufacturers. For
also affect gear ratios. On cars with quick- many years it was assumed that some sort
change ratios, this can be corrected relative- of grooves (circumferential) and sipes
ly easily. Conversely, on cars with more dif- (lateral) were necessary on a race tire, to
ficult differentials, a change in diameter can allow for deflections in the contact patch and
take the place of a gear change. Another to swallow dirt from the track. It has since
point to consider with really high- been proven that for dry pavement it is far
performance lightweight race cars is the more important to have the maximum The famous ‘‘gatorback’’ tread, design-
variation in tire roundness and diameter. amount of rubber in contact with the road, ed to channel water from the contact
Racing tires are usually more carefully built and slicks are now the rule. Apparently, patch.
than production tires, and there is less then, it would seem that whichever tread has

12
in rubber compounding cause race tires to
be obsoleted almost monthly, and even if
not, the rubber ages rapidly. A race tire that
is stored for a year or two will continue to
vulcanize on its own until the rubber is
harder and more brittle, and not only does
it have less adhesion, but it will tear more
easily.
Rain tires are a special case in both tread
and compound. According to the theory of
tire traction, there are two considerations
when moisture enters the picture. A light
film of water, such as fog, dew, or mist
coating the track, reduces the molecular
adhesion. But it is only when the water film
gets thick enough to prevent the mechanical
gripping of irregularities that a definite tread
is required. So there are stages of wet-track
requirements. As long as there is a light
film—which is likely to be rapidly
evaporated or thrown off anyhow—a perfect-
ly smooth tire tread may still be best.
However, the moisture will cause tire
temperature to drop, and a softer compound
would be advantageous. When the water
reaches a point where it fills in between the
irregularities, or it collects in puddles, it is
time to go to a tread with definite grooves. A new tire being mounted, showing the six retention screws in the rim
The grooves serve both to channel water out which keep the tire bead seated even when pressure is lost.
of the contact patch, and to reduce the total
contact area for higher unit pressures be-
tween the rubber and the track. For ex-
prevent sudden loss of the seat or keep the brakes. The conflicting problems in wheel
tremely wet conditions, the tread may be as
tire on the wheel even if it goes flat for other design are strength versus weight and
much as 1/2 inch high, and half its area
reasons. airflow through the center. For the weight,
devoted to grooves. These tires also have
A third important dimension on race car magnesium is currently the most common
successively softer compounds, since they
wheels is offset, which is the distance be- compromise, although filament-reinforced
work at lower temperatures. If, for some
tween the centerline of the rim and the ~ plastics show a lot of potential. Aluminum
reason, a rain tire has to be run on a dry
mounting face. Racing wheels are com- wheels are the street-racers copy of the real
track, there is a good chance of tread
monly available with offsets from +2 to —2 thing, and frequently weigh more than the
separation due to the heat, or at the least,
inches. The greatest offset would be best for production steel wheels they replace. The
rapid wear and poor laptimes.
minimum weight transfer and maximum strongest design in any material is a solid
cornering, except for a few other considera- disc center, although that allows very little
WHEELS tions. Even when it isn’t necessary to keep airflow to the brakes, unless inboard brakes
Racing wheels are apparently another off- the tires under fenders, more offset affects are part of the package. Where solid centers
the-shelf item that everyone buys simply the suspension geometry and ride rate, and can be used, they are made of spun sheet
because they look right, or because they’re also creates more stress on the suspension. stock or honeycomb laminates. The most
wider, or because they’re on a faster car. But On the other side of the wheel, offset may critical high-stress areas on a wheel are the
again, to optimize a car’s performance, and be necessary to keep the tire from rubbing, bead seat, where high cyclical cornering
to take full advantage of the rules and what or to provide room for brakes and suspen- stresses are resisted, and the hub flange,
is available, you need a deeper under- sion components. Some wheels are even where bending and rotational torques are the
standing. Size is truly an important con- designed specifically to clear disc brake greatest. Some well-known cast wheels have
sideration, as mentioned in the section on calipers. Spacers are sometimes used to in- had the design error of spokes that were nar-
tire widths. The rim diameter is fairly well crease offset, but this also increases bending rower at the hub flange than at the rim, and
limited by the suspension and brakes on any stresses on the lug bolts and allows more the centers frequently broke out. Lug nuts
particular car, but rim widths are limited misalignment. should be the ones designed for a specific
only by clearance, imagination, or racing The construction of a wheel can be just wheel, especially with cast wheels, which
tules. Most tires are designed for a specific as important as its size. A racing wheel has may require captive washers to prevent
rim width, plus or minus an inch or so, but to be much stronger than a production galling of the alloy face.
it’s safe to say—the wider the better. For a wheel, because the traction loads are so The total weight of a wheel and tire com-
rim wider than recommended, or for very much greater. Two-ton NASCAR sedans bination is important for a number of
low tire pressures, some sort of mechanical running on banked superspeedways, for ex- reasons. It is well known that a low ratio
retention is advisable for the tire bead. Some ample, are required to have reinforced steel of unsprung weight (wheels, tires, brakes,
racers use screws through the rim, or more wheels, which are very strong and very axles, and half the weight of springs and
sophisticated clamping devices, which either heavy, and allow little air to pass over the shocks) to sprung weight (everything else)

13
 timized. This allows the racer to know when
they are operating at their best, or if not,
why not. The second stage is to obtain a
number of tire designs or compounds, test
each one completely as in stage one, and
identify each as best for any particular
condition. And stage three is for a driver
to work with tire-company engineers to
produce more extreme examples of the best
tested.
The optimum temperature and camber
angles should be determined first, since they
have the greatest effect on the Cf. The tires
should be broken-in, if necessary, and then
allowed to cool. Assuming the front tires are
to be tested first, the car is set up to
understeer, preferably by using a very stiff
anti-roll bar in front. By following the
driving and measuring technique as describ-
ed in Chapter 13, the tire is pushed harder
and harder, with laptimes and temperatures
taken every few laps, until it’s obvious that
performance is dropping off with increas-
ed temperature or the car weight and cor-
nering capability aren’t great enough to heat
the tires any higher. In conjunction with
average tire temperature, the temperature
profile should be taken in at least three loca-
tions across the tread. If it is noticeably
uneven, the camber angle should be ad-
justed away from the hotter edge, to get the
contact patch rubber working more evenly.
A tire temperature card, which shows typical temperatures in three locations on Either camber or temperature can be studied
all four tires after a hard clockwise cornering test. independently, but it saves time and tires to
work on both together. Finally, with the con-
tact patch working evenly, the skidpad times
contributes to good ride, or in the case of 20 to 40 pounds to be transferred from the can be converted into lateral acceleration in
a race car, good traction. Unfortunately, the inside to the outside two wheels. In addi- g’s (approximately equal to the average Cf
ratio is worse for a race car than a passenger tion, this much more weight transfer can be between left- and right-side tires), and plot-
car because of the size and strength required added or subtracted on the front wheels ted against tire temperature to identify the
in unsprung components. Another reason alone, by a steering correction of one revolu- optimum point. If the Cf curve is still rising
for reducing tire and wheel weight is that tion per second at the steering wheel. at the highest attainable temperature, simply
the rotating mass has a rotational inertia Whether these forces work with the chassis pick an arbitrary temperature to use as a
which contributes to the effective mass of or against it requires a complex analysis, but constant for further tests.
a vehicle—which must be accelerated or they do have an effect on vehicle control, At this point, the proper camber angle can
braked. In the case of typical race car tire and should be minimized. Rotational iner- be rechecked. Evenness of the temperature
diameters, this can account for an approx- tia can be measured by swinging each across the tread is a good indication of the
imate additional 3/4 pound for every added rotating component from a torsional pen- proper camber in skidpad tests. But since
pound of tire or wheel weight. In other dulum, and gyroscopic precession can be there may be other influences, such as bet-
words, if a new set of tires weighs a total calculated from equations out of an ter cooling on one side of the tire, or the
of 10 pounds more than the old ones, weight engineering mechanics book, but the above fact that uneven temperatures can balance
and rotational inertia account for a total of estimates are true for most race cars in the out somewhat in just the few seconds it takes
17.5 pounds of additional mass that must be 1500- to 2500-pound range. to stop and take a reading, camber change
accelerated. Likewise, taking 10 pounds out should be compared to skidpad laptimes—
of the wheels reduces the effective mass by DEVELOPMENT but always at the same tire temperature.
17.5 pounds. There are three stages to the development Vary the camber angle by plus or minus a
Wheels and tires also have gyroscopic ef- of race tires, and all of them require a skid- degree, to see if either improves the Cf. If
fects that are commonly feared even if not pad or a simple, accident-resistant race so, take that as the true indicator rather than
understood. Obviously, the greater the rota- track. (The use of either is further covered temperature evenness. It’s worth mention-
tional inertia, the greater the forces on the in Chapter 13, but this section will include ing that the true camber angle—in action—is
chassis. For speeds of around 15Omph, with some basics on tires.) The first stage is to not known, nor is it important. All this test
10- to 18-inch-wide wheels and tires, the take a given pair of front or rear tires, not shows is the maximum capability of the tire,
forces can be noticeable. In a hard turn at four tires, and change all the available and the proper static camber angle setting
that speed, gyroscopic precession can cause variables until they are understood and op- for that suspension and tire.

14
 ni
naeA CAO

entanasnrsomarcennmerantenntrt ey

Comparison of the inner and outer tires under hard cornering shows the extreme lateral tread displacement
with respect to the rims.
Ultimately, what really matters is how fast
Once proper temperature and camber are on the diminishing average Cf between a the tires will go around a race track, or
resolved, the less measurable effect of tire more heavily and more lightly loaded tire. whether they’ll last the required distance.
pressure can be checked—again trying to If at any time during front tire tests, the But there are two ways to get to that point:
maintain a constant test temperature. If the front finally gains enough traction from the by sophisticated and controlled test techni-
first tests were run at some reasonable change of variables so that the car stops ques, or by cut-and-try methods. A
pressure—say between 20 and 40 psi—there understeering or begins to oversteer, that’s _ manufacturer may produce three tire
should be small effect on a racing tire above great! However, it does prevent further designs, in three compounds each, with
those figures. However, advances in tire measurement of the front tires. At that point varying tire pressures, and run lap after lap
construction and lightness in vehicle weights it is time either to go to a stiffer front anti- around a complex road course to try and tell
make pressures worth looking at. So some roll bar or to start improving rear tire trac- which produces the lowest laptimes—while
extremes—say plus or minus 10 psi—should tion until the car understeers again. The assuming that all other variables, such as
be tried to see if there is more to be gain- measurement of rear tires is more difficult the driver and car, remain constant. This
ed. At the very least it will show what is because it is harder to control a race car on may be the most foolproof way to determine
to be lost if tire pressure must be decreas- a skidpad in an oversteering condition. As the best tire—for that car, on that track, on
ed because of track roughness or increased always, the objective is to be as steady as that day. But it does little for the state-of-
to get more contact pressure on a wet track. possible in throttle application, so as not to the-art, or for a real understanding of what
There are other variables in the perfor- cause sudden breakaway. High engine rpm is happening, and how it is happening.
mance of a tire, but they are less significant and a very smooth throttle linkage help a On the other hand, some sort of real-life
or less controllable. The rim width can be great deal. test situation is valuable even after a long,
increased, with temperature, camber, and As a final note on skidpad tests, careful development session. The previous
pressure held constant (or varied again to remember that ambient conditions and the tests were limited to steady-state conditions.
verify the optimum operating conditions). track surface can change during a day’s hard For a given tire, rough track traction or
The question of maximum load effects on testing, so it’s a good idea to go back to the sudden transient response in or out of a
a tire are relatively unimportant, consider- original baseline configuration and see how corner could be impossible to live with. If
ing the fact that there is little that can be much, if any, it has improved on its own. a standard, repeatable transient maneuver
done about it. The basic philosophy of rac- Over a long period of testing, even tire wear were possible, such as a controlled J-turn,
ing is to have the lightest vehicle possible can have an effect on performance. Without it might indicate how sharp the breakaway
for any number of other reasons. Besides, intentional tire tests, this wear can even in- point was, or how sharp the Cf versus slip
it is practically impossible to isolate the ef- dicate, over the life of a tread, something angle curve was. But in the final analysis,
fects of total load in a race car tire test about the performance of a tire. If the wear a car is controlled by a very human driver,
anyhow, considering the interaction with pattern is grossly uneven, then it’s time to and what is comfortable for him can be
suspension geometry. This is also the case make adjustments from insight and ex- more important than the highest coefficient-
with the negligible effect of weight transfer perience or by careful development steps. of-friction characteristics.

15
 great enough, a small stone could do the
trick.
The high and cyclical stresses on a racing
wheel make it mandatory to inspect each
one frequently for fatique cracks, in addi-
tion to normal scratches, especially in the
case of lightweight alloy wheels. The most
certain and expensive method is by X-ray,
which may have been done by the factory
to detect casting voids. The more familiar
practice is to use a penetrating dye and
black-light inspection, which is more ade-
quately covered in Chapter ll. The most
common starting points to watch for cracks
are: the spokes at the hub root, the lug nut
faces if they are countersunk, and the inner
bead seat. It is possible that a cracked
casting can be welded by a certified aircraft
alloy welder, but he will refuse to guarantee
it because of the effect of heat on the sur-
rounding metal, and the further stresses it
causes. If a wheel cracked once, it will
probably break completely the next time.
Preparing for battle may mean mounting thousands of tires. Goodyear has pro-
bably invested more money in racing than any automobile company in the world. The balance of high speed racing tires and
wheels is important, but not so critical now
that construction techniques have made
to get the bead on the rim seat. Also, com- them more accurate and consistent in cir-
PREPARATION petition magnesium wheels have high cumference. Ordinary static balancing, or
Everything that has already been said casting porosity, and they can leak worse bubble-balancing, is probably adequate on
about tires can be ignored if the driver is than the tire does. Some wheel manufac- a heavy sedan, since there is so much mass
content to go out and circulate somewhere turers offer spray cans of sealant that should to resist small vibrations. Static balance is
down in the pack with the also-rans. But the be coated on the inside of the rim before also good enough for any solid axle rear
care and maintenance of tires and wheels mounting tires. end, because it’s only when a wheel can
can be a life and death matter. The first step An alternative to fighting a tire onto its pivot about its hub centerline that diagonal,
is in putting it all together right. rim is to build the rim around the tire, as or dynamic, imbalance will be apparent. But
Most race tires leak like sieves. In the in- truck wheels are with split-rims. Some even on lightweight open-wheeled cars, with
terests of low weight and low rubber masses, racing wheels are available in which the two four-wheel independent suspension, there
the sidewalls have just enough rubber to sides of the rim are bolted to a common may not be a need for dynamic balancing.
hold the cords together. If immersed in center section. This type of rim may be So little weight is ordinarily needed (it is
water, they can be seen to simply ooze air. heavier for the same strength and more added evenly to the inside and outside rims)
Air leakage shouldn’t matter a great deal, troublesome to assemble and seal, but it that dynamic balancing is seldom important
however, considering the fact that a race does have some advantages. Most of these unless the driver still feels an imbalance.
tire’s useful period of operation, or perhaps wheels are available with many different Most people who balance alloy wheels are
total life, can be measured in hours. Tubes widths of inner and outer rims, so not only aware of the stick-on lead weights, and also
are never used in a race tire (except on can the total width be varied, but the offset automatically cover them with racer’s tape
NASCAR superspeedway cars for safety can also. Another good point may be the as insurance. It’s up to the mechanic to see
reasons) because they store excess heat and lack of a drop-center in the rim, which is that inner wheel weights don’t get knocked
add to weight and inertia. If a tire can be otherwise necessary to drop the tire bead off by interference with the brakes or
shown to be operating below its optimum into during mounting. Having a relatively suspension.
temperature, a tube might be used to raise flat inner rim area provides more clearance Mounting new wheels and tires on a car
it, but anyone who has that knowledge pro- for brake assemblies and cooling airflow, may be the first opportunity to check for
bably also has the option of using a softer and helps keep the tire on the rim if air total runout, or out-of-roundness. If the tire
compound. pressure is lost. hasn’t been sitting with a lot of weight on
The mounting of race tires on alloy rims In any case, after the tire is mounted, it it, there should be no reason for it or the
is best left to the people who sell the tires, is a good idea to overinflate it at first by 20 wheel to have a total of more than .030-inch
although a persistent individual can do it or 30 psi, to make sure the bead is well total runout as they are rotated on their ax-
with padded tire irons, a rubber mallet, and seated, and to make sure the rim seat is not le. If there is more than that, there are three
lots of soapy water or rubber lubricant. The going to snap off, as sometimes happens possibilities. The wheel may not have a
consequences of a damaging mistake, with truck rims. This has been known to centering hole which locates it on a mat-
however, such as a cut wall, or a nicked or happen on racing wheels, so it’s wise to use ching surface on the hub, or spacers may
cracked rim flange, may not show up until an inflation cage. And finally, use valve core have moved these apart. The wheel could
it is too late. Then again, once the tire is caps. Natural centrifugal forces on a spin- have been machined out-of-round, though
on the rim, it can be nearly impossible to ning tire are in the proper direction to open even a production steel wheel should be
inflate without some sort of clamping device the valve, but even if these forces aren’t within .030 inch. Or the tire may be out of

16
shape. The quickest solution is to mark the
high spot, dismount the tire, and see how
it compares to any high spot found on the
rim. It may be possible to rotate the tire with
respect to the rim to get two small tolerance
variations to cancel each other out. When
all else fails, if the tire has been paid for,
and the out-of-roundness can be felt at high
speed, the tire can still be shaved true on
machines found in some recapping shops.
One last word about bolting wheels to the
car. What with increased hub thickness, and
perhaps even wheel spacers, lug bolt thread
engagement can be a problem. A good rule
of thumb is 1.5 diameters, and more couldn’t
hurt unless quick-changes are required. As
the bolt thread pitch should be known—say
itis a 7/16-20—then as the nut is screwed
on, it should go (7/16 x 20 x 1.5) or 13 com-
plete turns. The contact faces should be free
of dirt, and each nut should be evenly tor-
qued down. If spacers must be used, they
should be carefully machined and fitted to
prevent wheel wobble. Individual washers
The safety liner, designed as a ‘‘fail-safe’’ for two-ton NASCAR stockers corner-
should never be used, because they localize
ing hard at over 200 mph, has probably saved a number of lives.
stresses and can wear into the wheel.
Most new tires require break-in before
they can generate maximum traction forces.
in a lower gear. If the driveline and brakes When a front wheel locks up—usually an
This may be due to a hard skin on the sur-
are warm enough, or marginal in durabili- inside front in a turn—it is often felt as a
face of the vulcanized rubber, or to a sur-
ty, simply swerving the car back and forth slight tug on the wheel to the opposite side.
face that is too smooth, but it should be an-
across the track may get the tires up to (Racing movies to the contrary, a skidding
ticipated on most tires, and especially tires race tire seldom makes a sound that can be
operating temperature. A tire can cool down
with definite grooves and sipes. When spare heard above the engine.) If that isn’t enough
quickly at high speed with no load, or while
tires are going to be needed during a race,
sitting still for a pitstop, so the driver needs of a signal, or there is still some doubt, a
they should also be broken-in, or feathered- flat spot will create a noticeable vibration
to compensate for this when entering the
in or scrubbed, beforehand. If a skidpad can
next corner. It’s not a bad idea to spin the in the steering wheel on the next high speed
be used, laptimes will indicate when they _ straightway—assuming that everything was
rear tires freely when leaving the pits, as
are ready. On a track it will be more up to in balance beforehand. And finally, on the
this warms them up rapidly and at least
the driver’s feel, but probably not more than next lap, all that lost rubber can usually be
prevents a drastic oversteer condition.
a few hard laps will be needed. seen in a recognizable spot in one particular
When tire-company engineers are around,
DRIVING and pitstops permit, tires are frequently place. As for the driver identifying a tire that
The last thing to consider in race tires— checked during fuel stops. The wear rate, has been worn to the cord all the way
after everything else has been tried, tested, or total tire mileage for a given car and around, things are a little more difficult.
and understood—is finally getting them on track, ought to be known, but little things Drivers who have a lot of experience in this
the track. Temperatures and pressures are such as a change in ambient temperature or situation sometimes claim that the loss of
the only variables under control at this driving technique, or a spin, can change the traction can be identified when they get to
point, and these should be carefully wear drastically. Visual inspection can be the cords, and they can get to the pits on
watched. For long races, where tires will good enough for an experienced person, but just one or two cord layers.
have to run for several hours—or the dura- a tread-depth gauge is better, as it allows A slow leak during a race is less obvious
tion of their treads—it can’t hurt to take mathematical predictions. This, of course, and can be a little misleading. The first
pressures a few hours apart just to make just indicates average wear around the cir- signal is usually a slow degeneration of cor-
sure that no tire has a slow leak. Since cumference. Tires are frequently lost nering characteristics, most often a definite
pressures rise 10 to 20 psi as a tire heats because a wheel locks up in braking or the oversteer or understeer to only one side. A
up, and the ideal running pressure is known car slides sideways, and a flat spot is worn running check can be made in the next
from tests, make the proper allowance if in one small area. The tire appears to have straightaway. If, under hard acceleration, the
they are checked while cold. plenty of tread, and yet it wears to the cord car pulls to the side in which it was
Most race drivers are aware that tires in that one spot and blows out. In the mid- oversteering, that side likely has a soft rear
don’t have peak traction when cold, and they dle of a race the person best able to discover tire. If, under braking, the car pulls away
use the warm-up lap for just that. Since the a flatspot is the driver. Due to the practice from the side in which it was understeer-
brakes and driveline usually need to be of using greater front brake bias (explained ing, that side of the car probably has a soft
warmed up also, it is most efficient to ac- in Chapter 5), and the known fact that a front tire. Of course, abnormal cornering
celerate and brake sharply—if no other cars sliding tire has less traction, the driver can or handling can also be the result of broken
are around. Another technique is to ride the usually feel a flatspot at the steering wheel, or bent suspension components, which
brakes while maintaining moderate throttle if not see it happen on an open-wheel car. won't be cured by a tire change.

17
AVRIL ROBARTS LRG
 car. Even in a road race a car will, on the
average, make one more complete 360
degree turn around the track in one direc-
tion than the other. Since the inside tires of
the car will be doing less work, a softer
compound may be used to keep them at op-
timum running temperature.
When a track starts to get wet, the deci-
sion to stop and change to rain tires is best
made by computer and prayer. Some of the
factors are: wet-driving skill of the driver,
depth of the water, rate of evaporation or
squeegee action of the cars, localized pud-
dles and streams, and the immediate
forecast. But most important is the tradeoff
between: length of time for a complete tire-
change pitstop, the difference in lap times
between drys and wets, and the time remain-
ing in the race. Even if a driver can run five
seconds faster on rain tires, he can’t make
up for a fifty-second pitstop if only eight
laps are left. Changing tires is a difficult
decision, and the very best teams don’t
A tire flatspotted by brake lockup or a high speed spin requires an immediate change guess right more than half the time.
to avoid a sudden failure like this when the cords wear through. Although tires seem simple, they have an
all-important and complex influence on the
rest of the racing package, from the driver’s
A complicated trick that is often used by tires are running too cold, or are overheating technique to the selection of gear ratios. The
racers is varying the compounds. It is not and blistering. It is also possible to vary next two chapters show how tire
unusual in NASCAR or USAC oval races compounds from front to rear, to try and characteristics totally dictate the design,
to change compounds all the way around on correct a handling problem, or even to vary characteristics, and strength of race car
a pitstop, if it can be seen that the starting compounds on the inner or outer side of the suspensions and springs.

18
 3
SUSPENSION GEOMETRY AND ALIGNMENT
Once you understand the theory and func- 4. The longitudinal force Fx comes from ed at. Race car designers frequently use a
tion of racing tires, race car suspensions can braking or accelerating, the lateral force Fy factor of three g’s vertical acceleration, or
be analyzed and modified to take full comes from cornering, and the vertical force in other words, a safety factor of three times
advantage of them. The design and con- Fz is from vehicle weight and aerodynamic the maximum of the other known forces.
struction of suspensions, especially race car loading. Then there are the torques on the Other factors that enter in are the mass of
suspensions, is so complex and so full of hub: Tx, which is generated by cornering: the suspension components and the
pitfalls, that it won’t be discussed here. Most forces at the ground; Ty, which is generated spring/damper rate, which are covered in the
racers find it challenging enough simply to by bearing friction or brake torque; and Tz, next chapter.
try to optimize what exists, whether on a which is steering or aligning torque. It is Tire load transfer, commonly called
production sedan or a factory-built formula well known that in most suspension systems weight transfer, is the vertical force which
car. A total redesign is beyond their com- any force or movement in any direction will shifts from one end or side of the car to the
prehension. The broad subject of suspen- inevitably cause another force or movement other, and depends solely on the g’s of ac-
sion geometry is so vast that it would take in another direction. A vertical bump can celeration in any direction, the location of
a book (which only a Detroit engineer could cause a steering reaction or camber change, the center of gravity, and the wheelbase or
love), plus a dozen computer programs and a brake application can cause a suspension track width. If we assume the maximum g’s
a knowledge of advanced mathematics and deflection or steering pull, and so on. Each in any direction to be a function of weight,
engineering dynamics to cover it adequately. of the six will be covered separately — with aerodynamic downforce, and the maximum
At that, there probably aren’t a few hun- some mention of its relationship to the tire coefficient of friction, then: g=
dred race car designers in the world who others. In many cases, the most vivid way
(weight + aero downforce) xX (tire Cf)
would appreciate it. to understand a concept is to consider ex-
The best bet is to modify what exists, or treme examples of positions, forces, or weight
build and test, rather than try to design the movements. Simplified equations for maximum load
ultimate suspension from advanced theory. transfer can then be given:
Occasionally a designer has been convinc- FORCES AND DEFLECTIONS
lateral load transfer
ed that he has discovered “The Truth,” and Forces and movements in the vertical or
when his brainchild was built on a race car _ (g) (weight) (c.g. height)
z direction are known as tire load and Ib
it turned out to be too complex, or bump/rebound travel. The magnitude of the : (track width)
dynamically unstable — or even statically forces is a combination of vehicle weight, ~ longitudinal load transfer
unstable — because some minor effect was tire load transfer, aerodynamic forces, and _ (g) (weight) (c.g. height)
overlooked on paper. However, it is quite bump accelerations. The first three can be lbs
7 (wheelbase)
possible that a person with a thorough easily measured or closely estimated, but
knowledge of production suspension the last one depends on the track or the This load transfer figure in pounds is then
engineering could use the following specific bumps a driver hits, and can only be guess- added to one pair of tires and subtracted
information as it applies to racing, and
further references in the bibliography, to
come up with a major advance. There are
some new ideas that haven’t been fully ex-
plored yet, but most of them are proprietary,
and so can’t be covered here.

THEORY
The first thing that has to be understood
about any moving mechanism such as a race
car suspension system is that six com-
ponents of forces act upon it, and therefore
there are six components of potential move-
ment or restriction. Also, each of these six
has to be analyzed in three modes: static
position; dynamic displacement or travel;
and transient movements or accelerations
and direction changes. In addition, there are
highly complex interactions between various
combinations. For example, consider the Figure 4. Forces and Torques on Each Wheel and Suspension
single wheel and hub carrier shown in Fig.

19
 tions. As the car goes up, the roll center
rises, and as it comes down, so does the roll
center. For a description of how to locate
and change geometric roll centers, see Car-
roll Smith’s Tune to Win.
In the same manner, longitudinal ac-
celerating or braking forces can be put to
use in intentionally jacking up the front or
rear of a car. Under braking, tire load
CENTRIFUGAL FORCE JACKING FORCE
transfer to the front causes nose dive. But
REACTION FORCE if the braking reaction force point on the
chassis is above ground level, the pole-vault
Figure 5. Jacking Effect Due to Roll Center Being Above Ground Level effect can be great enough to cancel out all
nose dive. (See Fig. 6) This is called 100
percent anti-dive, and is the basis of
from the other two. Another load transfer the roll center. The car has separate roll measurement for lesser percentages. The
is found on solid-axle rear suspensions, centers for the front and rear suspensions, same principle works in reverse at the rear
when driveshaft torque tries to force the left which may be from a foot or so above the of a car when it is accelerating. Another way
rear tire down harder and lift the right rear ground, to some inches below it. The car of visualizing this is to think of it as a ramp
tire. The measurement or estimation of leans, or rolls, about those points — the roll at the axle, resisting the longitudinal force.
aerodynamic downforce is explained in axis — due to centrifugal force in a corner. As the axle pushes against the ramp, it pro-
Chapter 6. If the center of gravity happens to be low duces a vertical force to counteract load
What a race car driver wants, obviously, enough to fall on that axis, however, there transfer. Using this analogy, it can be seen
is to reduce vehicle weight and load transfer, will be no roll. On the other hand, the roll that the percent of anti-dive at the front, or
and to increase the aerodynamic downforce. axis should be close to the ground for other anti-squat at the rear, is defined by the angle
There are no limits to these particular goals, reasons. The higher these lateral force (from vertical) in which the tire patch moves
and few negative interactions. The bump ac- centers or roll centers get, the greater is the as it rises and falls. In the case of linkage-
celeration force should also be minimized, jacking effect. (See Fig. 5) This is analogous controlled suspensions rather than ramps,
and could be zero for a perfectly smooth to a pole-vaulter planting his pole at the the angle will usually change under deflec-
track. But since track (and off-track) bumps location of the outside tire, and using the tion, or have a moving resistance point, as
must be absorbed without structural failure, force and leverage to boost him into the air. in the case of lateral jacking. The subject
this force is largely a function of how much That is definitely not desired in automobile is covered in much more depth in an SAE
vertical movement can be tolerated. All race suspensions. But the lower a roll center is, paper by Winklemann.
car chassis are as close to the ground as the greater the lean in a corner due to the It may seem as though 100 percent anti-
possible, to lower the center of gravity and leverage of the center of gravity above it. dive and anti-squat would be desirable to
to improve aerodynamics, so bump travel Or, conversely, the stiffer the springs re- maintain bump travel and aerodynamic
has to be compromised. Typical figures vary quired to resist roll. As if that weren’t com- ground clearance on a very low race car.
from perhaps half an inch of bump travel plicated enough, with most independent However, there is another complicating
at maximum downforce on an Indy car, to suspensions the roll center moves up or factor at the front suspension. To get anti-
over four inches on a large sedan on a rough down and laterally as the suspension dive means that the tire patch must move
road course. Rebound travel is limited only deflects, and usually in the wrong direc- forward as it goes up. But to absorb road
by spring/damper travel or driveshaft
deflection.
CHASSIS LIFT REACTION
Maximum lateral and longitudinal forces,
Fy and Fx, are a direct consequence of the
vertical force Fz, and the maximum fric-
INERTIA
tion capability of the tire. Since a tire can FORCE
have a Cf of well over 1.0, this means that
suspension forces in the horizontal plane are
much greater than vertical forces. In addi-
tion, due to geometric leverage, lower
suspension links carry much greater forces
than the upper links. This is particularly
important to realize because lateral and
longitudinal movement of a tire should be
essentially zero. Not only would these
movements give the car a vague or loose
feel, but they can cause a reduction in tire
traction, or have major dynamic complica-
tions such as tire chatter.
Lateral forces have to be transmitted from
the ground to the chassis, and the single
point at which their resultant force vector Figure 6. Anti-Dive Suspension Geometry
acts on the chassis centerline is known as

20
bumps, the tire would ideally move
CENTRIFUGAL|
backward, giving with the blow. Many FORCE ,
passenger car suspensions, in fact, do just
that, and amplify the dive in the search for
softness. Obviously the higher the center of
gravity, and the higher the percent of anti-
dive, the harsher the ride. Few race cars can
use more than 50 percent anti-dive because
the harshness in braking is intolerable to LEFT TURN REACTION
drivers. Conversely, anti-squat in the rear
is doubly beneficial, in that the desired tire VERTICAL FORCE
patch movement is rearward, which absorbs
bumps. An excess, however, can cause
wheel chatter under hard acceleration.

TORQUES AND ROTATIONS

If suspension forces are straightforward,
BUMP REACTION
the torques and rotations are the secret of
winning races — or disaster. Perhaps most Figure 7. Suspension Camber Effects
important, and most difficult to control, are
the camber angle and torques about the x
axis. From tire characteristics described in ment. With a swing axle, the center of rota- a roll angle of less than one degree at 1.5
the previous chapter, it’s apparent that a race tion is obviously the inner pivot. When there g’s lateral acceleration. Everyone is realizing
tire should always be perfectly vertical with are upper and lower lateral control arms, that the cost of high roll angles — used-up
respect to the ground, or perhaps cambered however, the center of rotation is at the bump travel and undesirable camber
away from the direction of a turn by not intersection of lines projected from the changes — is not worth anything it may pro-
more than one degree. What complicates control arm pivots. The reason the word vide in softness or feel. Low roll angles also
this, is the fact that the car must usually turn “instantaneous” is used, is that the center reduce the value of camber-eliminating solid
in either direction, and it deflects both ver- of rotation may be anywhere from a few feet or deDion axles. A fixed roll center at the
tically and in roll. It’s practically impossible from the tire patch in the case of very center of gravity would eliminate roll, but
to design a suspension that will produce the steeply inclined control arms, to infinity for a roll center at ground level tends to improve
optimum geometry under all conditions, perfectly parallel control arms. But in any camber change and eliminate jacking. Get-
without its becoming too complex and case, the roll center of the chassis (center ting the center of gravity down to ground
heavy. If a wheel remains perfectly vertical of lateral cornering force application) is at level would cure everything, if it were
over a one-wheel bump, then it leans in roll, the intersection of this radius and the possible, but it would take away all the fun
and if it remains vertical to the ground as centerline of the car. and confusion.
the car rolls, then it cambers over a bump. What is desired in either roll center height _ The torques and rotations about the y
The suspension can’t distinguish between or swing arm length is not easily defined, axis, or axle, would seem to be irrelevant,
roll and a one-wheel bump. Suspension since there are so many interacting factors. since that is the plane of free rotation of the
designers have compromised back and forth A relatively long swing arm radius is good wheel. However, that is where some sizable
for years, finally settling on two basic solu- in that it reduces lateral deflection (causing brake torques — and perhaps some ac-
tions. Most simple, and most familiar at the lost traction) in the tire contact patch as the celeration torques at the rear — may have
rear, is the solid axle, which of course has tire moves up and down. It also reduces the to be resisted. Brakes themselves are
no camber at all — except perhaps half a dynamic camber change, which causes covered in another chapter, but the suspen-
degree in bending. In direct comparisons steering wheel kick from gyroscopic forces. sion reactions are important in themselves.
with independently-sprung rear wheels, in And finally, the longer the radius, the less If the brakes are mounted to the chassis,
smooth corners, on heavy cars, there is no the geometry changes during suspension with double U-jointed axles out to the hubs,
appreciable difference. movement. However, as the radius ap- the situation is simplified considerably.
On lightweight race cars on bumpy tracks, proaches infinity, and camber change in When the brakes are applied, all torque is
however, the independent rear is superior, bump approaches zero, the wheel is forced resisted directly in the frame, so that the
just as it is on almost every front suspen- to remain parallel to the chassis — even only extra forces carried by the suspension
sion. In this case the most common solu- during chassis roll. Since the chassis rolls control arms are directly rearward at the hub
tion to the compromise is to design the to the outside of a turn, the wheels camber center. This rearward force in itself was
locating linkage so that the wheel cambers in the wrong direction. This can be mentioned in a previous section, so now
inward as it goes up, and outward as it goes eliminated by reducing roll, through moving only the torque of a hub-mounted brake will
down, so that both the left and right tires the roll center close to the center of gravity be considered.
tend to remain vertical as the car rolls. (See or using high roll resistance from anti-roll If the brakes are mounted on a solid axle,
Fig. 7) This means that both also camber bars. For many years drivers demanded as at the rear of most sedans, the torque
when the car hits a bump while traveling some roll angle for an indication (feedback) reaction can be used to maintain a more
in a straight line, but traction is not usually of the limit of traction. But in recent years constant ride height in braking. This can be
as critical then. they have become aware that steering wheel totally independent of the ramp effect men-
Any suspension which pivots while it is angle and forces, and slow tire breakaway tioned previously to prevent squat under
deflecting always has an instantaneous center characteristics, are perfectly adequate in- acceleration, because it is dependent on
of rotation, which is used to define its move- dicators. Some current race cars may have pure rotation of the axle housing, and not

21
 at the rear, this can be separate from the
BRAKE TORQUE LOAD TRANSFER ramp effect mentioned for anti-dive. It may
DOWNFORCE LIFT AT REAR
be possible to have a travel angle that is ver-
tical at ordinary ride height to reduce harsh-
ness, with enough rearward control arm axis
convergence to provide the desired anti-dive
BRAKE CALIPER
TORQUE REACTION effect. However, it may have adverse effects
on steering angles and torques.
Torques and rotations about the vertical
Z axis are commonly referred to as steer-
ing. But they aren’t as simple as the mere
act of intentional steering, since there are
LOAD TRANSFER
so many other movements and deflections
SQUAT AT REAR that make other steering contributions. Even
ordinary steering mechanisms include a
number of factors such as steering ratio,
DIFFERENTIAL TORQUE ACCELERATION TORQUE
LIFT FORCE REACTION IN DIFFERENTIAL
variable ratios, total angle, force required,
and feedbacks. In recent years, it has been
discovered that very little steering angle is
required — much less wanted — in a well
Figure 8. Anti-Squat and Anti-Lift with a Live Axle
set-up race car. When a car is balanced and
driven correctly, total steering wheel move-
necessarily fore-and-aft movement. For forward from the right side of the housing, ment may not be over one quarter turn
example, consider a short arm extending its action under acceleration will be to in- throughout an entire road race. Of course,
forward of the axle housing, with a link crease the load on the right wheel. However, the steering ratio has a lot to do with this,
attached to the underside of the chassis. (See under braking it will cause the right rear but the driver is an amazingly adaptable
Fig. 8) As the brakes are applied, forward wheel to try to lift off the ground. There’s machine, and a wide range of ratios are
load transfer causes the rear to want to lift. no easy way out. acceptable. Variable-ratio steering gears
But at the same time, the brake torque is There are no reaction problems with in- have been used in race cars with slow steer-
trying to pull the rear down. Depending on dependently sprung wheels with outboard ing in the center for precision in control at
the amount of axle rotation that can be brakes — as long as the upper and lower high speeds, and fast ratios on either side
allowed, enough downward force can be control arms are parallel. If the hub cannot to provide quick steering in emergency
generated to make the rear end actually rotate, it causes no upward or downward situations. It’s worth noting that almost any
squat under braking. The linkage is not that forces on the chassis. Although there are steering geometry has some inherent
simple, of course, because of other control considerable forces in the links themselves, variable-ratio, due to the conversion of
arm requirements and necessary bump they all cancel out. But consider control arm rotary to linear to rotary motion, but in this
travel. axes of rotation which converge toward each case the effect is generally negative: the ratio
Solid rear axles with integral differentials other as in Fig. 9. This means that the hub is quicker in the center.
also have the opposite torque reaction under will rotate as it rises and falls — or con- More important than steering ratio is the
acceleration, due to the input gear (pinion) versely — that as it is forced to rotate by amount of force required at the steering
trying to climb up the axle gear (ring gear). brake torque, it will try to make the chassis wheel on very heavy cars and the kind of
If the axle is totally restricted from any rise or fall. The drawing shows that this con- reactions the driver can feel at the wheel.
rotation with respect to the chassis, there vergence point forms another instantaneous As the driver approaches the front tires’
will be no lifting or downforce. But if it has center of rotation, which is the location of limit of adhesion (understeer), perhaps his
an instantaneous center of rotation in front resultant upward brake torque reaction. As best information comes from the amount of
of it, where the longitudinal forces are
transmitted to the chassis, an upward force
will also be generated at that point which
can be great enough to cancel out accelera-
tion squat. This is further covered with
mathematical examples in an SAE paper by
Winklemann. It may be desirable to have EQUIVALENT TORQUE ARM ——>-
100 percent rear anti-squat and anti-lift in — —
<a a
braking, but as usual, there are other com- BRAKETORQUE =
plicating factors such as the effect these esp Nee
movements have on the steer angle of the
rear axle, which will be covered later.
, Another point which should be mentioned BRAKE TORQUE
LIFT FORCE
here is that a solid rear axle torque-resisting
arm may be located off-center. This can help
cancel lateral load transfer under accelera-
tion, when driveshaft torque tries to rotate Figure 9. Anti-Dive from Brake Torque at Front Knuckle
the axle housing to the left, and unload the
right rear tire. If the torque arm projects

22
force required to keep the wheel turned and
how the car reacts when he moves the wheel CASTER ANGLE
the slightest amount. At the total limit of a STEERING AXIS
tire’s lateral adhesion, the steering force 1 aaa CD hhahahahnhahnd »
O
approaches zero and the car responds less Q
0)
and less to steering input. And that is what 4
)

makes an understeering car so much easier

to control than an oversteering car, in which
the driver is less able to feel the approaching
limit of rear traction.
The factors which affect steering feel —
besides overall steering ratio — are caster, OOd
Oe
Ole
ae
Mihee

kingpin inclination, kingpin offset, tire

width, tire pressure, and weight on the
steered wheels. The last three are beyond CFT IITIAIL LALLA

modification with respect to feedbacks, but

eeeKINGPIN OFFSET reat DISTANCE
the first three are all-important in optimiz-
ing control. (See Fig. 10) Caster is already
well known. It places the tire’s center of drag
behind its steering pivot, for self-alignment
at speed. Lateral kingpin inclination pro- Figure 10. Front Suspension Alignment Considerations
vides self-alignment even when sitting still,
as it causes the chassis to rise when the
Then there are the factors of roll steer. in a suspension geometry is that it’s so hard
wheels are turned in either direction so
In almost any suspension geometry there is to eliminate completely. Due to all the
vehicle weight tends to keep the wheels
some degree of unavoidable steering angle previously described desirable conditions
straight ahead. Caster and inclination also
created as the wheel travels vertically. If it that are designed into a suspension, such as
provide some degree of camber change as
were merely a matter of symmetrical bumps caster, camber, anti-dive, anti-squat, roll
the wheels are steered, although the effect
during straight-line travel, it wouldn’t matter, center, swing arm radius, and so on, a
is negligible for the small steer angle of race
because the left and right wheels would steer suspension designer may be hard-pressed to
tires. Kingpin offset is the distance between
in opposite directions and cancel each other get roll steer anywhere near zero. Besides,
the center of the tire patch and the center
out. At least it wouldn’t matter as long as quite often it’s impossible to get suspension
line of the kingpin at the ground. The off-
the wheels didn’t go into toe-out, which is components in the right location for precise
set is necessary to provide some steering
a relatively unstable condition. But as the geometry because of space considerations
feel, from the balance of longitudinal drag
car leans in a turn, it is important to have such as a tire, brakes, or engine being in
between the two front tires. Under braking,
the roll steer working with the driver and the way. At any rate, if it can’t be avoided,
this can be felt as a pull away from the side
not against him. it should be developed to produce roll
that is losing traction or sliding a tire. In
For example, a driver knows the approx- - understeer instead of roll oversteer. If the
the case of a hard corner, where most of the
imate steering angle required in a given chassis is going to surprise the driver, it
load is carried by the outside wheel, the un-
turn. But as he steers into a steady-state ought to be with a more stable state, or
balanced pull created by that wheel’s offset
condition, the car’s roll angle lags behind. understeer, even if it is likely to reduce total
provides a strong understeering force at the
When he finally reaches what he believes cornering power.
steering wheel. A few inches of offset isn’t
uncommon, but when it gets too large — to be the final steady state, the car continues COMPLIANCE
usually due to wide tires — the driver will to roll more (or may even hit an asym- The final consideration in the theory of
have a constant wheel fight on his hands. metrical bump), and it may provide still suspension forces and deflections is uninten-
Ackerman steering is the geometry which more steer angle at the front and/or rear. tional stress deflections or compliance. Until
causes the inside front wheel to steer at a If it steers into the turn more at the rear than the seventies, bending or compression of
greater angle because it is traveling in a at the front, that’s not too upsetting, because suspension components and rubber bushings
tighter radius than the outside one. This is it gives the rear tires more cornering power, was considered insignificant and was
important in low speed maneuvers such as or understeer. But if the rear steers less into ignored. Design for a soft ride was assumed
parking. However, in the case of race cars, the turn, or even out of the turn, the driver to lead inevitably to soft handling. But now
where at first glance it might seem desirable is suddenly surprised by a loss of traction engineers have begun coming up with ways
to have both front tires operating at about at the rear — oversteer — just when the car to make compliance work for them. General
the same peak slip angles, Ackerman steer- appeared to be stabilized. Motors’ “leading steer arms,” which were
ing may be a disadvantage. On the other There is no significant reason to have any introduced on the Nova, are one of the most
hand, scientifically-obtained tire test data intentional roll steer in a race car, assum- interesting applications. Traditionally,
tends to indicate that the lighter the tire load, ing steer angle is parallel on the left and steering arms are mounted to the rear of the
the higher the slip angle required for peak right sides. The driver can control the front suspension, and the steering linkage is
cornering power. This would indicate that or rear tire slip angles up to their peaks usually stiffer in lateral compliance than the
Ackerman is in fact useful in race cars. As simply by basic chassis setup and the steer- front suspension. In a hard corner,
in the case of caster and inclination, though, ing or throttle application in a turn. Any therefore, the suspension deflects inward
there probably isn’t enough steering motion uncontrolled steer angle from rolling or more than the steering linkage, causing the
to have a very significant effect. Only your deflections simply complicates his task. The wheels to steer inward, which contributes
skid pad or test track will tell for sure. most common reason for having roll steer to oversteer. Conversely, if the steering arms

23
 decelerating free roll decelerating

Conventional Suspension Weissach-Axle

Figure 11. The Porsche 928 ‘‘Weissach’’ rear suspension arms are designed to flex intentionally, to create a stabilizing toe-in
during braking. In effect, it works like a trapezoidal 4-bar linkage rather than a semi-trailing arm. (Porsche)

are mounted to the front, the effect will be the wheel to steer inward (See Fig. 11). ahead of the rear axle, across the top of the
to provide more limit understeer. These two effects tend to cancel each other pinion, and call it a"J-bar." It's a result of
For a given amount of rubber bushing out for a more neutral response. the fore or aft offset direction of the Panhard
softness, the further apart the mounting The concept of exploiting rear suspen- bracket creating a steer torque on the rear
points are, the less the angular deflection of sion compliance to produce intentional axle. (See Fig. 12) And active control of
the suspension will be. The MacPherson understeer or oversteer in also seen on the rear compliance (actually rear steering),
strut, for example, may not have camber location of the Panhard bar on oval cars. has appeared on a few production sedans
characteristics as desirable as double A- The familiar location behind the rear axle and race cars, without so far demonstrating
frames, but it was soon discovered that it promotes understeer, as rising cornering any great value aside from parking ease.
reduced the problem of camber compli- forces cause the rear axle to try and steer As rubber suspension bushings are elimi-
ance. With these struts, the upper and into the turn. Conversely, on short dirt nated from race cars, compliance comes
lower suspension mounts are about as far track cars, where they want oversteer to mostly from the takeup in joint tolerances
apart as you can get them -- from just under kick the rear out, they put the Panhard bar and deflection in assumed rigid metal com-
the hood, to the bottom of the chassis.
General Motors has used a live rear axle LEFT TURN
with a driveshaft-long “‘torque arm’’ as a
similar solution at the rear of the Camaro.
Of course, the torque-tube rear axle, witha PANHARD BAR |
single driveshaft U-joint at the rear of the
transmission, goes back past the Model T. oe
rca < —————
———
But NASCAR rediscovered it, calling it
"truck arms," to provide more positive
location control on superspeedway cars.
Porsche has even used controlled com-
pliance to control rear suspension over-
steer. For years they were famous for their
“‘trailing throttle oversteer’’ in the 911,
which means that as driving power is re-
duced during cornering, the more heavily
loaded outside rear wheel tends to steer
outward -- an oversteer condition. But
starting with the 928’s ‘“‘Weissach axle’’
the trailing control arm is designed to bend
intentionally -- in braking only -- to allow Figure 12. Panhard bar contributing to understeer, J-bar contributing to oversteer.

24
 COMMON SUSPENSION TYPES

CLASS TYPE SUB-TYPES COMMON USE ADVANTAGES DISADVANTAGES

live axle, or various rear- low cost poor space

Hotchkiss locating most large fair handling poor ride
linkages | FE/RD cars poor weight

rear- low cost poor space

beam or some large fair handling
rigid axle FE/FD cars fair ride
(front - good weight
Dead axle some trucks)
(non-driving) low cost
trailing rear- good handling
torsion many good ride
INTERCONNECTED beam axle FE/FD cars good weight
good space

deDion sliding tube, rear- good handling high cost

(Sprung rotating tube rare, on fair ride poor space
differential) FE/RD cars poor weight

front- fair cost

Double or rigid arms, most large good handling poor space
unequal link arms, FE/RD cars good ride
A-arms some FE/FD
driveshaft good weight
as a link rear-racers
fair handling
low cost
good ride
good space
good weight
Arm and strut
rear- low lost
Chapman many good ride poor handling
FE/FD cars good space
good weight

rear- low cost

INDEPENDENT many small good ride poor handling
trailing arm FE/FD cars good space
(front V/W bug) low weight

Single semi-trailing rear - many low cost

swing arm arm FE/RD cars good ride poor handling
(some RE/RD) low weight poor space

lateral or rear- low cost awful handling

transverse rare now good ride poor space
(front - low weight
some trucks)

Table 1. A summary of most basic suspension types, and their characteristics.

25
 Outer Front Wheel in Curve Straight Ahead Position Inner Front Wheel in Curve

P = Virtual Steer Center

Figure 13. BMW’s ‘‘virtual center’’ front suspension uses two lower ball joints to create a more beneficial but extremely com-
plex steering geometry, in which the steering axis actually moves longitudinally and laterally when steered. (BMW)

ponents — but it is definitely there. Even duction sedan or race car that doesn’t have aware that most of the characteristic points
if a suspension is designed on paper to have some non-optimum condition. There may — such as instantaneous centers — are not
precisely the right camber angle in all posi- be as many different types of suspensions fixed, but vary with suspension movement.
tions, lateral force can cause an additional as there are classifications in the Patent So perhaps the greatest disadvantage is in
unwanted camber of over one degree. The Office, so this section will deal only with technical complexity.
only ways to cope with it are: anticipation those which are likely to be found on a race It is also a very efficient suspension, from
in original design (very difficult), stress the car: the upper and lower control arm in- a strength/weight standpoint. The natural
components on a bedplate and measure the dependent, the vertical strut, the swing axle, triangulation of upper and lower arms, with
deflection (difficult), or develop the running the solid axle, and the deDion. Most of an apex near the center of the wheel, is a
vehicle on a track until everything seems to these have been well illustrated and de- nearly ideal structure. The potentially wide
be working the way it was intended (less scribed in great detail already (See summary attachment base on the chassis has excellent
difficult). in Table 1), so this section will deal primar- resistance to forces and deflections both
At this point it ought to be clear that the ily with their racing characteristics or laterally and longitudinally. Because of this
design of a total suspension package is best potential. efficiency and compactness, it is probably
handled by a computer. Fortunately, few The first system, the upper and lower con- the lightest suspension possible, particularly
production race car suspensions today are trol arm linkage, is the most common layout in unsprung weight, and particularly when
very bad, but unfortunately, few production in race cars of any size or purpose. Perhaps the brakes are mounted inboard. However,
passenger car suspensions are any good for its greatest value is the almost unlimited whenever a flexible driveshaft must be
racing — except the solid rear axle} which range of geometry variations which can be provided between the chassis and hub, extra
is very difficult to foul up. If the fanatic race designed into it. Practically all of the complexity and weight is added in universal
car builder/mechanic must make radical theoretical considerations previously joints and some provision for plunge, or
changes from what exists, the safest route discussed can be controlled — if not com- changing length in the shaft. In some cases
is to make some models first, out of some pletely independent of the rest. The inter- the geometry is simplified by using a fixed-
deformable plastic. This may help in seeing actions can be either an advantage or an im- length driveshaft as the upper control arm,
where the weak points are, and if it’s pos- possible nuisance, depending on the although this limits geometry possibilities
sible to exaggerate the angles and deflec- brilliance of the designer. Virtually any roll and can have critical failure consequences.
tions, it might even be possible to see where center or swing arm radius or anti-dive/anti- There are some other disadvantages that
the inevitable interactions converge into con- squat geometry is available, and it lends should be mentioned. At the front, the
fusion. Otherwise, the rest of the chapter itself perfectly to steering control linkages. desire to have good geometry and large, stiff
is devoted to making the best of what exists. At the rear, for example, it’s possible to have control arms is limited by space inside the
practically zero roll steer — including moving wheel. By the time a wide-rim
HARDWARE
perhaps zero compliance steer — by having wheel is vertically deflected and steered in
TYPES OF SUSPENSIONS a lower control arm with absolutely parallel full travel, there isn’t much room left for
If the ultimate suspension geometry has inner and outer pivot axes. Non-linearity of control arms — especially when brakes are
been built, it’s only because half of its the geometry, like interactions, is also either also in the way. At the rear, rim width also
mechanical components are variable, and an advantage or disadvantage, depending on makes it difficult to find space for the ideal
controlled in action by an onboard com- particular needs, so it is important to be pivot points. In addition, this sort of suspen-
puter. There isn’t a suspension on any pro- aware of them. It is also important to be sion is mechanically complex and requires

26
 a large number of precision parts, such as
hub castings, control arms, extra suspension
linkages, and dozens of adjustable pivots.
The next suspension, the MacPherson/
Chapman strut, in which a vertical strut
replaces the upper control arm, is relatively
uncommon in racing, although it is used on
a number of production sports cars which
have been quite successful. This system has
the upper locating point moving essentially
in a straight line, while the lower locating
point travels in an arc. Depending on the
angle of the strut, and the angle and length
of the lower control arm, it can be designed
to approximate the geometry of a double
control arm system, although it doesn’t have
quite the range in variability. This suspen-
sion is difficult to analyze in a short space,
much less make generalizations about. Not
' that it matters a great deal anyhow, since Figure 14. Live Axle Locating Linkages
modifying its geometry by moving points
around is very difficult. The best that can for it in racing, except that it is light and as possible and always perfectly horizontal
be said about strut-type suspensions is that simple. The camber change curves work in (as it can’t be with suspension travel) it will
they are relatively inexpensive in mass pro- the right direction, causing both the inside change the roll center height differentlyin
duction because of the reduced number of and outside tires to lean in the proper direc- left- and right-hand turns. On the other
components — especially when a coil spring tions. However, at the rear, where the inner hand, it seems reasonable that this asym-
and damper are incorporated in the strut. pivot is commonly at the inboard differen- metry could be put to use in balancing out
But there are some definite disadvantages tial universal joints, the short swing arm the inherent left rear wheel loading under
for racing. At the front, where the strut is radius causes excessive camber change for acceleration with a live rear axle. If the
the steering axis, greater tire width produces wide racing tires. But by far the greatest Panhard rod is anchored to the chassis on
extreme kingpin offset. And at the front or disadvantage is that the roll center is very the left, the roll center will rise as the
rear, there can be wear problems because high, and there isn’t enough freedom in chassis rolls to the right in a left-hand turn.
of the sliding surfaces, and it may be heavier geometry to get it lower. Swing axles have This causes more load transfer to the right
than double control arms for the same a drastic jacking effect, which is highly rear tire when accelerating out of a left turn
strength. undesirable in race cars, if not all cars. At than it would to the left rear tire when ac-
BMW’s new MacPherson strut, however, least they aren’t complicated by having a lot celerating out of a right turn.
is sO innovative that it causes other engineers of interacting effects, and they are quite in- For perfectly symmetrical lateral location
to wonder “Why didn’t I think of that?” One expensive, as few components are necessary. there is the Watts linkage or the roller-and-
good reason is that while it looks simple in The solid axle seems to have little going guide. These provide very definite fixed roll
hardware, it is incredibly complex to analyse for it in racing, and yet it is used on so many centers. One word of caution, however, is
in its multiple interactions between steer production cars that it is practically that the roller or center pivot should be fixed
angle, changing caster, changing camber, unavoidable. Its use is limited to the rear of to the chassis and not the axle. Bump travel
changing trail, changing offset and anti-dive. front-engine cars for a number of reasons. is perpendicular to the control links when
BMW calls their design the “double link It takes up a great deal of space for ride they are parallel, and if they are attached
strut,’ and they almost explain it in a travel, and at the front it has almost insur- to a chassis that has some lean, the chassis
22-page technical paper, from which Fig. mountable ride, steering, and gyroscopic — or axle — will be moved sideways as it
13 is excerpted. In essence, the lower A-arm effects. The greatest advantage a solid axle travels vertically in bump.
is replaced by two double-ball-jointed links has is in keeping both tires flat on the And finally, there is a more complicated
which have a projected intersection “‘P’’ ground most of the time. There can be a lot method of lateral location. If four
This “P” is the actual (instantaneous) center of slop and a lot of travel between the longitudinal control arms are used, as on
of rotation, which moves both laterally and suspension and the chassis, without any many General Motors coil-spring rear axles,
longitudinally as the wheel is steered. In ad- undesirable camber effects. Because it is the convergence of any two to a point on
dition, by varying the relative heights of the such a long component, it is easy to locate the vehicle centerline defines a roll center.
ball-joint ends, it is possible to add anti-dive in any direction, by a number of possible However, this is uncommon to race cars for
to the confusion. Suffice it to say that linkages. a number of reasons, primarily excessive
BMWs with this suspension have superb The lateral locating mechanism of a solid stress in the components.
handling and response. However, this is axle defines its roll center, which is rela- The ideal longitudinal locating
probably because the angular movements tively limited by ground clearance. The mechanism for a solid axle is a little harder
are never great enough to make large most common method — with the excep- to resolve, because of more interactions.
changes in the geometry. It is conceivable tion of leaf springs — is the Panhard rod, Even if it were possible to get perfectly
that someone with a lot of time could ex- or a simple transverse link from the chassis vertical movement of the axle, anti-squat
pioit this system in a race car — even ap- to the axle. (See Fig. 14) This is simple, requires a wheel to move longitudinally a
plying it to double A-arm linkages. light, and easy to adjust. The disadvantage slight amount as it travels vertically. And
The swing axle doesn’t have much going is that it is asymmetrical. If it isn’t as long yet, if both left and right wheels don’t move

27
 strenuous uses their vehicles may be put to
— whether as police cars or taxi cabs, in
rallying or racing — and they often have
optional heavy-duty suspension components
available. Chevrolet, Ford, Chrysler, and
American Motors all have separate catalogs
or listings of these pieces, and as expensive
as they may seem, they’re hard to beat in
value per dollar. Even if stronger com-
ponents aren’t readily available, racing rules
usually allow the reinforcement of any
chassis parts in the interests of safety. As
long as there are minimum vehicle weight
restrictions, any added weight for strength
in the suspension is a good investment in
survival.
The selection of suspension pivots is also
important, and not just from the standpoint
of ultimate strength. To get rid of excessive
compliance, or deflection in components, all
rubber should be eliminated from any pro-
duction suspension. This includes control
arm pivots, anti-roll bar mounts, and
damper mounts. Rubber pivot bushings
reduce ride harshness, but they also reduce
Production sedan suspension modified for racing, with solid eccentric the accuracy of geometry.
bushings in the A-arms, and solid bushings and ball joints on the anti-roll bar. The purpose of a suspension pivot is to
rotate freely, of course, but all too often, the
the same amount when the chassis rolls (one locating linkages. wrong type of pivot will bind as if it were
wheel up and the other down) the axle will One problem, of course, is that a deDion welded. For example, two inner control arm
have a steer angle. Since this is practically axle ordinarily requires two double- rubber bushings may be replaced with
impossible to avoid, designers at least at- universal-jointed drive shafts, with some bronze bushings. But if there is any distor-
tempt to produce roll understeer, which is provision for plunge — as with most in- tion along their axis under high stresses, the
more stabilizing. dependent suspensions. In some cases, misalignment may lock them up solid, pro-
The strength/weight ratio of a solid rear however, the need for changing driveshaft ducing an essentially rigid suspension. At
axle must be the worst of any suspension. length is eliminated by making the dead axle best, it will cause a great deal of friction
Unlike an independent suspension, the solid variable in length, and absorbing lateral and high wear, which results in looseness.
axle has significant forces and torques to be forces through fixed-length axles into the A spherical bearing should be used
absorbed and transmitted to the wheels differential. The drawback in this, is that it wherever there may be the slightest
through its own unsprung structure. In ad- places the roll center up where it would be misalignment. But what is not commonly
dition, it is in the shape of a beam, rather with swing axles. With the acceptance of known is that even the best spherical rod
than a space frame, which gives it poor lower roll angles in the ideal desired suspen- ends produce a lot of friction when sub-
resistance to bending from vertical load and sion geometry, the need for connecting the jected to the kinds of loads a race car tire
cornering forces at the tire. And, of course, left and right wheels for vertical stabiliza- can generate. When a long control link is
the last criticism of a solid rear axle on a tion has diminished. What advantage that under compression — for example a rear
race car is its high unsprung weight and had over double control arm independent radius rod during acceleration — loads in
gyroscopic interactions between left and suspension is now generally overruled by the both rod ends can be so great that as the
right wheels. disadvantages of higher weight and dynamic suspension deflects, the rod will bend rather
The deDion axle eliminates most of the instability. than pivot at the joint. For the most critical
worst features of the solid axle with integral and highly loaded bearings, nothing is good
differential. By mounting the differential to MISCELLANEOUS COMPONENTS enough but teflon-lined spherical bearings.
the chassis, and routing a dead axle beam There are a couple of miscellaneous sub- All-metal spherical bearings — with no
around it, many suspension forces are jects that ought to be mentioned in the teflon lining — are adequate only where
reduced. Unsprung weight may be cut con- general category of suspension hardware. misalignment rather than high load is the
siderably, which reduces spring and damper Whenever larger tires are fitted to a car, problem, such as throttle and shift linkages.
requirements; the differential doesn’t need potential suspension stresses go up — Some catalogs list not only the loads that
to be supported between the wheels; and maybe a little, but more probably a lot in a bearing will carry, but the kind of rota-
there are fewer torque problems at the wheel the case of a production car taken racing. tion torques to be expected under given
hubs (practically none if inboard brakes are It may be of critical importance to loads. When selecting a bearing size for a
used also). In addition, as with fully in- strengthen the components which have to new application, these catalog figures
dependent suspension, there is no lateral absorb lateral and longitudinal forces. On should be carefully considered, using other
wheel load transfer due to torque under ac- superspeedway banked tracks even vertical known load factors explained in this book.
celeration. Otherwise, the deDion is much forces may be doubled. Most manufacturers At that, the bigger the better, within the
the same as a solid axle, particularly in of production cars are aware of the more limits of overall weight.

28
 When mounting spherical bearings or rod
ends, some thought must be given to all the
forces they may have to carry. They are
designed primarily for radial loading, and
another force — say the spring load halfway
out on a lower control arm — can cause an
axial load on the ball. Catalog figures show
that spherical bearings are only about 20
percent as strong along this axis, and can
be pulled apart relatively easily. In some
cases a high-misalignment bearing may be
necessary, which has an axial strength of
only about 10 percent of the radial strength.
Where axial loads can’t be avoided, an over-
size washer should be provided on the
mounting bolt to retain the rod end if it pulls
apart. This isn’t necessary, of course, if the
ball is mounted in double-shear, where it
is supported on both sides — a structurally
* more efficient design anyhow.
Using a set of electronic computer scales to check left/right, front/rear, and diagonal
MAINTENANCE AND SERVICE
cross weights. Ballast has been added to simulate appropriate fuel and driver weight.

At this point, without other absolutely time, then you havea reference for dynamic even when the shocks are off, the car should
necessary information — track development camber development on the track or skidpad, be rolled forward and backward while
procedures, spring rates, aerodynamic using tire temperature profiles. Of course, bouncing it lightly up and down from its
loads, brake dynamics, handling, etc. — it that doesn't apply to oval track cars, which natural balance height before any critical
is practically impossible to define the may start with many degrees negative cam- weights or measurements are taken. Caster,
desired suspension geometry for a given car. ber on the right front, and many degees camber, and toe angles are all likely to
However, some idea must be given of how positive on the left front. change with any change in ride height, so
to align and modify geometry. Later, with The most basic alignment tool, for solid it’s mandatory to have an accurate reference
more information about other areas, the rear-axle cars, is some sort of front camber point.
principles here can be better applied. For measuring device. Professional bubble Wheel load balance for a road-racing car
the time being, some very broad general gauges which also read out caster angle are should be as equal from side to side as
assumptions will be made about desired available for about a hundred dollars. For possible, keeping in mind that the center of
alignment. If the chassis builder didn’t pro- the beginner, a bubble-level protractor will gravity may be off-center when the engine
vide some basic static alignment specifica- give rough figures for a fraction of the _ or the driver’s seat is offset. Oval track
tions to start with, the following estimates price. To align a four-wheel independent racers frequently use intentional asym-
probably aren’t far off. suspension professionally and frequently, metrical loading since they only turn in one
Ride height, if not defined by racing optical alignment gauges are available for direction, but the same desired effect could
regulations, should be just enough to avoid two hundred to three hundred dollars. probably be created by other modifications.
scraping the track under the worst condi- Finally, but most important, you need a Asymmetrical loading is certain to result in
tions. Two to three inches ought to be good truly level and flat surface on which to set an instability under some condition or com-
for a starter, with more for heavy sedans and the car. If the floor isn’t flat enough, four bination of cornering, accelerating, or
large aerodynamic down forces. Caster spots mustbeshimmed up to within 1/8 inch of braking. Ride height symmetry and wheel
angle is perhaps least important, as the perfect horizontal, using a carpenter’s level. load balance are interrelated, so it is impor-
driver can live with an uncomfortable More harm than good will be done if car is tant to work on them at the same time, with
amount for a short time, and it has small aligned on an uneven surface. constant rechecks.
effect on vehicle performance. Anywhere A set of scales will be necessary to
from two to six degrees ought to do it. Toe- BALANCE adjust wheel loads. Since most tracks have
in should be just enough not to allow toe- The first step in static alignment is to set their own electronic wheel scales at in-
out under any condition of deflection or the desired static ride height and wheel load spection, some racers just rely on their
braking stresses — from 1/8-inch toe-in on balance. Both should be adjusted with the figures. However, scales should be used so
heavy, flexible cars to zero on formula cars. anti-roll bars disconnected, as there is likely often that having even an inexpensive set
For extremely fine tuning, more or less toe to be some asymmetrical pre-load in the bar. isa good investment. Except for matching
is sometimes used to create an artificial When the bar is finally reconnected, it the official scales, where total weight of
Ackerman effect, but this is a hard to should have no preload or torsion when the the car is critical, accuracy isn’t as impor-
measure bandaid for improper geometry car is level. The shock absorbers should also tant as equality between the four scales.
development. Camber angles are really be disconnected if they aren’t the spring This can be checked by frequently weigh-
unknowns at this point but tend toward zero mounts, in which case they should be ad- ing the same object, such as yourself, on all
to negative one degree (in at the top) at static justed to full soft action. Good shocks have four. Unlike the alignment surface, where
ride height, with perhaps two degrees for enough static friction to hold a car a half it is mandatory that all tire platforms be
heavier cars. The point here is that after the inch off the true spring balance point. Since perfectly level with each other for camber
car is set up with specific figures for the first suspension pivot friction can be significant gage reference, it's only necessary that

29
 tolerance of one degree might be adequate
at first. Before beginning, both camber and
toe-in should be set roughly where they
ought to be, so that their own later align-
ment won’t get the caster too far off. If an
ordinary caster bubble-gauge is used, the
standard procedure is adequate: turn the
wheel 20 degrees in either direction, and
read the caster angle. An alternate pro-
cedure for the front, and perhaps the only
solution for the rear, is to use a bubble-level
protractor between two ball-joints. On most
front suspensions, caster is adjusted by
moving the upper kingpin pivot fore or aft.
If there is a choice, it should be moved by
whatever method least affects the camber
and toe. Record-keeping is particularily
important here, in caster change per adjust-
ment rotation, because it will probably have
to be redone when everything else is
correct.
Camber angle and toe-in are both critical
Rear ride height adjustment screws in the back of a NASCAR sedan, enough to require precise measurement, and
showing extreme weight bias for left-hand turns on speedway banking. they usually interact to a degree that makes
frequent rechecking necessary. Gross
measurements can be made with a tape
scales all bein the same plane. Thatis, you Careful reasoning ought to indicate which measure and bubble-level protractor — they
can be on a slight slope if necessary, as spring ought to be adjusted in which direc- would probably be adequate for amateur use
long as the front and rear or left and right tion, but if not, experience is also a good on a heavy sedan. But for greater accuracy
pairs of scales are perfectly horizontal. educator. If, after trying all alternate solu- on light, responsive cars, better equipment
Before wheel load balance can be set, tions with the greatest care, the chassis is is necessary. Toe-in can be set by comparing
the car must be loaded with a typical driver still not balanced and level, it is probably the parallelism of the left and right wheels
weight, full oil and water, and a known due to unequal-rate springs or a twisted to each other, although it is also important
fuel load--full, or inspection standards-- chassis. Once it is finally correct, variations — especially at the rear — to reference the
and the tire pressures must be equalized. in ride height or pitch angle can easily be parallelism to the centerline of the chassis.
Roll and bounce out the suspension fric- made by carefully adjusting left and right If the rear wheels are perfectly parallel, and
tion, and gently roll the car on the scales. springs the same amount in the same yet both steer off to one side by a few
Ifthe wheel loads aren’t what they should direction. degrees, the handling will vary in left and
be, one or more springs must be adjusted, right turns. The centerline of the chassis is
depending on the desired height and atti- ALIGNMENT best located from lower control arm inner
tude of the car. The method of spring When the chassis height and balance are set, pivot points, and should be permanently
tension adjustment varies, depending on a few alignment accessories can be made. marked on fixed crossmembers. These
the type of springs, from threaded coil With them, the car can be aligned while points, and subsequent measurements, could
spring mounts, to lever-arm screws on sitting on jack stands, without even having also prove to be useful later in straightening
torsion bars, to variable shackle holes on tires mounted on its wheels. a damaged chassis. A string stretched be-
leaf springs. If there is no adjustment, For alignment while the car is on the tween front and rear center points
some must be made, as cutting or heating floor, a set of ride-height blocks are useful. establishes a centerline from which other
springs causes other serious problems. These are cut to just the right heights and parallel reference lines can be located. The
The spring to be adjusted will not only located in just the right positions to hold the parallel reference at the outsides of the
correct imbalance, but it can correct varia- car the proper distance off the ground while wheels can be straight aluminum beams or
tions in ride height. Say, for example, that making static suspension measurements. It string stretched between jack stands. A
the right front load is 50 pounds greater than is also useful to have a set of dummy shocks typical racer’s ingenuity can produce any
the left front. Possible options are to made from rigid pieces of steel, with the number of other rectangular frames of
decrease the tension in the right front or left correct shock mounts on each end. These reference.
rear spring, or to increase tension in the left hold the suspension in precise position, even Toe-in and camber angles can be
front or right rear springs, or a combina- if the car is on jack stands or the springs measured from any truly flat machined sur-
tion of all four. If ride height is about correct have been removed. An even better idea is face on a rim (check the runout by rotating
and equal from side to side, tighten the left to make the links adjustable to three definite it on the axle), or from a fine line scribed
front and loosen the right front the same positions: full bump, full rebound, and ride on the tread of a rotating tire. A machinist’s
amount. If the front is too high, and yet height. scale is accurate enough to read the toe-in
level, loosen the right front and left rear. If The next static alignment step in the se- from front and rear reference points to the
the car has less ride height on the left, quence probably ought to be caster, and for parallel surface. For front alignment, the
loosen only the right front. There are dozens double-control arm rear suspensions, rear steering rack or gearbox should be centered
of combinations of odd wheel loads, chassis upright inclination. The effect of these and held fixed while each tie rod is adjusted,
tilt angles, pitch angles, and ride heights. characteristics is slight enough that a so that there will be equal travel in both

30
directions. Camber angle is read directly
from the camber gauge. In an emergency,
angles can be obtained by measuring the
relative distances from a vertical plumb line,
and converting to angle by the following ap-
proximate equation: degrees=
(57) X (inches from vertical)
(distance between measuring points)
This also applies if there is any reason to
convert toe-in from inches to degrees. As
with most other suspension geometry, toe-
in and camber changes affect each other,
and it will probably take a lot of going in
circles to get both to the desired point.
With a solid rear axle, toe-in or camber
can’t be easily adjusted, of course, but a
quick check will indicate whether the axle
is in straight, or whether the housing has
been bent. It is possible for construction
tolerances to allow for a half-degree of
misalignment. Some racers attempt to
take advantage of this by intentionally
bending-in a degree or so of camber, but Application of a camber gauge (left hand) and caster gauge (mounted to hub)
excessive wear on bearings and differen- to check and align both simultaneously.
tials counters any cornering advantage,
unless special hub and axle splines are used.
Independent rear suspensions are align- Therefore, this section will be limited to To measure these curves, some specia.
ed much the same as the front, except that shop development. preparation and equipment are necessary. A
the alignment of axle shafts must also be Assuming that the racing regulations pair of inexpensive dial gauges will improve
considered. The goal is to have all four allow suspension modification, there are accuracy and speed up the process, although
universal joints on a perfect axis when the enough pivot points, and enough ad- a machinist’s rule will do. If dial gauges are
car is accelerating (all deflections allowed justments, and enough directions of reloca- used, they should be mounted on a portable
for). tion to make the number or possibilities but rigid stand, and positioned so that they
Ideally, toe-in would be zero under all almost infinite. Camber change curves can’t can contact a wheel face about a diameter
conditions of bump, roll, acceleration, and be improved without a firm knowledge of apart, either horizontally or vertically. A
braking. Any amount of toe-in causes some what they ought to be, but it will help later large flat plate is bolted to the wheel or hub
degree of tire scrub and drag, while any to know what they look like. On the other flange which the dial gauges contact as the
amount of toe-out, especially at the rear, hand, there is a general consensus that roll wheel is moved vertically throughout its
will cause instability. Therefore, even on steer ought to be nearly zero, or at least tend total travel. The chassis is set firmly on
relatively rigid, light-weight cars, it’s a safe toward roll understeer. blocks or jack stands, the springs and shocks
bet to lean toward a slight amount of toe-
in. Some toe-in also absorbs any toe-out
caused by vertical travel, if it cannot other-
wise be eliminated through procedures in
the next sections.

DEVELOPMENT
There are two separate areas of develop-
ment toward better suspension geometry.
The first is to analyze and modify the
geometry while the car is held fixed in a
shop area, and the second is skidpad or
track development. Each has its advantages,
but neither is enough in itself. Knowing
geometry theory and the approximate goals
to be achieved permits a great deal of design
and adjustment to be done in the shop. At
the same time, the proof of the theories, and
the last final degree of adjustment, can be
found only in a real-life situation. Since
track development has to include the areas
of springs, aerodynamics, and test equip- Using an ART laser-plane alignment head to set toe, with a custom fixture attached
ment, it will be covered in another chapter. to the hub, which can be rotated 90 degrees to also set the camber angle.

31
 When the rear toe is as close as possible,
the front toe steer curve can be approached.
The kingpins and steering hardware com-
plicate the picture, but allow more options
to correct a poor curve. If the rear curve
had zero change, it would be nice to have
zero change at the front also, but it at least
ought to have Jess of the characteristics
designed into the rear, as shown in Fig. 5b.
Possible options for changing the curve in-
clude: steering rack or gearbox height and
longitudinal location, steering arm height or
angle, tie-rod length, and steering rack
length. The effect of each of these changes,
depending on the type of suspension and
shape of the curve, is worth a chapter in
itself, and better learned by
experimentation.
Changes in control arms will also change
roll steer, though there are fewer interactions
in steering linkage change, which should
only affect the static toe-in setting. Another
important factor is whether the steering
arms are ahead of or behind the kingpin
axis. All else being equal, leading steer
This bump steer fixture is designed to read out toe change on the dial gages as arms are more stabilizing at the front and
the suspension is moved through its total travel. trailing steer arms are more stabilizing at
the rear (as exemplified by the Corvette)
because of lateral force suspension com-
are removed (if they are not part of the appears bad enough. The rear is the first pliance. Finally, where a compromise must
locating linkage), and the suspension is problem to solve, since it is usually less be made in the toe-steer curves, the bump
moved vertically with a hydraulic jack. At complicated due to its lack of intentional side is more important, since that wheel is
the front, the steering gear is locked in its steering geometry. If the lower control arm obviously the more heavily loaded. Con-
perfect center position. Of course, the axes are horizontal and perfectly parallel, versely, when a wheel goes into full
suspension geometry has to have already it probably won’t steer. But if not, the goal rebound, it approaches zero loading.
been statically aligned to the desired is to get the steer near zero, or at least so Although not enough is known at this
specifications. that the wheels toe in as they rise, and toe point to modify camber change curves,
Each wheel is then jacked vertically out as they fall. The means of achieving this some mention is due. The curves can be
through its total travel, while dial gauge or are as varied as the number of suspension obtained in the same way as the roll steer
scale readings are taken at each inch. When systems available, but it should rapidly curves, only with the dial gauges aligned
plotted as shown in Fig. 15a, they give a become obvious when major changes are vertically, or with the use of a camber
vivid picture of the unwanted steer angles made in radius rod lengths and mounting bubble gauge. The way to find the proper
that come from suspension deflection. Then points. Then, however, the static alignment camber angle is to test the car on a skidpad
it is possible to improve the situation if it will probably be out of adjustment again. and measure tire temperature evenness or
DESIRED laptimes, as discussed in Chapter 2.
POINT AT However, that produces only the correct
MAX. LATERAL
BUMP 3”. BUMP 3” camber at maximum roll angle. The final
static camber angle required to give that roll
Qu camber may be ridiculous for straight-line
running. If a static camber of three degrees
1”. is required for correct cornering camber,
then it would be of value to try and change
STATIC TOE OUT the curve. In this case, the extreme bump
SUSPENSION
HEIGHT 1° 1° point of the curve is known, and it is
He “" desirable to get the ride height portion of
CHANGED / 1” /oHANGED the curve to pass through zero camber, as
CAMBER / STATIC
CURVE / >” | CAMBER shown in Fig. 1Sc. Of course, the curve
! | probably can’t be changed without also
3” REBOUND 3” REBOUND | gr! REBOUND changing swing-arm radius, roll center
1la REAR BUMP STEER 11b FRONT BUMP STEER 11c FRONT CAMBER CHANGE height, toe-steer curves, and so on. The
complexity of the geometry increases ex-
Figure 15. Dynamic Suspension Alignment Curves ponentially. In the next chapter, springing
and damping are added to the muddle.

32
 a
SPRINGS, ANTI-ROLL BARS, SHOCKS
A spring is a spring is a spring. It doesn’t cause unbelievable complexities in a race wheel rate, since both force and dis-
matter whether it’s made out of steel, car’s suspension. It’s worth noting that placement change by a factor of two. If this
titanium, rubber, silicone, air, or jello, as variable spring rates usually only rise under ratio is carefully measured in one-inch in-
long as it has the required spring rate in deflection, which fortunately is the desired crements over full suspension travel, the
pounds per inch of deflection. Of course, property in a suspension system anyhow. variability in rate can be plotted. It isn’t
there are a few other considerations of im- What is not commonly known is that the uncommon for this changing rate to
portance which have fairly well limited the geometry of the spring mounting system decrease with deflection, due to undesirable
selection for race cars to metal alloys. The usually gives a vehicle a slightly variable spring geometry. The leverage ratio also ac-
spring has to have a high value of energy spring rate even if constant-rate springs are centuates the effect of offset wheels on ride
. Storage per pound. It has to be durable used. If a coil spring could be mounted rates with some non-parallel independent
under repeated stress cycles. And it is also directly above the wheel, and if it deflected suspensions. The further out the wheel gets,
frequently used as a locating mechanism. in a perfectly straight line, the wheel rate the greater the leverage it has on the spring,
Over the years, racers have learned that to would be constant and equal to the spring and therefore, the lower the ride rate, the
make a production car handle better, they rate. But since springs have to be mounted lower the chassis sinks, and the lower the
have to install stiffer springs. They also inboard of the wheel, and the connecting roll rate.
believe in installing stiffer shocks, and stiffer links travel through odd combinations of The ride rate is also linked to the chassis
anti-roll bars. Those are all over- arcs, the spring operates at a disadvantage roll rate from lateral forces in a turn.
generalizations that may be applicable in the in leverage. (See Fig. 16) Given a set of Because the chassis center of gravity is
conversion from street to track, but which engineering drawings, the leverage, or usually above its center of roll, something
can easily cause a loss in performance if mechanical ratio, could be calculated. For must be provided to keep the roll to a
there is no real understanding. an existing vehicle, however, the ratio can minimum. If single suspension springs were
be found by a more straightforward method, located in the center of solid front and rear
THEORY due to the fact that the force ratio between axles, the chassis would literally fall over.
SPRINGS the tire and the spring is the inverse of the Obviously, then, the further apart the
displacement ratio. Say, for example, that springs are, for a given spring rate, the
The whole idea is to keep the tires in the
when the wheel is deflected 2 inches, the greater the roll resistance. Since it is
firmest contact with the road as long as spring deflects 1 inch. The displacement
possible. As a tire bounces over ir- desirable to have a race car’s center of roll
ratio between spring and wheel is 1:2 and ‘very low, more roll resistance is required
regularities, its contact force averages out
the force ratio is 2:1. In this case, the spring than can be provided by the relatively soft
to be the load on that corner of the car. But
force is twice the wheel force. However, it’s spring rate and narrow spring base. This is
the average doesn’t count in a race car at
important to distinguish that the spring rate the reason for anti-roll bars.
the limit of traction. When the tire bounces
in pounds per inch will be four times the With the ability to control roll angle
up and unloads itself, that is the traction
limit of the car. So spring rate is of far less
importance than control of the tire, which SPRING : TIRE
is also strongly influenced by the damping DISPLACEMENT RATIO 1 2
action of the shock absorbers and the ratio FORCE RATIO 2 1
of unsprung to sprung weight. Too high a SPRING RATE RATIO 4 1
spring rate or damping rate is just as bad
as too low a rate.
The rate of a spring in pounds per inch,
sometimes just referred to as pounds, is
determined by design or test, and usually
is a constant for a given spring. A
200-pound spring will deflect one inch if
200 pounds is placed on it, or 2 inches if
400 pounds is placed on it. If the rates of
a car’s springs are unknown, they can easily
be determined in this manner. In addition
to ordinary constant-rate springs, variable-
TIRE FORCE SPRING FORCE
rate springs are sometimes available. With AND DISPLACEMENT, AND DISPLACEMENT
these, the spring might deflect 1 inch with
a 200 pound load, and yet only 1.5 inches Figure 16. Geometric Consideration in Spring Rates
with a 400 pound load. These are un-
common, usually quite expensive, and they

33
 BARS
An anti-roll bar cannot rightly be called
a sway bar since it neither causes nor
reduces sway. Nor can it be called a
stabilizer bar since it may very easily cause
the car to be unstable. All it is supposed to
do is resist vehicle roll angle in a turn.
However, the distribution of roll resistance
between the front and rear wheels has a
great effect on the cornering stability of a
vehicle. The explanation has usually been
dependent on a previously mentioned
characteristic of tires: the total cornering
power of two equally loaded tires is slightly
greater than the cornering power of one tire
with twice the load. In other words, the end
of the car with the greatest load transfer to
Coil and leaf springs for a racing sedan. Note that two coils have been the outside tire has slightly less cornering
close-wound and ground flat, and the eyes are centered on the main leaf. power. This is true, and it works in practice
— but there is also another significant factor.
The major effect of anti-roll bars may be
seen from a consideration of torques on the
relatively independently of the vertical a number of factors to consider besides roll vehicle about its center of gravity, as shown
ride rate, race car spring rates have gotten and bumps. Wheel load transfer under in Fig. 17. The diameter of the circle around
a little lighter over the years -- except where braking and accelerating will use up wheel each tire represents the load on that tire, and
aero downforce is involved. A complex travel too, if 100 percent anti-dive and anti- therefore its traction capability in any direc-
analysis is necessary to study the effect of squat is not designed into the geometry. In tion. If the front and rear cornering forces
spring rates, but experience has shown addition, a lateral force jacking effect will and the left and right longitudinal forces all
that softer is better--within limits. The cause a change in wheel travel as the chassis balance perfectly around the center of
lower limit of spring rate and/or travel is rises. The interaction of all these variable gravity, then the car is perfectly balanced
the point at which the chassis hits the track forces and displacements is beyond in that condition. However, the circles in this
on hard bumps or high banks, or where the reasonable analysis, which means that the example indicate that there is greater load
moving suspension components contact final selection must be made from track test transfer at the front, reducing front cor-
each other. Although it is possible to provide experience. Theoretically, it may be possi- nering power, and resulting in a counter-
relatively soft contact points for suspension ble to combine the geometry effects of anti- clockwise torque, or understeer. But we
components (to prevent stress failures), most dive in braking with lateral jacking as the must also consider the cornering drag of the
racers prefer to have the absolute limit be car enters the corner and anti-squat as it two front tires, and enough forward thrust
the chassis hitting the ground. Whichever accelerates out, to maintain a relatively from the two rear tires to balance that drag.
part hits the ground — body or frame — constant suspension travel for bumps. The As front lateral load transfer increases the
is probably less critical than the highly problem is that these geometry requirements drag on the outside tire and decreases drag
stressed suspension, and it can safely take have other more serious negative on the inside tire, there is again a counter-
some wear. In addition, in the search for consequences. clockwise torque about the center of gravity.
minimum ride height, positive stops on the
suspension may eliminate some valuable
REAR TIRES
travel. For example, positive suspension BALANCED THRUST
stops which just keep the chassis off the FRONT TIRES
UNBALANCED DRAG
track in pure bump may restrict the bump
travel too much in roll. Finally, and possibly
most important, when the suspension
bottoms suddenly, its spring rate rises
instantaneously to infinity. This causes that
particular tire to have a sudden loss in
traction capacity (assuming a less than
glassy-smooth track), and therefore a sudden
handling change — which at the rear creates
critical oversteer. The chassis striking the

(©)
track at the same rate will unload a wheel
somewhat, but the effect is far less. That is CIRCLE SIZE PROPORTIONAL
not to say that either is a desirable condi- TO WEIGHT ON TIRE
tion. Suspension travel and spring rates
should be selected so they just approach that
point but do not allow contact under normal
Figure 17. Anti-Roll Bar Effects on Chassis Balance
track conditions.
When selecting travel and rates, there are

34
 At the same time, increased front load
transfer reduces rear load transfer and ca RESISTANCE CURVE
causes forward thrust to be more centered
on the chassis centerline and contribute less
to a clockwise torque, or oversteer. This is
PEAK REBOUND FORCE
especially true of cars with an open, or
unlocked, differential, in that forward thrust
is supposed to be perfectly balanced be- i deflection
tween the left and right wheels, regardless in inches

of the load carried. An accurate mathe- force

pounds
in
matical analysis of both tire depreciation
effects and tire drag effects depends on a
large amount of other variable data on tires, TOTAL TRAVEL
load transfer, roll resistance distribution, dif-
ferential type, and so on. But it is safe to
say that both effects are important, and
fortunately both work toward the same Figure 18. Shock Absorber Force-Deflection Curve at an Arbitrary
result. A distinction between the two is not
Cycling Speed
irrelevent, however, since the latter helps to
explain other stability effects from power ap-
. plication, braking and differential action. SHOCKS Some manufacturers of racing shocks can
At either end of the chassis, the effec- Shock absorbers are more appropriately provide graphs with a series of curves show-
tiveness of an anti-roll bar is dependent on called spring dampers, since they reduce the ing force as a function of travel for various
a number of geometric factors. Of course, suspension spring oscillation to zero cycling speeds. The more that compression
the diameter of an anti-roll bar is all- relatively rapidly. The ideal spring has prac- and extension forces increase, the faster they
important to its roll resistance, all other tically zero friction in its movement, which bring a suspension oscillation to zero. This
linkages remaining equal. But for the in- allows it to keep oscillating for a long time. can be overdone, however, to the point of
stallation or relocation of a bar, the So some sort of controlled friction must be “critical” damping, where the suspension
mechanical force ratio between tire travel supplied. Mechanical friction, as between doesn’t cycle at all but slowly returns to its
and bar lever-arm travel must be considered. two solid surfaces rubbed together, is not original ride position without overshooting.
As with finding spring rate ratios, this can desirable, because it is relatively indepen- Even if this were desirable — it tends to
be done from engineering drawings, or ex- dent of velocity or displacement. It treats hold the tire off the ground too long — the
perimentally on an existing suspension. The little bumps as firmly as large, high-speed forces it would transmit to the chassis would
force ratio is the inverse of the displacement bumps. On the other hand, viscous or be intolerable. As it is, shock absorber
ratio between tire travel and bar-end travel hydraulic friction increases as a function of forces are probably many times greater than
— and is likely to vary with movement. velocity, and its force/deflection curves can spring forces and should be respectfully
Another complication to be aware of is be modified precisely to any desired per- considered in bracket design.
deflection in the anti-roll bar lever arms. If formance. The problem arises in deciding The rise of the force/travel curve from
the bar has relatively rigid arms attached to what that performance ought to be — a zero is an indication of absorption ratio bet-
its ends, and they are close to the bar pivots, complex and little-understood area. Some ween small and large bumps. A gradual
there is probably little unwanted deflection. of the factors to be considered in damper curve avoids harshness over minor bumps
But if the arms are merely bent out of the design or selection include: total travel, such as tar strips, without reducing the max-
same material as the shaft, then the arms maximum resistance force, rise of the imum bump absorption capability. However,
will have some bending which will reduce resistance force curve, total energy to be a gradual rise also decreases the total area
the effectiveness of the bar. absorbed, ratio of bump/rebound force, in- inside the force/travel curve, so that total
The lowering of a race car center of ternal mechanical friction, jerk, and weight. energy capacity is decreased.
gravity toward the roll center, and the use Since few racers can build their own shock Bump/rebound ratio is the comparison
of a very stiff anti-roll bar, can bring the absorbers, the factors will be merely de- of peak forces in either direction. For
maximum roll angle down to one or two scribed, as relating to selection or adjust- comfort, most passenger cars have tradi-
degrees on a modern racing car. At one time ment. (See Fig. 18) tionally used a 30/70 ratio, meaning that
some drivers claimed that a significant roll Total travel is a simple factor, but often only 30 percent of suspension travel is
angle was necessary to provide information ignored. It is very important that suspen- resisted inbump, and 70 percent resisted in
about cornering forces. But high roll angles sion travel not be limited because of bot- rebound--over a series of cyclesto zero, On
have so many other disadvantages — taking toming in the shocks. Adjustable shocks in the other hand, racing shock absorbers
up bump travel, causing camber changes, particular may have delicate valves which may havea ratio closer to 50/50, with most
and causing roll steer — that other feed- can be pounded to pieces. As with springs having the capability of adjustment to other
backs have replaced roll in importance. On and anti-roll bars, the shock travel and force ratios. It has not been conclusively proven
the other hand, a stiff anti-roll bar has a is seldom equal to that of the wheel but has that one damping ratio is better than an-
noticeable drawback in one-wheel bumps, some mechanical ratio. Even with a solid other in racing, all other factors being
as it increases the single-wheel spring rate rear axle, inclination of the shock absorber equal. However, a shock absorber with an
as though there were an additional torsion- reduces its travel and, therefore, its uneven ratio has a feature that can be a
bar spring in the suspension. As in effectiveness. definite advantage on a race car. Over a
everything else, there has to be some trade- Maximum resistance forces are generally series of bumps, a 30/70 resistance ratio
off between the advantages and referred to as stiffness and are always related will have a tendency to let the wheel re-
disadvantages. to a specific test velocity of piston travel. bound alittle less oneach cycle. In other words,

35
 | | he
Be i me REP Se

Beaee
a? . Us ae Bal
==.
aaa
aaeae
fee
| ‘mi! TD Pas | ea
Ceearsere Ati aera} ine cee reBI
ae byaa ies geal alt ie
Gn am on vA APRA NN
eeSe
53 a :
Ae
a Po
n\n
Se
a, A A” A,
eae at
ese
Bee nd BS
eesa
et | Ca ON Fa ee a ET NT
hh aor
= -]eres atsidae oihoc
on ica i ee
ui i. od -1 XV) de
|S a OD Yon —2 —

) Spe ees cSeoe

| eet ae ee re
= ia | Sooeo aE iii
Bie E ale Coleg tata
i ese maifees
| -T-* Fe =ssac=
3 ra
a | f enSOR Oo) Clean tovecgeed
See
rm Ra
Le ae EET mown
tosh toll
af ) EreSs a

Figure 19. Actual front shock travel for hard (solid line) and soft shocks on a race track at 100 mph. Note at A, the harder
shocks pull the nose down almost 4 inch, and at B, they eliminate two overshoot cycles after a severe bump. Note also how
each bump on the track repeats identifiably on successive laps.

instead of the wheel falling, the chassis will on automotive suspensions, this is another not be used because of space limitations or
fall, decreasing suspension travel (See Fig. step toward computer-controlled, actively the accessibility of mounting structure. The
19). It might be said that the suspension is modified suspensions. coil has shape and displacement
being pumped down. If the effect is known There are other performance characteris- characteristics similar to a shock absorber,
and anticipated, a desired height change tics of shocks that have less application to which makes them convenient to mount
might be possible, such as a 70/30 shock racing. Jerk, or the instantaneous change in closely, if not concentrically. On the other
absorber which pumps the suspension up. acceleration, is an important consideration hand, the torsion bar takes up far less
In fact, this sort of action is designed into in ride comfort. It has not yet been applied volume, especially when it can be mounted
some kinds of automatic load levelers. to the improvement of tire traction but there parallel to, or inside, frame members. The
One new suspension trick that should be are possibilities. Heat absorption and leaf spring doesn’t have much volume in
pointed out, has been introduced on motor- dissipation are also factors. A shock may itself, but it often requires a great deal of
cycles first. Because of steering geometry generate a great deal of heat over a long, space for bounce travel.
requirements, ordinary motorcycle front rough surface, and if it can’t get rid of it More important in a lightweight race car
forks have negative anti-dive, meaning they fast enough, the fluid may boil, or aerate, is the ability to also use the spring as a
have geometrically unavoidable pitch in and lose its damping ability. suspension locating device. In this case the
braking. Therefore, some Japanese bikes coil spring is practically useless, since it has
HARDWARE
now combine hydraulic brake forces with little structural rigidity and in fact may have
the hydraulics in the front shocks to provide SPRINGS a problem in buckling. The torsion bar,
more compression damping when the front Current race cars are most commonly which requires a lever arm and pivot, is a
brakes are applied. When this happens, supported by coil springs, torsion bars, or natural selection to double as a suspension
common road roughness will have a ten- leaf springs, plus the occasional air/oil control arm. And the leaf spring, of course,
dency to “pump-up” or raise the front spring and sometimes even rubber. The ma- does a fairly decent job of locating a solid
suspension because of the asymmetrical jor criteria are space requirements and the rear axle in practically every direction.
damping. While Yamaha and Suzuki use use of springs as locating mechanisms. A Transverse leaf springs have also been used
hydraulic brake pressure directly, Honda coil spring and a torsion bar are essentially as control arms on the independent sus-
uses caliper rotation forces to push a shock the same in performance, because the pension of lightweight cars, though the
absorber valve closed, with an adjustment former is no more than a torsion bar which combination of loads can cause a lot of
to vary the degree of anti-dive. While has been wound into a different shape. trouble.
geometric anti-dive may be more practical However, there are places where either can- If spring friction ride harshness is a prob-

36
es
iea—orB
 lem, coils are the best selection, with
virtually no sliding or rotating surfaces. CHASSIS HEIGHT CHANGE CHASSIS ag RATE CHANGE
Torsion bars are next best, with a couple of
easily lubricated pivots. And leaf springs are
the worst, with a large amount of sliding
friction between leaves. In some cases, this
has been eliminated with single-leaf springs,
but they are not common because of con-
struction and stress problems.
Weight, and especially unsprung weight,
favors the torsion bar. Some part of every
spring has to be considered a part of the
unsprung suspension mass. In the case of
a torsion bar, less than half the weight of
the lever arm alone is considered unsprung,
while half the total weight of a coil spring Figure 20. Choices in Spring Adjustments
is. Leaf springs come out worst of all, with
the largest mass usually being centered components in a steel spring system. Rubber duce a variable-rate suspension in normal
about the axle. There is further potential for may be used in series with the spring or as travel instead of just at the limit of bump.
’ coil or torsion springs, in using hollow wire, spring mounts to absorb minor vibrations But this is undesirable from two stand-
or tubing, for the spring structure. For a before they reach the primary spring. This points. The rate change is very much
cylinder in torsional stress, very little of the is in effect a variable-rate system that is unlike a variable-rate spring--which is al-
core material is used effectively, so by common to production cars, but which ready too complex for accurate suspension
hollowing the wire and increasing the reduces desired spring control on race cars. analysis--in that it has more sudden
diameter slightly, the same performance is At the extremes of suspension travel, bump changes. And the durability of bump
available at a lighter weight. rubbers are used to prevent metal-to-metal rubbers can’t compare with a steel spring.
A final consideration which can be very contact by providing a very rapidly in- If one of them fails or drops off, the han-
important in a highly tuned race car chassis creasing spring rate instead. dling will become unlivable.
is the ease of spring adjustment or replace- As simple as bump rubbers appear, they In recent years, complex spring link-
ment. There are two separate variables here can have a large influence on the suspen- ages have been developed for open-wheel
that must be distinguished. Any suspension sion and can be put to good use if cars. (See Fig. 21) Rocker-arms have been
can be adjusted in either wheel spring rate understood. Although production bump around for a long time, primarily to get the
or ride height, or both. For example, if an stops are usually mounted between the spring/shock out of the sidepod airflow.
independent front suspension coil spring frame and axle or lower control arm, race But now, because of the beam stresses
mount is moved vertically along its axis, the car bumpstops are commonly mounted on involved, and sometimes to get an inten-
ride height will change, but not the ride rate. the shock absorber shaft. Some manufac- tional rising-rate, more designers have
On the other hand, if the chassis mounting turers provide production bump rubbers to - gone to push- or pull-rod suspensions.
point is moved outboard, the ride rate will fit their shocks. However, they are simple Since any motion ratio is possible with
increase without necessarily changing the enough that a racer can modify the stock either, the design choice may be more
ride height. (See Fig. 20) On race cars, bumpers or build his own. The material dependent on other factors such as aerody-
where efficiency of space and chassis used affects the force/deflection curve and namics, packaging in the footwell, service
material is important, it is often more con- the impact absorption or rebound, and convenience, or C.G. height. Spring/
venient to change springs, rather than the varies from butyl rubber on passenger cars damper location has generally gone to
mechanical leverage, for a change in rate. to polyurethane on race cars. When upper longitudinal positions with push
Ride height, however, is often easily ad- available, different material densities can rods, to reduce frontal area. Pull rods are
justed by screw mounts on a coil spring also be used to get different characteristics. sometimes used in a similar way, and have
base, by anchor screws on a torsion bar, or But by far the greatest effect comes from less potential for buckling. Having the
by relocating holes in leaf spring shackles. the shape of the rubber. Production shock- springs close also makes it easier to in-
Variable rate springs should also be men- mounted bumpers usually are of a hollow terconnect them, as with third spring systems.
tioned, although on a race car their action conical shape, which gives a smoothly rising
is too complex to be understood easily. In rate as they are compressed. The rate curve BARS
addition to the potential for rate change in can be found by slowly compressing the An anti-roll bar is a fairly straightforward
the suspension linkage (which is intention- rubber in measured steps on some sort of device that doesn’t allow much choice in
ally provided on a few race cars) only leaf a scale, such as a valve spring compression shape or layout. For passenger cars, which
and coil springs are easily made variable- tester. If the rate change seems too steep, aren’t expected to need much adjustment,
rate. The leaf is obviously and easily it can be modified by carving material off they are bent out of a single length of bar
changed by varying leaf thickness and the circumference and re-testing. This is an stock, with fixed lever-arm holes in the
lengths in the stack. The coil spring has to area for considerable experimentation, both ends. However, since the lever-arm length
be built intentionally — and expensively — in the desired curve and in how to obtain as well as the shaft diameter can have a
to have a variable rate. It can be done either it — but the materials are cheap. For those major effect on race car handling, arms are
by using tapered wire or winding the spring who want to start out with some engineering often made easily adjustable. Lever-arms
with varying coil spacings so that coils data, Kruse and Edwards have the best may be provided with a number of holes,
bottom progressively instead of all at once. available information. or have a sliding mounting block. They may
This leads to the consideration of rubber Bump rubbers are sometimes used to pro- also be made removable, so that they can

S77,
 ROCKER ARM response preferences. Generally speak-
ing, damper cycling rates below 10 Hz
affect chassis motion or handling charac-
teristics, while rates above that control
wheel motion and traction over rough
surfaces. Factory representatives are ex-
perienced in this area and can quickly
produce a modified damping action to suit
whatever a driver asks for.
Mounting a shock absorber is a simple
task, but a few principles should be kept in
mind. First, the end forces are quite high,
and mounting brackets should be in double
PULL ROD shear and well reinforced. Rubber grom-
mets may be good enough for passenger
cars and sedan racers, but but for good
wheel control you must use the highest
quality spherical bearings. And it's worth
mentioning that some shocks, whether
mounted inboard or outboard, have to be
mounted rightside up. There may be an air
reservoir inside which won’t work if it is
mounted horizontally orupside down. One
last word about shock absorbers for the
ultimate racer--some are available with
light alloy components, for those last few
pounds of saved-weight at any cost.

DEVELOPMENT
SPRINGS
It is practically impossible to explain the
racing development procedure for springs,
anti-roll bars, and shock absorbers without
a knowledge of many other areas of vehicle
dynamics. But chassis development has to
Figure 21. The three basic suspension types used when the springs or shocks are be broken down into manageable bits,
mounted inboard. Some geometries now mount the spring horizontally. recognizing that each section may be in-
complete with respect to later chapters. At
the end of the book it should all fit together.
be fitted to any number of different diameter added to allow for a change in ride height Spring selection can be approached fairly
shafts. The bars themselves must be without bottoming. Damping characteris- closely in original design if the engineer is
mounted in low-friction bearings which are tics of a shock absorber may be a purchase willing to make the calculations. All he
as Close to the arms as possible, to minimize option, with manufacturers often tailoring needs to know is the leverage ratio between
two-wheel ride rate effects or bending of the force/travel/velocity curves to fit a particu- the wheel and the spring, total travel
shaft. All linkages should use spherical rod lar vehicle. Most racing shock absorbers desired, total loads on the chassis, and an
ends to eliminate any possible binding under have adjustments which vary the bump or estimate of maximum vertical accelerations
any extremes of suspension travel. the rebound force without the necessity of expected. From there it is a relatively sim-
SHOCKS disassembling the shock. Methods of ad- ple matter either to select the spring from
Shock absorbers are like tires, in that a justment vary from a single knob on the a catalog or design one from scratch.
few manufacturers have tied up the racing body, to disconnecting one end, compress- However, even the best designers usually
game so securely that there isn’t much ing it totally, and rotating the shaft, to have to make some later modifications in
choice. But there also isn’t much chance for modern passenger car adjustable shocks spring rate or height on the running vehi-
anyone to have a far better shock absorber which are controlled via an internal motor cle so this section will concentrate on
than everyone else. Unlike tires, the dif- and a remote switch. changing characteristics rather than original
ference between the best shock and a Other characteristics, especially the design.
mediocre shock isn’t great enough to have difference between high frequency and The essential criteria for springs, with
a major effect on ultimate performance. But low frequency bump damping, can be respect to race car handling, is that they be
that doesn’t mean that the best shock, varied by complete disassembly, modify- as soft as possible, without allowing bot-
properly tuned to a car’s suspension, won’t ing valves, springs and/or orifices--as- toming, and that they be relatively well
improve it’s handling. suming a person knows what effect he balanced from front to rear. Preventing
The most important selection criteria are wants to produce. There is no consensus bottoming means that the suspension should
length and total travel. These figures are on what characteristics make up the per- not hit the frame, the chassis should not hit
easily measured from chassis drawings or fect racing damper, it being mostly depen- the ground, the shocks should not collapse
the vehicle itself, and a tolerance should be dent on the track condition and the driver's completely, and the spring coils should not

38
eS
S
e
 stack solid (in the case of coil springs).
Bump rubbers will protect things at the last
fraction of travel but they shouldn’t be con-
sidered a part of normal spring action. With
all possibilities for bottoming, especially
considering the variation from smooth
tracks to bumpy or hilly tracks to high-
banked ovals, the only way to know is by
testing on each particular track.
In the absence of sophisticated instrumen-
tation to record suspension travel, any
bottoming will have to be felt in the chassis,
or heard as parts hit. Clay indicators may
be attached to suspiciously close surfaces,
or wear strips attached to the bottom of the
chassis. If nothing is even close to touching,
then either the chassis or the spring rate
aren’t low enough. Any given suspension is
probably set up not to bottom under the
worst anticipated track conditions. But if
’ those conditions aren’t encountered, ground Two types of anti-roll bar ends. The perforated end is more rigid than
clearance is being wasted. As mentioned the bent bar, which can bend in addition to its normal twisting action.
previously, the chassis bottoming on the
ground is probably the least undesirable
condition. Spring height is the easiest factor deflection, and that may be the most ac- common for a spring to be stressed beyond
to change, at least as a test. If lowering the curate method. But some springs would its yield point, and to sag to a different ride
suspension by an inch causes metal-on- require a great deal of weight, and it makes height. Most spring builders are aware of
metal contact, remove or relocate those things a lot easier to know some of the the proper steel alloys to use, the calcula-
pieces and try again. When the chassis just simple mathematics of spring rate. tion of stress limits, and the best method of
barely touches the track under worst The basic equation for coil spring rate is heat treatment after winding. A final sug-
possible conditions — full fuel load and a function of the coil diameter (D), the wire gestion for the selection or construction of
maximum aerodynamic downforce over the diameter (d), and the number of active coils coil springs is the type of ends. Passenger
hardest bump — that’s low enough (assum- (n). The rate is found by: cars usually have spring seats designed to
ing there are no minimum ground clearance 1,500,000 x d+ accept an inexpensive “open-wound”’ spring
rules). At this point there is the option of Ibs per inch = end. Race cars, however, usually require
no D3
raising the ride height slightly and trying a ends which are “‘close-wound” and ground
lower rate spring. There are no definite rules flat, to better distribute their load on the
Using this equation, a new spring can be
to follow to decide between minimum - spring seat — particularly in the case of
designed, or it can be seen how a change
ground clearance and minimum spring rate. shock-mounted coils.
will affect its rate. For example, if the wire
There has to be a trade-off between better The basic equation for torsion bar spring
diameter is increased by 10 percent, then the
aerodynamics and lower center of gravity rate is a function of the bar length (L), the
rate will increase by: (1.10 d)* or 146 d*
with the former, to better ride and tire bar diameter (d), and the lever arm length
or 46 percent. Similarly, it can be seen that
traction with the latter. Only testing will tell (1). Its spring rate is derived by:
a 10 percent increase in the number of coils
— if all other factors such as alignment and (1,120,000) x (d*)
will decrease the rate by about 10 percent, Ibs per inch =
roll stiffness are optimized in both cases.
and a 10 percent increase in coil diameter (L) x (’)
At least it is possible to estimate some will decrease the rate by 1/(1.10 D)? or 25 Using this equation, it can be seen that a
things about a spring change. For example, percent. In practice, the coil diameter is 10 percent increase in bar diameter will in-
if the front wheel ride rate is 200 pounds usually difficult to vary without modifying crease its rate by 46 percent, just the same
per inch and the chassis just barely touches spring mounts, so wire diameter and as with a coil spring. Similarily, a 10 per-
the ground with three inches of ground number of coils are most often altered. Free cent change in length will change the rate
clearance, that means there is a bump of height, ride height, and solid stack height by about 10 percent, and a 10 percent change
about 600 pounds force. Therefore, if the are also important to consider in the in arm length will change the rate by about
ride rate is reduced to 150 pounds per inch, winding of a spring. Free height may end 20 percent. Torsion bars don’t have the same
600/150 or 4 inches of ground clearance will up shorter than full rebound travel allowing space and height problems as coil springs,
be required. Of course, this is a gross the spring to become loose during rebound. but material and stresses are just as im-
simplification, which ignores a lot of other Even worse is a condition where the bump portant. For those who want to build their
factors such as shock absorbers and travel is limited by all coils bottoming at own bars, nickel chrome-moly (4340), or
front/rear roll resistance, but it gives once, causing destructive stresses in the chrome-moly (4130) steel are satisfactory.
ballpark figures for faster changes. spring and suspension components. Use 4130 if there is any welding. After
It is also valuable to know how to iden- The material a spring is made from can welding, or heating for bends, bars must be
tify springs by their rate, how to modify have a slight effect on rate, but the previous heat-treated to a Rockwell C hardness of
them if necessary, and how to select a equation is applicable for most spring steels. about 35. In making any calculations, it is
different rate. Of course, it is possible to More important is the effect of material on important to note that the above equation is
load a suspension spring and measure its durability and yield strength. It isn’t un- for spring rate at the arm end in pounds per

39
 effects of modifications ought to be
sufficient. Since an anti-roll bar is simply
an unanchored torsion bar, the same sort of
equation applies, at least in geometric
relationships. Of course, even the exact
desired change in bar stiffness is unknown,
but the equation gives some relative
magnitudes:
(constant) x (d*)
bar roll resistance =
(L) x (1?)
Length (L) of the bar is relatively difficult
to change, but it can be seen that changing
arm length (1) is more effective, and that the
slightest change in bar diameter (d) can have
a major effect. Because most of the chassis
roll resistance is ordinarily provided by the
suspension springs, however, the percentage
in roll change will be far less than the
percentage change in the bar’s roll resistance
alone. For cars that need an extremely stiff
Two types of anti-roll bar adjustments, the sliding end to vary roll resistance,
anti-roll bar, or where ultimate lightness is
and the variable length link to balance wheels loads.
a factor, hollow anti-roll bars are often used.
Installation of an anti-roll bar is
straightforward — with a few notes. All
inch — which is not the same as pounds- will naturally affect the front/rear roll rate, rubber should be eliminated from
feet of torsion in the bar. The spring rate and therefore the oversteer/understeer mountings and linkages, to provide the most
is also not the same as ride rate, which will characteristics. However, the handling can direct-acting roll resistance. Linkages should
be some fraction thereof, based on the usually be brought back to balance with also be checked for binding throughout total
square of the mechanical force or deflection properly selected anti-roll bars. Without a travel. All anti-roll bars should be checked
ratio between arm end and tire travel. definite knowledge of what the amount of for parallelism between the arms before
Leaf spring calculations are a mess roll resistance ought to be at the front and installation, and the links adjusted for zero
because of the possible variations in the rear, it is useless to try to pick the proper preload in the bar with the chassis sitting
length, width, thickness, and number of anti-roll bar for a car without road testing. on a perfectly flat surface and all four
leaves, not to mention spring arch and Therefore, a simple consideration of the wheels at ride height. If the handling
location of the axle. Any calculations would Sh _— .
be far more difficult than actual physical Bc
measurement of the rate, with changes made
by adding or subtracting leaves. However,
to give a rough idea of how changes will
effect the rate, a basic relationship can be
shown, where leaf width (w), thickness (t),
and length (1) are considered:
(constant) < (w) xX (t®)
Ibs per inch =
(aya
This isn’t enough information to design a
spring from scratch, but for a symmetrical
leaf spring, with evenly spaced leaves, it in-
dicates the effect of changes. If the width
of all leaves is increased by 10 percent, the
rate will increase by 10 percent. However,
if the thickness or length of all leaves is
changed by 10 percent, the rate will change
by 25 to 33 percent. The selection of
materials and heat treating, with respect to
stressing and sag, are the same as for coils
and torsion bars.
BARS
When selecting or modifying spring rates
<
in a race car, it isn’t enough to merely con-
sider ride rate and ground clearance. Any One method of altering anti-roll bar resistance from the cockpit is to use
change in the ratio of front/rear spring rate a sliding bracket with manual control cable to vary the arm length.

40
 develops an asymmetry between left and
right turns after some time, the bar may
have twisted. A recheck of arm parallelism
will quickly answer the question. Unless a
bar is overstressed, it shouldn’t twist to the
point of yield, especially if the right material
and heat-treat are used. As with springs,
chrome-moly 4340 steel (or 4130 if welded)
is best, with a heat-treat to Rockwell C 35
hardness.
There is so little interaction between anti-
roll bar stiffness and any other chassis com-
ponents or characteristics that its selection
or adjustment is usually the last step in
controlling oversteer/understeer. As was
explained in the Theory section, changing
the ratio of anti-roll bar stiffness from front
to rear usually has a great effect on hand-
ling. Ordinarily, whichever end of the car
- has a reduced roll stiffness will have greater
lateral traction. A stiffer front bar produces
more understeer, and a stiffer rear bar pro- Ajustable roll resistance mechanism created by mounting rotatable blade ends on
duces more oversteer. Usually. In some the suspension rocker pivots and connecting them with a rigid link. The vertical
cases the effect has been indistinguishable, position shown is full hard, to minimize roll.
or even worked the opposite. This could be
dependent on the type of differential, the
braking distribution, or even driving style. developed race cars are so sensitive to affect oversteer/understeer fora short time.
It may even be that the chassis is so weak changing conditions that they can have a This can have a positive or negative effect,
in torsion that the springs and anti-roll bars critical change in handling over the length but the most desirable case is probably to
have no effect at all. Only skidpad testing of a race. Reduced fuel loads or a track have zero effect. Next best may be to have
— confirmed by track testing — will tell surface which becomes oily can cause a car a decrease in understeer as the vehicle
for sure. to oversteer or understeer excessively. If approaches steady state. This would imply
The current consensus in racing is that pitstops are long enough, the anti-roll bar that the front damping rate shouldbe greater
the best handling comes from a very low can sometimes be adjusted — if a tire com- than the rear. In other words, the rear roll
spring rate with high roll stiffness — pound change won’t produce the desired resistance increases as the chassis rolls and
whether from stiff anti-roll bars, a wide result. Oval track racers frequently use the the rear shocks deflect more slowly. It
spring base, or by keeping the center of technique of diagonal weight-jacking for probably takes an extremely aware test
gravity and roll center close together. Soft really quick and rough changes. More driver to make this distinction, but on the
springs, however, require some degree of recently, the variable anti-roll bar has other hand, an intolerable inverse condi-
anti-squat and anti-dive geometry to keep become popular. By providing a hydraulic tion may be cured by the application of this
the car off its bump stops. or mechanical linkage to change the length knowledge. Especially since it is so easy to
The primary reason why variable-rate of the arms or the spring rate in the arms, adjust the damping rate on some racing
springs are so difficult to deal with on a race the driver can modify the handling at any shocks. This also makes it important to be
car, is that they also provide a variable roll- time or place. It’s even conceiveable that the certain that in road racing the rate is equal
rate. With linear-rate springs and anti-roll handling could be changed from corner to from side to side at both ends of the car.
bars, the front/rear distribution of roll rate corner, if any driver could tell exactly what A final point on shock absorbers is that
is relatively constant, making the handling he wanted and where he wanted it. they may be used to control other than
fairly constant and predictable. However, vertical oscillations. Some dampers are
when the roll rate changes with suspension SHOCKS designed for the steering system--to absorb
deflection and/or roll angle, anything can Shock absorbers don’t ordinarily get unwanted steering wheel kickback. This
happen. If the suspension design has the much development time. They’re either too tends to be a quick fix for poor design, but
springs totally removed from any roll hard or too soft, or right or wrong for the it may come in handy if the kingpin offset
resistance — if their center of support is at driver’s taste. However, it isn’t commonly is excessive for other design reasons. Damp-
the roll center — and the anti-roll bar takes known that they can have a direct effect on ers are also used to control other solid rear
care of all roll resistance, then variable-rate handling characteristics other than ride. axle oscillations, whether in rotation or a
springs may be more tolerable. In fact, such Shock absorbers may not provide any roll longitudinal direction. Many passenger cars
suspensions have been used, but not too resistance in a steady-state condition such have the shocks staggered fore and aft on
successfully, for unknown reasons. It may as a long curve or accelerating or braking. the rear housing. By having one mounted
be that variable-rate suspension springs But they do resist suspension movement in ahead on one side and the other mounted
aren't that desirable under any short-term transient conditions. As the behind, high-frequency housing rotation
circumstances. chassis rolls to one side in a turn, the roll can be reduced. Specially-designed hori-
Once the car is developed and on the race is resisted not only by springs and anti-roll zontal dampers are sometimes used to re-
track, it would seem that the suspension was bars but by some damping rate. If the rate duce any longitudinal movement which
beyond consideration. However, very finely is not equal at front and rear, the shocks will may be the cause of rear brake hop.

41
 forces are available, as with brake boosters,
but race engines just don’t produce enough
vacuum. Electric screw jacks have time lag,
reliability and weight problems, although
they may allow simpler control systems.
The sensing and control system is the real
problem. For either force or position
devices, the sensor should probably be sen-
sitive to position, since that is the quantity
to be corrected and it is usually easier to
measure than forces. The advantage of a
suspension travel sensor on each wheel is
that it can distribute a corrective change to
each wheel separately, whether the deflec-
tion is due to load transfer, cornering forces,
fuel consumption, or aerodynamic loads.
The difficulty comes in trying to filter out
the deflections that shouldn't be corrected,
such as ordinary bumps or dips in the road.
An improper sequence or frequency of
bumps could cause the suspension to get out
of synchronization, causing the chassis to
leap up and down.
A lightweight adjustable racing shock absorber, with adjustable spring
perches on the threaded housing, and a urethane bump stop (top right). ACTIVE SUSPENSION

In the early eighties, Lotus tried to

LOAD LEVELERS force balancing systems use air or hydraulics
totally eliminate passive springs and ac-
as either primary or booster springs. In this
tively control their suspension (see p. 163).
Load levelers are a special case in sus- case, the spring rate can be changed rapidly
Active suspension is actually a high-fre-
pensions which can include the areas of and at any time by simply changing the
quency load leveler which forces the
springing, damping, and roll resistance. The static fluid pressure in the spring. The
wheels up as well as down. After a few
idea is to balance any changes in load, or primary disadvantage in this is the same as
for variable-rate springs. Changing the
seasons, Lotus abandoned it due to reli-
changes in suspension travel, by some
spring rate is apt to change the roll-rate ability, weight, and power problems. Fi-
separate and controllable force. The
distribution and, therefore, the handling
nally, by 1992, advances in computers and
principle is fairly common to the rear of
characteristics. logic reached the point where the Wil-
passenger cars and trucks but it presents
some difficulties to finely balanced race car Deflection balancing systems would seem liams team could re-introduce it, and give
handling characteristics. Another important to be simpler, and adaptable to any type of it a lot of credit for the Championship.
distinction is that a race car cannot be springs. In this case, the chassis mounting One key to making it work, is to sus-
stopped and adjusted for changing condi- points are made vertically adjustable, and pend the chassis with very low rate passive
tions as a passanger car can, but must can be moved by hydraulic or pneumatic springs, and only use the active to move
respond very rapidly and automatically. cylinders, or even by electric power through the mass of the wheel and hubassembly up
Still, a load leveler is highly desirable in motor-driven screw jacks. This method and down. This creates immense com-
racing, considering the large changes in fuel reduces the effect on handling, since for puter logic problems, as any time any
load and aerodynamic downforce. It would constant-rate springs, the ride rate or roll wheel load changes, the handling may
simplify the design of all suspension rate distribution shouldn’t change with change, creating the potential to alternate
geometries if some reliable system could be deflection. The primary disadvantage is that between understeer and oversteer many
built in to provide constant suspension travel a great deal of extra spring and shock times per second. Reliability is also criti-
and chassis attitude under all conditions. absorber travel is required. cal, with precise sensors needed to sepa-
Many of the important problems of roll Then there is the problem of a power rate true signal from the noise, and valves
center height, anti-pitch, lateral jacking, and source, as it does take some power to raise operating at high speed in an incredibly
aerodynamic loading could thus be reduced a chassis on its suspension in a short time. harsh environment.
or eliminated. In fact, it would take half a horsepower to What was frequently overlooked in the
The solutions fall into two categories: raise a 1650-pound car 2 inches in | second. early years, was that active suspension had
force balancing systems, in which the spring While that doesn’t seem like much com- less to do with wheel control, than with
rate is increased, or deflection balancing pared to engine output, it isn’t a simple aero control. In spite of the ban on “move-
systems, in which the springs are jacked up. operation to tap it from existing power trains able aerodynamic surfaces,” controlling
Force balancing systems are quite difficult or to generate it independently. Possible the front and rear ride height is equivalent
to design for use with metal springs, since options are: existing oil pumps for hydraulic to changing the pitch and ground interfer-
their rates are largely fixed functions of power, an auxilliary engine-driven pump, or ence of a wing. For maximum downforce
original spring design. The most probable the pumping action of the shock absorbers. in a turn, the body can be lowered and
possibility would be adjustable leverage on The latter device actually exists, and seems pitched down, and for minimum drag in
the spring, perhaps by making the chassis to be a source of free power, but it is the straights, it can be raised free of the
mounting points move horizontally. Most undependable on smooth roads. Pneumatic ground, reducing aero and tire drag.
42
 D
BRAKES
From a safety standpoint, the brakes by two friction surfaces transforming energy square of velocity, making it far more ef-
are just about the most critical system into heat, and if the wheels are not rotating, fective at high speeds. It is not uncommon
in a race car. And yet, paradoxically, then the friction is generated exclusively be- for air drag on a race car to provide 0.5g
there is little to be gained in lowered lap- tween tire and ground. This works, of deceleration at 140 miles per hour with no
times by development or engineering course, but not more than a few times before brake application at all. And at 200 miles
improvements. Brakes are generally the tires get flat spots and/or blow out. per hour, the deceleration will double to
bought as a package, and there is little Meanwhile, the brakes do practically no 1.0g. Race cars intentionally have the lowest
choice or variation from race car to race work at all. In addition, as discussed in possible drag, which further adds to the load
car. If the racer follows all the rules of Chapter 2, a tire has its maximum traction on the brakes.
setup and maintenance, the brakes are at a point just short of lockup, or at about The rest of the energy has to be absorbed
about as close to their maximum 5 to 10 percent slip. So the car can by the four brakes and stored until it can
’ theoretical performance as any part of theoretically stop faster with unlocked be dissipated into the air stream. The fact
the car. Stopping performance is cur- wheels — if the driver can modulate them that even a race car takes far longer to ac-
rently limited by tires and aerodynamic that closely. In practice, race tires are so celerate than to brake means that the brakes
downforce, not by the state of the art sensitive that a race car may stop in a shorter have some reasonable time to dissipate this
in brakes. Of course, developments in distance with all four wheels locked up. heat. Still, there is a limited amount of metal
automatic anti-skid brake systems — That’s worth keeping in mind for the critical in the brake discs in which to store heat,
covered briefly at the end of the chapter — emergency when loss of steering control is and a typical racing application can raise
promise some gains in performance. How- no problem — or is already lost. their temperature 200 to 300 degrees in a
ever, even if there were a lot to be matter of seconds. If that energy can’t be
gained, so little time is spent braking that dissipated before the next application, there
improvements don’t have a great effect on HEAT ABSORPTION will be a net temperature rise. The average
total laptimes. The energy that a brake system must operating temperature will rise until it
absorb is mostly a function of the weight stabilizes where braking input equals
REQUIREMENTS of the vehicle and the change in speed. The cooling output (see Fig. 22) and can be
The desired characteristics of a race car equation for this energy is: foot pounds = anywhere from 1000 to 1500 degrees at the
brake system are: 1. reliability or fail-safety, disc face. At those temperatures the friction
(wt) (top speed? — min speed”) (.0335)
2. ultimate stopping power, 3. balance or material decomposes rapidly, not to men-
proportioning, 4. feel or modulation, 5. For example, a 2000-pound race car . tion the changes in the metal at the disc face.
durability or fade resistance, 6. pedal braking from 200 to 50 miles per hour must If the mass of the discs or the cooling air
pressure, 7. life of the friction material. As absorb and dissipate 2,500,000 foot-pounds are not adequate, temperatures will rise to
long as the driver can hold all four wheels of energy. Even more amazing, it will prob- the point of lost friction capability (fade),
at the verge of lockup from top speed to the ably generate it in a period of less than seven where the brake material burns, the brake
minimum speed, and the brakes will main- seconds. Fortunately, aerodynamic drag fluid boils, or the discs themselves crack
tain that consistency for the length of a race adds some braking force, and the faster and and fly apart.
— both of which are reasonably possible — less streamlined the car is, the more it helps.
then there are no further gains to be had, Depending on the average speed and shape FORCES
unless perhaps a reduction in weight. But of a race car, from 10 to 20 percent of The ultimate one-stop force that can be
there will always be a trade-off between one- braking energy comes from air drag on the produced is a function of a number of
lap performance and durability. Anything body. The drag force is a function of the
that improves performance — whether in- 1500 APPLICATION PERIOD
creased tire traction, increased downforce,
or decreased brake mass — will increase the
demands on the system and decrease
durability. Decreasing the mass of the entire
COOLING PERIOD
vehicle, of course, is all gain, but that is
taken for granted in race cars. As with
wheels and tires, the rotating components
of a brake system have a double contribu-
tion to mass. Their rotating inertia can add
about another 50 percent to their total
disc
brake
degrees
in
temperature
effective weight which must be accelerated
or braked.
2 3
The old idea of testing brakes by locking
time in minutes
up all four wheels and sliding to a stop isn’t
too sensible. Stopping power is generated Figure 22. Brake Cycling Temperatures
43
| MOORES UNIVERSITY
LIVER LEARNING SERVICES
 vacuum reservoir. But power brakes also
cause some loss of response and feel be-
tween the pedal and the wheels.

BALANCE
When reliability and stopping power are
adequate, balance or proportioning must be
optimized. Distribution of braking forces to
each wheel is fixed. But the load on each
wheel is not, so there comes a time when
one wheel is more lightly loaded and will
lock up more easily. The most common ex-
ample is the inside front tire when braking
into a turn. When the load transfers away
from it, the limit on braking will be that tire,
unless it is to be skidded and flat-spotted.
This is less of a problem at the rear, where
a drive axle — and usually a locking dif-
ferential — connect the inside and outside
wheels.
The effect of longitudinal load transfer is
even greater since maximum braking is
ordinarily done in a straight line. It is well
known — often from personal experience
— that if the rear wheels lock up first in
braking, the car will be highly unstable. So
it’s important that under all circumstances
the front wheels should lock first — and yet
This plastic ‘‘hubcap’’ has spokes designed to pump cooling air through the wheel not so much sooner that the rear brakes
from the brakes. The wiring is for an electronic test speedometer. don’t do their full share of stopping also. At
the same time, neither the left or right wheel
should lock before the other, which could
mechanical design factors — and the force the travel ratio are inversely proportional. cause swerving or instability. But that is
available from the driver’s foot. The design In other words, to get a force ratio of 1 to more often a maintenance and proper func-
factors are friction material coefficient (k), 70 between the pedal and the brake discs, tioning problem than one of design or
tire radius (R), brake disc application radius the travel ratio must be 70 to 1, or 7 inches development.
(r), and mechanical force ratio (f). of pedal travel for every 0.1 inch of pad Proper longitudinal braking distribution
Therefore: stopping force = travel. It doesn’t matter whether the force is an important design consideration and is
(pedal force) x (f) < (K) x (r) x (1/R) ratio is from mechanical linkages, different a function of center of gravity height (h),
hydraulic cylinder sizes, or a combination the wheelbase (b), the total weight (W), the
Since most of the factors are relatively of the two. When pedal force or travel limits front and rear weights (Wf and Wr), and the
fixed, the stopping force is most easily con- are reached, an assist will be necessary. For- coefficient of friction (Cf). The equation for
trolled by the mechanical force ratio be- tunately, light race cars are usually below the ideal front/rear ratio is:
tween the pedal and the braking device, the point where a force boost is required, Ff _ Wf + (W) x (Cf) x (h) x (1/b)
although the others may have some small but 4000-pound sedans with tire coefficients
positive or negative effects, and must be of 1.2 can be marginal in required pedal
Fr Wr — (W) x (Cf) x (h) x (1/b)
considered. The maximum stopping force effort. When necessary, force assist is The tire coefficient will probably be
required for a race car will be the point at available in two ways — power booster or slightly less at the front for smaller front
which the tires are on the verge of sliding servo-action. Servo-action brakes, whether tires, but this complicates the balance equa-
— with peak aerodynamic downforce. For drums or discs, use a wedging action to in- tion and the effect is less than the accuracy
a 2000-pound car, with 500 pounds of crease the force of the friction material of other quantities. This simplified equation
downforce and a tire Cf of 1.4, the stopping against the rotating surface. The reason why also assumes that aerodynamic downforce
force required is (2500 x 14) or 3500 they aren’t usually applied to racing or even is balanced the same as the weight from
pounds. Assuming a peak pedal force of 100 to disc brakes in general is that the action front to rear, that Wf and Wr are constant,
pounds, a friction material coefficient of 1.0, is relatively hard to control. The force and that tire diameter is equal front to rear.
a brake application radius of 0.5 foot, and multiplication is non-linear and tends to In fact, this is far from reality. Varying tire
a tire radius of 1.0 foot, then the force ratio cause wheels to lock — and to be hard to diameters can be an important problem or
(f) will be found by: unlock. Vacuum power boosters are rare on solution to front/rear brake balance. Also,
3500=(100) x (f) x (1.0) x (0.5) x (1/1) race cars but are used on some large Wf and Wr can change considerably, not
American sports cars and sedans. The first only from fuel consumption, but from the
or a force ratio of 70 to 1. drawback is the lack of vacuum generated common practice of having higher
The problem is that the force ratio and by racing engines, which requires a large aerodynamic downforce at the rear — which

44
changes the load distribution with velocity.
It is important to remember that the tire
coefficient varies with the track surface and
is not a fixed quantity — as the brake
distribution is. For example, when the track
becomes wet, the coefficient can drop to a
fraction of the dry track value, changing the
optimum balance by 10 to 20 percent. For-
tunately, the change in Cf works in such a
way that the slipperier the surface, the more
likely the front wheels are to lock first. In
other words, if the balance is set for perfect
brake proportioning on the maximum trac-
tion surface, the car will always brake
stabily on a wet or oily surface — although
with reduced efficiency, since the rear
brakes are under utilized.
Obviously the optimum brake balance is
a very complex problem. Although an equa-
. tion can be written to include all the
changing factors, it doesn’t make much dif-
ference as long as the balance can’t easily A driver-adjustable brake pressure limiting valve helps prevent early lockup of
be changed in action. The problem is the rear brakes by compensating for changing fuel load or traction.
minimized by placing the fuel load at the
center of gravity, by keeping the center of
gravity low, and by distributing front and cooling air. For a given race car with limited contact with the caliper and pistons, which
rear aerodynamic downforce properly. But power to get it back up to speed, however, can be over 500 degrees themselves. When
the only solution to variable tire coefficients if the brakes don’t overheat and drop off in the fluid boils in the caliper, air bubbles
is to have an electronic anti-skid system or 15 minutes of hard racing, they’ll probably form, and being compressible, they use up
an easily varied proportioning valve. last the life of the pad material. When all-important pedal travel. The pedal may
A final important influence on front/rear brakes do go away, there are three possible go all the way to the floor while producing
brake proportioning is the effect of engine reasons or types of fade. only half the pressure required at the caliper.
drag at the rear wheels during deceleration. First is mechanical fade, where the The best solution for this, aside from bet-
At one time, when brakes were less durable, temperature increase causes metal com- ter caliper cooling, is to have a number of
that was a valuable contribution to braking. ponents to expand or deform excessively. insulators between the disc and fluid.
But now that brakes can work at their peak Drum brakes could expand away from the Usually, a Bakelite insulator on each piston
for the length of a race in many cases, shoes or warp the friction surface. Discs will suffice. Insulation also comes in handy
engine drag simply upsets the balance. A have less of a problem in this area, but they ‘at other locations, such as the disc mounting
5-liter engine with closed throttle at high can still warp out of a flat plane. And the flange. The thinnest layer of chemical or
rpm can add over 200 pounds of brake force expansion in calipers, cylinders, and lines composition insulation, or even a firmly
at the rear tires — which will cause lockup displaces fluid that would otherwise be bolted interface, can make a difference of
and skidding if the balance was otherwise applying some force. 100 degrees. This can mean life or death for
ideal. On the other hand, at low rpm, if More common is the chemical fade in wheel bearings where the disc is bolted to
engine inertia is very great and the car is friction materials. There is a coefficient of a front wheel hub.
decelerating very rapidly, the brakes may friction curve for each different material,
also have to drag the engine down, which which usually rises with temperature, has HARDWARE
subtracts rear brake force. The actual a relatively constant value over a wide Because of the emphasis on racing ap-
figures require some specific dyno work range, and then drops rapidly at a critical plications, this chapter will concentrate on
and/or heavy calculations but it is apparent temperature. The peak friction may never disc brakes only instead of including drum
that the effect is undesirable in either case come back again if the material has been brakes. Although drums are used on the rear
and should be avoided. If declutching during permanently glazed. But there is also the of some production sports cars and racing
braking isn’t practical, leaving the transmis- possibility of vaporization in the bonding sedans, they are not highly stressed in such
sion in the highest gear as long as possible material. In this case, some of the bonding applications. When drum brakes approach
will minimize the problem. material in composition pads is vaporized, their limit of racing performance, they are
and it forms a layer of high-pressure gas be- inevitably replaced with discs. The advan-
FADE tween the pad and the disc. This is the type tages of disc brakes in a racing application
The theoretical consideration of brake of fade which occurs on new, unburnished are: less flex or expansion problems from
fade — or total loss — can’t compare with pads, and is referred to as green fade. braking forces and heat, perfectly linear
the actual experience on a race track. At Preventative measures — beyond careful force multiplication (no servo-action), less
such times, the importance of durable and break-in — include air grooves in the sensitivity to water, ease of maintenance and
reliable brakes can make a definite, life-long friction material or holes in the disc surface. replacement, and most important, far bet-
impression. Any brake system can be faded The last kind of fade can come from ter heat dissipation characteristics. The
if the energy input is great enough or boiling brake fluid. The fluid comes very relatively minor disadvantages — higher
frequent enough, or if there isn’t adequate close to the red-hot disc, and is in direct forces required and disc warpage — are

45
 is the major factor in pedal force multiplica-
tion and front/rear proportioning. With all
else being equal — master cylinder, tire, and
disc size — the ratio of piston area from
front to rear is the braking distribution ratio.
Using previous equations, the total required
force ratio between master cylinder and
wheel cylinders can be calculated. This
defines total caliper piston area. Then,
calculation of the proper longitudinal force
distribution will give the ideal distribution
of that piston area between front and rear
calipers. For example, assume a required
force ratio of 80 to 1 between calipers and
pedal, a 1-inch-diameter master cylinder, a
mechanical leverage at the pedal of 4 to 1,
and 60 percent of the vehicle weight on the
front wheels during peak braking. The
piston force ratio will need to be 20 to 1,
requiring a total caliper piston area of at
least 16 square inches, with 60 percent or
9.6 square inches at the front and 6.4 square
inches at the rear. It isn’t likely that calipers
are available in exactly those sizes but it
provides a first approximation in selection
and the ratio can be further modified with
et. master cylinders or tire diameters.

These inboard brakes will receive more heat from the engine and the exhaust pipe The greater the piston area per caliper,
below than the simple duct can possibly dissipate. the lower the pedal force and the hydraulic
line pressure will be. However, as men-
tioned previously, this requires more travel
tolerable or easily overcome. A discussion ability, and cost have made cast iron the at the pedal. Another disadvantage is that
of the hardware breaks down into six most practical choice. Although beryllium more piston area, and/or more pad area,
categories: discs, calipers, master cylinders can do the same job at a quarter the weight, means less disc exposed to the air, and less
and lines, brake fluid, pads, and accessory as it wears it generates a poisonous dust. cooling. In some cases, two pairs of calipers
devices. The most exotic and promising material is are used on the front discs, but they usu-
sintered carbon, which is covered in more ally cause such severe overheating that it is
DISCS detail at the end of this chapter. necessary to go instead to higher hydraulic
There usually isn’t much choice in discs In cost per pound of weight reduction, it pressure with single calipers. When
for a given car. The racer takes whatever makes sense to stick with ventilated iron hydraulic pressures get well over 1000 psi,
comes bolted to the hubs. If there is a discs. A halfway solution is to replace the then caliper stiffness becomes critical. Any
choice, it is probably between a solid disc non-wearing center of the disc with an spreading of the caliper, even a few
and a ventilated disc, and perhaps also the aluminum hub. This cuts weight and yet thousandths of an inch, can use up valuable
thickness. A ventilated disc has far better provides enough cast iron in the friction and pedal travel. The deflection is easy enough
heat dissipation capabilities per pound than ventilated areas to absorb and dissipate the to check with a dial gauge and a heavy foot,
a solid disc and can run a few hundred heat. A further advantage is that the bolted but it may be impossible to reduce without
degrees cooler, although a solid disc may interfaces prevent some heat from passing adding material which further shields the
be perfectly adequate for relatively slow and to the hub bearings. One innovation which disc.
lightweight race cars. All else being equal once appeared on a European disc was fric- Due to the natural flex and movement be-
— in ventilation and dimensions and tion surfaces with intentional deep V- tween caliper and disc, something must
material — the heavier the disc, the more grooves. When the brake pads were grooved float, or have a built-in allowance for
heat it can store without building up to to match, this provided much more contact relative movement. If the caliper has inter-
destructive temperatures. When everything area and cooling surface for the same size connected pistons on both sides of the disc,
else fails to improve brake life, the last step pad and disc. While it was never positively they can float from side to side. Otherwise,
is to increase disc thickness or mass. proven superior, at least it suggests that either the caliper or disc must be mounted
The most common material for brake accidentally scored discs aren’t necessarily loosely enough to move laterally and yet
discs is cast iron or Meehanite, although a ruined — as long as the pads are burnished- have no freedom in any other direction.
number of other alloys show some poten- in to match. Sometimes the caliper slides on bushings on
tial in better heat conductivity and lighter the mounting bolts, and sometimes the disc
weight. Bendix (SAE paper 700137) studied CALIPERS is located freely on drive pins.
aluminum and copper alloys, and beryllium The choice of caliper can be all-
has even been used for brake discs in some important, from considerations of strength, MASTER CYLINDERS
aircraft applications and a few race cars. piston area, and cooling. The primary Although it is possible to get by with a
However, the problems of strength, dur- criterion is the total piston area, since that single master cylinder, there are some very

46
good reasons for using two, or at least, a
common production tandem cylinder. The
advantage in fail-safety is obvious and in
most cases required. If one system loses all
its fluid, the other independent system will
still slow the car — although at a drastically
reduced rate. The tandem cylinder does this
by having a floating piston isolating the two
fluid systems in the same cylinder bore.
When fluid leaks out of one system, the
piston bottoms at one end and allows the
other system to keep functioning. A pair of
parallel master cylinders has the same ef-
fect, except that the two pistons are con-
nected by a variable balance bar. This can
be used to modify the force ratio to either
one, an adjustment not possible with the two a
equal tandem pistons. A brake balance bar, with clevises and rods leading to front and rear
Previously discussed force ratios between master cylinders. As the central pedal pivot is moved toward either cyl-
‘the pedal and the caliper must take into con-
inder, that end of the car receives more braking force.
sideration the mechanical leverage between
the pedal and the master cylinder. This is pedal travel. All lines and hoses should be extreme pressure and temperature. But some
the ratio between the measurements from carefully routed and shielded to protect them pads derived from passenger-car designs
pedal to pivot and from master cylinder rod from any bending, impact, heat, chafing, have relatively weak steel backing plates and
to pivot. The ratio may vary anywhere from crimping, or vibration. They are very may therefore be reinforced with a flange.
6:1 to 3:1, and is relatively constant with definitely the driver’s lifelines. Heat insulation between the disc and the
travel. The fact that the ratio is not perfectly caliper pistons is perhaps as important as
constant makes it possible to use the FLUID rigidity. There is a very small distance bet-
variability to advantage. By careful location Brake fluid is usually taken for granted ween the disc, which may be over 1200
of the pivot points and travel arcs, the ratio — until it boils from excessive heat degrees, and the brake fluid, which must be
may be made to vary from large travel and causes brake loss. Actually, it is hard under 550 degrees. One type of brake pad
multiplication with low force multiplication to find a fluid that is practically incom- once had an air gap between two thin metal
at the beginning, to take up the slack, and pressible and that will boil at over 550 backing plates, but extreme forces on some
little travel with high force at the end. Just degrees, although a few racing brake fluids heavy cars caused the plates to bend and
as at the caliper, even the most minor flex will meet these criteria. There is also a li- collapse. While passenger car brake pads
from extreme pedal forces must be avoided. quid silicone fluid available that has an even have a relatively constant coefficient of fric-
And it is important to check that the pedal higher boiling point, but it hasn’t been con- tion up to perhaps 1000 degrees, they
isn’t unnecessarily limited in travel by bot- clusively proven in racing yet. decompose or fade rapidly above a certain
toming on the floor when everything Moisture is a critical problem with brake temperature. Racing pads have somewhat
deflects under pressure. fluids, since they have a strong tendency to the opposite effect in that their friction coef-
Master cylinders and their fluid reservoirs absorb it from humidity in the air and the ficient is constant at much higher
are sometimes mounted with an extreme water has a drastic effect on the boiling temperatures but drops off below 800 to
rearward rake. When they are being used, point. An absorption of even one percent 1000 degrees. On some race cars there isn’t
the deceleration will cause the fluid to flow of water can drop the boiling point over 100 enough friction to lock the brakes until they
forward, and settle at an angle of over 45 degrees. It is apparent that the brake system have been brought up to operating
degrees from the horizontal. So it’s im- — and all brake fluid containers — should temperature.
portant that the reservoir angle and/or be kept well sealed at all times, especially
capacity be adequate so the drain hole into in humid or rainy weather. A leak in the VALVES
the cylinder isn’t uncovered. The reservoir brake line is bad enough in that it loses fluid A number of brake proportioning valves
should also provide enough capacity not to and pressure. But at the same time, air leak- have been designed in an attempt to
run dry when the pads are completely worn ing back into the system has the same ef- provide a better, or variable, brake
and the pistons are at full travel. Reservoirs fect as boiling the fluid, and will probably balance between front and rear wheels.
must be vented to the atmosphere to allow be noticed much sooner than fluid loss dur- Most common is the pressure relief valve
the fluid to drain. But free access allows the ing a race, if the leak is very small. for the rear brakes on cars with a tandem
fluid to slosh out, and allows destructive master cylinder. There is no way to vary line
moisture to enter. The best solution is a PADS : pressure between the front or rear brakes
rubber diaphragm or bellows to seal the vent Developing the friction material for with such a cylinder. If the rear is slightly
and yet allow atmospheric pressure into the brake pads is a black art beyond the racer’s over-braked and the next smaller calipers
reservoir. capability to modify. Selection is simply a are too small, then a valve must be provid-
All brake plumbing must use the highest matter of buying and trying until the best ed to limit the brake pressure to the rear.
grade steel tubing, preferably with aircraft is found. Of course, it is important to keep This is simply a spring-loaded piston install-
nuts and fittings. Flexible hoses should be all oil off the pads because once it soaks in, ed in the line, with the spring load ad-
of the braided-stainless-wire and teflon type the pad is ruined. Most racing pads are very justable for the desired pressure limit. There
to prevent pressure expansion from using up stiff, to prevent deflection or warping under are two disadvantages to such a valve. First,

47
 advent of ground-effects tunnels. Now it
appears that brakes must often be left out
in the wheels if only to allow cleaner airflow
in the tunnels.)

DEVELOPMENT
PROPORTIONING
Since most race car brake systems are
already optimally developed to handle any
forces that current cars and tires can
generate, the only areas of potential develop-
ment for the racer are in proportioning and
cooling.
Proportioning is the first consideration,
since it has such a great effect on stability
and on the generation of heat between front
and rear brakes. The idea is to have the
distribution of forces such that the rear
brakes will never lock first even under the
most severe possible racing conditions. For-
tunately, those conditions are also the best
racing conditions, or in other words, max-
imum braking on the highest coefficient
surface — and with minimum fuel load if
An early production-type brake proportioning valve (lower right corner). Rear the fuel tank is toward the rear. Because the
brake force can be limited by screwing out the plunger on its right side. previously mentioned aerodynamic down-
force bias to the rear is greatest at high
speeds, it is also wise to test the brake pro-
the rear braking force is non-linear — is seldom enough cooling airflow when the portioning at relatively slow speeds — say
although that characteristic might be put to brakes are mounted deep inside. It is also under 80 miles per hour.
good use if carefully thought out. And difficult to find adequate space for suspen- With the above fixed conditions — typical
secondly, the valve uses up some valuable sion or steering travel with the brakes in the coefficient road surface, minimum fuel
fluid displacement or pedal travel. way. Also, unsprung weight is reduced by load, and minimum aerodynamic downforce
Another alternative, particularly for cars having the brakes inboard of the suspension — the test procedure is as follows. First the
in which the rear load changes drastically and the stresses in suspension components brakes and tires are warmed up to typical
due to fuel consumption, is a proportioning are reduced by not having to resist brake racing temperatures by practicing pedal
valve in which the limiting spring pressure torque. When double universal-jointed axle modulation with many high-deceleration
is controlled by axle deflection. The less the shafts are already present, as with an stops. Wheel lockup can be determined by
axle deflects because of reduced loading, the independently-sprung rear end, it seems to using electronic instrumentation, as dis-
less the pressure spring force and the less be an obvious advantage to mount the brakes cussed later, or by an outside observer to
the brake line pressure at the rear brakes. on the differential housing. Even at the watch the tires. Sometimes it isn’t easy for
Conversely, the greater the downforce at the front, it may be desirable to add the weight the driver to detect which tires are locking
rear — say from aerodynamic loading — the and complexity of the shafts and universal first without a long, tire-destroying skid.
greater the line pressure and braking force joints just for the inboard brakes alone. At The driver accelerates to 80 or 100 miles
available at the rear. Although this sort of the front the only drawback is the added per hour, shifts into top gear, and then
valve is available on some passenger weight, but at the rear there are some gradually applies the brakes harder and
vehicles and trucks, Mac Tilton has de- problems. First, the brakes, the differential, harder until he can feel or hear a lockup —
signed one which can be easily adjusted by the transmission, and the engine or exhaust hopefully right in front of the observer —
the driver during a race. pipes all contribute heat to each other. Their and then he immediately backs off. At this
Finally, on cars with front discs and rear closeness can make cooling as difficult point, all that has been found is which end
drums, some sort of residual pressure valve inboard as it is out in the wheels. Second, has the greater braking force. It is still
may be installed to hold pressure in the rear oil could get on the brakes from the engine necessary to find the force distribution at
lines. Otherwise the rear brake return or differential. And finally, there is the which the other two wheels lock up.
springs would apply a reverse force in the possibility of axle failure causing loss of the On cars with parallel master cylinders, it
hydraulic lines, which could cause excessive brake force at any one wheel. At the front, is simply a matter of adjusting the balance
drag on the front brake pads (which have axle torque acts only in one direction: bar between the two pistons and the pedal
no return springs), and would require extra braking. But at the drive axles, there is a arm. Moving the pedal arm pivot closer to
brake travel on the next application. complete and severe stress reversal with the front or rear system cylinder increases
every change from braking to acceleration, the force distribution toward that end of the
INBOARD BRAKES
and axles have been known to snap. The car. On cars with a single tandem master
There may be advantages in having the consequences of losing one brake can be as cylinder, the adjustment is a little harder.
entire brake system mounted on the chassis, critical as the loss of two because of the In the first place, if there is a pressure relief
with flexible axle shafts out to the wheels. effect on braking stability. valve in the rear brake system, it should be
Modern racing wheels are so wide that there (This was originally written before the plumbed out, to learn just how far off the

48
basic proportioning is. Then, whichever end
of the car locks up first is ballasted with
known weights, and retested until the
opposite end begins to lock up first. The
proportion of ballast weight compared to the
dynamic load on that end of the car is the
approximate percentage of overbraking at
that end. Say the front wheels lock first and
it takes 200 pounds of ballast directly over
the front wheels to make the rear wheels
lock first. If the static front weight is 1500
pounds and there is about 500 pounds of
calculated load transfer to the front, then the
front brakes have approximately
(200)/(1500+500), or 10 percent too much
braking force. At this extreme, or if the
balance bar on parallel master cylinders is
too far off-center, it will be necessary to cor-
rect the brake balance by some major com-
ponent changes.
There are many possible options for cor-
recting a misproportion in braking, such as
changing weight distribution, tire co-
efficients, disc diameter, center of gravity
height, or wheelbase. But the only practical
means of modification on a given race car
An air scoop and duct leading to the eye of an inboard rear disc, which
are tire diameter, caliper piston area, or, in
the case of parallel master cylinders, a varia- would otherwise be shielded from any cooling airflow.
tion in piston diameters between front and
rear systems. Even changing tire diameter or suspension components; radiation to on the front of a front-engine sedan. This
can have an effect on other vehicle nearby components; and convection to the means that the front brakes of a sedan can
characteristics and a variation in master moving airstream. Only the last is either generate about twice as much heat as the
cylinders is also complicated. So the most very effective or very easily improved. rears and are usually the limiting factor. For-
professional solution is to change the brake There have been attempts to water-cool tunately, they are closer to the frontal air
caliper piston area. brakes with cooling passages plumbed into flow and are easier to get cooling air to.
In the example of the front being over- the shoes or pads, but such systems have For a measure of the problem in degrees,
braked by 10 percent, this would require a been too heavy or inefficient compared to it is useful to record brake temperatures
corresponding reduction in front piston the relatively free use of natural airflow. during the most severe applications and
area, or an increase in rear piston area, Occasionally someone will add a water therefore be able to measure the effec-
depending on whether the over-all pedal spray device to lower the temperature of tiveness of cooling improvements. Even
force was desired to be greater or less. cooling air in the ducts. However, this is without complex electronic instrumentation
When the front/rear brake ratio has been merely a temporary fix for a poorly and thermocouples, it is still possible to get
developed as closely as possible by major developed system. peak temperatures by applying thermo-
component changes, then it can be fine- Before any effort is made to improve changing materials. Welding-supply shops
tuned with other devices such as the balance cooling efficiency, it is a good idea to know carry wax sticks and paints which melt or
bar or rear pressure relief valve. Since the whether there is a problem. For open- discolor at certain specific temperatures. A
relief valve doesn’t have the correct action wheeled race cars there may be adequate air more accurate — and expensive — method
for early lockup at the front, it is important flow. If there is no fade and the brake pedal is stick-on indicators called Temp-Plate
that tandem master cylinder systems have a remains firm throughout a race, then there which are available in various increments
basic over-bias to the rear before fine-tuning. is nothing to be gained from adding brake between 100 and 1100 degrees. If they are
The balance should then be altered back and scoops or ducts. But on a heavy race car attached to a brake caliper or piston and the
forth between front and rear, until it is ever with full fenders and wide wheels, durability dots blacken up past 500 degrees, then big
so slightly greater on the front on the highest is sure to be marginal without some directed trouble can be expected.
traction surface. Still, it should always be airflow. At best, high temperatures will Cooling air is most effective on the disc,
remembered that any change in tires or track probably cause rapid pad wear. In such a since that has the highest temperature dif-
coefficient will probably upset the balance case, it is hard to get too much brake ferential and the effective airstream velocity
again. cooling, even though the air scoops and is greater than at the caliper. It is not enough
ducts could increase air drag. When it rains, just to have a lot of air — it should be
COOLING of course, it is possible for the brakes to run distributed evenly to both sides of the disc.
The brake discs and calipers have a great too cool and ducts should be closed off. A disc which consistently runs hundreds of
deal of heat to get rid of in a very short time The energy input to the brakes is largely degrees cooler on one side than the other
between brake applications. The three proportional to the load on the tires during will have destructive thermal stresses if not
means of doing so are: conduction through deceleration. That can be from about 50:50 permanent warpage. The situation is poor,
the mounting surfaces to the hubs, wheels, for a rear-engined race car, to 65 percent considering the fact that the natural airflow

49
inside a wheel is usually in one direction, gauge should be used to measure the runout micro-bubbles to consolidate. When no
depending on the shape of the wheel spokes at the friction face. If it is over .010 inch, foam or bubbles can be forced out of the
and the air pressure under the chassis. It is then it will be necessary to rearrange or bleed hose, it will be practically impossi-
possible to increase the natural airflow, and grind components until it is the absolute ble to tell whether there is flow at all, except
in the proper direction to reduce pressure minimum. It is sometimes possible to for the rising fluid level in the jar. At that
under the body, by designing a centrifugal eliminate runout by rotating parts with point, all pistons should be blocked in the
blower as part of the wheel spokes or as a respect to each other until production fully retracted position. It is very important
hubcap. This has a pumping action which tolerances cancel each other out. Play in the that bleeding be slow and careful and steady,
is proportional to velocity but which does wheel bearings also appears as runout, so with no let up on the pressure until every
cost some power in the process. At worst they must be adjusted with some slight last air bubble pops out through the bleed
it might amount to a few horsepower at top preload to insure zero play under extreme screw and it is closed. Otherwise, air may
speed and yet it does improve the hub heat and stresses. Live rear axles are be drawn back into the lines or calipers.
aerodynamic downforce. especially prone to end-float which can push With the pistons blocked in their fully-
When air is ducted to a solid disc, the best the disc and pistons around. It may be retracted position, the caliper is fitted over
solution is to run it into a fabricated can necessary to open the differential and use its disc. As a last measure it can’t hurt to
which directs it equally to both faces. For spacers under the axle keepers to pre-load check the system by very careful pressure
ventilated discs, with a natural centrifugal the axles in position. bleeding, if the pressure bleeder is proven
airflow due to the internal vanes, the best Mounting of calipers is also critical, since to have clean moisture-free fluid.
duct location is to the center or eye of the the braking force at that radius is more than When one end of the car is finished, it
disc. The best source of cooling air is any twice the braking force at the tire patch. might be a very educational experience for
positive pressure area at the nose of the car. Mounting brackets can’t be too stiff, and all the driver and/or mechanic to see how the
On sedans this is often the former location mounting bolts must be of the highest grade, car stops with only one brake system
of headlight lenses, and on sports cars with carefully torqued, and safety wired. The working — with rear-only being the worst
headlights the highest pressure area is usu- caliper should be shimmed until the disc is possible condition. In the first place, it will
ally the under-nose spoiler lip. The greatest centered and parallel. The caliper may be show whether there is enough pedal travel
difficulty is in getting large duct hoses — toed-out a few thousandths, to allow for for that system to function at all with no
the bigger the better — from the nose to the torque deflection in the direction of rotation fluid in the other system. But even with full
disc, with all the suspension hardware in the during braking. For proper pad wear it is force available at the rear, it can be a
way. Although smooth fabricated aluminium important that the pads be perfectly aligned shocking discovery to learn how much
or fiberglass ducts are the most efficient, with the caliper pistons, although they can worse and how much more unstable the
wire-reinforced flexible hose (such as 4-inch be shifted in or out radially, or rotated car’s stopping performance is. It makes
dryer duct) passes a lot of air and yet will slightly with respect to the disc without any perfection worthwhile.
give way if the wheel contacts it. Front harm. It can’t hurt to file and polish the With all the previously discussed prepara-
spoilers prevent a lot of useful air from edges of the pad backing plates and the stops tion and precautions, the actual racing per-
reaching the front discs, but as a last resort, on the caliper where they contact, to insure formance is mostly a matter of treating the
what air does pass under the nose can be smooth sliding movements under high brakes with some respect. First, as was
directed up toward the discs by shields forces. Before filling the system with fluid, explained, new pads or discs must be
fastened to the suspension arms. When all it is a good idea to check piston and pedal broken-in carefully before they can be
else fails, it is time to find heavier discs and travel at the master cylinder — and to allow pushed to their limit. This is especially true
calipers. some extra pedal travel to make up for stress of spare pads if replacement is necessary
There are a few other areas in which deflection. during a race. One recommended procedure
braking performance can be improved, such Brake system bleeding can be one of the is to slowly run the pads up just to their
as reduced component weights, increased messiest and most difficult jobs to do maximum operating temperature, and then
aerodynamic drag during braking (when properly. Especially with four-wheel disc let them cool slowly. But check with the
allowed by the rules), and perhaps, anti-skid brakes, balancing valves, and a tandem racing representative for each particular
systems. But the payoff in reduced laptimes master cylinder. The calipers may be such make. For long-distance races, it is im-
is hardly worth the cost or effort. It is far a maze of pistons and passages that the only portant to have good records as to how many
more practical to try and increase tire co- positive way to get every last air bubble out, miles a set of pads will last under the most
efficients and aerodynamic downforce so is to bleed them before mounting. If the extreme conditions. It also helps to know
that the existing brakes will have more trac- caliper has four pistons, it helps to have how far this range can be extended, if
tion to work with. blocks of wood cut the right thickness to necessary, by reduced braking effort. The
hold the pistons in the caliper, in either full wear is non-linear with temperature (See
ASSEMBLY AND RACING compression or full extension. The caliper Fig. 23), which means that backing off just
Since the brake system is so critical to the is first filled by pumping fluid down through a little can save a lot. Or conversely, a small
life of a race car and driver, its assembly the master cylinder and out a bleed screw increase in demands on the brakes may
and maintenance should be handled with hose into a glass jar of clean, fresh, shorten pad life a great deal — not to
more care and intelligence than any other moisture-free brake fluid. Using the glass mention the effect that extreme temperatures
area. Mounting discs is more than a simple jar makes it possible to see when air bubbles can have on caliper seals.
bolt-up job, as any wobble or runout will are no longer being expelled. Having the Reading the wear profile on a set of well-
cause the caliper pistons to be knocked back caliper off the car makes it easy to rotate used brake pads can be as informative as
at high speed, using up pedal travel. The it into every possible position of “air-up”’ reading tire wear, if something in the brake
disc, flanges, and hub mounting faces in even the most torturous passages. When system is out of alignment. To get maximum
should be clean, free of burrs, and perfectly the air bubbles diminish, a little rest or life, ideally the pads would wear down
parallel. When the disc is bolted up, a dial tapping with a mallet allows the remaining perfectly parallel with the backing plates.

50
However, if there is taper-wear in any direc-
2500
tion, or rocking-wear, then something is not
positioned quite right or is flexing. Taper 90-60 mph
wear may be reduced by shimming one end
of the caliper to counteract stress deflections 2000
in the mounting brackets or by moving the
pads with respect to the caliper pistons in
the direction of maximum wear. This
1500
repositions the application force to where
it is more needed. Relocation is relatively
130-90 mph
easy, by grinding away locating faces or
redrilling the mounting pin holes. Rocking- 1000
wear is more apt to be caused by flex in the
pad backing plate and is hard to cure number
of
pad
snubs
per
without the manufacturer’s help.
If pads must be replaced during a race, 500
teams spend a lot of time and effort to
reduce the time required. The biggest
problems are heat, retainers, and piston
displacement. The easiest brake job can be 0 400 600 800 1000
practically impossible when the components temperature - degrees F.
are between 500 and 1000 degrees. Asbestos
gloves can really be cumbersome. The re- Figure 23. The importance of cooling air is shown by representative brake pad
tainers have to be fail-safe in operation, and life as a function of temperature for two materials under two stopping speeds...
yet quick-release when necessary. And the
pistons have to be displaced back to their
original position before new pads will slip is to install a proportioning valve next to the characteristics in action. With race car
in. One possible solution is a vacuum source driver. In the first place, it is more accessi- operating temperatures of up to 1700
connected to the master cylinder vent, which ble if it is necessary to quickly change the degrees, they tend to glow, literally bright-
the driver can switch on to retract all pistons brake balance for wet or oily track condi- red hot. This creates some concern among
at once. Of course, it must be precisely tions. But it is also possible for the driver spectators and other drivers, and in fact can
regulated so that air is not sucked into the to change the balance himself while driving be a problem if not allowed for in system
system past piston seals. — to account for fuel consumption or rain design, which will be discussed later.
When the brakes do fail, it can be either — if some sort of simple lever is provided.
slowly and with some warning — or with However, the correct lever position for cor- The History and Theory
a bang when a line fails or a disc cracks rect balance under any condition must be In the seventies, the Hitco division of
apart. Small, beginning disc cracks can well-known in advance, because a race is Armco Steel began fabricating carbon/
sometimes be felt as a definite vibration in no place to experiment. carbon brake discs for military aircraft,
the brake pedal, and are a strong warning where every pound saved is worth its weight
that the disc may be ready to go. At best, CARBON BRAKES in graphite. While the basic properties of
any crack will probably be shaving material At this writing, sintered carbon discs and carbon canbe obtained from any engineer-
off the brake pads at a rapid rate. If the pads were just starting to become common ing text, the exact method of fabricating
failure is slower, as with fade or fluid in Formula One, starting with Brabham in carbon/carbon composites is usually pro-
leakage, disaster may be delayed a little by 1983. Although such discs and pads are off- prietary. The United States military estab-
pedal pumping. The pumping can build up the-shelf items that cannot be easily lishment once classified carbon/carbon as
pedal height by bleeding some of the air up modified, they have valuable and peculiar astrategic material and restricted process-
through the master cylinder reservoir or by characteristics that are important to ing knowledge to American citizens. How-
making up some pedal travel with more understand. ever, it is possible to reveal the commonly-
fluid per bite. Still, it’s not conducive to con- Oddly enough, carbon/carbon (so called known basics. First, a disc is layed-up
fidence or low laptimes for the driver to be because it is carbon filaments sintered in a from woven layers of graphite fabric--
pumping with the left foot while the right carbon matrix) discs are not blatantly dif- which may actually be woven ina circular
foot is hard on the accelerator down a long ferent in appearance from conventional shape, with fibers oriented radially and
straightaway with no escape road at the end. metal discs. If it were not for the fact that circumferentially. Then the layers are
At the end of a race or even after a hot carbon/carbon rotors are currently only bonded tightly together with a phenolic
practice session, the cool-off lap should be available as solid discs, instead of the com- resin in an autoclave. Next, the composite
used for cooling the brakes the same as mon vented iron rotors, they might pass un- is placed in a furnace and carefully fired at
anything else. If the car is stopped with noticed. They look like metal, and they feel about 4,000 degrees, which turns it into a
blazing hot discs, the exposed portion cools like metal; it is only when you lift one that solid block of carbon. Finally, the rotor is
much faster than the part covered by the the most valuable attribute becomes finish-machined, using diamond tooling.
pads. This can cause heat-checking or obvious. With a density of only 1/5 that of The properties of this finished material
cracking of the disc — not to mention the steel, they feel like they are made of anti- are extremely interesting to a brake
effect it has on the pads and calipers. gravity unobtanium. (If you don’t keep them engineer. But to be able to relate the
A final trick, which’is occasionally used bolted down, they tend to float away.) properties to design, it is necessary to
on otherwise well-developed brake systems, Another giveaway is their telltale color review the fundamentals of braking. The

51
 3700
keep in mind that this is heat storage per
pound, and not per volume. In brake rotors
of equal dimensions, the carbon/carbon
rotor would weigh 1/5 as much as iron, and
therefore have only 1/5 X 150 percent or 30
percent of the heat storage. However, for a
given storage requirement, the carbon/
carbon rotor can be 2/3 the weight.
A comparison of “thermal conductivity”
is hard to evaluate without serious analysis,
because of conflicting requirements and the
extreme differences. Conductivity is ex-
pressed in heat per volume per second per
degree, but the concept of heat transfer
through the material is easy to grasp.
Carbon/carbon, with a conductivity only 1/3
that of iron, is a very good insulator. This
means that over the short term, less heat will
be transferred into mounting attachments.
However, this also means that it can take
longer for the material to heat-saturate to its
core, and longer for deep-stored heat to be
dissipated off the surface. In the case of
vented rotors, this could also be a problem
in transferring heat from the friction face
into the internal air gap. While aluminum
and copper are good in this respect, iron
rotors can have a temperature difference of
a couple hundred degrees between the rotor
face and the internal vanes. This can create
Relative Heat Relative Relative Coeff. of Relative Melt serious stress cracking over a large number
Storage Conductivity Thermal Expans. Temperature
(calories per square cm.
of braking cycles.
(calories per gram (micro-cm. per cm. (degrees C.)
per degree C.) per cm. per second per per degree C.) Melting temperature is not a great con-
degree C.) cern, since other structural or functional
problems will normally occur first. But it
Figure 24. A comparison of engineering properties shows the extreme theoretical is interesting to note that while aluminum
advantages of carbon/carbon rotors in race car brakes. appears marginal at typical race car braking
temperatures, carbon/carbon is essentially
indestructable. (And at its limit, it actually
vaporizes rather than melts.)
idea is to mechanically convert the race cars’ Materials Comparison The coefficient of thermal expansion is
kinetic energy into heat, and then dissipate Now, let’s consider how the basic proper- usually only important to engine builders,
that heat from the hardware. Over the very ties of different materials compare in terms but it can also be a serious problem for
short term — during braking — most of the of braking performance. Although discs brake designers. While the coefficient for
heat is stored in the brake materials: the disc have occasionally been made of copper, iron is small, the temperatures are high
and pads. During braking, and especially in beryllium, and cast or metal-sprayed com- enough to create large stresses in parts with
the intervals between brake stops, the binations, the most common materials are nonuniform temperatures. This is what
greatest percentage of the heat is dissipated iron and aluminum, which are compared to creates heat checking, thermal cracks, and
into the cooling airstream by a process of carbon/carbon in Figure 24. Structural warpage in conventional iron rotors.
“convection.” A smaller amount is properties such as density, strength, and Carbon/carbon has essentially no thermal
dissipated by “radiation,” or the transfer of strength-to-weight ratio need not be dis- expansion compared to iron or aluminum,
heat (in terms of light), to nearby surfaces. cussed because they are not a major factor meaning that this is a design problem which
This is undesirable simply because those in brake rotor design. can be ignored, countering the poor con-
nearby surfaces — such as tires or transmis- The first and most interesting thermal ductivity characteristics.
sion — often have heat problems of their comparison is “specific heat,” or the abil- Brake life or durability can be considered
own. Finally, a small amount of heat is ity of materials to absorb and store heat. from two standpoints, the life of the fric-
transferred by “conduction” through This is important because there is no way tion material throughout a race, and the life
contacting hardware into the caliper that the rotors can dissipate all the heat from of the rotors in terms of the number of races
mounting brackets and wheel. Here again, one brake stop into the air in the few before they check, crack, or score and must
this must be minimized due to its effect on seconds spent braking. Specific heat is ex- be replaced. There is no textbook answer
hub bearing life and tire temperatures. In pressed in energy per weight per degree of to this question; it has to be determined in
fact, it is common to thermally isolate the temperature rise. While carbon/carbon has the field. An early Brabham test report in-
rotor from the hub using air gaps and 50 percent more heat storage capacity than dicated that carbon/carbon brake life was
thermal insulating pastes. iron (but less than aluminum), you should seven times that of conventional brake

52
systems. However, it was not obvious
whether that referred to pad life or rotor life.
When I first heard about carbon (or
graphite, see Ch. 10) as a brake material,
my immediate thought was, “Hey, isn’t that
the stuff they use as a lubricant in engine
oils? How can it have a good coefficient of
friction in a brake pad?”
As it turns out, a‘direct answer to my
question was not available. In the first place,
there are a large number of combinations
of materials (even between pads and rotors)
and the coefficient of friction will vary with
each. Secondly, the coefficient of friction
varies considerably with temperature and is
hard to express as one given number. Fin-
ally, no one has actually tested for the co-
efficient of friction on real race car brakes.
Goodyear (as designers/marketers of aero-
space brakes) merely report that the coeffi-
cient of friction is about the same as cast
iron (0.3 to 04), dropping off slightly at This massive carbon/carbon brake rotor has radially-drilled cooling passages
higher temperatures. But to them, higher and thermo-sensitive paint stripes. The hub is slotted to allow for expansion.
operating temperatures means 4000 degrees
after an aborted takeoff. mass-produced automobiles. While there 200-400 degrees hotter, better insulation
Nelson Piquet was once quoted as saying, are definite advantages, the transfer won’t and caliper ventilation is mandatory. Rub-
“Less pedal pressure is required,” and be easy. The requirements for aircraft are ber piston seals, in the calipers and hub or
“They have better feel.” However, this can somewhat different, usually consisting of transimission bearings may also suffer from
mean two things. Either the coefficient of one-shot stops with a long heat-soak and the increased heat, especially with all the
friction is actually better than asbestos on cool off period during which there is no aerodynamic enclosures necessary for
iron, or they were able to use greater high speed air cooling. In this case, banks proper ground effects.
mechanical leverage. The upper limit to of solid discs and plates are the most A final practical consideration to car-
mechanical or hydraulic leverage in a brake appropriate solution. (Aircraft tires have bon brake components is their cost, which
system is in getting enough travel to take up been known to blow out from the trans- may be many times that of cast iron. Some
the air gap and any force deflection in the ferred heat, many minutes after a particu- sanctioning bodies have restricted their use
disc, pads, calipers, etc. If the carbon/ larly hard stop.) from this standpoint, even though they
carbon components have less force deflec- On race cars, however, efficient air often turn out to be cheaper in the long run,
tion or less disc warpage due to thermal circulation has proven to be the best solu- ‘dueto a longer life. Also,when you consider
deflection, then a greater mechanical ratio tion. For the same mass and material, a the weight savings, which might be about
is possible, requiring less pedal force and solid disc rotor may reach a peak steady- 30 pounds, times about a 1.5 multiplication
giving a better feel. Without a known co- state temperature over 300 degrees hotter factor for rotational inertia, that works out
efficient of friction it is impossible to design than a vented rotor. To improve on this to be a real bargain in dollars per pound.
a brake system on paper, but if we can
performance, Mac Tilton once marketed a
believe Piquet, there is no problem here that cast iron rotor with cones (or matching ANTI-LOCK BRAKING
cannot be worked out in development.
stalactites/stalagmites), separating the Although anti-lock braking systems
There are two other factors which are im- discs, for even better air circulation. This
portant in brake heat dissipation: the co- (ABS) are currently making a big impact
sort of design freedon is difficult, though in new production cars, development has
efficient of emissivity (which predicts the
not impossible, in carbon/carbon fabrica- been going on for a very long time. Large
amount of heat radiation) and convection,
tion technology. So it may be some time passenger and military jets demanded such
which are simply functions of exposed area,
before we see the ultimate in carbon/car- systems in the mid-fifties, probably starting
temperature difference between the hard- bon brakes, though they certainly have
ware and the air, and air velocity. However with the B47 in 1947. Then Detroit came out
more potential than other metals. with some half-hearted unrefined systems in
important these are to cooling, they do not
An occasional mention has been made 1969-1971. First they were made available on
vary with the type of materials used. But
that, ‘“some alterations in the braking the rear only, (SAE Paper No. 690213) and
one important point to note, is that thermal
system may be necessary if conditions then some came out on all four wheels (SAE
radiation (though a small part of total
dissipation) increases as the fourth power
dictate.’ What this means (in addition to Paper No. 710248) — but they were soon
a force ratio change) is that the hotter abandoned for cost, legal, and technical
of temperature, or will double with about
operating temperatures can create thermal reasons. It seems that all of them were great
a 300 degree rise in peak temperature.
problems in other components. Most ob- in a straight line on a constant coefficient
Practical Aspects vious is the brake fluid. The best glycol- surface, but road testers began discovering
Many American and European com- based racing brake fluids will boil at about that braking in a turn or braking on a “‘split
panies are attempting to transfer the 550 degrees. There is perhaps an inch of coefficient’? created more problems than
technology of carbon/carbon brakes from material in pad and piston separating that ABS could solve. In addition, there were
military aircraft to race cars, and then to fluid from the rotor. With a rotor running limitations in the basic analog circuits of the

53
 of Fl computer-controlled injections. At the
very least, the system should revert to
normal manual operation in case of failure.
Those are the relatively simple,
straightforward obvious development
N problems. The transient dynamices of a race
— — WHEEL SPEED \
car are a whole different picture. Even in
production cars, it has not been resolved
mph-—=
what the optimum split or balance between
wheels is. The first systems simply had a
single driveshaft sensor with a single
modulator for both rear wheels. Others now
use a diagonal split to modulate left front
with right rear and vice-versa. And the most
elaborate, of course, is four separate sensors
ABS VALVE RESPONSE TO WHEEL DECELERATIONS with four separate modulators — though it
is not necessarily the best. (SAE Paper No.
brake pressure
——t
cylinder
830481)
Again, even in race cars, straight line
braking is a simple problem for anti-lock
0.1 to 0.5 sec. — or for skilled drivers. But in a turn, they
both have difficulties. For example, consider
Figure 25. Typical computer-controlled ABS force cycling at the verge of lockup.
the ultimate 4-wheel anti-lock system in an
There are times in racing when this may not be appropriate.
extreme case — a narrow car with a high
center of gravity which transfers almost all
period, plus reliability problems both in following inevitable development problems. of its weight to the outside tires in a turn.
electronics and the severely abused wheel First, the computer logic of these systems Braking into that turn, all four wheels are
speed sensors. By 1979, however, digital is designed for production tire retarding the vehicle at their own maximum
microprocessors and the next generation of characteristics, wheel/tire inertias and rate, based on their individual loads. But the
sensor hardware had reached an acceptable deceleration rates. Race car characteristics outside tires are therefore providing almost
level of sophistication: four separate wheel can be different by a factor of two in each all the drag. Meanwhile the forward inertia
sensors, choice of modulation distribution case, requiring a major rewrite of the pro- is still at the center of gravity on the
(two front with one rear, two front with two gram. This could be a major undertaking, centerline of the vehicle. That can account
rear, diagonal, etc.), allowance for brake- but if done by the factory, a field modifica- for a significant torque about the center of
in-turn, variable cycling modes, and self- tion might be as simple as replacing a single gravity which is trying to straighten the car
diagnosis (SAE Paper No. 790455). computer chip. These controller decisions, out — or understeer it off the outside of the
The first attempt to use ABS in a race car which determine the duty cycle or stepping turn. This can be great for straight line
was in 1972, when Mark Donohue tested a logic to “pump” the brakes at a high rate, stability — but that may not be where you
system for Porsche. Even though he was can create other problems. If suspension want to go.
enthusiastic about its potential, they felt frequency is in resonance with modulation The existing manual system is inherently
(wisely, it turns out) that it wasn’t suf- cycling, destructive vibrations or wheel hop better in this respect. When the driver
ficiently fail-safe for racing yet. Now, could reduce braking effectiveness — if not perceives inside front lockup, he feathers the
however, with a number of improved pro- damage the suspension. (See Fig. 25) brake — and at that point the retarding force
duction systems available, and a few racers There may also be difficulty in finding is equal left and right, although perhaps less
experimenting with them (especially in enough power to modulate the brakes than the ultimate since the loaded outside
rallying and showroom stock), some men- properly. Both vacuum and _ hydraulic tire is operating well below its capacity. The
tion should be made of possible racing systems are used in production, but since driver feels no yaw understeer. Granted,
applications. In a race car, with a skilled little engine vacuum may be available in a natural forward weight transfer has the op-
driver, in a straight line, on dry pavement, race car, only hydraulics should be con- posite effect in lightening the rear, which
the potential of ABS is limited. On the other sidered. Also, racing brake forces and generates braking oversteer in the turn. But
hand, braking deep into a turn, where the cycling frequencies often tax the brake fluid this is usually corrected by front/rear
limit is inside front lockup, or on a wet capacity, which cannot be ignored as a braking balance systems.
track, or in off-road or rallying, it seems potential problem. So what is the perfect 4-wheel brake force
that there could be some gains available. Reliability is a major concern — just as distribution? It’s a problem for a computer
“Could be” is to be emphasized because of in production cars — though the environ- analysis. An onboard computer should be
the three stage nature of innovation: ““Who ment can be much worse in some respects. able to handle it — given a constant update
needs it” to “Hey, that could be a winner”’ The wheel speed sensors will be subjected on what exactly the driver intended —
to “Uh oh, there are a few bugs in the idea to far greater heat and vibrations, not to straight or turn. And on top of everything
after all”. mention abuse from wheel changes and else, a skilled driver may have good reason
The core of these systems (sensors, logic, frequent teardowns. And the race car engine to intentionally lock up the brakes, so there
and controllers) is too complex for anyone has far more powerful RFI (radio frequency should be a pressure override. With that list
but the factories to do original designs. interference) to confuse the logic unit, with of potential problems, it may be a long time
However, some production systems might little or no steel body shielding — although before there is much more improvement in
be adaptable to race cars, if we consider the this has been dealt with already in the case existing manual brake systems on race cars.

54
 6
AERODYNAMICS

sidered, particularly open cars and open- much faster rate than lift (or downforce)
THEORY wheeled cars, the total area touched by the does. On race cars with wings or with large
Nothing has had as great an impact on airstream has an effect. That includes tires, downforces in general, induced drag can be
race car design in the last few decades as wheel wells, cockpit, radiator ducts, and considerable.
aerodynamics. And it seems that it has been even the smallest odd obstructions. As with Since profile drag is the most immediate
the subject of more technical articles than any other moving object, there are six com- problem, the basic over-all shape should be
any other area besides engines. But because ponents of aerodynamic forces or torques considered first. Generally speaking, bet-
you can’t see or touch aerodynamics, very on a race car. But since lateral force and roll ter results can be expected from streamlin-
‘few automotive people have a firm inputs are insignificant on a race car, only ing applied to the rear of blunt objects. This
understanding of the subject. Perhaps that’s drag, downforce and its front/rear distribu- is apparent from the realization that it is
why so many people write about it. It is very tion, and lateral stability will be considered much easier to push air than it is to pull it.
difficult to prove someone’s theories wrong here. That is, air which is trying to compress at
— until a faster race car comes along with the front of a body tends to follow smooth
a different approach. The basic principles DRAG streamlines, while at the rear, airflow.
of air flow can be picked up from a fluid Air drag is broken down into profile or separates from the surface and becomes ran-
mechanics textbook, which will give an shape drag which is a function of frontal dom, without a solid object to push it back
engineering explanation of such terms as area; skin friction drag which is a function into place. When air can no longer follow
viscosity, turbulent flow, laminar flow, of surface roughness; and induced drag due the body, it breaks away in a wave which
boundary layer, density, and so on. to lift. All of them have a varying relative curls under and rolls like an ocean wave,
However, for this chapter to make sense all importance depending on the speed and creating a series of alternating high-drag
you need is a basic awareness of the flow of shapes considered. With the shape that race vortices. Therefore, disregarding all other
air and the fact that air forces generally cars are in today, profile drag is by far the effects, it would seem that the maximum
increase with the square of velocity. In most significant. When profiles start im- body volume should be moved well to the
other words, when speed is doubled, the proving, then skin friction drag may become front to permit the most gradual con-
forces increase by four times. For a better important enough to consider, primarily in vergence angle at the rear of the body.
background, the best hardcover references controlling turbulence in the layer next to Various theories and experts suggest that the
are Fluid-Dynamic Drag by Hoerner, and the surface (the boundary layer). Induced convergence of the body should never ex-
Race Car Aerodynamics by Katz. drag depends on lift, and it increases at a ceed 10 to 15 degrees to avoid separation of
A great deal is known about aircraft
:
aerodynamics, from all the government- |
financed work by NACA, and a lot of it is
applicable to race cars. However, there is
a major difference in free air at aircraft
heights and air in the earth’s turbulent boun-
dary layer where natural obstructions cause
uneven airflow. Another problem unique to
race car aerodynamics is the ground effect,
which causes objects to have radically dif-
ferent airflow characteristics as they
approach the ground. Therefore, most air-
craft data, with the exception of some airfoil
tests run close to the ground, must be used
with caution.
To assist in understanding airflow, a few
general comments should be made. First,
at current racing velocities, the air is seldom
compressed to any great degree, and ac-
tually flows very little. What happens is that
the moving body displaces each air parti-
cle perpendicular to its surface, and in the
absence of skin friction the particle would
return to its approximate original position
after the car had passed. Second, when the Development of the first ‘‘winged’’ race car in 1966, using tufts, an electronic speed
aerodynamic profile of a race car is con- recorder, and telemetry from the track to engineering offices in Detroit.

55
 Shape Drag Coefficient prevent loss of pressure.
Since almost all braking power is
Flat plate dissipated in heat to the air, and mostly
Indy car while the car is accelerating between brake
Typical American sedan applications, a smooth passage of air both
Corvette to and from the brakes ought to be as im-
Sphere portant as radiator ducting. More care in the
Land speed record car weightless air-cooling of brakes permits
Typical airfoil lighter discs and calipers and should reduce
wear.
The engine and driveline radiate a great
Table 2. Examples of air drag coefficients deal of heat from their walls — especially
from the exhaust pipes. An engine compart-
ment can’t be totally enclosed without rais-
the flow. However, this figure depends on quarter of the tire, the airflow is trapped bet- ing radiator requirements. So it’s a good
the condition of the boundary layer and flow ween the tire and the ground. This, plus the idea to have some circulation around these
around and under these relatively narrow vortex created by the spinning wheel and components regardless of the cost in air
shapes, so it can be found with any accuracy brake assembly, can upset the airflow all drag, or seals, gaskets, and wiring can suf-
only experimentally. along the side of a car and can be the fer heat damage.
This airflow breakaway at the rear of a greatest contribution to drag. All else be- Engine induction requires the coolest,
body which can’t converge gradually ing equal, a well-enclosed race car has fastest moving air available. Theoretically,
enough, or can’t have a long enough tail, much lower air drag than an open-wheeled a horsepower increase of about 1 percent
is the reason for the abruptly cut-off rears car. Where race sanctioning bodies prohibit can be expected for each 10-degree drop in
on racing cars — the Kamm theory. The fenders, even the most modest fairings ahead inlet temperature and a 2 percent increase
idea is that if the airflow has separated at of and behind the tires probably can help. is available for an inlet airstream velocity
some point, there is no reason to continue When a complete vehicle is designed, air of 150 miles per hour. However, an inlet
the body any further, regardless of the scoops and ducts are usually an after- velocity that varies with speed may disrupt
shape, as it will have little effect on air drag. thought. The only really necessary high- intake ram tuning pulses which could be
However, if the body is smooth enough that pressure areas are on the nose of the body more valuable, and an unusually large inlet
streamlines are still attached all the way to and downward on the tail. The creation of may cost more in air drag than it is worth.
the trailing edge, as with a well-profiled positive pressures for air induction Driver comfort is every bit as important
wing, then the cut-off will have a noticeable anywhere else on the body is a waste of air as vehicle performance. If radiator or brake
effect. drag and apt to be an upsetting influence on heat spills into the cockpit, it can easily cost
With a really clean shape, even the air flow elsewhere. The currently popular a race of any great distance. Most heat
volume-forward generalization can be NACA submerged-type duct inlet, however, sources have been moved to the rear on race
disregarded. If the contour is correct and has good over-all efficiency. It is designed cars, but where the radiator or engine are
there are no body seams, lips, or projec- to counteract the slow-moving boundary still up front, cool air ducting to the driver
tions, a fully attached flow may be possible layer by pulling in high-momentum free- is critical. Also, the intakes ought to be
in which the air has a much greater tendency stream air without disrupting external placed as high as possible to avoid pavement
to follow the body. But in practical applica- airflow or creating a high-pressure area on heat, dust, and exhaust fumes from the car
tions, dents, nicks, dirt, or uneven airflow the body. It is primarily used when an air ahead — especially where a lot of close
from another car encourages the flow to inlet is needed at a specific location on the drafting is expected. Wherever auxilliary
separate. The individual drag from mirrors, body where there is no natural positive ducts and scoops can’t be avoided, it is im-
struts, or other protrusions is insignificant pressure. If there is room for ducting, portant not only that they have a well located
when considered separately. However, when however, a full-frontal nose inlet is inlet and smooth duct flow, but that the con-
placed on or near any body surface, not only somewhat more reliable. trolled exit be in a negative (or relatively
are they in a much faster airstream, but they Internal airflow is needed to cool the less positive) pressure location.
can critically affect the airflow on the vehi- engine and drivetrain, the brakes, the driver, The mathematics of aerodynamic forces
cle as a whole by their turbulent wakes. The and for carburetion. Although a water- are quite straightforward, even if measur-
wake is assumed to spread out like a boat cooled engine may be heavier, it is better ing the necessary quantities is not. The in-
wake at about a 15-degree angle behind the for aerodynamics in that heat can be taken fluencing factors are frontal area (A) of the
disrupting object. Even major bumps, to the air instead of vice-versa. If properly body in square feet, the air velocity (V) of
scoops, and bulges are likely to break up located, the radiator has a small effect on the body in miles per hour, and a non-
the airflow if their trailing surfaces aren’t drag, and in some cases could theoretically dimensional air drag coefficient (Cd) which
converged gradually enough. reduce drag. A heat exchanger in a duct can is based on the shape of the body. For stan-
Wheels and tires are a special drag pro- be considered as a ram jet. A tapered dif- dard air temperature and pressure, the equa-
blem on race cars, especially on open- fuser duct slows the air and builds pressure; tion is:
wheeled cars. Not only because they are the heat exchanger (radiator) expands the
relatively blunt and have a tremendously air, and a smooth discharge nozzle ac- force = (.0025) (Cd) (A) (V)?
large frontal area, but because they rotate celerates the air with some added thrust
and the front ones must be steered. Drag energy. To prevent the radiator from block- Some typical values for Cd will give an ex-
increases with the square of velocity, and ing the flow excessively, however, the dif- ample of what to expect. (See Table 2) The
the top of the tire has an air speed of twice fuser must slow the incoming air smoothly equation shows that a reduction in frontal
the speed of the car. And at the bottom front and evenly. And it must be well sealed to area is as valuable as a reduction in the

56
shape coefficient, although the latter isn’t as
easy to measure. The corresponding equa-
tion for horsepower required is:

horsepower = (.0027) (force) (V)

So it can be seen that forces increase as the

square of velocity, and horsepower required
increases as the cube of velocity.

DOWNFORCE EXTERNAL WING

In recent years aerodynamic downforce
has proven to be of far greater value to
reduced laptimes than low drag on high-
powered race cars. In some forms of racing,
however, fuel economy has recently gotten
more emphasis, so the compromise between
drag and downforce has become harder to
make. And while there is no question that
the lowest air drag is best and the greatest
downforce is best (assuming that the suspen- OPEN-TAILED BOX
sion can handle it), the proper distribution
of downforce between front and rear wheels
can be hard to resolve. Since the downforce
increases with velocity, it is possible that it
can cause the handling to change from
understeer to oversteer, or vice-versa, as the
speed increases.
Downforce can be achieved by at least
four distinct methods. (See Fig. 26) The
first, of course, was inverted wings either
SUCTION — TRACTION
stuck to the body or extended from it on
struts. While this method is almost as old
as race cars, the G.M./Jim Hall Chaparral
improved on the idea by passing the loads
directly into the suspension (now outlawed).
The second method might be called the
“open-tailed box,” in which the front and
ae
sides of a full body are sealed at the ground
by skirts or dams, and the rear is left open
to utilize natural negative wake pressure
from the rear. Method three, the “‘sucker,”’ POD WITH EXTERNAL WING
seals the rear also, and uses a separate pump
or fan to generate even more negative Figure 26. The four basic ways of creating low air pressures (shaded areas) under
pressure as the G.M./Jim Hall Chaparral a race car for maximum aerodynamic downforce — depending on the rules.
did in 1970. Unfortunately, that system was
too complex to be completely developed
before it was outlawed. There are a number race car body has very little underside raked forward at an extreme angle, the effect
of other methods which can be used to airflow because it is so close to the ground. is minimized, because the pressure will
reduce underside pressure — exhaust extrac- A large part of the air passes over the top decrease all the way back from the
tor systems, filtered engine intakes, or fan of the body, creating extreme negative minimum clearance point. In fact, ex-
blades on the wheel spokes — but they are pressures which try to lift the body. At one periments by Cornish have shown that com-
much less effective. The skirt which is time it was assumed that raising the nose pletely blocking the underside flow at the
necessary to seal off the underside would and smoothly contouring the underside front allows the naturally negative pressure
have to be even more novel and efficient would create a venturi, which would at the rear to act forward under the entire
than the one which was used on the Chapar- generate low pressure and draw the car vehicle. This more than counteracts the lift
ral. And method four is today’s hot setup, down. This works quite well with airfoils generated by flow over the body and creates
the “ground effect’’ principle. — which have low profile drag, and don’t considerable downforce. Total blockage is
Finding or increasing downforce is not have wheels. But the combination of high impractical on a vehicle which requires
simply a matter of inverting lift theory and pressures in front of the nose of a race car suspension travel, although race cars have
shapes from studies of aircraft in free air. and wide tires which block part of the closed the gap until the nose drags over
A non-lifting airfoil in free air has negative ground flow, can actually create high air bumps and during braking. There is a
pressures acting over much of its surface — pressures under the nose, which can migrate problem, however, in that the downforce
both top and bottom. But a relatively blunt well back under the car. If the underside is generated by a surface close to the ground

57
 The simplest solution is to mount a ver-
MODERATE DIFFERENCE tical lip at the extreme rear edge of the upper
body surface. This has the effect of
generating a high pressure area ahead of it,
which acts downward on the body. But sur-
prisingly, it doesn’t necessarily add drag to
the body as a whole, unless it adds to the
frontal area. If there is a rearward-sloping
body surface ahead of the lip, the positive
pressure area can also act to push the car
forward. So the lip contributes both to in-
foes
—
creased downforce and decreased drag until

SEE.
DIFFERENCE
LARGE
it reaches a height where the negative
pressure at the rear cancels out those effects.
FOOL GG The second and more effective answer —
VM
hhddddsdédd where it is permitted — is the separate wing.
SMALL DIFFERENCE Modern wings have amazingly high ratios
<- —
of downforce compared to drag. So much
Figure 27. Wing Performance — Free Air versus Surface Interference so that it seems like a real something-for-
nothing situation to get 500 pounds of extra
downforce at a cost of less than 150 pounds
is unstable. In other words, the closer the the pressure is really low, as at the crests in drag. However, it is necessary to know
nose gets to the ground, the more downforce of fenders, it can help to ventilate the how to get all the theoretical force that a
it generates, until theoretically the force bodywork — to reduce the surface area wing is capable of. Before a lot of time is
would be great enough to grab the ground which pressure acts on, and to let any high wasted in finding the most efficient airfoil
and bottom the springs. Some tests have pressure out from below. On the other hand, shape, the wing had better be mounted a
shown the force to increase as a function if a relatively well blocked underside has long way from the nearest body com-
of velocity cubed. negative pressure, the larger and better ponents. If there is any obstruction within
Considering the entire body, however, sealed it is, the more downforce it will pro- one or two chord lengths (distance from
downforce is generally a function of velocity vide. It is all a matter of knowing where and leading to trailing edge) of the wing, there
squared, the same as air drag. Therefore, what the pressures are. will be pressure interference. As Fig. 27
it can be seen that total horizontal area is At the rear it is not enough to simply con- shows, free flow is needed on both sides of
as important as the shape coefficient of verge the body at the proper angle to a surface to make it a wing. If the wing is
downforce. Identical airflow over a body minimize drag. At best that would merely placed near the body’s boundary layer or if
that has twice the surface area would have reduce the upper negative surface pressures it is tilted to a high angle and stalls out, a
twice the downforce at the same speed. The until they were neutral at the trailing edge. piece of plywood would work just as well.
problem is in getting the airflow to do what Instead, it is again a problem of generating Another problem in mounting a wing
it is supposed to. high pressures above and low pressures closely above a body surface is that the
At the nose, any available positive below any surfaces, without having high pressures can cancel each other out.
pressure (dynamic head) is helpful in forcing frontal and low trailing pressures. A large Negative pressure on the bottom of the wing
the body down, but at the same time it is board mounted at 45 degrees off the tail creates a negative pressure on the surface
harmful in increasing air drag. It should be would have a very high pressure differen- below it (see Fig. 28), and the wing might
apparent that the problem is not so much tial between the top and bottom, but it would as well be a huge spoiler mounted on the
a matter of where the air flows as it is one be just about as great in the forward direc- body. If the wing is stuck way out in back
of what the air pressures are on any given tion. It would have almost as much drag as with nothing below it, the negative pressure
surface. For example, NASA studies show it did downforce. acts directly at the ground. Of course,
that a wing works very well when close to
the ground, in having lower pressures
beneath it. But if it is mounted just ahead POSITIVE PRESSURE
of a blunt nose or tire, the wing may end
up with positive pressure behind and below 7 “|NEGATIVE PRESSURE
it, canceling any downforce it would have
|FORCES BALANCED
had by itself.
Moving back across the body, the prob-
lem of eliminating lift becomes essentially
unsolvable. If air must pass over the body
at a high rate, it is going to generate low
pressure. There is no way to avoid it without
creating a lot of air drag. The best that can
be done is to reduce the cross-section area,
so that high-speed airflow is minimized, but
the effect is hardly worth the effort. Where
the pressure is negative on top of the body, Figure 28. Interference of Body-Mounted Wings
the less body the better. In locations where

58
mounting the wing high enough above the
body would accomplish the same result and
provide smoother air to the wing. But even
if sanctioning bodies hadn’t ruled that out
for safety reasons, the height created prob-
lems of high center of gravity and high drag
pitch forces trying to lift the nose.
(A) GREATEST UNDERSIDE NEGATIVE PRESSURE — HIGH DRAG
TUNNEL CARS
Where allowed under racing rules, “wing
car” side pods are the most efficient way
to develop aerodynamic downforce.
Granted, they are just one part of the total
race car aerodynamic system, but because
of their highly desirable isolation beneath
the car, we can look at their aerodynamics
separately. And we can see how to optimize
performance by first maximizing negative
pressure with the body shape, and then
sealing it off with the tunnel sides.
Generating a lot of downforce through (B) LESS UNDERSIDE NEGATIVE PRESSURE — LESS DRAG
negative pressure over a large area, is com-
plicated by the tradeoff with air drag. If air
drag were of absolutely no concern, a
cascade of vanes and wings could be used
to exit the air almost vertically. It might be
possible to increase the downforce by
another 50%, but the drag penalty would
be unacceptable, considering top speed and
fuel efficiency requirements.
But with existing one-piece wing pods
(neglecting separate radiator ducts in the
upper surface), how is the proper shape
determined to maximize the average (C) SAME UNDERSIDE NEGATIVE PRESSURE — LESS DRAG
negative pressure over the total length? The
first consideration is the longitudinal or pro- Figure 29. Three basic pod configurations illustrate the tradeoff between air drag
file shape. Without a smoothly contoured and downforce, since the negative pressure pulls rearward as well as down.
underside, such as the case of the open-
tailed box, the negative pressure everywhere
will merely be equal to the base drag upward. Nor does typical venturi theory, best low drag configuration might be as in
pressure at the tail. This might be a max- where the goal is to create a very low Fig. 29b, which has peak localized
imum of about 1/3 the frontal pressure, or pressure at just one location and the rest of downforce near the front and the maximum
25 psf at 175 mph. the length is considered transition. Diffuser length in which to decelerate the air. In
However, when air is allowed to flow theory, as applied to wind tunnel diffusers, either case, the shape of the nose is critical,
smoothly underneath a body, the negative tail pipe megaphones, and jet engine outlets, to avoid pressure buildup or early airflow
pressure may be increased temendously. isn’t too useful either, as it usually assumes separation.
One well known test driver admitted that he axial symmetry and/or straight walls. Vanes could be used in a short diffuser
had seen a peak negative pressure of 9 The goal is to speed up the tunnel air flow to avoid separation, as shown in Fig. 29c.
inches of water at 100 mph (1 inch of water as much as possible (to generate the negative However, this is only useful in reducing drag
= 5.2 psf). If we convert this to 140 psf at pressure), and keep it that way as far back and not increasing downforce. The only
175 mph and average this down to the as possible. At the same time, positive negative pressure which can be utilized is
minimum of 25 psf at the tail, the overall pressure must be avoided at the nose, as it that which is directly above the road. Other-
average under the car might be 65 psf. With could migrate under the nose and destroy wise, it simply tries to pull duct walls
a typical underside area of about 30 square the downforce, and too rapid a divergence together.
feet, that adds up to almost 2,000 pounds must be avoided at the rear, which can cause The most important determinant of the
of downforce at top speed — a familiar high drag airflow separation or stall. The underside shape is the local divergence
figure in the rumor mill — not to mention best downforce configuration might be as angle. For simple constant-angle straight-
the force from positive pressures on the top in Fig. 29a, where the air is accelerated at walled diffusers, the optimum angle
and from external wings. the curved nose area, kept at high speed in depends on the ratio of minimum ground
Now, how is it possible to get those the straight area, and then breaks loose at clearance to pod length. As an example, for
negative pressures? Traditional free-air wing the blunt high drag tail. The straight area a ratio of 6 foot length divided by 0.3 foot
theory does not apply so close to the must have some small divergence angle to clearance (20:1), the diffuser angle should
ground, because only a limited volume of allow for boundary layer buildup, assum- theoretically be less than 4 degrees to safely
air can pass under this wing and diffuse ing a skirt-sealed tunnel. Conversely, the avoid separation all the way to the tail.

59
 Harvey. The primary problem is that the
frontal air pressure wedges itself against the
bottom front quarter of the rotating tire and
is squeezed into the tire contact patch. If
there are grooves in the tread where the air
can escape, there is little effect on traction.
But perfectly smooth racing tires may be
losing some traction by having a direct
stream of air acting on them. Fortunately,
the upper portion of the tires has an opposite
effect. In moving forward at a high velocity
(dragging a boundary layer with it), it has
a reduced aerodynamic lift at the top. Not
that it makes much difference, as little can
be done for open-wheeled cars anyhow.
In spite of the fundamental problem of
exposed wheels, in recent years designers
have found a way to take advantage of at
least some of their frontal pressure. At the
rear, where there is some side pod
bodywork close by, the body is waisted-in
just ahead of the tire, and a flat plate is
extended out from the bottom. This uses
A “‘short-track’’ CART wing in the foreground for maximum downforce, and an some of the unavoidable positive pressure
alternate ‘‘superspeedway’’ rear wing for minimum drag at continual high speed. at the bottom quarter of the tire to create
extra free downforce.
However, if clearance is doubled to 0.6 foot, free air stream. The earliest technique of As a sidelight to aerodynamics, it should
the safe angle increases to 6 degrees. sliding or flexible skirts did the best job of be mentioned that innumerable suction-cup
Because race car side pods are a three- sealing. They were eventually outlawed (for tread designs have been invented to reduce
dimensional problem, theory is only a both operational and safety reasons) but the air pressure in the tire contact patch. The
guideline, and experimentation is the only there are still a number of legitimate tricks advantage is obvious — a potential increase
reliable method. The shape must be which will produce almost the same result. in contact pressure of over 14 pounds per
modified until the most rapid divergence is First, the ground clearance gap can be square inch (and some race cars have over
reached, with no separation under the worst analysed as an orifice. With a given pressure 100 square inches of tire contact) without
conditions. difference between underside and outside, an increase in weight or inertia. The pro-
To keep the flow attached as long as there will be a given inward leakage rate. blems aren’t all solved yet, though — like
possible, it is desirable to have slightly But the larger the container, or plenum, or what to do with water and dirt, the high rub-
turbulent airflow. If suspension components tunnel air mass flow rate, the less effect this ber contact loading, how to compensate for
and other cars don’t create enough leakage will have on the negative pressure. wear, and most important, how to get the
turbulence, vortex generators may be Obviously the tunnels cannot be made much tire to let go enough to rotate.
necessary, aS were rumored to be in Lotus wider, but raising the wings inside them Radiator ducts can hurt or help
pods at one time. might help more than it would hurt the aerodynamic downforce. The duct inlet
The lateral profile, or change in cross- advantage gained from ground effect. must be in a positive pressure area to get
section area, may be just as important. The Second, the leakage might be minimized enough airflow, or a positive pressure area
most obvious peculiarity seen under many by any number of orifice tricks, such as a must be created. When the radiator is in the
cars, is the sudden obstruction caused by the curled-out edge design. And leakage might nose, the inlet is obvious and the outlet is
rear tire. One car, in fact, had telltale not be as great as expected due to the fact the problem. If it is ducted underneath the
leaking-oil streaks which showed the airflow that the outside “‘free air’’ is actually in the body, the air pressure creates lift, and if it
to be making almost right-angle turns in this wake of the front tires, and may therefore is ducted upward to generate downforce, it
area — almost certainly causing a pressure _have a slight negative pressure of its own. cooks the driver. Most nose outlets are
rise. The use of narrow transverse transmis- This leads into the third possibility for directed laterally around the cockpit
sion cases and smoothly extended bell- sealing. If the air pressure just outside the whenever possible. Centrally-mounted
housings (See Chapter 9) must help a great skirt can be made as low as the pressure radiators appear to have undesirable fron-
deal. At the very least, the cross-sectional under the skirt, then there will be no loss. tal pressure areas, although a submerged
area should be constantly increasing all the And in theory, the lateral or outside wall entry on the top of the body may create
way to the exit. This expansion also allows pressure can be made almost as low by valuable downforce there. But there is still
for some boundary layer buildup on the essentially the same method — a side a problem of what to do with the outlet air.
three sides of the tunnel. In the same light, airfoil. Some cars have very obvious If it is directed up for downforce, it
generous fillets in the upper corners will “external tunnels” which wrap laterally interferes with airflow to the rear wing, and
help keep adjoining boundary layers from around the tub just aft of the front wheels. if directed rearward it may return some of
compounding each other. the heat to the engine. An extreme rearward
Once this negative pressure has been OTHER EFFECTS radiator location may be best for low drag
generated in the pods, then some method Exposed wheels and tires are also a prob- but then there is the problem of getting air
is necessary to isolate it from the outside lem in lift, according to Fackrell and into the inlet duct.

60
DOWNFORCE ABSORPTION
The last and perhaps greatest considera- REAR
tion of aerodynamic downforce is what to
do with the force. It usually increases as the
square of velocity, which means that as the
speed quadruples from 50 to 200 miles per
hour (common for a road race), the
downforce will increase sixteen times. With
aerodynamic systems capable of 2,000
pounds of downforce at top speed, it is very
difficult to design a suspension that will
work right at the highest and lowest speeds.
The simplest and most practical solution
Figure 30. Effective Location of Aerodynamic Forces
— attaching aerodynamic load surfaces
directly to the axles — has been arbitrarily
outlawed. Therefore, we see all manner of
expensive or relatively unsafe alternatives. legend, when aero loads began to rise, table to the driver in terms of stability is one
The most common solution is just to in- Formula car handling refused to respond to where the rear force is proportionally much
crease the spring rate so that the chassis traditional spring and anti-roll bar changes greater than the front, so that as the car went
doesn’t bottom at top speed, and let the — until the designers realized that chassis faster it would have a greater tendency to
driver and chassis suffer the incredible road flex was allowing aero loads to dominate the understeer. However, this wastes potential
impacts at lower speeds. The reverse alter- front/rear roll resistance. front cornering power, limiting the cor-
native of providing three times the suspen- Finally, it may be important to analyze the nering speed to whatever the front tires can
sion travel with the original low spring rate, effect of downforce on rolling drag in a handle, and subsequently creates un-
is unacceptable from a center of gravity corner. In some high speed corners, a point necessary aerodynamic drag from the excess
height, skirt seal, and camber change may be reached where speed is limited not rear downforce. So the ideal downforce ratio
standpoint. by traction, but by power available to over- from front to rear might be approximately
Springs or linkages with a rapidly in- come tire slip angle drag. At Indy, for equal to the weight distribution from front
creasing ride rate were once thought to be example, when a driver can take all the to rear, or have a net force vector acting at
the trick answer, and they might be — on corners flat out, then he will probably get the center of gravity.
a 2-wheel motorcycle. But one of the most less of a dropoff in speed in the corners by But the problem has other complicating
important ways of balancing oversteer/ reducing downforce. In fact, it is fairly factors. First, as the car goes faster and air
understeer on a 4-wheel car, is the distribu- obvious in the difference in rear wings on drag increases, more tractive effort is
tion of front/rear roll resistance. If that CART cars. When they are at a fairly slow required at the rear wheels just to maintain
distribution changes with pitch, roll, and street course they use monster rear wings, a constant speed. This tractive effort detracts
jounce — because of variable-rate springs while at the superspeedways, wings are from cornering force available at the rear,
— then it may be impossible to maintain a fairly small and flat. Talk about too much creating a potential for oversteer. However,
constant handling balance. (See Chapter 4) of a good thing! this may be balanced by another effect from
“Active” suspensions, which change air drag. This drag acts at a point some
spring rate or height in response to in- STABILITY distance above the ground, which causes an
creasing speed or load (such as commercial Two primary sets of aerodynamic forces overturning torque above the rear tire con-
load-levelers) have been tried. But in addi- affect stability in race cars — the ratio of tact patch. (This is true only under power.
tion to the balance problem, these either downforce from front to rear and the ratio When the car is coasting, the drag torque
require some onboard power to raise the of lateral gust force from front to rear. In acts around the center of gravity.) For
chassis, or some suspension oscillation time taller passenger cars and trucks a rolling example, if a total drag force of 400 pounds
to “pump up” hydro-pneumatic spring- force from gusts can cause instability, but is centered at a point on the body 2 feet
shock units. (See Chapter 4) race cars are usually low enough to avoid above the ground and the wheelbase is 8
In the absence of automatic load-leveling this problem. Downforce stability is of feet, then even without downforce, (400) x
systems, racers have tried a number of fixes. major importance because of the high forces (2/8) or 100 pounds of tire load will be
They have designed suspension systems for available and the fact thay they change transferred from front to rear. (See Fig. 30)
minimum camber change with deflection, drastically with speed. Assuming that the For a fixed drag force height and vehicle
or tried to compromise between high- and car is mechanically stable, or that it would wheelbase, these two air drag effects can
low-speed setups, or simply pre-loaded the have the same degree of steady state only balance each other at one specific
springs to take the maximum load at speed. understeer at low speed and high speed if speed, since they don’t increase at the same
None of these solutions are without serious there were no air drag or downforce, the rate with an increase in speed.
faults, so aerodynamic downforce is problem is to make it aerodynamically Another problem is the fact that the
sometimes limited in potential until some stable also. In other words, mechanical in- mechanical load on the tires changes with
means is found to balance the forces at the stability should be corrected with load transfer under braking and accelerating.
suspension. mechanical solutions which are relatively During acceleration the rear tires gain load.
If that weren’t enough, chassis stiffness independent of velocity, and aerodynamic But if the center of aerodynamic downforce
has been an increased problem with these instability should be corrected with (which may be much greater than load
loads. Varying the front/rear roll resistance aerodynamic solutions which are a function transfer forces) stays at the center of gravity,
ratio to balance handling assumes that the of the square of velocity. The distribution then the rear tires will be too lightly loaded
chassis is a rigid block. According to of downforce that would be most comfor- and the car will have a tendency toward

61
 rear of the wheels. Each by itself not only
adds leverage loads to its own end of the car
but subtracts weight from the other end. So
it is apparent that downforce stability can
be changed considerably by moving wings
or spoilers fore and aft.
Even on a ground effects car, with a
somewhat fixed side pod center of
downforce pressure, the best mechanism for
trim, or front/rear downforce balance, is
still external wings. Some cars have tried
to eliminate them for their weight and air
drag penalty, but they keep coming back.
Lee Dykstra once commented that canard
wings actually disrupt pod airflow efficiency
to a noticeable degree, but they pay off not
only in trim adjustment, but in a little extra
useful downforce on slower tracks.
Ground effect cars may also have a unique
problem in side pod dynamics, called “por-
poising.’ This is due to a wing’s natural
tendency to suck itself right down into con-
tact with the ground. If that happened, of
course, the flow would be blocked, and
downforce would be lost, allowing the
springs to push the wing back up again,
where it could start the cycle all over again.
If the center of pressure is at the center of
gravity, a simple flat bounding might be
expected. However, when the gap closes,
frontal pressure will likely cause the nose
to come up first, then the tail, creating a
combination of pitching and bounding, or
porpoising.
Although no one has reported the chassis
actually contacting the ground, it probably
doesn’t need to, to start this action. The
boundary layer on the body is probably
thick enough to choke the flow when
clearance gets down to some fraction of an
inch. In theory, however, this squeezing
situation just might cause the opposite
effect, in causing flow separation to the rear
first, allowing the tail to come up and start
Two examples of rear spoilers, from the outrageously unrestricted and imagina- a reverse porpoising.
tive IMSA wing, to the restrained NASCAR version which is carefully defined In fact, diffusers with a marginally great
in both height and angle by the regulations. (Chevrolet) divergence angle have a tendency toward
pulsating separation. This means that the
slightest provocation, such as a change in
aerodynamic oversteer. During braking even is approximately what happens when the inlet clearance or inlet turbulence, or even
more load is transferred forward, and if the nose lifts under acceleration (reducing aero a resonant body vibration from the engine
center of downforce doesn’t also move downforce at the front) and dives under or tires, can cause the airflow to separate
forward, the front tires will be more prone braking (increasing aero downforce at the locally and then reattach itself periodically.
to lockup under braking at higher speeds. front). But when these conditions aren’t ade- As for cures, the best seems to be in
At least in this case it is a stabilizing condi- quately known and allowed for, it is safer keeping the side pod airfoil surfaces from
tion, but there is still lost efficiency in to have a somewhat higher downforce getting too close to stall or too close to the
braking unless there is some sort of variable toward the rear to prevent any possibility of ground. Dykstra, in fact, mentioned that
brake proportioning. oversteer at top speed. stiffer springs once made porpoising go
If it were acceptable under racing rules, Another complication is in the fact that away. But whether this changed onset speed,
it might be possible to provide variable aerodynamic downforces aren’t generated ride height at speed, damping, pitch or ride
aerodynamic surfaces to move the net force exactly at the front or rear wheels. To get frequency (porpoising may also be seen as
vector forward under braking or rearward better airflow and efficiency, the front an aerodynamic frequency in phase with a
under acceleration to maintain a more spoiler or wing is mounted well ahead and ride frequency, each amplifying the
neutral aerodynamic stability. In fact, that the rear wing or spoiler mounted well to the other). . .maybe no one knows for sure. If

62
you change something and the problem goes
away, why question it?
It is possible for any number of these
stability factors to work together, for or
against the desired result. When everything
goes against the driver, the results can be
disastrous. Say, for example, there comes
a time when the front load is lightened
from: acceleration rear load transfer, body
drag rear load transfer, rear wing downforce
behind the rear tires, reduced downforce at
the front from a nose-up attitude, the car
goes over a rise and turbulence from another
car combines with a wind gust — and up
it goes. It is not that unlikely, either, as in
the past over a dozen race car drivers have
learned from direct experience when their
cars flipped over on their backs. The same
sort of philosophy in downforce applies to
the center of lateral pressures due to side
winds on the body. The center of pressure
should be at the center of gravity and yet
vary with the change in.load transfer from
front to rear. The lateral forces from gusts Front aerodynamic spoilers are made low enough to just touch the
are usually far less than downforce, and yet ground under the worst conditions of cornering and braking, and as
at the limit of cornering traction at high wide as possible—except where limited by the rules, as in NASCAR. (Michael Lutfy)
speed, they can have a critical effect.
Unfortunately, the lateral center of pressure
on a high-speed body is about one fourth HARDWARE bumpers. Many production trunk spoilers
of the overall length from the front. With The basic hardware component of and undernose spoilers are stylish and do
race cars tending to a rear weight bias, then, aerodynamics, of course, is the body, which a good job as far as they go, but they don’t
there is a great torque trying to turn the car in many cases can’t be altered very much do anything that a similar-sized flat plate
away from the wind. Ideally, to minimize because of production rules or structural wouldn’t do in the same position. The ob-
lateral displacement (off the edge of the considerations. Even when it can be ject in such cases is to stop the airflow, and
track), perhaps the lateral pressure should modified, there isn’t a great deal that can the contour of the device doesn’t make much
turn the car into the wind so as to give it be done between the front and rear wheels. difference.
a corrective lateral force from the tires’ slip A certain amount of air has to pass over a At the front, production under-nose
angle. In practice, however, the driver would volume that has already been reduced as spoilers usually don’t come close enough to
feel that as oversteer and it could be tragic much as possible, so it is simply a case of the ground, because they have to contend
at the wrong moment. So it may be best to keeping airflow interferences at a minimum. with curbs in parking stalls and steep
have the center of lateral pressure a little Improved streamlining in this area will do driveways. In addition, they may be so
forward after all — but less than it usually little more than improve the airflow to effective, compared to rear downforce
is because of excessive understeer effects. whatever downforce devices are used at the devices, that an unwary person could easily
The most effective methods for moving rear. Otherwise, the hardware can be broken produce a high-speed oversteer condition.
the center of pressure rearward are to in- down into front or rear spoilers (for full- Where the maximum amount of downforce
crease the lateral area at the rear or to im- fendered race cars), front or rear wings, and can be tolerated at the front, however, the
prove lateral streamlining at the front. This ductwork. idea is to build whatever device is necessary
fits well with the practice of low, rounded to block airflow as completely as possible
noses and large tail sections with huge fins SPOILERS and as far forward as possible. For produc-
and end plates on wings. For maximum The function of front or rear air spoilers tion sedans, the best materials have been
efficiency, fins must be located in relatively has already been covered: they keep air urethane, a molded thermoplastic, or Lexan,
clean air flow and not in the wake of a from passing under the nose and develop a practically unbreakable translucent
“dirty” body or spinning tires. But since fins pressure on the upper rear bodywork. thermoplastic, while in limited production
add to the total side force, the ideal design Anything that accomplishes these ends and race car applications, Kevlar/epoxy seems
would be a very small fin at a long distance satisfies the race rules is advantageous, no to be the best bet. As for clearance, racing
out behind the body for clean air flow and matter what it looks like. Many devices have spoilers are often made slightly low and then
high torque leverage. Another aspect of fins been built and marketed for passenger cars worn to the proper clearance by contact with
is their effect on lift. As with end plates on with appearance as the primary object, and the ground. However, even Lexan can be
wings, laterally mounted fins will tend to in many cases they do make a contribution broken by a high velocity impact on its edge.
increase whatever pressure differential ex- to reduced drag and increased downforce. If it is necessary to angle the spoiler for-
ists above and below the body. Downward But understanding the previous theory will ward, it is better to leave more clearance.
pressure on the front or rear body will be indicate the more useless designs, such as If the spoiler is made from a mold, it
increased by fins as they prevent vertical trunk-mounted airfoils, hood-mounted could help to give it a scoop shape which
flow over the sides. vortex generators, and airfoil-shaped curves around the sides ahead of the tires.

63
488 THEORY OF WING SECTIONS APPENDIX IV

om
tS kewacas
*
EEE T
EEEPeppy EEE eee Coe
BEERSER ER tye Lo BESRRERESS
11 11 | ; 1 | ne | |
a2. 4444 stot date telolebetel 1
EEEEEE E E EEE EEE
EEE
Tri | ao wl
SSSR REEE it penal HAH z=
“ee
toot 7
SEBReEe t | —
Titi || ij
Eases Ci aed ia
BESSSRSEESSEEE Pin i 1] O24
2 ilt)| iii
(| tBESRDEE
1siat e-28| ERBSSERS
[Ep TI Littl ] | a
[i] [tl i {| i] | | | tI 1 a NACA 4412
EGnERSEE | LL EREERERSEGA 20 (Stations and ordinstes given in
2d EEEEEOnDEE ee per cent of airfoil
chord)
REERSUROSSS | EREn BE =
rT | £§
{ = 0/6
a
8
>
S
. £012
=

rs $
== &.
>
°
v

=
§
F rs

Oe: or ——

: ri i

-/ oA
= TT ;
r ¢ ane t
= ;
§~2
x
$24
8 &
¥ 3 GERSaae
& Tee 2 ee Slope of radius through L_E.: 0.20
ee e Peet tl togal |
$s g i | |12601 | | | se drougmess|
||
a Bea! 60°) _|
1 1020 simulated split flop Geflected
§
qo | rveoTT it faz Tieiict
Se =| - a BESELER tli rit ptt ;
——— pi} dui) a
| petty | rT | |/ iB!
B tyiyi i | EE! lil
+ al a OlOe |
= I rit i BERBD em { 1 eA
a =32 -24 -I6 -8 o a iT 24 2 a6 a2 -8 4 o 4 2 t2
Section angle of attack, x, deg Section lift coefficient, e,
NACA 4412 Wing Section NACA 4412 Wing Section (Continued)

Figure 31. A sample of lift and drag data and dimensions for one typical wing section from Abbott and vonDoenhoff’s
Theory of Wing Sections

This helps to channel some of the deflected most complete sources of data on airfoil simplified. What is most important is to find
air to the side of the body instead of over shapes and performance is Theory of Wing the maximum ratio of downforce to drag (or
the top where it contributes to lift. However, Sections by Abbott and vonDoenhoff lift to drag ratio, as it is reported for air-
a rigid plastic spoiler had better be very (See Fig. 31). For data on short wings close craft) at the highest values of downforce.
strong to cope with inevitable contact against to the ground, NASA Technical Note D-926 After the most efficient shape is found, then
the track or the ground during some off- by Fink and Lastinger does a pretty good the wing area can be selected to produce any
course excursion. A final note is to be wary job. amount of force for a corresponding cost in
of sharp, forward-projecting bottom edges. Technical reports from NACA or NASA
This particular shape is highly sensitive to also provide a lot of valuable data on wing All lift and drag coefficients, however, are
ground clearance and the closer it gets to flaps and slots. Their purpose is to keep presented for an ideal wing or one which
the ground, the harder it pulls the car down. airflow attached to the wing at high values is infinitely long and therefore has no tip
In other words, the downforce will be of downforce (or lift). By very careful effects. As a real wing becomes very short,
unstable and sometimes unpredictable. design and construction of multiple-flapped its aspect ratio (AR), or span/chord ratio,
The rear spoiler lip is pure simplicity as and slotted wings, it is possible to bend air must be considered. The actual drag co-
long as it’s strong enough and has been a long way around a curve without its efficient is:
proven not to contribute too much drag. separating and causing a high-drag turbulent (Cp) + (C,)* (chord)
wake.
WINGS With all that information available, it still (3.14) (span)
The subject of wings has been researched takes some interpretation before the desired Obviously, the shorter the chord and the
very thoroughly and with great precision profile can be selected. One point should wider the span, the better. Where the span
because of all the years of aircraft develop- be made clear, however, because NACA/ is limited by the rules or the width of the
ment. There is enough test data on every NASA assumes it is acceptable to have a body, the drag due to lift is by far the
shape and size of wing (even some in close moveable wing that can be trimmed out to predominant effect. But as is obvious from
relationship to the ground), to make the zero downforce and minimum drag. With every car on the track, wing end plates can
proper choice a simple, scientific matter — the necessity of running wings at a fixed be used to keep airflow from falling off the
if the requirements are known. One of the angle, the selection requirements are end of the wing. The “effective” aspect ratio

64
can be more than doubled if the end plates
are large enough, (see Figure 32), with a x2.0
corresponding improvement in both lift and
drag.
One wing shape that probably won’t be
found in government publications is the
Liebeck, or banana wing, which has been
quite successful on the rear of many race
cars. This profile is based on the fact that XES
race cars don’t have to face the aircraft com-
promise between minimum cruising drag
and takeoff/landing requirements. All that
is important in race cars is to have minimum
drag at the greatest downforce condition.
The Liebeck airfoil does this by essentially ratio
aspect
apparent
having no upper aerodynamic surface, x1.0
which would normally have some negative 10 20 30 40 50 60
pressure acting on its forward portion. In-
percent end plate height/wing span
stead, it acts as a tremendous high-pressure
air bucket with great concentration on main- Figure 32. The efficiency of a wing can be more than doubled when the end plate
taining attached flow on the highly curved height is at least 60 percent as great as the wing span.
lower surface. Also, by keeping the leading
edge as high as possible, little of the high-
pressure air on the top can act in the forward
or drag direction. A computer analysis of It would appear that the construction of turbulence. And yet, if the wing is in a high-
boundary layer theory and potential flow a wing is fairly simple, using familiar lift condition and airflow has separated from
theory gives the optimum shapes, relative riveted aluminum construction as is com- the bottom, intentionally created skin
size, and location of the airfoil and its flap. mon to most race car builders. But there are turbulence can cause the airflow to separate
The result is a claimed record coefficient a few hazards to be aware of, such as precise further back. This is the purpose of the
of lift of over 4.0, with a low (but unstated) shape, smoothness, and strength. To main- vortex generators or vertical tabs mounted
coefficient of drag. Incidentally, the Liebeck tain the proper smooth airflow, the cross- on the top of some aircraft wings. The need
airfoil has been patented in cooperation with section profile must be held to very close for, size, and placement of intentional
the McDonnell-Douglas Corporation. tolerances with the given airfoil coordinates. roughness is a job for an aerodynamicist —
Because of the precision in design and At the same time, surface roughness can. and a lot of experimentation.
construction, airfoils such as these are par- either hurt or help, depending on its location The strength of a wing can be critical to
ticularly sensitive to external disturbances. and the airspeed. For example, a wing with the driver, and yet some constructors have
It is possible that surface roughness (from laminar flow may have no airflow separa- been unaware of the loads that a wing may
tape, dents, paint nicks, or excessive dirt) tion at low angles and generate no be carrying. It isn’t unreasonable for a large
anywhere on the lower surface could upset
the entire flow and destroy over half of the
downforce capability. Although wing strut
fairings and end plates are necessary, they
are also major contributors to disturbed
airflow, with the standard 7- to 15-degree
wake in which the downward-acting
negative pressure is destroyed. The struts
themselves don’t add much to over-all vehi-
cle drag, but the interference effect in an air-
foil’s high-velocity area can be critical. It
may be very important to have well-designed
airfoil shapes, smooth seams and well-
filleted intersections in these locations.
Further improvements have been made by
Pershing, in using non-constant airfoil cross-
sections to take advantage of local airflow
variations.
One other related piece of aerodynamic
hardware is the Gurney flap, or the vertical
tab at the upper — or high pressure —
trailing edge. According to Dr. Liebeck, this
adds a great deal to the downforce with an
immeasurable increase in drag — as long
as the flap is not much more than one These wing supports are designed to improve airflow on the deeply curved and
percent of the chord length. critical lower surface. The tubes are for measuring pressure distributions.

65
 average speed would have to be known. But
since most are sized intuitively, or to match
a given duct or radiator core, the following
figures should produce an approximately
efficient NACA opening. Using a recom-
mended airfoil-shaped upper lip, the
opening should be from 3 to 5 times as wide
as the open depth, and the leading flat ramp
should be from 6 to 10 times as long as the
open depth. The most ideal ramp opening
shape can be obtained by scaling up the co-
ordinates from Fig. 33, using sharp edges
everywhere but at the top lip. Practically
speaking, however, a very precise NACA
duct is perhaps more cosmetic than
measurably effective on most race cars,
where drag is usually relatively less
important than downforce and is over-
whelmed in comparison to internal duct
losses.
Even after high-speed air has been drawn
into a duct, it must be carefully routed.
Rough, corrugated ducts can reduce the
iil SI flow, and sharp turns can almost totally
destroy it. Sometimes turns are necessary,
In all series, especially F1, you will see unusual shapes which may not seem such as from horizontal flow to vertical
theoretically optimum. They are often “rules-driven” configurations adopted carburetors, or to distribute the flow evenly
to circumvent a set of regulations that were intended to slow the cars. (Benneton) across a radiator face. In such cases, one
carburetor or one edge of the radiator core
may get most of the ram air effect, which
race car wing to be able to generate over wing isn’t mounted in a good location. At starves the rest of the carbs or core. A
a thousand pounds of downforce above 200 the front, the only consideration is to avoid common solution for engine inlets is to
miles per hour. And yet some are con- high-pressure areas (or body surfaces) im- make the inlets bell-shaped and to surround
structed as though all they had to support mediately below or behind the wing. The each inlet with a filter or porous foam to
was their own weight. Just the fact that the farther ahead the wing is, the better it will even out the flow. Removing this element
load is distributed evenly over the surface work. And the closer it is to the ground, may increase total flow but it has been
and yet has to be carried by two or four the harder (and more unpredictably) it will known to cause some of the cylinders to
specific points is enough to create structural press down. At the rear, it is hard to find misfire at high air speed.
problems. Most wings are constructed with clean, non-turbulent, free-stream air. Airflow through radiator cores can be
the standard aircraft technique of riveting Upstream disturbances include the cockpit, distributed more evenly by guide vanes in
a sheet aluminum skin over stamped roll bar, engine inlets, turbochargers, a large radiator duct. These ducts usually
aluminum spars and ribs. Plastic wings, mirrors, and so on. A wing will work in have a small opening which must rapidly
however, have proved to have smoother turbulent air but nowhere near as efficiently expand in one or more directions to a much
surfaces with greater strength for a given as it will if some thought is put to stream- larger core size. If any wall slopes away
weight. Most of them are constructed with lining — or eliminating — everything ahead from the free-stream airflow direction by
a fiberglass skin bonded over a rigid of it. Again, the farther it is from the body, more than 5 to 10 degrees, the air will
urethane or honeycomb core, which may be the better it works. probably keep going straight. A vane, or
reinforced with some sort of spar to provide group of vanes, perhaps extending all the
stronger attachment points. It ought to be DUCTS way to the core, will be necessary to keep
apparent that a fairly good wing can be Ducts will be considered aerodynamic any airflow divergence angle below the
designed and built by a competent race car devices — although they are internal and separation point.
builder, but to produce a really professional consist largely of empty space — because
wing takes a lot of knowledge and they are most often a last-minute add-on. DEVELOPMENT
experience. To get the maximum air into the duct with The development of aerodynamic race car
Mounting a wing can be just as impor- a minimum of frontal pressure or external shapes should go all the way back to original
tant as its basic design, considering the loads interruption, it is necessary to keep the vehicle design since the arrangement of
involved. The failure of a wing or support slower boundary layer out. The entire duct components is so closely related to the ex-
strut at speed can cause an instantaneous opening can be raised above the body sur- ternal form. In such a case, the scale model
loss of traction and stability. It pays to know face with a splitter plate below to deflect the wind tunnel is the most valuable tool in that
what the loads are and to test-load the boundary layer out, but the submerged it allows cheap and rapid experimental
assembly well over what is expected, to NACA duct opening may have less effect on changes to be made. (See Chapter 13) If
account for vibration and bump acceleration vehicle drag. anyone really knew how all the particles of
overload. To properly size a duct opening, the air would flow about a body, that wouldn't
Of course, the loads won't be there if the required mass air flow rate at a given be necessary. Since many designers think

66
they know — and don’t use wind tunnels —
it becomes necessary to develop the final
shape and various accessories by on-the-
road testing. Full-scale road testing isn’t all
bad, however. The forces are usually much
greater and much easier to measure, and
there is no question whether they can be
directly scaled up to the real car from a
model.
MEASUREMENT
Aerodynamic development is usually con-
centrated on lowering air drag, increasing 8% 16% 24% 31% 39% 47% 62% 77% 92% 99% 100%
downforce, and improving internal airflow.
In the reduction of drag on a full scale body,
the primary means of measurement is the Figure 33. Recommended dimensions for an efficient NACA duct, with an open-
coastdown, which is explained in more ing height/width ratio of 1:3 to 1:5, and a ramp slope ratio of from 1:6 to 1:10.
detail in Chapter 13. By applying the
theories previously discussed, making in-
tuitive changes, and recording the rate at been able to keep up with “dirty” winged ment it can be an educational experience to
which the car decelerates, it is relatively and spoilered cars. find out just where the car starts showing
easy to find some improvements. It isn’t Downforce is usually easier to measure high-speed oversteer characteristics and to
even necessary to get the actual drag figures than air drag if precision is not important try to keep the force balance just under that
im pounds or horsepower as long as the tests — and with the size of the figures in pounds, point. The problem is that the driver may
are accurate and repeatable enough so the it usually isn’t. The forces are most useful find the car tragically uncontrollable at that
data can be believeable. It usually takes a when measured directly at the front
and rear point.
great change in configuration, however, to axles, and with a constant rear tire thrust Another problem is that it may not be
show up in reduced air drag, and a little air to include air drag rear load transfer. For- easy to get enough downforce at one end to
drag doesn’t count for much in reduced lap- tunately, that is also about the easiest way balance the other, and therefore the other
times except on the high-speed banked ovals to measure them — by using the suspension will have to be reduced. Obviously, an in-
of USAC and NASCAR. spring deflection and holding the car at a crease in downforce is preferred, at any cost
When drag figures are calculated, one of constant high speed while recording data. within the rules. Using the previous
the most useful means of reporting them is Even if an elaborate electronic or theories, this may mean ridiculously large
im pounds at 100 miles per hour, instead of mechanical recorder (as discussed in spoilers and/or wings or minimum ground
in drag coefficients, which neglect the possi- Chapter 13) is not available, it is often possi- clearance must be tried at the lesser end of
ble change in frontal area. A figure of 280 ble to mount visual suspension deflection the car. It can even be worth while in
pounds drag at 100 miles per hour is a real indicators where they can be seen by the development to try things that are not legal,
figure that means something to everyone. If driver at speed. The deflection, of course, just to find out what extremes might be
someone needed to know the dragat 150 is irrelevant except that it is related to the necessary. In NASCAR, where bodies and
mph, all he would have to do is multiply number of pounds it takes to deflect the spoilers are very explicitly regulated, about
the drag by the speed imcrease fraction suspension an equal amount at rest — and the only thing that can be changed to affect
squared, or (150/100) , to get a figure of also that total suspension travel should not the front/rear downforce is the rake angle
2.25 x 280 = 630 pounds at 150 mph. The be used up. of the chassis. By lowering the nose and
same applies to any aerodynamic downforce raising the tail, downforce can be increased
Measurement in pounds at the front or rear BALANCING FORCES at both ends. Where rear spoilers are not
wheels — as long as suspension deflection So there is no need to know the exact regulated, it is more efficient to keep the
doesn’t change the increase to some other figures, except to know their balance be- entire chassis low and raise just the rear
than a square function. tween the front and rear wheels. With the spoiler.
The most important figures to get are the static weight distribution known, it is fairly The optimization of wings is a little more
aerodynamic downforces, and if there are easy to balance the front and rear complicated, even assuming a given airfoil
any compromises to make between them aerodynamic downforce in pounds in about profile and size has been selected. The first
and drag, downforce usually is the more the same proportion, with a slight bias to step at the rear is to get clean air to the wing.
valuable — up to a point that can only be the rear. The most accurate data available, Cotton tufts mounted to wire uprights can
determined by laptimes. A reasonable however, won't account for dynamic and show which directions the air is flowing over
estimate for high-powered race cars on road transient conditions during acceleration and the body, if it isn’t apparent from the loca-
courses, however, is that a 1 percent in- braking, and the final balance must be deter- tion of upstream interferences. If the car has
crease in downforce is more valuable than mined by the driver’s acceptance. Still, it been raced, dirt and oil streamers on the
@ 10 percent decrease im air drag. For is important to remember that an aero- body, wing, and supports often give the
example, reducing drag from 280 pounds dynamically stable (understeering) race car picture even more graphically. Long
to 250 pounds is less important than in- is limited in cornering by front wheel trac- streamers of a colorful lightweight thread
creasing wing forces from a total of 280 tion, and at that point no increase in rear can be seen from the cockpit and can give
pounds to 285 pounds at 100 miles per hour. downforce will have much effect — except a good visual indication of turbulence or
As ndiculous as that seems, it shows why perhaps to make it feel better and go slower flow detachment. If it exists, there are only
highly streamlined road race cars have never due to increased air drag. During develop- two choices: eliminate whatever is

67
 speeds. This was discussed in Chapter 4, but
another technique might be mentioned. On
high-speed USAC ovals, where vehicle
speeds are relatively constant, the springs
may be pre-set for those speeds. In other
words, the springs are designed to hold the
suspension at a specific ride height with a
large amount of aerodynamic downforce —
even though the springs will raise it against
the rebound stops at very low speeds.
PRESSURES
0 20 PSF The development of airflow, whether on
bodies or wings or in ducts, can be
Figure 34. Pressures on Body at 100 mph in Pounds Per Square Foot simplified by the measurement of pressure
distributionsand velocity profiles. Pressure
distributions come from recording air
upstream, or find a more aerodynamic maximum forces that were predicted, itmay pressures at many points on a surface at a
shape for it. be necessary to change the flow conditions
For both front and rear wings of any on its surface. If it isn’t smooth enough, a a larger area. The most common method
design, the maximum downforce occurs just new wing is called for. On the other hand, of measurement is to drill tiny holes in the
before stall, or the angle at which airflow a glassy smooth wing surface might be im- surface, bond in tubes connected to a multi-
can no longer stay attached to the bottom proved by a boundary layer tripper strip tube manometer, and read the relative water
surface and separates. Downforce drops consisting of no more than a span-wise level heights at speed. More equipment
sharply at that point, and drag rises at an length of tape. However, the need for, or detail and procedure is given in Chapter B.
even faster rate. Standardairfoil data cannot location of, a boundary layer transition can The data can then be plotted on a profile
be relied upon to give the maximum wing only be determined by long and careful of the body in pounds per square foot, as
angle at the rear because the angle of the testing. shown in Fig. 34, based on the relationship
oncoming airstream cannot be accurately Another problem at the rear is that the that 1 inch of water level change equals
determined. The stall point should be ob- struts will always have some amount of flex about 5.2 pounds per square foot of air
vious in performance, especially if it due to loads, and the flex is in a direction pressure. If the pressure were known over
happens at the rear wing first and causes to trim the wing, or reduce its angle and every square inch of body surface, it would
an oversteer condition. But it ispossiblethat downforce. This can be an advantage if the theoretically be possible to average them all
a race Car wing may stall gradually due to wing loads become too great at peak speed. out — both positive and negative, on top and
other interferences such as bodywork, At least it should be considered, as it could bottom — to get the total aerodynamic
airstream turbulence, or attachments caus- cause aerodynamic oversteer at high speeds. forces in any direction. It is much more
ing some wing surface areas to have early If after the most careful development a wing useful, however, to apply the knowledgeof
airflow separation. In such a case, still doesn't generate enough downforce at pressures to locate ducts, vents, deflectors,
downforce instrumentation or tufts attached either the front or rear, or both together, and undesirable interferences. For example,
to the lower wing surface will show the con- then it is time to find a better profile or to itis useful to know the pressure on the body
dition at any wing angle. It is a good idea build a bigger wing. close to a wing. If the pressure changes very
to allow a couple of degrees of marginto Since current racing rules prevent the ap- much when the wing is removed, then the
allow for gusts or the deflected airstream plication of aerodynamic loads directly to wing and body may have harmful pressure
from another car. When a wing is accurately the wheel hubs, it is necessary to allow for influences on each other. At the front of the
tested and doesn’t seem to be producing the the loads in the suspension springs at high
pressure areas will show which surfaces and
airflow in feet per second curvatures contribute to downforce and
Q 200 which should be modified. Once the
pressures are known, any number of pro-
ductive changes can be found.
Velocity profiles are the measurementof
the airstream velocity at various distances
from the body and across air duct cross-
FREESTREAM
sections. The same multi-tube manometer
can be used, except that it is connected to
rows of parallel tubes, or a “rake,” which
is placed in the airstream to be measured.
When placed behind a protuberance or on
the rear of a converging body shape, it can
show at what level the airstream may have
separated. The most common use is to
measure the airflow in a duct, of which a
Figure 35. Airflow Over Body and in Radiator Duct
typical measurement plot is shown in Fig.
35. In this case the data is better represented

68
by velocity m feet per second. which is
equal to the square root of the manometer
Water level in inches times 66. If the average
velocity across the duct is estimated and
multiplied by the duct area in square feet,
the result is flow rate in cubic feet per sec-
ond at 2 given vehicle speed. That in itself
doest mean much. unless it can be related

ing the flow as necessary, if there isnt

enough coolmg or if air drag is wasted in
too much airflow.
Once the optimum rate of airflow is deter-
mined for a cooling duct — and i mayvary
somewhat with changing temperatures— it
iS important tp let nothing inferfere with that The technique of drafting was practiced long before it was scientifically explained
flow. Many race cars have overheated and by wind tunnel studies, which quantified drag as a function of relative car positions.
destroyed their engmes when the radiator
duct picked up trash, newspapers, or even
enough leaves to block the radiator core. A Drafting has been used by racers for years difficulty in staying within a few car lengths.
heavy-mesh screen placed im front of a duct but it wasn't until a study by Chrysler im1971 If two cars are perfectly equal in accelera-
(SAE Paper No. 710213) that any real data tion and they leave a 50-mile-per-hour cor-
yet can be angled to deflect trash away from was available to show how it worked. They ner bumper-to-bumper, the gap will natural-
the duct. A finer mesh screen is necessary also demonstrated the operation of the ly increase to two car lengths at 150 miles
to keep smaller particles out of the engine slingshot passing technique, im which the per hour. The reason can be seen by com-
inlet ducts, as gravel is not only hard on the sega aa ae eel eee: paring their spacing in seconds. If the lead
engme, but can cause the throtiles to The results of the Chrysler study showed car is 15 feet long, then it is 0.2 seconds
jam open. that the air drag of the trailmg car was ahead at 50 miles per hour. But 0.2 seconds
reduced by about one third as it approach- at 150 miles per hour is equal to 45 feet,
RACING ed the rear of another car. The drag on the or a gap of 30 feet between cars. So the trail-
Back m the days when movable leading car is reduced also, but not quite as ing race car must have a slight advantage
aerodynamic devices were allowed on race much. Therefore, two or more cars work- im acceleration to keep closed up until
cars, there was more that could be done ing together can increase their top speed by aerodynamic effects rise enough to help, or
@ considerable amount, as long as they else wait to draft out of a faster corner.
remain im tandem. When two cars pull
abreast, the air drag mcreases on both. The STABILITY INTERFERENCES
slingshot effect comes from the fact that the Another serious interaction is the reduc-
Wings with the maximum downforce angie trailing car can accelerate into the low drag tion in aerodynamic downforce on the trail-
during cornering and braking. and to trim wake ahead. and use that momentum to pull ing race car. The Chrysler study showed the
them out durimg acceleration. Now the out and pass. When the pass is completed. values for large sedans, while the effect on
driver is limited to simply driving the car. the other car is then at the advantage. and open-wheel race cars with wings was
Of course, it is possible to change wing demonstrated in a report by the Jim Clark
angles durmg pit stops. When a stop must Foundation in 1969. In both cases the results
be made anyhow, for fuel or tires, & may relatively equal cars may maintain their show that there are considerable changes in
be an advantage to be able tw rapidly vary positions until the last lap, when both try downforce on both cars, not only in tandem,
Wing angles to adapt to changing track con- to be the trailing car to slgshot at the finish but when they are parallel. What is worse,
ditions. However, it is necessary to know Ime. the changes vary greatly from front to rear,
exactly what wing position is desired ahead But even when you know the science of producing sudden changes in stability. For
of time, for a required change in handling. drafting and slingshoting. there are some example, as the trailing car approaches, its
Since a race is no time to rum experiments. details to keep m mind. First, the trailmg front downforce diminishes, creating
car may have a serious cooling problem, understeer, while the other car’s rear
because not only is the frontal airflow downforce diminishes, creating oversteer.
reduced but itis filled with heat and exhaust There have even been cases where the front
from the leading car. Second is the ques- of the trailing car began lifting, until the car
must be aware of while driving. since there tion of gearing, since a pair of race cars may was flipped over on its top. As long as race
are other race cars on the track to cause ma- end up running hundreds of rpm over what cars require large amounts of downforce,
jor changes in the upstream air. If any two seemed like the correct gear ratio in prac- there is no way to eliminate this effect. But
Cas are within one or two car length of each tice or qualifying. drivers should at least be aware of the pro-
other. they will have a considerable in- The technique is not quite as valuable on blem and possible consequences, and be
fluence on each other's drag (the drafting @ road racimg course because of the prepared for the change in handling.
effect) or aerodynamic lift (stability effect). diminishing effect at lower speeds and the
7
HANDLING
Once the previous chapters are said about an understeering race car at this stability response include wind gusts, sud-
understood, it is possible to combine them point is that it is usually more controllable denly changing track surface conditions,
all into one manageable whole. Tires, than an oversteering race Car. bumps, and the corresponding shifts in
suspension geometry, springs, anti-roll bars, The first distinction to be made is between vehicle equilibrium. The driver's control
shocks, aerodynamics, and other less ride and handling (see Table 3). Ride, or inputs that require a predictable vehicle
significant areas are all related to each other forces and motions in the vertical direction, response include steering, throttle, and
in their contribution to handling. Where are relatively unimportant here. The value brake applications.
some of the factors have both positive and of proper geometry, springs, shocks, and
negative aspects, sometimes undesirable unsprung weight in keeping the tires on the CONTROL
results have to be accepted. But it may be ground has been covered already. With Vehicle controllability, or response to con-
possible to counteract some problems with modern knowledge and technology, it is trol input, is the handling characteristic most
solutions from another area. possible to keep ride somewhat independent associated with race cars, since that is a
Up to this point it was assumed that the of handling, which concerns forces and quality which can be sensed at low levels
terms “understeer” and “‘oversteer” were motions only in a horizontal plane. of performance — even though it is difficult
understood. In everyday language they are The previous explanations of vehicle to evaluate objectively. It is often referred
accepted to mean that either the front or rear components have only considered their to as “feel,” “nimbleness,” “response,” or
wheels lose cornering power first, or that effect on maximum lateral acceleration or “quickness,” and is a function of a car’s
the front or rear slip angle is greater, or that comering power. However, that in itself is suspension, tires, compliance, steering, and
the car either “plows” or is “loose.” To a not necessarily related to good handling. It mass distribution or polar moment. When
vehicle dynamics engineer, however, a com- is quite possible for handling to be so poor it comes to measuring response time,
prehensive explanation would take a chapter that another vehicle with less cornering however, the basic question is: “from what
in itself. The problem is that such terms as power can be driven through a turn faster. to what?” One of the standard methods is
understeer, oversteer, stability, and control The driver is an incredibly adaptable con- to provide a “step” input, such as a sudden
are all highly conditional. You can’t proper- trol system and can usually overcome some steer angle change from zero to exactly “X”
ly say a race car is understeering unless serious handling deficiencies — for a short degrees, and record the time taken for the
many specific conditions are defined. The time. But over the length of a race, it is prac- car to settle out into a steady-state tum.
stability can change drastically between high tically impossible to avoid making a critical However, even the most mushy cars seldom
and low speeds, straightaways or comers, mistake under such circumstances. have a response time of more than 0.5
steady state or transient turns, high or low Handling must be further broken down seconds, while really quick cars can get
loads, and can vary with inputs from steer- into stability and control, or the vehicle down below 0.1 second. This means that the
ing, braking, acceleration, or the wind. response to external disturbances and driver measurement must be very, very precise.
Generally speaking, about all that can be inputs. External disturbances that require a Another complication is that the start and
stop of response is seldom sharply defined,
regardless of what is measured. Turn rate
(or yaw rate) in degrees per second is most
common, although yaw acceleration or the
STABILITY RESPONSE CONTROL RESPONSE rise in lateral g’s are perhaps better sensed
(EXTERNAL DISTURBANCES) (INTERNAL INPUTS)
by the driver. Fig. 36 shows some of the
most common examples of turn rate in-
STEERING eee crease after a step steer input. (Steering is
shown as a ramp increase, because the
ASYMPTOTIC FEEDBACKS wheel cannot in fact be tured instantan-
eously, even if yanked over hard against a
stop.) Curve A is a nicely behaved, slow
responding, well-damped large sedan.
STEADY-STATE DYNAMIC Curve B is a more responsive sports car that
reaches its steady state condition in about
TRANSIENT DYNAMIC AERODYNAMIC one third the time. And Curve C is a car
that is somewhere in the middle, but with
an unrelated problem of not enough
VEHICLE CHANGES
damping. It overshoots the steady-state con-
Table 3. Breakdown of handling terminology dition, and oscillates a few times before
settling out. Obviously it is not easy to
define the end point of these transient

70
 =
— a” == s Maes s
=

= = | "4a 3
Pat 2

= = | z
= ‘ =

= 7 . =
=
= :
|
=
=
=- | z
;

g ZS Fi = z pati
ime
ft ===

S ess

ee See to ie aly oe ee se ie 2 ee Te a a
im == mee Beas See oe wees Ee oie 292 ier mee of ee a Sie
fe seme we ee ne wee Mn SOE ee zie eS eT Ce ee i eT
ene al of ies ieee ae ee eee

Tae. All takes is some aicoation aut

alpeas Suc & oe — 2 fic ese eee te es ee.
ae el cae SE a os ve wae

amet Gages Ov tie dias foe ec a= Of 2 ey est pela ea we

ot hom ane Ter aie Samet 0), ihe resi
or 4am aves a0 a Sess i ave=naer woe fe os =
ap & ie wees For commie wt:
2M wasn Gti = me ot LO

afcoms= Ge lowes
ite cee Be Geehe
=
sami Ger a usualy les eapeve
imo ee. For cee tis Soe

Weenr wie qt

is: ae he =f

=—_a rid 2a Lae

Geutweess @ cme be oe ees ei ba 3 Tene aie
Qemaily. With ite ned meee ae wa a :a pes
pS Gein ae he bw ees we :7 010g nae
ae E mie
125 a] =o és
ee ae oo) § Tar

7
 desirable because it reduces the possibility
DIVERGENT INSTABILITY that the driver’s control could get out of
phase with natural oscillations and increase
instability. A good example is steering cor-
rections in a skid on ice. As the car swerves
back and forth and the driver continues to
correct and then counter-correct, he may get
ASYMPTOTIC STABILITY behind in his response and end up steering
the wrong way at the wrong time — which
time can cause the car to spin.
In addition to the various types of stability,
OSCILLATORY there are four conditions under which
STABILITY
stability must be analysed in vehicle opera-
PATH
INTENDED
FROM
DISPLACEMENT tion: static, steady-state dynamic, transient,
and aerodynamic.
Static stability is important only to
Ye OSCILLATORY INSTABILITY something like a two-wheeled cycle, which
will fall over as it approaches a static, or
Figure 37. Representation of Four Types of Stability rest, condition.
Steady-state stability is most often con-
sidered in the analysis of vehicle dynamics
since it is relatively easy to comprehend and
realized (except with turbos) just how im- signals are apt to be analysed consciously, correct. It assumes that the driver has
portant it is to have a very responsive engine the more hours of experience a driver has, already controlled the vehicle into a desired
reaction to throttle application. But ideally the more these signals become related un- condition — such as a constant-radius turn,
it would be as precise and predictable as the consciously to vehicle performance. Some a straightaway, or constant braking force —
steering linkage. If it were, it would make of the various feedbacks that seem to be of and the vehicle is supposed to maintain that
the rear wheels of a race car as “‘steerable”’ value to a race car driver are: steering wheel condition with minimal corrections by the
as the front wheels. When a car is balanced forces and position, vehicle heading or tire driver. Any external disturbance — such as
on the fine edge of stability during accelera- slip angle, position of the car on the track, suspension deflection in the turn, road
tion out of a corner, it can be disastrous for lateral acceleration and lateral jerk, yaw camber in the straightaway, or one-wheel
the engine to either stumble for an instant rate, ride or vertical accelerations, vibra- lockup in braking — should be auto-
or suddenly turn on and produce a step in- tions, and sounds from the engine and tires. matically corrected for within the dynamics
crease in thrust — which equals a step As yet no one has determined any relative of the vehicle itself.
decrease in rear cornering traction. Engine order of importance, but it is assumed that Even if a race car is developed to have
torque curves should be studied and tailored the greater and more consistent the feedback steady-state stability, as is usually done on
with this possibility in mind. It was just such — within the limits of comfort and tolerance the constant-radius skidpad, it is still possi-
development that Mark Donohue used to — the better a driver can control his race ble for it to be unstable in a transient
make the first turbocharged Porsches so suc- car. Apparently the only feedbacks which maneuver. Applying throttle or brakes, or
cessful. Some drivers have even developed can be modified independently of vehicle turning the steering wheel, may cause
their throttle linkages to provide a slower performance are steering force and position, vehicle response to overshoot the desired
ratio for better feel at the upper rpm, which but it may be possible to improve the com- condition until it goes beyond the point at
gives better control of the axle torque. munication between the vehicle and the which the driver can still correct it. The
Brakes have less effect on the control of driver in other ways. most common example is the race car that
a vehicle, primarily because the, ratio is appears to have perfectly neutral cornering
invariable between all four wheels and STABILITY capability on the skidpad. When driven on
because they are designed for inherent Before stability can be related to the race a race course it will probably be quite
braking stability. But brakes can have an im- car, it must be defined in some of its various _unstable (divergent, with no oscillation)
portant control function in critical situations. types. (See Fig. 37) A stable vehicle can upon acceleration out of a turn because it
When all four tires are locked up solid and have either asymptotic stability (returning loses rear tire cornering power from forward
skid, the race car will tend to continue to its original condition without any over- thrust.
sliding in a straight line regardless of its shoot in the opposite direction) or Finally, there is aerodynamic stability,
direction, rate of spin, or the surface it is oscillatory stability (cycling back and forth which was briefly discussed in Chapter 6,
on, or its steering and supension or failure but converging to the desired condition). An and which can add or detract from steady-
thereof. One less important control aspect unstable vehicle can have either divergent state or transient stability at any time de-
of brakes is a momentary increase in front instability (a constantly increasing response pending on speed of the vehicle and velo-
tire traction. A sudden stab at the brake in an undesirable direction) or oscillatory city of any side winds or gusts. The prob-
pedal may give a helpful load transfer to the instability (a small disturbance creates con- lem of stability as a function of front and
front tires or drop the nose enough for better stantly increasing oscillations until total con- rear downforce was explained as well as
aerodynamic downforce. trol is lost). The condition that primarily possible, considering its complex interaction
To be able to control a race car at the fine determines whether stability will be with load transfer. But lateral gust stability
edge of traction, a race driver needs many oscillatory is the damping in the system. seems, oddly enough, to have inverse re-
precise signals, or feedbacks, from as many Within the limits of acceptable control quirements from mechanical stability. By
sources as possible. Although none of the response, increased damping is felt as previous definition of asymptotic stability,

72
 =
it would be assumed that a side wind should
THRUST AND CORNERING FORCES
have a greater force at the rear of the car,
steering it back toward its previous path. DRAG AND CORNERING FORCES
However, this is felt by the driver as an
unstable or oversteering condition, in which
sudden, violent steering corrections are
usually required. The preferred practice is [Ss SSS /

to have the center of lateral pressure slightly

forward of the center of gravity, so that the Su OVERSTEER TORQUE
car must be steered into the gust. = C.G. 5 ee © SF Coed 4
OVERSTEER/UNDERSTEER UNDERSTEER TORQUE
From the previous analysis of stability
and control, it may be easier to compre-
hend oversteer/understeer in race cars at
the limit of performance. The general
subjectiscovered in somewhat more depth
by the Millikens, although they concen-
trate more on the mathematics than prac- Figure 38. Tire Forces Contributing to Oversteer and Understeer
tical racing applications. For our pur-
poses, we will limit the analysis conditions
to cornering maneuvers--primarily tran- drive, because of so many other factors such torque about the center of gravity either into
sient--at speeds from 50 to 200 miles per as lateral and longitudinal load transfer, the turn (oversteer) or out of the turn
hour, and in response to control inputs by suspension deflection, tire characteristics, (understeer). A breakdown of each set of
the driver. Straightaways, disturbance path requirements, and so on. The analysis forces shows that:
inputs, and steady-state turns may be im- is so complex, in fact, that it is beyond the A. Rear tire thrust detracts from the tire’s
portant. but not so much to the race driver. comprehension of any human mind, and ultimate cornering capability, contributing
The mechanical setup of a race car will there are only a few computer programs to oversteer.
determime its basic oversteer/understeer which can adequately predict transient cor- B. The rear tire thrust required to main-
characteristics in a steady-state turn at the nering behavior. A joint venture between tain constant velocity or to accelerate the
limit of traction. But however finely the car General Motors and Calspan Laboratories vehicle can have either a neutral effect (with
is balanced, it still requires constant instan- has produced a program which can ac- an open differential), or contribute to
taneous steering motions to imcrease or curately simulate a vehicle and a driver. (See oversteer when greater thrust is available at
decrease the front cornering power, which Chapter 15.) But such programs tend to be the outside rear tire.
corrects for normal inconsistencies in the so large and unwieldy, and require so much C. Tire load transfer to the rear increases
track surface. Braking forces may also upset input data, that they are usually impractical rear tire cornering capability, contributing
the cornering balance, although the brakes for real-life predictions of race cars. The to understeer.
may be balanced so as to simply cause the quickest and most reliable results still come D. Tire load transfer off the front tires
vehicle to lose lateral traction evenly be- from actual road testing. decreases cornering capability, contributing
tween the front and rear wheels. This Stull, it is useful to know what effect some to understeer.
increases the comering radius without individual vehicle components and E. Front tire cornering drag forces con-
necessarily requiring further steering characteristics have on transient behavior tribute to understeer because of the greater
corrections. under acceleration. Chapter 4 (springs and load carried by the outside tire.
The balance of a race car is apt to be anti-roll bars) showed how the distribution F. Smaller front tires have a smaller coef-
upset completely however, by applying large of vertical loads on the tires and a ficient of friction than the rear tires, con-
acceleration thrust forces at the rear driving knowledge of tire characteristics could be tributing to understeer.
Wheels. This may seem obvious, but it used to understand steady-state cornering. G. Greater lateral load transfer at the front
points out the fact that most high-powered The use of springs and anti-roll bars, plus tires diminishes their combined coefficient
Tace cars can be made to oversteer at lower the effects of changing tire sizes and com- of friction more than at the rear, con-
speeds by merely applying enough throttle pounds, plus variations in suspension tributing to understeer.
— which has practically nothing to do with geometry make it relatively easy to produce It would appear that the majority of fac-
the chassis setup or basic oversteer/ any degree of steady-state oversteer/ tors contribute to understeer. However, the
understeer characteristics. It is not un- understeer desired. If maximum cornering magnitude of each factor is more important
common for a race driver to get in a car speed on the skidpad were all that was to consider, as each is added for an over-all
with a great deal more power than he is necessary, itwould be easy. But the car must net torque about the center of gravity. It is
accustomed to and claim that it oversteers be balanced so that it can be accelerated out also important to know how the magnitudes
simply because it is easier to break traction of that state without becoming unstable. change with an increase in rear tire thrust
at the rear tires. If the car had four-wheel To understand the transient problem, it as the car is accelerated. To find this, a
drive, it would appear to be relatively easy will be necessary to return to a tire load given vehicle is assumed to be at maximum
to have the application of thrust at front and illustration (see Fig. 38) and make a further lateral acceleration at some given speed
tear wheels balanced — as with the brakes analysis which can be varied with rear tire (which would affect only aerodynamics and
— to have a neutral effect. thrust. By careful consideration of all the perhaps tire characteristics), and calcula-
The problem is nowhere near that simple, forces and their locations and directions, it tions are made to find the actual torques
however, with either two- or four-wheel can be seen that they all contribute to a generated by each factor at a number of in-

ux
 the effect ofaerodynamic downforce to each

“|
A(+) tire’s cornering capability as speed in-
creases, especially if the front/rear ratio
changes significantly. But this example is not
1500 intended to be absolutely accurate even for
OVERSTEER the particular race car used. All it doesis
illustrate trends of the different affects on
1000 transient vehicle response to control inputs.
Even by using this technique it may not be
vehicle
ound possible to develop a perfectly neutral con-
500 trol response in a two-wheel-drive race car.
but it can probably help to delay the
oversteer phase as long as possible with a
minimum of steady-state understeer.
turning
torque
ar
foot-pounds
gravity—in
of
center
FOUR-WHEEL DRIVE
500 Although four-wheel drive is not currently
popular on high-speed racing cars, it does
1000 UNDERSTEER
if tire traction seems adequate to handle the
Figure 39. Dynamic Stability Graph engine output and four-wheel drive seems
to be unnecessary added weight, it may add
a great deal to stability — if not control. By
creasing steps in tire thrust. These points understeer. It is necessary to begin the reference to the plot of oversteer/understeer
are then plotted as shown in Fig. 39 along analysis at a total rear tire thrust of about torque versus throttle application, it can be
with the net oversteer/understeer torque. 200 pounds or (100 pounds-feet of torque seen that there will probably be no sudden
The plots shown are for a typical mid-engine at each axle for 24-inch tires), because it leap into the oversteer phase — at least if
trace car of about 1600 pounds with typical takes some thrust to maintain a constant unlocked differentials are used at front and
Tace tire characteristics. Vehicle speed in the comer. At this point it can be rear. In fact, the theory indicates that it
characteristics that must be considered are seen that there is about 300 pounds-feet of could be difficult to keep such a race car
wheelbase, tread, center of gravity location, net understeer about the center of gravity. from understeering. This brings up a great
roll couple distribution, and differential type This means that the front tires must be potential disadvantage: the removal of one
(locked in this case). steered into the turn until their slip angle of the driver’s most familiar control func-
It can be seen that some factors are generates the necessary balancing force. If tions — the ability to create throttle
relatively linear, if not constant, from they are 5 feet ahead of the center of gravity, oversteer. All the driver is left with is
steady-state cornering to full throttle. The they will have to produce an extra 60 pounds steering control, which is changed a great
effects of the front tires, whether from cor- of cornering force. deal by the front-wheel driving effect.
nering drag forces (E), smaller coefficient As tire thrust is increased, the understeer At any fate, it is possible that power-to-
of friction (F), or their greater lateral load decreases gradually, until it reaches a fairly
transfer (G), are hardly changed by rear tire constant level at almost perfect balance, or increase someday to the point where four-
thrust except as diminished by the load neutral steer, where the necessary steering wheel drive becomes practical. Without
transfer to the rear under acceleration. On angle is essentially zero. This is the area of going into a great deal of analysis, it appears
the other hand, the direct consequence of the old familiar “perfect four-wheel drift’” that wheels should probably carry the same
rear load transfer steeply increases the Above this point, however, or at about 1200 Static weight, and roll rates and tire
understeer effect. It is greater at the front pounds of thrust in this example, the rear diameters should be the same front and rear.
(D) because the transferred load is a larger thrust effects rise so rapidly that the hand- If some degree of steady-state understeer is
proportion of static load than at the rear (C), ling characteristics suddenly cross over into desired, it probably should come froma
and the non-linearity comes from the fact the oversteer phase. It may be possible to narrower front tire. However, it is difficult
that load transfer increases with the attitude maintain a balance by steering out of the to predict just what sort of handling is
change of the chassis. But what really upsets turn — if there is enough road to allow for required when the primary feedbacks to the
the balance, as is familiar from practical ex- the increased turn radius — but the more driver are changed so drastically.
Locked or
perience, is the rear cornering loss from productive control input at this point is to limited-slip differentials might not be re-
thrust (A) and locked axle forces (B). The back off on the throttle slightly. Ifnot, the quired, but there should be a drive torque
locked axle oversteer effect doesn’t really increasing oversteer will eventually over- bias toward the rear wheels since they still
come into play until the thrust is so great come the reduced steer angle at the front carry an increased load transfer under ac-
that the inside rear tire capability is totally until the car spins out completely. celeration. The exact torque bias can even
used up. At that point, increasing forward Of course, thrust available is going to be be predicted mathematically, depending on
thrust at the outside tire rises dispropor- limited by engine capability and will change such factors as static weight, tire
tionally. Finally, by far the greatest effect with rpm, gear ratios, and vehicle speed. characteristics, wheelbase, and center of
comes from the actual cornering traction At high speed in a high gear, the thrust gravity location. In fact, I had to make those
loss, (A) which contributes to oversteer available may be less than 1200 pounds, and calculations back in 1968, to resolve
from the very beginning. therefore the driver may go to full throttle whether Chevrolet would try to design a
But it is the net torque around the center in the middle of the turn without creating four-wheel-drive Can-Am car. More re-
of gravity that produces oversteer or oversteer. It would be a good idea to add cently, Porsche re-worked the problem in

74
 (on full-fendered cars), the weight
is in- have been tried — especially the differen-
clusions are that the ideal front/rear torque creased by greater strength requirementsin tial type — then it may help to increase the
split is 2 variable which depends on m- the suspension, and the cornering turn static front weight.
Stamtancous load distribution and road co- radius and passing ability is made worse When it comes to changing the weight
for a given race track width. It would distribution, some race car components are
Porsche is able to utilize a controllabledif- appear that — within reasonable limits — easier to relocate than others — but the
ferential bias- the net effect still balances out that wider easier ones to move are usually the less ef-
With current race car power, weight, and is always better. However, the geometric fective ones. To get an idea of the effect of
tires, these mathematical predictions m- relationshipsare such that increasing track a relocation, the center of gravity will move
dicate that on dry pavement four-wheel width has a dimimishing advantage. In other a distance which is equal to:
drive is negligibly faster m a steady-state words, making it wider by one inch won't (component weight) x (move distance)
turn. And although it does accelerate a car help as much as making it narrower by one (vehicle weight)
out of the turn much faster im low-speed inch will hurt. For example, moving a 20-pound battery
The same sort of relationship is also true 100 inches from the engine room to the
of center of gravity height, since it is the trunk on a 2000-pound sports car will move
other major factor in lateral load transfer. the center of gravity (100)x(20)/(2000) or
In this case, of course, lower is always one inch to the rear. Another way to change
better, except where there is too great a cost weight distribution on a race car is to move
im ground clearance, weight, or driver the front wheels forward or rearward, but
straightaway, four-wheel drive would have comfort. If there is a conflict of interests unless done in original design, too many
to produce a very strong advantage m cor- here, center of gravity height may be the other effects can be upset.
meting acceleration. It also adds work to less significant consideration. That is not Wheelbase seems to be more closely
be done by the brakes and has a complex- to say it should be ignored, because it also determined by component locations than
ity which afiects reliability. With the has an important effect on longitudinal load vehicle dynamics considerations. The lower
transfer, but some priorities must be kept limit is the necessity of fitting the engine,
in order. In recent years, it seems that driveline, and driver between the axles, and
be some time before it becomes popular in racing tires have become less load sen- the upper limit is the increasing weight of
road or track racing. sitive, so that lateral weight transfer is less chassis material needed to resist beam and
The fact that 4WD is currently doing of a problem, and therefore center of torsional loads. Within those limits there
well under low-traction conditions such as gravity height and track width are less doesn’t appear to be any significant control
rallymg however, indicates that it may not important. or stability effect, except as related
to center
be long before someone tries it m road Proper longitudinal center of gravity of gravity height and longitudinal load
racing again. Tt has certainly become location, or front/rear weight distribution, transfer. Some aerodynamic regulations,
popular om the street, especially as a is a little more difficult to analyze. Based however, have created an artificial problem
relatively mexpensive addition to front- on the consensus of most experienced race on open-wheeled race cars. In order to
wheel drive cars, even though the typical car builders and drivers, the optimum for reduce the interference of rotating tires on
driver may not really need it more than a any rear-wheel-drive race car seems to be wing aerodynamics with a limited-length
few times a year. As the analysis above anywhere between 60 and 70 percent rear car, the wings are extended well ahead and
showed, more important than its ability to weight bias. It should be apparent at this behind, while the wheelbase is shortened.
pull out of the occasional low traction pomt that “rear weight bias causes When the downforce is concentrated at ex-
crisis, is its value im stability during sud- aversteer™ is a total myth, considering all treme ends, the effect of an aerodynamic
den surprises, where practice isn’t possi- the other factors that have an overwhelming imbalance on the shortened wheelbase is far
ble— such as rallying. influence on control and stability. In fact, more significant.
rear weight bias partly reduces oversteer,
HARDWARE im that it mereases the acceleration traction DEVELOPMENT
Most of the components or systems that capabilities at the rear tires. It also in- Optimization of total vehicle handling in
affect race car handling have been dis- creases Steady-state cornering capability in control and stability is a relatively
cussed im previous chapters. But there are that it decreases the cornering drag due to straightforward matter of following all the
some over-all vehicle design factors that load on the front tires. The upper limit to basic sub-steps from previous chapters and
don’t fall im any of those categories — such rear weight bias is the point at which load putting it all together as a package. The
as track width, wheelbase length. and transfer to the rear under acceleration desired result is to have a vehicle that is
center of gravity location. Each
of these is makes the steering unresponsive. As the asymptotically stable in both steady-state
@ very important consideration im basic wheelbase gets shorter, acceleration trac- and transient cormering and one which also
has quick and consistent response to driver
higher and farther rearward, there comes control inputs. It would also be desirable to
@ point when the front wheels will be lifted have this degree of stability and control
@ number of positive and negative contnibu- clear off the ground. remain constant with increasing speed, but
tions, which makes it hard to predict the The amount of dynamic load required on that may be asking too much, considering
the front tires under the worst conditions the effects of aerodynamics. However, it is
As vehicle track width grows, lateral load depends on a large number of other race car theoretically possible, as discussed
transfer in cornering is reduced, which in- factors. But when it is difficult to steer a previously, by having mechanical fixes for
mechanical stability problems and
Sion deflection problems. But at the same aerodynamic fixes for aerodynamic stabil-
ity problems.

75
 be throttle-oversteered at low speed, but
what is being tested is the stability at cons-
tant throttle. To eliminate steady-state
oversteer, it will be necessary to increase
cornering power at the rear or otherwise to
change the torque about the center of grav-
ity. Rear comering power may be increased
by larger tires, softer compound, higher
pressures, Or more camber. Assuming that
those have already been optimized, the next
Step is to change the roll resistance distribu-
tion. By using a stiffer anti-roll bar or stiffer
springs at the front, more load transfer is
carried by the outside front tire, tending to
rotate the car out of the turn. The car should
begin to understeer well before the front roll
resistance is so great that the inside front
tire lifts at maximum lateral acceleration.
If not, then it will be necessary to reduce
front tire comering power as a last resort.
Any car which is to be driven on a road
course should be frequently tested in alter-
nate directions on the skidpad. If it isn’t
equal in left and right turn performance,
then something is out of symmetry. Original
suspension alignment and balance should
have taken care of obvious problems, but
there is also the possibility that the effect
of a live rear axle, or driver weight, or even
fuel movement could affect the dynamic
balance. If so, it may be necessaryto
slightly pre-load the anti-roll bars diagonally
on the chassis until the maximum capability
is equal in either direction.
It is unlikely that the car will ever have
truly neutral steer characteristics on a skid-
pad. If itappears to, the driver should speed
up slowly, with minimum throttle change,
to a new constant speed — without moving
the steering wheel. If the car really does
have perfectly neutral response, it will move
out to a larger radius with the same center
and without any change in attitude.
If a race car understeers on the skidpad,
it is still important to know how much.
Assuming that front comering capability is
7a s
at the optimum from tire and suspension
The same car with different anti-roll bars, set up to oversteer (top), or development, the rear anti-roll bar rate
understeer (bottom) at a constant speed on a 100-foot radius skid pad. should be increased in steps to determine
Extreme steer angle in the latter case indicates excessive understeer. how far the handling is away from neutral
steer. The steering wheel angle is also a
rough indicator of understeer. If the wheel
Unless a racer has access to a computer proven to be more effective in most cases, is at a large steer angle and it can be tumed
simulation of total race car dynamics, such the older step style will be disregarded. even more with a negligible change in vehi-
as the expensive and complex General The first stage in development is to set cle attitude, then the understeer is too great.
Motors/Calspan model, a skidpad and race the car up with some degree of understeer But the ideal amount of understeer can only
track will be necessary for final handling on a relatively low-speed skidpad, or a track be determined by hard transient accelera-
development. It will also be necessary to with a long, constant radius corner. The tion out of this steady-state condition. If the
distinguish between two styles of driving: track and driving requirements for steady- skidpad has tangential access roads, lane
the step style, in which braking, cornering, State tests are further detailed in Chapter 13. markers can be set up to simulate a typical
and acceleration are all done in separate If, at maximum speed on a constant radius, turn. Otherwise, it will be necessary to go
discrete steps; and the blend style, in which the car oversteers in either direction, then to a race track and try the car under actual
either accelerating or braking is blended it will be practically undriveable on a race conditions.
into the comering phase. Since the latter has course. Of course, nearly any race car can Changing the car’s oversteer/understeer

76
TAREE SE he tedoad wil sve
Ciamiidive tae mn he He of vations
mgs. Disease ages a the race
Wack temid heatiore se ere orice
if Saset on these tet figeues. t preverts

etigurazion hat is now mh werstess

Tea etine = cemsive aid aivety
ia7m-rms. Bt heomidSe lated wo Om
oman Sf vxioms awed degrees of

& cognition
of drives tt and accurate.

Be casei oi alow nat secclecation ont

fron the apex of 2 omnes, ad yer mot
pimmese
ieet ee car cat Ne Seecextims the
COTES Or Teka = ett love? of seat eae:
iad ceceration_ Whe esi Gaia wil

commses— Tent one walt 2 cumier

of igh & Sadly sverstesrime car going ont of control at the aitry to 2 form
Ted CREASE TS i we = Even nil nex comnter-tesring fv 2 ddillet driver comidn’t ave it
cumier of Sow comers ad lone
maemawaes. The immer wonid oropabiv
TES 2 SS Closex eh ected steer tires gnngiv dependent on sismension firmness
aig tire stiffness that if racing mspeisons
ered 1 tewid te qentified or fore Saiancet io fave 2 reiaive?ay pi mig mes tres ae ised_ © is ant os ope
fees on Tadime, aud only 2c z very last he veticie response. If for some reason z
resent tamid hes ee useciee comrect for Tateroe isdesred_ aout ail that can
Ramil deficiencies: Teedo iS i Merease the steering ratia,
Terps: on the past me cheagiess: as TETAS Ca ie, vee merrol agi.
mid Sitten the tock asarhers. Fverstinns
Of come. tus: amaivers: of Gadi oniy Se af is open
eve? of Handle perio cise Tas: Neen determined
fir optimus vein
CEES Exits ames ant a de caer, a tnty these for Hevors cast
Saar. See Se Gb CEE ipa Ges mer Ges midified ir 4 tangs it ramsient fai
CveTE if deqeit Gere Gecause that is seit taces of the driver wiheur upsetting vei
ae a aa me 2 Dmbient i macs ce develommat. fis a ae tie nea tess
8
ENGINE SUPPORT SYSTEMS
Although the race car engine may be middle of a race is no bargain. However, shops can properly develop an engine for
one of the most important components to probably 90 percent of all race engine maximum performance. There probably
be engineered and developed, it will be failures are due to owner ignorance of the aren’t a few dozen such race engine builders
discussed here merelyas a purchased item. proper use and maintenance. Every driver in the United States, including the people
The development and assembly of all-out and mechanic owes it to himself to know who develop components at the speed-
racing engines has reached the point of everything possible about the care of such equipment manufacturing companies. A
scientific art where it is impossible to do a highly stressed component. Unless the racing engine runs on such a fine line be-
justice to the subject in less than a book. In engine is to be used once and then returned tween ultimate performance and total failure
fact, the subject is already well covered in for a rebuild, the builder should supply a that no science is precise enough to beat
anumber of books and countless magazine lot of maintenance information: ignition trial-and-error development. It will always
articles from the enthusiast press. The timing, advance curve, cam timing and lift, be possible for someone to get more power
best hardcover engineering textbooks will valve gap, spark plug heat range and gap, out of anyone’s racing engine — but the big
be found in the SAE Bookstore catalog, internal clearances, bolt torques, oil type, question is for how long. It is very difficult
although they will be oriented toward nar- and oil system layout. It is also a good idea to compromise performance and reliability,
row specialties in production engines, and to get the dynamometer curve for that par- which makes development a long, slow pro-
the Classic Motorbooks catalog. The best ticular engine, for use in selecting gear cess of try, break, and fix. You raise the
examples of practical race engine assem- ratios. Once all this data is at hand, it is power or the peak rpm and run the engine
bly, tuning, and maintenance for Ameri- possible to tear an engine down to check until something fails. Then you strengthen
can stockblocks are probably the paper- clearances and inspect components a that part and run until something else fails.
back “‘How to...’’ books by H.P. Books, number of times before a complete rebuild When all the components are strong enough
and the manufacturer’s race preparation is required. One over-all tip not mentioned to last a race or two, then it is time to search
publications. Although written for spe- in most engine texts is that the pistons, rods, for more power. It becomes a continual em-
cific engines, they are full of information crank, and block all need to get “‘friendly” pirical struggle until every part is working
applicable to most racing engines. with each other before they will develop at its absolute limit. It costs a lot of broken
Because of the finesse and precision maximum horsepower. The longer a short- engines to get there, and it is better to do
necessary to build a race engine, not to men- block assembly can be kept alive—while the it on the dyno than on the track.
tion the equipment required, most racers heads and other external components are Most production engines will run forever
will end up buying an assembly complete rebuilt—the more powerful the engine will be. if you keep oil and water in them — and
from the induction system to the oil pan. There are a number of shops that can do keep the rpm low enough. Unfortunately,
The finances available will probably deter- a fairly good job of building or race- racing usually requires the highest rpm
mine the source and quality of the purchased preparing an engine, but only the most possible to produce the maximum power
engine, but one that self-destructs in the experienced and professional dynamometer output. Component stresses rise exponen-
tially with the increase in rpm, so an engine
might last 100 hours at 6000 rpm, 10 hours
at 7000 rpm, and less than one hour at 8000
rpm. That’s why it pays to have an accurate
tachometer and to keep the rpm as low as
possible to win the race. But there are a lot
of other causes for failure that aren’t peculiar
to racing.
It may help to list all the possible reasons
for engine failure and to try and avoid it with
warning devices or fail-safe systems. The
most common reasons for failure probably
are: overheating due to high ambient con-
ditions or an inadequate cooling system; oil
starvation due to loss of oil or a poor oil
scavenge system; detonation due to im-
proper timing or low-grade fuel; fatigue
from vibration or excessive stress reversals;
and electrical component overheating or
breakage. There are of course an infinite
A smoking engine is one of the most discouraging sights for a car owner, and number of other possibilities, but this
sometimes a mixed blessing for the driver behind. chapter will concentrate on the optimization

78
of five categories of engine support systems:
the fuel system, the oiling system, the
cooling system, the electrical system, and
the exhaust system.

FUEL
The fuel system for a race car is not quite
as simple as running a hose from the tank
to the fuel pump. Fuel starvation can mean
decreased performance or total stoppage, or
even destruction of the engine if it leans out
and detonates badly. When restricted fuel
capacity is a problem it can also be impor-
tant to be able to pick up the last quart of
fuel remaining in the tank.
During early development driving, the
fuel pressure gauge should be frequently
checked, especially during acceleration out
of hard corners with a light fuel load. The
gauge ought to read a minimum of 4 psi,
and if it fluctuates or drops to zero even for
an instant, big problems are sure to follow.
The most likely sources of trouble are the
pickups in the fuel cell. There should be a
minimum of one pickup in each rear corner a

of the tank, with each hose leading to a Complex plumbing in a formula car, with twin radiators, twin fuel
separate electric fuel pump. This way, if cells, fuel reservoir, dry sump system, and automatic fire extinguisher.
either runs dry under lateral acceleration
and starts pumping air into the system, the
other will overpower it. It is even a good
idea to wire each pump to a separate switch leaks or breaks in the line. Steel tubing is really long hoses may require a larger
on the dash, so that an offending pump can satisfactory anywhere except to the pumps diameter.
be cut off if necessary. Check valves ought and to the engine, where vibration and flex A final word on the fuel system: find the
to be added to the system to prevent could cause a rigid hose to crack. Needless absolute best grade of fuel available or at
pumping fuel back into the tank through a to say, all connections should be shielded least the same grade that the engine was
failed pump. and protected. If adequate pressure can be dynoed with. Due to exhaust noise, it will
If, after all that, the pumps still pick up maintained all the way to the carburetor be difficult to detect detonation — or
air, it will be necessary to install pickup (where the pressure should be measured), pinging — before the engine is badly
reservoirs in the tank. A trap can be built then 3/8-inch I.D. hose is large enough, but damaged.

1 4eES
around each pickup tube which allows fuel
to flow in during lateral or longitudinal
acceleration but prevents it from flowing out
through flapper valves. Even a quart capa-
city in each corner will maintain fuel
pressure until the fuel level evens out again.
Race cars with multiple cells must also be
equipped with surge tanks. These are
designed to collect a gallon or so of fuel
through one-way valves from whichever cell
is upstream at the moment. In such case,
a single pickup at the bottom would prob-
ably never run dry until the last ounce of
fuel had passed, and one outlet pump would
theoretically be adequate. But where weight
is no great problem, it can’t hurt to have a
spare second or third pump plumbed into
the system. Mounting the pumps as close
to the tank outlet as possible improves their
performance (it is easier to push fuel than
to pull it) and helps prevent vapor lock from
low pressure in the lines.
With that great length of pressurized fuel The necessary fuel plumbing for an ordinary sedan includes multiple pickup hoses
line to the engine, it is important to prevent to the bladder, pumps, filters, check valves, and a surge tank.

JOHN MOORES UNIVERSITY 79

ul VERPOOL. EARNING SERVICES
 more subtle advantage is that the stock oil
pump is removed. The pump has a tendency
to wear its drive gears and thereby retard
ignition timing, since the distributor is also
driven off of it.
The disadvantages of a dry-sump system,
in addition to cost, are all the external com-
ponents that can fail or be damaged. The
external pump requires a toothed gearbelt
which, while more reliable than a V-belt,
can still be broken. And there are a number
of extra hoses, fittings, and containers re-
quired, all of which have leak potential.
The design of a proper dry-sump oil pan
is the responsibility of the engine builder if
no pan is commercially available. The pan
will have at least two scavenge pickup fit-
tings that must be routed to separate stages
in the scavenge pump. The oil is then routed
to the reservoir. The reservoir should have
a capacity of perhaps twice that of the
previous wet pan and be as tall and narrow
as possible, to improve de-aeration. If the
inlet enters tangentially at the top, the swirl
will centrifuge out most of the air. Oil level
Dry sump pump and toothed drive belt on a sedan engine. Note in the reservoir should be higher than the
that the water pump belt has no idler, and rides in deep-groove pulleys. pump to aid in priming it, and cone-shaped
to the outlet.
Oil leaving the reservoir goes back to
The Aeroquip flex hoses with braided the rocker covers under long, hard lateral another stage in the oil pump stack and is
stainless wire sheathing that are common accelerations. In that case, external drain- pressurized for the engine. Only at this point
to all professional race car plumbing are backs will be necessary between the valve should oil coolers and filters be plumbed
fairly expensive if bought new. Fortu- covers and the pan. Check valves will prob- into the system. Again, it is easier to push
nately, they are sometimes available ably be required to keep the oil from a fluid through restrictions than it is to pull
through surplus houses at prices that are flowing up the hoses at times. it. Some engine builders recommend plug-
reasonable for amateur race car builders. Vent hoses and catch cans are also af- ging any filter bypass valves, which will
They add such security in strength, dura- fected by lateral accelerations. It is very easy require careful monitoring of the oil
bility, and leak tightness that even consid- for the lateral forces in a 1.1- to 1.4-g corner pressure gauge (which must measure oil
ering the cost, nothing else comesclose. A to tilt the oil level until it starts siphoning pressure at the engine inlet). When the
catalog from Aeroquip or Earl’s Supply in out a vent. Or it may clog the vent until pressure starts to drop, a clogged filter
Lawndale, California, will give all the blowby pressure starts pushing oil up the should be considered first. Filter cans
information necessary on types and sizes line. Most sanctioning bodies require the should also be carefully taped or safety-
for each application, varieties of connec- vents to run to oil catchcans to protect the wired, as they have been known to vibrate
tors, and method of assembly. While such track and other cars. The best vent line loose with disastrous results. The oil cooler
hoses are a good idea for protection in the routing for V-8’s is to run all hoses to the should be designed for minimum restriction
case of low-pressure fuel lines, they are top center of the engine, and then run a and be able to take the kind of pressures a
mandatory for high-pressure oil lines and single large-diameter hose from there to the racing oil system can generate.
flexible teflon brake hoses. can or the dry sump tank. This will allow The wrong kind of oil can destroy an
most of the oil in the lines to return to the engine just as effectively as no oil at all. The
OIL engine by way of another vent. first consideration is to use an oil with
The oiling system in a race car is seldom The problem of tilting oil levels can only ashless additives such as designated aircraft
restricted to the inside of the engine. At the be properly solved by using a dry-sump or racing oils. Some additives in passenger
very least, it will include an external oil system. But there are a number of other ad- car oils create deposits in the combustion
cooler and perhaps remote oil filters. Since vantages to a dry sump. The oil pan can be chambers which can cause preignition. The
oil pressures and temperatures are quite high made much shallower, which allows the distinction between aircraft and racing oils
and the consequences of loss are a ruined engine to be lowered. It also permits a large is that the latter usually have anti-foam
engine, great care is required in plumbing. remote reservoir which can de-aerate the oil additives to help keep the oil from aerating
Half-inch I.D. hoses are the minimum, and and which has a more positive single pickup at high rpm. If an engine is properly
the more filter area provided, the less point. But most important, it can add as clearanced and not running too hot, straight
pressure lost. It may be necessary to tape much as 20 horsepower by keeping the oil 30 weight oil ought to be adequate. As the
up the oil cooler at times to maintain at least level away from the crankshaft. The oil in bearing clearances wear open slightly or ex-
200 degrees, but if things get too hot, a pair a wet sump is constantly whipped around treme temperatures are encountered, it may
of coolers should be used in parallel — not by the crank, which aerates the oil and be necessary to go to 40 or 50 weight oil
in series. Some V-8 engines may trap oil in causes hydraulic power losses. Another to maintain pressure. However, heavyweight

80
oils require careful warmup as they may be
too viscous to circulate in a cold engine. The
only time that supplemental oil additives
should ever be used is during break-in. New
components with very high contact
pressures may require a coating such as the
molybdenum disulfide compounds, but the
oil should subsequently be changed before
racing.
WATER
The water-cooling system of a racing
engine must have a capacity that is propor-
tional to the horsepower output. If a given
system is adequate for a production engine,
running twice the power through it will
require about twice as much effective
cooling. However, since there is much
greater airflow, and a lot of cooling is
handled by convection and radiation around
the engine itself and through oil coolers, it
Engine room plumbing on a sedan includes top-centered crankcase vents,
may not require a doubling of the water
a coolant recovery can, and steel braided hoses for all high-pressure fluids.
radiator size. But it’s worth remembering
that the cooling system is as dependent on
power output as it is on ambient ally required to help keep water off the the kind designated for high-rpm racing use,
temperatures. track, it might as well be plumbed in as a with reinforcement in the vee to help pre-
Almost any radiator a car comes equipped coolant recovery system. Running the vent turning over. Deep-vee pulleys should
with ought to be sufficient if the airflow to overflow tube to the bottom of the catch can also be used to prevent the belt from walking
and from the core is adequate and carefully and fastening the pressure cap closed pro- off, and it is a good idea to fit the water
routed, as discussed in Chapter 6. But the vides a reservoir for water which may ex- pump belt tightly around only two pulleys,
radiator is a very vulnerable component pand or boil out past the cap. As the system with no idler pulley which could loosen or
which should be carefully protected from cools, the excess water will be siphoned get out of alignment.
puncture and vibration. A wire mesh screen back into the radiator.
(and no sharp projections) ahead of the core Pressure caps and anti-freeze solutions IGNITION
and rubber-insulated mounts are mandatory. won't improve the cooling any, but they can The electrical system is often overlooked
All hoses should also be isolated from con- be a valuable aid in preventing water loss. because it offers little potential for making
tact with the ground or sharp edges and A 21-pound pressure cap will raise the atace car faster. But it can very easily cause
insulated where they may pass the driver. boiling point from 212 to 260 degrees, and the car to run poorly or not at all. The first
The relatively low pressures in the cooling a 50 percent solution of ethylene glycol will consideration is the engine’s ignition system,
system don’t justify braided-wire hoses raise it further to 275 degrees. Of course, consisting of plugs, wires, distributor, and
except where protection is worth the cost. an engine allowed to run that hot for very coil. Ordinarily the engine builder is
But the effects of water loss are serious long is going to have other serious prob- responsible for most of that — particularly
enough to require great care in all connec- lems, not to mention the power loss at such the distributor advance curve and the proper
tions of any type. Only the best non-ribbed temperatures. But it is better than losing spark plug heat range. However, since the
hoses are durable enough and have low water. The best example is an unusually long spark advance can easily deteriorate over a
enough restriction for a race car. For long, pitstop where the temperature may rise into race weekend, it should be checked
straight lengths, rigid aluminum tubing pro- the danger zone for a few seconds and yet frequently against the builder’s
vides increased cooling surface, but it must cool down rapidly when the car can get specifications.
be isolated from chafing and vibration by moving again — if all the water hasn’t been There may be a choice in the ignition
flex hoses at each end. Only worm-gear dumped out. If a 21-pound pressure cap is system between: ordinary breaker points and
hose clamps are secure enough, and even going to be used, it is essential to pressure- coil; a transistorized system; a magnetic
then they should be tightened while hot and test for leaks in the system before the car pulse distributor; and/or a capacitor-
safety wired. All metal tube-ends should is taken to a track. discharge system. (Modern computer-
have a bead to prevent the hose and clamp Water pump belts are as important to controlled ignition systems are discussed in
from slipping off. Where long cooling water finishing a race as the tires are, and yet Chapter 15.) When all are working at their
hoses are routed through a car, it will receive far less attention. The most efficient optimum, there is very little difference —
simplify filling if air bleed valves are pro- and durable belts are the toothed gear belts for a short time. But only the magnetic pulse
vided at the high points, but they must also ordinarily used for external oil pumps and distributor can avoid rapid deterioration due
be fail-safed. A commercial leak-inhibitor timing drives. They are so effective that they to point bounce at high rpm and rubbing
solution might come in handy in the middle may be narrower and lighter than the block wear which causes spark retardation.
of a race, and a corrosion inhibitor is equivalent V-belt required to handle the The magnetic pulse distributor will probably
worthwhile for an aluminum radiator or same horsepower. Their cost, however, has stay in adjustment for the life of the engine,
aluminum engine components. kept them from completely replacing the V- while breaker points should be checked for
Since a water overflow catch can is usu- belt. At the very least, all V-belts should be proper dwell and advance as often as

81
 The vehicle wiring system can be as
simple as an on-off switch, fuel pump
switch, and brake lights, or it can be as
complex as the wiring in a production car.
In an endurance racer it will be. But no
matter how simple or complex, all wires
should be numbered or color-coded during
assembly and a complete wiring diagram
sketched up for future reference. The
minimum wire size should be 14 gauge,
with a heavier gauge for high-load com-
ponents such as fuel pumps. All wires ought
to have thermo-setting insulation if they
come in contact with other wires or hot
surfaces, to prevent multiple melting and
shorting out. For large cars with lots of
wiring, or cars with lights on removable
body panels, multiple-pin connectors will
make it easier to disassemble the harness.
To keep them from being accidentally
yanked apart, they should be backed up with
a solid wire cable. Every wire connection
should be protected in a number of ways.
Eye-type terminals are both crimped and
soldered to each wire end and the exposed
metal protected with vinyl shrink-tube
An excellent example of wiring, with silicone-sealed terminals, tie- sleeves. The terminal is held onto its stud
wrapped wires, looms for plug leads, and easily changed components. with a lock-nut, and the whole connection
coated with liquid silicone sealer. The seal
not only insures that the wire doesn’t vibrate
possible. On the other hand, a magnetic particular cylinder with specific troubles loose or break off from metal fatigue, it also
pulse system with capacitor discharge may such as oiling, wire failure, or valve protects the terminal from short circuiting
be prone to sudden and total failure, and it problems. against other metal components.
is not easy to diagnose problems or to repair Assuming that the engine electrics are Just in case a wire gets pinched or
them rapidly. It is wise always to carry a well-selected and carefully tuned, they still crimped by accident and shorts out against
complete replacement setup. must be protected from their hostile en- the chassis, it can be a great help to have
Spark plugs can tell more about the vironment. As the saying goes, “‘electricity circuit breakers in the system. It won’t repair
engine than the driver could possibly detect. never fails — it’s always a mechanical the wire, but it is an immediate signal of
Reading plugs is an art that must be ac- failure’’ The most common problems are a problem and an indication of where the
quired by direct observation. It is practically caused by heat, vibration, and corrosion or problem may be. It also prevents the entire
impossible to illustrate in photographs just shorting-out from moisture. Most tran- wiring system from going up in smoke, so
exactly what the porcelain tip looks like sistorized components and magnetic devices that once the short is found, no other wires
when everything is running right, or when are vulnerable to heat and should be located have to be replaced. Fuses would do the
disaster is imminent. A spark plug engineer in a cool area or have cool air ducted to same job, except that an instantaneous
at professional races can probably give a them. It is also wise to have such black-box overload out on the track could blow one
good demonstration, or the following ap- components easily accessible and replace- — say in the fuel pump system — and pre-
proximate description may help. able in case of failure. All electronic devices vent the car from getting to the pits unless
Normal: the porcelain has a light tan, or — except the distributor — should be rigidly a spare fuse is at hand. Needless to say,
gray, or very light brown color, particularly mounted to the chassis to reduce vibrations. there ought to be a healthy amperage
at the base, with no specks or deposits. In many cases, all internal electronic com- margin, even in circuit breakers.
Overheated (or lean mixture): the ponents can be “‘potted”’ into a solid block When all else fails, and the electrics give
porcelain is chalky white and the electrode of epoxy, which eliminates vibration fatigue up in the middle of a race, it might help to
is apt to be noticeably eroded or melted. and improves heat transfer. have an emergency backup system. The ex-
Fouled (or rich mixture): the porcelain The spark plug wires should be carefully tra wire weighs next to nothing, and even
and electrodes are coated with black, sooty routed through non-conductive and rigid spare onboard coils, amplifiers, and voltage
carbon. wire mounts, but never in contact or parallel regulators could be worth their weight in
Any reading will be meaningless unless with each other for any distance as cross- getting the car back to the pits. The only
the engine is first run hard with a heavy load firing could result. The best all-metal plug component that can’t easily be backed up
at full throttle and then ‘“‘cut clean,” or im- wires should be used in place of carbon is the distributor. Of course, a master shut-
mediately turned off and declutched. The resistance wires since radio reception won't off switch is usually required, and it can also
reading should indicate any serious problem be a problem. Each wire should be fail or be turned off accidentally. Therefore,
with fuel mixture, spark advance, fuel numbered at both ends to speed up trouble- the duplicate wiring system should begin at
grade, or plug heat range. Comparison of shooting. All connections can be protected the battery, with another normally-off switch
all the plugs will also indicate any one from moisture by liquid silicone sealant. in the line. The backup system probably

82
doesn’t need to pass current to anything
more than a fuel pump and the engine —
assuming that it can be cut in before the
engine dies and needs restarting. Still, a
backup system will never substitute for
adequate preparation in the first place.
EXHAUST
The exhaust system should be completely
designed, built, and tested by the engine
developers. Header pipe size, lengths, and
layout have such critical effects on power
and the torque curve that they can only be
properly developed on an engine
dynamometer. However, there are a number
of areas where the race car mechanic is
responsible for the installed system. The
first point is that there will have to be a lot
of clearance around all pipes. Engine move-
ment due to g loadings and torque reactions
can cause a half-inch of movement at the
headers. Temperature clearance is also a
problem if at least two inches of cool air
space isn’t available between pipes and
critical components such as brake lines, in-
Headers designed for equal lengths, durability, and ease of installation.
sulated wires, and lubricated joints. Other-
wise, aluminum or asbestos heat shields will Note the reinforcements, safety wiring, and slip-fit into the collector.
be necessary. Engine movement causes
other problems because the tail pipes are
supported by the chassis, so there must be
some sort of loose connections in the matter to start it up — and ruin it before it And here is the most important sentence in
system. If the entire header is one piece and even gets warm. A race engine is a fairly this entire chapter: Do not rev up a race
mounted rigidly to the engine and the reliable hunk of machinery, but it still re- engine — or any engine, for than matter —
chassis, it will quickly crack due to the con- quires a reasonable amount of care. Because until it has had a chance to warm up! The
flict in stresses. The most common solution of the necessary clearances, it has to be engine should be held at the minimum
is to have loose slip joints at the collector treated gently until temperatures and smooth running speed possible until the oil
pipe, held together with bolted tabs. Even pressures get up to operating condition. and water temperatures have gotten up to
the bolts should be left loose (with lock nuts A fresh or cold engine should be cranked 140 or 160 degrees. Race cars with dry
at their ends), and yet kept from rattling over with the ignition off until the oil sump oil tanks can be helped along by
around too much with compression springs. pressure comes up on the gauge. If it hasn’t wrapping the tank with electrical heating
For ease of assembly, it may be more run for some time and fresh plugs are re- tape, and pre-warming the oil for an hour
practical to break the head pipes apart at the quired, it is easiest to crank up oil pressure or so before the engine is to be started. After
flange and slip them into the collector in- while the plugs are out. Plug gaps, valve the engine is good and warm, it can be shut
dividually. This also tends to reduce stresses lash, water level, oil level, and throttle ac- down and given last-minute checks: oil and
on the header due to uneven thermal expan- tion should have already been checked. The water levels, hot valve lash, hot bolt torques,
sion. The combination of heat and vibra- fuel pumps should be turned on until the belt tensions, leaks, and the last visual in-
tion is so severe that some professionals feel pressure stabilizes, and the fuel system spection to insure that everything is tight and
the only way to avoid cracking is to use should be checked for leaks. safety-wired.
stainless steel or titanium pipes. The header Each engine induction system will have Aside from that, the only race tuning that
flange bolts must be safety-wired, because it’s own peculiar starting requirements, but should ever be attempted at the track —
nothing else will keep them tight under it is better to try and start with too little fuel assuming the engine is already developed
those conditions. At that, they should be pumping than too much. It isn’t uncommon — is fuel mixture, spark advance, and plug
retorqued to seal the flange after the engine for a cold race engine to backfire through changes. Local atmospheric conditions may
has been warmed up, before wiring them in the inlet and start a small fire. A large, clean require a minor change, but running too
place. rag should snuff it out, but a CO, fire ex- lean, too far advanced, or too hot can be
Once the engine is installed and tinguisher may also be necessary. (The disastrous. It’s better to have a slightly weak
everything is hooked up, it is a simple chemical types are a mess to clean up after.) engine than none at all.

83
9
GEARING AND DIFFERENTIALS
Transmissions we tend to take for granted. RATIO THEORY tions, high gears need not be plotted at low
Since few of us have the tools to make our speeds, and since all curves will have similar
own, we buy them off the shelf — like By this stage the racer should know nearly shapes, some can be sketched in from a few
engines — and plug them in. We seldom all that is necessary about a given car to points. A pocket electronic calculator or
question them as long as they are strong make some good mathematical predictions. home computer/plotter simplifies the job
enough to last a few races, and have enough The necessary data includes: an engine immensely.
different ratios available. ““Gimme five and torque curve, all numerical gear ratios The upper and lower limits of considera-
a stick to row it with.’ available, tire rolling radius, and vehicle air tion are indicated by two other curves. At
But only after a race car’s handling, drag. It is also useful to have a rough idea the top, forward thrust is limited by the
brakes, and aerodynamics have been sorted of the rear tire coefficient of friction, and traction available at the rear tires, which is
out and the optimum engine is selected, is the center of gravity height. We will ar- a function of rear weight distribution, rear
it possible to make the final subtle changes bitrarily assume the total driveline efficiency load transfer, tire coefficient, and
in gearing. Gearing should be just about the to be about 85 percent at high torques. aerodynamic downforce at high speeds. At
last stage in race car development. Not only Using this information, it is possible to plot low speeds, where it is the greatest problem,
does it have practically no influence on any a series of curves on a graph to show rear the equation is: thrust limit =
of the preceding considerations, but it is tire thrust at all speeds for each gear ratio. (Cf) (static rear weight)
relatively easy to change and highly de- (See Fig. 40) There are two equations that 1 —(Cf) (cgi height)/(wheelbase)
pendent on all of those other factors. Proper must be used for each plotted point:
gearing is a direct consequence of corner- The lower limit is the point at which total
ing capability, braking capability, _ (.072) (rpm) (rolling radius) vehicle drag is greater than the thrust
i
aerodynamic drag, and engine torque (trans ratio) (diff ratio) available in any given gear. This drag curve
curves. Also, with the abundance of ratios can be roughly approximated by data from
available, especially with the popularity of
5-speed racing transmissions and various
thrust = (85) (torque) (trans ratio) (diff) aerodynamic development. If drag in
pounds was found at 100 mph, drag figures
(rolling radius)
tire diameters, the number of possible at higher speeds are found by- multiplying
combinations is staggering. It is possible to In the first equation the assumption of zero that force times (speed/100)2 . If a more
find the optimum gearing at a race track by tire slippage is close enough, considering precise drag curve is desired, it must include
guess and test, but the number of choices the balancing effect of tire growth at high a Static force component and a rolling drag
and the scarcity of track time make it im- speeds, and in both cases rolling radius is force which increases in direct proportion
portant to use some rational method. in feet. To reduce the number of calcula- to speed, or: vehicle drag =

(K,) + (Ky) (velocity) + (Kg) (velocity)?

When the thrust available curve crosses the

thrust required curve, that is the approx-
imate top speed of that particular
aerodynamic shape with that particular
engine — regardless of gear ratio.
The use of these curves should be ap-
parent. An arbitrary gear is selected and run
up to its maximum speed in rpm (or the
S8 TOTAL VEHICLE DRAG point at which thrust intersects the next ratio
curve), and it is then shifted down to another
higher ratio. Of course, the number of shifts
pounds
in
thrust
tire
is limited by the ratios available, so each
ies SPEED POINT should be carefully spaced out between
ACCELERATION SPEEDS minumum and maximum speeds. The
BETWEEN TURNS
9 AT RIVERSIDE optimum desired selection will maximize
the enclosed area under the choice of ratios.
0 20 40 60 80 100 120 140 160 180 200 220 240 If it were possible to have an infinite number
speed in miles per hour of gears — or an infinitely variable ratio —
then the rpm could remain at peak torque
Figure 40. Gear Ratio Selection Curves and the thrust available curve would be a
smooth curve. In practice, each shift
requires some time — say 0.1 seconds — Main Cluster
Tooth Count
in which no thrust is being generated, so Shaft Shaft
24/32 23/32 23/33 22/34
there must be some balance between the Gear Gear
magnitude and the duration of thrust.
If these curves don’t provide enough range Tooth Count Gear Ratio :1
in ratio selection, it is always possible to
raise or lower them by changing the over- 39/16 3.39 3.50
all ratio. A change inthe differential gear 39/17 3.19 3.29
ratio may be difficult (or impossible with 38/17 3.11 3.21
some transaxles), but it has less effect on 38/18 2.93 3.05
vehicle handling than a change in tire 37/19 PTA 2.80
diameters. It may also be possible to raise 36/20 2.50 2.58
the upper rpm limit, for whatever gains are 35/21 2.32 2.40
available at a corresponding reduction in
engine life. Unfortunately, an over-all ratio
change requires a new set of calculations
and curves, although the upper and lower
limit curves will remain the same.
It should also be apparent from these
curves that it is important to have a relatively
flat engine torque curve over as broad an
rpm range as possible. If engine torque falls
off sharply or has noticeable dips due to
Ww
intake ram tuning and exhaust tuning effects, 2)
it will be more difficult to fill the holes in z
<
the thrust curve by changing gear ratios. It fa
makes the common practice of rating race a
4
engines by peak horsepower look relatively =
RATIO
iva
meaningless. Of course, the smaller the wi
>
range between upper and lower speeds and [e)
the greater the number of gears available,
the shorter the critical rpm span is.
TRANSMISSION HARDWARE
Even if a car isn’t equipped with a racing 1.00 1.00 1.00
transmission and easily changed ratios in
five gears (See Table 5), that doesn’t mean Table 5. Available ratios for the 5-speed Doug Nash racing transmission.
it is limited to the ratios it came with. Any
manufacturer who has a competitive car
probably offers a selection of ratios for the Gear Muncie trans Borg-Warner T-10 trans
standard transmission. For example, any
front-engine rear-drive American car that Ist 2.20 Diy) Megas 2.43 2.43 2.64 3.44
is equipped with a four-speed manual can 2nd 1.64 1.88 i 1.76 1.61 1.75 2.28
be fitted with any one of seven combinations 3rd 127 1.47 1235 1.47 3} 1233, 1.46
of ratios in the first three gears (See 4th 1.00 1.00 1.00 1.00 1.00 1.00 1.00
Table 6).
Table 6. Available gear ratios for a 4-speed transmission
It would be theoretically possible for a
team to carry one of each set and bolt in
whichever one seemed best suited to a par-
ticular track, along with a corresponding synchronization. Passenger car transmis- probably receive more driver abuse than any
change in differential gear ratio. However, sions usually have friction surfaces which other part of the car, and are more often the
a speed chart for most road-racing tracks contact first, to match rotation speeds cause of “transmission failure’? than the
would show that there is no need for a low smoothly. In racing use, this type of gears themselves. Many well-known drivers
gear ratio below about 2.2 to 1. And for a synchronizer wears rapidly or is prone to either “power shift”? without letting up on
reasonably flat torque curve, there won’t be more instantaneous failure. In pure racing the throttle, or do not use the clutch pedal
much difference between the remaining gearboxes the gears are alternately con- except for standing starts. And unless they
options in second and third gears. nected to their mainshaft by stepped cogs, are very good at it the dogs get ground
Gears have some rotational inertia when or “dogs,” which are designed to grab the round and either won’t go in, or stay in,
freewheeling in the middle of a gear change, next gear very rapidly and firmly, an action gear. Either type of synchronization is
which means that the next two gears to be too harsh for most street use. The undercut helped a great deal if the driver learns to
engaged will seldom be rotating at the exact angle and corner radii of these dogs are match speeds between gears through
proper speed. If they are forced together, carefully designed to catch without wear- modulation of the accelerator pedal — by
there will be a destructive gnashing of the ing and to lock firmly together as long as a split-second letup during the upshift or a
engaging teeth unless there is some sort of torque is being transmitted. These dogs slight blip during the downshift. Otherwise,

85
 ALFA VI2 INDY PORSCHE

Figure 41. Four possible case configurations to allow for different engine crankshaft heights (h) and vehicle wheelbases, all
using the same basic Weismann transverse transaxle gears, shafts, bearings, and differential.

the synchronizers or lugs may be rapidly dynamics, but was a need for raising their Bobby Unser’s Eagle at Indy in 1978, and
worn or broken, causing the driver to have flat-12 engine enough to get smooth header was later incorporated in other cars.
to match speeds even more closely, or to pipe flow beneath it. Rotating the two In 1980, Brabham’s Gordon Murray
hold the shift lever in gear. transmission gear shafts from longitudinal decided to switch from the Alfa V-12 to a
The final significant racing transmission to lateral, gave the designer freedom to Cosworth V-8, and he wanted a transverse
difference is in the gear cuts. To reduce easily raise or lower the engine centerline gearbox at the same time. Because the new
noise, passenger car transmissions use angle with respect to the rear axle centerline as engine was shorter, he asked Weismann to
or “helical” cut teeth, and differential gears shown in Fig. 41. But because they were not build the gearbox case with a long bell-
use “hypoid” cut teeth (also used to lower concerned with underside aerodynamics, housing, so that he could keep the same
the driveshaft centerline). But for reasons they simply stuck the change gears in bustles wheelbase chassis. He also requested that
of strength and efficiency, race cars use only on both sides of the case, making it wider the shifter mechanism be raised to the top
straight cut “spur” gears, which have than a conventional transaxle. from its usual position on the side, and the
primarily rolling contact on the teeth instead Roman Slobodynskyj was impressed with case be made as narrow and smooth as
of sliding contact. this new way of looking at transaxles, and possible, to aid in airflow at the back of the
The strength of a transmission may in 1974 he went to Pete Weismann, ground effect tunnels. With this new criteria,
become critical in the transition between a America’s only racing transmission builder, Weismann relocated the narrowest possible
production car and a race car. But it isn’t to see if one could be built for Indy. change gears to the very center of the case,
so much a factor of engine torque as it is Weismann suggested that he lay an Offy and got it down to 8 inches with five speeds.
of tire traction and driver manhandling. down on its side, offsetting its crank to the When Murray took the new car to the wind
Torque in the driveline is limited by the trac- left, and use a bevel gear drive into the tunnel, he was impressed with the improved
tion of the rear tires, which may be doubled transverse transmission. So the box was flow between the transmission and the tires.
by an increase in tire size and rear weight designed and built just in time for the Offy Suddenly the transverse gearbox was the
bias. As the transmission is overloaded more to die of old age. Even though McCluskey aero trick of the year.
and more, it becomes necessary for the did fairly well with the car in 1977, it was Since then, layout varieties have pro-
driver to reduce shock loads as much as soon outclassed by the Cosworths. liferated. With fairly common internals,
possible by making smoother, more precise In the same period, John Barnard also such as gears, dogs, and bearings, the case
shifts. adopted the transverse concept for his new and its shaft locations change like the
Parnelli-Indy chassis, and Weismann began weather, since everyone wants a different
‘THE TRANSVERSE TRANSAXLE
to see commercial possibilities. So he engine centerline height or different case
The modern transverse gearbox was first designed the Transverse 78, a gearbox with length. And of course there are different
seen on a Formula One Ferrari in 1974. Ap- an emphasis on practicality and simplicity, torque requirements. Five different widths
parently their design requirement of that and one which was not tied to a specific of gears are available, with the widest for
period had nothing to do with aero- chassis. This gearbox first appeared in turbo-charged Indy cars. When the Indy

86
 5TH GEAR SET DRIVE PINION DIFFERENTIAL CASE
4TH GEAR SET 3RD GEAR SET RING GEAR
|

BY /// 1FG ZsGi,LZ Z7

ETI
LF SF LE LF SYSEP
MF IF LPP FL DF 1

Ax 223 Yipy
ISO Zed/ ee
+
DRIVE

‘ez7077
BEVEL

ae

iL EEE
a
ae/
29Ge y,
es) LB AES Sor os 4 caeeaen eel
3-4 DOG CLUTCH rach BEVEL 2ND GEAR SET | AXLE FLANGE
1ST GEAR SET LOW DOG CLUTCH
Schematic of the Weismann 5-speed transverse transaxle shows the narrow package, and how having the bevel gears at the
relatively low torque input shaft eliminates thrust bearing problems at the differential.

cars need more speeds for road courses, a drag. But in Weismann’‘s transverse trans- the speed is due to a dry-sump oiling
*“bustle’’ sidecover is stuck on to house the axles, the right-angle gears are at the lower- system which avoids draining and refilling
extra set. Even for a common shaft ar- torqued input shaft, and even so, Weismann oil each time. With a quick-release cover,
Tangement, everyone wants different uses ball thrust bearings to take the axial you might beable to change gears ina racing
mounting bosses and shifter extensions loads. pitstop on a road course and not lose a lap.
and bellhousing lengths and shapes. So One of the most-liked features is the speed The dry-sump oiling system is almost a
each one is almost custom-designed. In with which change-gear ratios can be story in itself. There are two pumps on the
Formula One, transverse transaxles go in modified. With the removal of 10 side-cover input shaft--similar to an automatic trans-
and out of favor as rules change regarding bolts (but no suspension hardware) all of the mission--one to drain the sump into a
car dimensions and aerodynamics. change-gears drop out, in addition to the reservoir, and the other to supply oil via
drive pinion. Some racers report they can nozzles to each gear tooth contact area.
Recent Innovations change a full set in less than six minutes, Because of all the loose pieces of ground-
Although we can easily take them for avoiding cool-off time in practice. Part of off shifter dogs floating around in the oil,
granted, Weismann transmissions are full of
clever little surprises that even non-
engineers can probably appreciate. The first
shock is in the final (differential) gear ratio
change. Just a few years ago, gear design
was such that to change ratios, you had to
either change both gears or change one gear
and the center-to-center distance. This, of
course, is a real pain when it requires
changing the differential ring gear. Now,
however, Weismann uses computer-designed
gear teeth profiles, with a sophisticated
technique which allows just the drive pinion
to be changed by +1 or=1 teeth. For exam-
ple, a typical transaxle might have a normal
final drive of 17:52, which can be quickly
changed to 16:52 or 18:52.
The second surprise upon opening the
box, is that there are no tapered-roller
bearings. Normally these are necessary to
take the side loads on the highly-torqued dif-
ferential ring-and-pinion, and they are a real Both of these 18-tooth and 19-tooth drive gears will mesh perfectly with this 52-tooth
problem in preload adjustment and rolling differential gear, thanks to computer design.

87
 transmission practice of constantly driving
the lower counter-shaft, and shifting gears
back up to the mainshaft. In this case, as
in racing transaxles, the mainshaft is con-
stantly driven and the countershaft changes
speed with gear changes. Then, at the very
end of the case, rotation is taken back up
to the engine centerline via a set of overall
quick-change gears. The main advantage of
this setup is that all change ratios, including
the overall ratio, can be rapidly switched,
just as in transaxles. No longer is there any
need for quick-change axles in modified-
production sedans. On top of that, these
“standard” boxes incorporate most of the
tricks Weismann uses in his transverse trans-
axles, including the gear selection avail-
ability, dry-sump oiling, and the use of
engine oil.

DEVELOPMENT
Weismann’s standard 5-speed quick-change transmission for front-engine race cars Since a transmission is a relatively fixed-
is actually installed on its side like this for lower engine mounting. design mechanism, development is limited
to optimum ratio selection. To do so, a
thrust/speed chart will be necessary. The
(which are mostly filtered out between the obvious question — why not magnesium? one already shown is a typical example for
two pumps) these pumps must be very The answer is unexpectedly more than a full-bodied race car with a 5-liter engine
forgiving (have loose tolerances). Instead money. First, the stiffness of magnesium and Hewland transaxle over-all gear ratio
ofthe familiar gear-rotor pump, Weismann castings is much less than aluminum selections. So far, the chart takes care of all
uses a sort of vane-type pump, in which the castings. Second, Weismann claims that the relevant vehicle data, but some track
vanes are replaced by rollers. These rollers American magnesium castings creep, or data is necessary also.
are not even spring-loaded, but are pressed actually ‘“‘grow”’ over a period of a year or The ideal information would be a
outward by centrifugal force. Weismann so. This plays havoc with the most carefully- recording of speeds all around the specific
calls them his ‘‘trash pumps,”’ but they can designed tolerances. Also, apparently track for a similar vehicle, from which
still put out 100 psi at 10,000 rpm. One OSHA (The U.S. Office of Safety and minimum corner speeds and maximum
advantage to a dry sump systemis that with a Health Administration) has determined that straightaway speeds could be obtained. The
transparent cover, you could fire the engine magnesium is a hazardous material. Not second most accurate source is a driver who
up in neutral, and study the distribution only does it burn with fierce intensity when can note — and remember — tachometer
nozzles spraying the gears. ; ! ignited in powder or shaving form, but it can readings at each of these points, from which
Another jaw-dropper is the oil itself, actually explode when molten for casting. the speeds can be calculated. The last
which is ordinary engine oil! The old hypoid So between regulations and practicality, not choice, for a track that has never been raced
differential gears were so critically loaded many fabricators want to use it any more. on, is an estimate from an accurate course
that in addition to 90-weight gear lube, they A question about the strength of these map. Knowledge of the car’s approximate
required “E.P.” (extreme pressure) additives cases revealed a recent change in priorities. cornering capability in g’s and the minimum
(which stink like fish emulsion).;
But with Apparently the gear torques and spread path radius in each corner, will give
straight-cut gears and no sliding friction, forces are no longer the significant factor. minimum cornering speed from the
about all the oil is needed for is cooling and With the state of aerodynamic chassis equation:
carrying away the “trash.” loadings and stressed engine/gearboxes, the
The otherwise wasted space in the greatest case concern is torsional stiffness. mph = (3.88) V(g) X (radius)
bellhousing extension between the gears and In other words, the case wall thickness is
the engine is often used for the engine oil determined by the required chassis stiffness. The maximum speed will have to be even
reservoir, with foam-separation towers And Weismann points out that designers are more roughly approximated by the distance
above. In fact, surprise upon surprise, some fooling themselves unless they test the stiff- to the next assumed braking point, com-
teams have tried using a common reservoir ness at a typical case temperature of 200 pared to a straight line speed/distance curve.
for both the engine and transmission oil! degrees Fahrenheit, where the stiffness Computer programs have been designed to
While the distribution systems are separate modulus drops drastically. do this operation with some precision, but
— with even more careful filtering for the Almost as a sideline, Weismann uses the that explanation is worth a chapter in itself
engine oil — it all goes through the same same technology in a relatively (See Chapter 15).
sump and cooler. Rightly or wrongly, conventional-looking front-engine trans- The minimum and maximum ranges of
engine builders have been complaining, mission. This is a truly quick-change acceleration for each straightaway can then
though, and the practice is being transmission that has been needed for years, be presented as a bar chart at the bottom
discouraged. although because of its sophistication and of the gear chart, as shown in Fig. 40.
The cases themselves are cast in a com- cost, it is used only in professional high- Assuming that there will be no standing
mon aluminum alloy, which leads to the dollar sedans. It avoids the standard manual starts, except perhaps for a pitstop, these are

88
the only areas of consideration in gear selec-
tion. The highest gear must just allow the
engine to peak out at the maximum speed
bar. A few hundred rpm to spare at that
point, however, will prevent over-revving if
a “‘tow”’ is picked up, without a great loss
in acceleration capability. At the bottom
end, there is no reason to have a low gear
ratio which provides thrust greater than the
traction limit up to peak engine rpm. On
the other hand, it will be necessary to
accelerate the car from a standstill, so one
gear should be low enough to prevent
lugging the engine or slipping the clutch too
much. At any rate, four ratios ought to be
adequate for most tracks. The next, or sec-
ond, gear will probably be the lowest one
used in the race, so it should have a low
enough ratio not to be too far off the torque
peak at minimum speed. Still, the driver
won't be able to apply full throttle until he
has accelerated well out of the apex anyhow.
The first selection of the two intermediate
gears might be in even numerical steps be- At left is a dog engagement ring, designed to mate with matching dogs on the.
tween second and fifth gears. Say second straight-cut gear next to it, and at right is a conventional gear with synchro surfaces.
was 6.10 and fifth was 3.13, then third would
be 5.07 and fourth would be 4.10. But by
selecting higher ratios than that, say 4.58
and 3.70, it can be seen that more area on and Overhaul Manuals are a must for trans- early downshifting may reduce the load on
the graph is picked up at the higher speed axle owners, with additional valuable tips the brakes, but it upsets the brake balance
than is lost at the lower speed. As it turns by Carroll Smith in Prepare to Win. For and causes other unnecessary stresses.
out, the ideal steps for maximum net ac- adapted passenger car transmissions, all Stress reversal in the driveline adds to heat
celeration will diminish in size as the factory information should be reconsidered input and almost doubles the total stress
transmission is shifted up. with the knowledge that racing will produce input. In addition, it is easy to downshift too
Once the optimum ratios are selected for almost twice the stresses. The most care- early, causing the engine to over-rev with
maximum acceleration, a check should be fully checked gear tooth alignments and pat- low cylinder pressures — a sure means of
made for each short straightaway. The ratios terns will change a great deal due to deflec- destruction. In some cases it may be most
should be spaced so that it isn’t necessary tion under stress, not only in the gears but practical to downshift by two gears at once
to shift while coming out of a corner or just in the case itself. Where possible, tolerances — after braking — so the linkage should
before getting on the brakes. Ideally, each should lean in the direction to best neutralize have a definite reverse lockout.
short straight should fit the rpm range of these deflections. Shifter forces are also Even with the best of assembly and
only one ratio. This is where the com- likely to be especially great, so the linkage driving care, a gearbox and/or differential
promises must be made. It is worth keeping and detents must be adjusted with these can still self-destruct from inadequate
in mind that ideal computer-simulated laps, force deflections in mind. lubrication or cooling. Ordinary gear lube
driven by a mathematically perfect driver, The technique of shifting will vary often isn’t good enough for the extreme
may not decrease more than 0.2 percent in somewhat with the type of transmission, contact pressures and speeds in racing.
laptimes when ratios are changed by two or type of clutch, type of racing, and life ex- Some racing lubricants have additives to
three steps. Since the driver’s performance pectancy of the gears. A good knowledge take the higher pressures and to help reduce
will cause more variation than that, it is of what happens and how it happens inside foaming. A gearbox which is 95 percent
probably a good idea to select ratio steps the transmission should improve the driver’s efficient will still have to absorb 5 percent
that make his job easy as possible. Of performance. The perfect driver would be of the power it transmits, or often more than
course, even the most accurate predictions able to shift quite well — up or down — 30 horsepower. That should give some in-
won't produce the ideal ratios for a track. without using the clutch at all, by carefully dication of the cooling required. If the oil
Slightly increased power or cornering force sychronizing rpm and shifting speed. In fact, temperature ever gets much over 240
may require a higher gear at any location. many imperfect drivers can do it fairly well, degrees, an oil cooler will be necessary.
But the chart should improve the estimates but the clutch pedal provides a margin for Otherwise, natural air circulation may be
and help avoid more than one gearbox error. The wide-open-throttle power shift adequate. It also may help a great deal to
disassembly at the track. may be a shade faster for a few drag-strip break in new gears with less than racing
runs, but it doesn’t promote longevity in a loads, especially hypoid differential gears.
RACING AND MAINTENANCE road race. The shock on the driveline is far A final detail point is that high temperature
Because of the variations in gearboxes, greater than ordinary racing loads, and the does increase performance up to a point by
the best advice for assembly and engine has a tendency to over-rev by a few reducing the lubricant viscosity, so it may
maintenance will be found in the factory thousand rpm in that brief disconnected be an advantage to start a race with pre-
service manual. The Hewland Maintenance instant. As was mentioned in Chapter 5, warmed fluids.

89
 stick-slip or chatter in a corner, which will
cause variations in the thrust at left and right
wheels — an unstabilizing effect at a critical
time. The limited-slips are worst in this
respect, while a fully locked differential will
allow a constant — if not always desirable
— balance under most conditions. Still, if
one wheel gets on some slippery stuff, the
longitudinal forces will be unbalanced, but
then the lateral consequences are far greater
any way.
One of the most popular types of locking
differential in racing has overrunning
clutches for each axle (Weismann and
Detroit Locker), which allow either wheel
to overspeed the differential. This is sup-
posed to allow a speed or distance difference
between the inside and outside tires in a
turn. It works perfectly in very sharp turns
at low speeds — say in the pits — but an
analysis of vehicle dynamics gives some
doubt whether it has any effect at all on a
high-speed race track.
First, the theoretical overspeed at the out-
side tire is inversely proportional to the
radius of a turn. For example, if the vehi-
cle track width is 5 feet and the turn radius
is 100 feet (a tight corner in racing) then the
outside tire will need 5/100 or 5 percent
greater speed than the inside tire. On a
500-foot-radius turn, the overspeed will only
be 1 percent greater. But as was explained
in Chapter 2, tires already have a slip rate
when generating thrust. This slip rate,
especially under hard cornering, will be
over 5 percent, so it appears that the high-
speed cornering overspeed required is
negligible. If it were true that thrust was
directly proportional to the net percentage
These cutaways show 5-speeds with conventional helical-cut gears and synchronizers of slip, then the inside tire would always
(top), and straight-cut gears with dog engagement (bottom). generate more thrust, producing an
understeering force which would decrease
with increasing speed. However, lateral load
DIFFERENTIAL THEORY nering breakaway by reducing thrust losses transfer seems to have an overriding effect.
AND HARDWARE on the outside tire. It would take a computer or a great deal of
Straight line stability is a relatively minor experimentation to separate all the effects.
The purpose and operation of differentials problem on a race track, but even there a Braking stability should also be con-
ought to be well known already, but their locked or limited slip differential can cause sidered in the selection of a differential type.
application to total vehicle dynamics, instability. For example, suppose one Only a locked differential will positively
especially in racing, is still debatable. Con- driving wheel suddenly loses thrust traction prevent one-wheel lockup at the rear,
sidering the torque that must be transmitted on gravel, water, or oil — or in the worst although other vehicle stability interactions
while a vehicle is cornering hard, some sort case, due to a broken axle. The remaining may be more important to handling during
of control is necessary to limit or prevent forward thrust at the other tire will have braking. At any rate, the overrunning clutch
inside wheelspin. The most common such a tremendous leverage about the center differential appears to have the fewest of all
methods can be broken down into the locked of gravity that it will be practically impossi- the possible evils.
or locking type of differential and the fric- ble to react fast enough to keep the car from Lubrication and cooling are as important
tion limited-slip type. Either type will allow spinning. At best, it steers the car toward to differentials as transmissions, so the same
greater acceleration out of a slow turn, the more slippery surface. Then there is cor- advice applies. But a differential can have
although both of them have a number of nering stability, which if not reduced by a other friction effects to raise its temperature
stability advantages and disadvantages. The locked axle, will at least have more sudden even more. The limited-slip types are
open differential can’t be faulted in this breakaway characteristics at the rear under especially prone to high temperatures, and
respect, as it always provides perfectly acceleration. In addition, some types of dif- if overworked, can literally burn themselves
balanced thrust forces between left and right ferentials have a relatively uneven force out and revert to a relatively open differen-
tires, and it limits the severity of rear cor- balancing mechanism, and can develop tial. Even some locking differentials with

90
friction over-running clutches can slip,
which causes abnormally fast wear —
especially on really low-traction surfaces.
If locking differentials seem to be the best
choice, they are also relatively expensive
and perhaps not even available for some off-
brand race cars. In such a case, a solidly
locked differential is a’fairly good second
choice in any race car with enough power
to spin the inside rear wheel. If the rules
don’t allow a locked differential, the spider
gears may be accidentally buggered until
they work poorly, or a limited-slip differen-
tial can be tightened by wedging the pre-
load springs almost up to a solid stack. The
quickest — and most easily reversible —
method of locking a differential is to replace
the spider gears with ones that have been
specially machined or welded solid. On
some limited-slip differentials the easiest
method is to weld the friction clutch packs
The Weismann/Brabham transverse transaxle designed for the ‘‘laydown’’ For-
together on one side. But the cleanest
mula One BMW, showing the change gears and spur-gear differential.
technique is to replace the entire spider gear
cage with a specially machined solid spool,
which also reduces weight and rotating
corresponds to maximum acceleration. insurmountable problem is the horsepower
inertia.
Sliding splines are usually necessary with loss in the converter, which is about 3
DRIVELINE universal joints that have no provision for percent even at the highest rpm where its
As simple as they seem, driveshafts and changing length. These splines have tremen- efficiency is greatest. Modern converters
universal joints are such important links in dous torques acting on them when they most have a “lockup” feature for greater fuel
the chain that they shouldn’t be neglected. need to slide, so they should be well efficiency. A mechanical clutch is incor-
lubricated with the best grease available. porated to lock the converter solid, or
Any excess torsional strength here is well
spent, as failures are usually sudden and exactly 1.00:1.00 at a specific rpm or
AUTOMATIC TRANSMISSIONS hydraulic pressure. But the lower rpm
total. Assuming that the shaft has been
designed to take all the torque the tires can Although automatic transmissions in slippage still creates unacceptable heat
put to the ground — with continual stress road racing go back to the Chaparrals in and power loss. fe
reversals — race preparation can be concen- the sixties, they still offer some potential _ For many years, Porsche made some of
trated on the universal joints and splines. advantages to race cars and are at least the most advanced road racing transmis-
One last word about the drive shaft: it worth a mention. The obvious advantage sions, such asatwo-clutch auto-shift gear-
spins at engine rpm in high gear and any is ease of driving, so more concentration box, in which each clutch controlled alter-
imbalance can be significant. Its trueness can be put into aiming the car. That nate gear ratios. That is, first gear is
is fairly important also, so it should be assumes a fully automatic shift, however, disengaged at the precise instant that sec-
checked for runout before balancing. when in fact most often shifting is done ond gear is engaged — if all goes well.
There are four types of universal joints manually, but without a clutch. Depend- Electronic-shift manuals are becoming
common to race cars: the cross-and-trunion, ing on the system, it is also possible to popular in some series. These use conven-
the Rzeppa caged-ball, the ball and trunion have faster, more reliable, and less dam- tional spur gears, plate clutches, and shift
pot, and the rubber donut. The most com- aging shifts. mechanisms, but with external servo-con-
mon and inexpensive is the cross-and- At one time a torque converter was felt trol devices. However, precise timing
trunion, which unfortunately has the least to be an advantage, from the increased demands superior servos and onboard com-
constant velocity in shafts which meet at any torque multiplication capability at lower puter logic. Under the best conditions,
angle. However, if the trunions at each end rpm. With the broad torque curves of shift time can be reduced from about 0.20
of a shaft are perfectly parallel and the larger displacement orcomputer controlled sec. down to about 0.05 sec. The shift
driving and driven shafts are perfectly engines, however, there is less to be gained speed limit is not in the mechanism, but in
parallel, then only the intermediate shaft over the ratios available in a six speed the driveline torque shock limit. Even
will have rotational velocity fluctuations manual transmission. The greatest disad- with engine modulation via electronic
with each revolution. Constant-velocity vantages of a torque converter are weight throttle or spark control, the best shafts
universal joints eliminate this problem and efficiency. Even if fewer mechanical and gears can only take so much. Given
entirely and are mandatory for large angular ratios are required, a torque converter is the electronic clutch and engine control,
deflections. The smoothness, stress, and quite a bit heavier than an equivalent some teams have also experimented with
power loss will all be optimized by laying clutch. That combined with the greater wheelspin control and automatic standing
out all joints to have the minimum angular diameter necessary, means that rotational starts, giving some truly great launches --
deflection at the suspension height which inertia will be even worse. But the most when the rules allow it.

91
10
FRAME AND BODY
Few racers build their own chassis from once a component or an entire race car is Steel selection for race cars is usually
the ground up. Not only is it an inefficient constructed and seems to be too heavy, the limited to three common alloys. Common
use of time, energy, and dollars, but it can cost of labor may prevent switching to exotic carbon steels (1010 to 1020) are inexpensive
be rather dangerous for someone without a lightweight materials. There has to be some and easily fabricated but have a relatively
lot of experience in either engineering stress consideration of the cost involved in re- low ultimate strength. More expensive
analysis or first-hand race car chassis con- ducing weight through a material change, chrome-moly steels (4130) and nickel-
struction. A thorough strength analysis up to the point where it becomes inefficient chrome-moly (4340) can be heat-treated to
during the design stage is an incredibly in dollars per pound of weight removed. the highest tensile strength of almost any
complex task, even for an engineer with the And without an extensive test program, an material. Heat-treating is mandatory for
most sophisticated computers. (See Chapter engineer can’t determine the precise strength highly stressed components such as springs
15) Even where it is economically feasible, requirements for a component anyway. At and anti-roll bars, and normalizing, or
say for a limited-production race car, the any rate, material costs vary so much with stress-relieving, is recommended for all
input loads have to be estimated or deter- purchased condition or configuration, and welded parts. Each of the hundreds of race
mined by survey at all race tracks. In addi- with economic situations, that they won't be car components will have different require-
tion, an arbitrary safety factor is needed for considered in great detail here. ments that should be matched to a particular
collisions or off-course excursions. For- Due to the basic nature of racing, all alloy or heat treatment by an engineer or
tunately, most existing race car chassis are structural material properties must be experienced fabricator. The most important
fairly strong, but at that they are designed and related to density, or the material weight per considerations are usually: stiffness,
built primarily from intuition and experience. volume required. And yet, for any com- ultimate strength, yield strength and failure
But even with a strong and rigid vehicle, there parison, material properties vary con- mode, fatigue life, and formability. In
are a number of possible modifications that siderably with different compositions and addition, it may be necessary to consider
the racer should be aware of. hardnesses. Table 7 indicates general ranges high temperatures, wear, and cost. Armco
for the most common race car materials. Steel once sponsored a race car project to
MATERIALS Of course, with all the different composi- demonstrate the uses of steel alloys in
Before getting into the actual hardware of tions available in each material, the selection racing, and some of the data was presented
frames, bodies, and components, it is will be difficult without a comprehensive in SAE Paper No. 700056. The large steel
necessary to consider some of the basic knowledge of component requirements and companies can provide invaluable assistance
structural materials as they relate to racing. other specific material properties. Engineer- through consultation and handbooks on the
With all materials there is a wide range of ing assistance can save a lot of money over latest alloys.
costs per pound, ranging from reasonable selecting simply light weight or a name with Aluminum alloys are even more diverse
for low-grade steels to astronomical for a nice sound. It is hard to beat the security in their properties and ease of fabrication.
titanium and pure filament fibers. However, per pound of using whatever material a Alcoa has a series of handbooks for the
for one unit of anything, especially a race component was originally designed around. engineer to use in alloy selection and com-
car, the cost of a material is usually in- Still, a few generalizations can be made ponent design. The same considerations for
significant compared to the cost of labor in about application to the specific area of racing apply as with steel alloys except that
design and fabrication. On the other hand, racing. temperatures may be more of a problem,

Alloy Aluminum| Fiberglass |Magnesium |Titanium Graphite Kevlar

Steel / Epoxy | Epoxy /Epoxy

Density
(pounds/cubic inch) 0.3 0.1 0.05-0.08 |0.07 0.15 0.05 0.05
Ultimate tensile strength (1000 psi) |50-200 20-80 40-100 30-40 100-160 150-220 40-190
Strength/ Density ratio 1000) 200-700 {200-800 |500-2000 |400-600 700-1000 2500-3700 {800-3800
Stiffness(1,000,000 psi) 29 10.5 3-4 6.5 16 32 14-8
Stiffness/ Density ratio(1,000,000) 100 105 40-80 95 105 400-700 80-160
Formability good good excellent |poor good good good
Machinability excellent |excellent excellent fair
Weldability excellent |good poor poor

Table 7. Characteristics of race car materials

92
especially in welding. The more common
alloy designations and their chief attributes
are: 3003 — inexpensive and easily
fabricated; 2024 — strong and formable;
6061 — more easily welded; 7075 — highest
strength and most easily machined. Letters
and numbers following the material designa-
tion, such as T3 to T6, indicate material
temper or heat treatment.
Laminated composite materials like
fiberglass have an even wider range in
composition, especially since both the resin
and the fiber cloths can be varied in-
dependently. The choice of resin is usually
between a polyester and an epoxy, while
cloths can be fiberglass, carbon, or Kevlar,
the last two of which are discussed in the
aerospace materials section. Due to all the
applications for fiberglass, there is as wide
a range in cloths as there is in resins. For
common body components, ordinary woven
boat-cloth is an inexpensive material, but for
ultimate strength and light weight, the more A few varieties of reinforcement fabrics. Unidirectional fiberglass and carbon
sophisticated materials are worth their extra filaments (left), and woven fabrics (right), including a Kevlar/graphite weave.
cost. Most of the “pre-preg”’ cloths require
external heat to cure them, and an oven pro-
vides an evenness that is lacking with heat in absolute strength per pound. In addition, proven to have strength-to-weight ratios far
guns or lamps. The greatest advantage of it has excellent fatigue resistance and main- greater than even titanium. Of course, their
laminates is the ease of forming them into tains its strength at very high temperatures. cost is also comparatively great. But when
large, smooth, compound body shapes. And This makes it ideal for reciprocating engine formed into components, such as a replace-
when the surfaces are too large to support parts and exhaust headers. Unfortunately, it ment for glass fibers in an epoxy binder,
themselves, it is relatively easy to reinforce is practically impossible to cast and very their strength and stiffness is incredible. As
them with sandwich panels of foam or difficult to machine, weld, or lubricate, material production increases and race car
honeycomb or with integral ribs or which limits its usefulness primarily to construction costs rise, they are becoming
stiffeners. forged connecting rods. It has been used for more and more common. ~
Other plastics are frequently used in race race car chassis and suspensions, but most oe

cars, although seldom for structural pur- often only when cost was absolutely no AEROSPACE MATERIALS
poses. Urethane foams are available in either object. By now, the legendary advantages and
flexible form, as in bump-stops, or in rigid Spin-offs from the aerospace industry are disadvantages of carbon filament materials
form, to be used as a stiffening filler or making other forms and combinations of are familiar to anyone who can read. These
impact absorber. It is sometimes easiest to materials available to racing people. Where filaments are far stronger and stiffer per
buy the foam as a two-part liquid, which large flat areas are required to support a lot pound than any other material available for
foams up to fill a cavity when the two are of weight without buckling, as in fuel cell race cars — and they are also many times
mixed. Thermoplastic materials aren’t nor- containers or aerodynamic surfaces, metal more expensive. Currently, however, more
mally used because of their sensitivity to or plastic sandwich materials are being and more teams are justifying their use in
heat and their general lack of structural used. These can be in any combination of race cars, from Formula One to IMSA. The
strength. The exceptions are acrylics (Plex- materials that can be welded or bonded most dramatic introduction, of course, was
iglas), used for transparent windscreens, and together. The most familiar combinations the 1983 McLaren MP4B. This graphite-
polycarbonates (Lexan) for spoilers. are probably fiberglass skins over a core of filament epoxy chassis was of a sophistica-
Magnesium is most common in racing urethane foam or aluminum skins over a tion that it could be flown into outer space.
wheels and other castings where a large core of aluminum honeycomb. The chief There is a whole continuum of levels of
given volume is required to resist deflections drawbacks are that they are difficult to join technology below this, however, so that
or buckling. In that case, its density being at their edges and practically impossible to everyone down to the club racer could be
70 percent of aluminum overrules disadvan- form into compound curves. taking advantage of these amazing material
tages such as higher cost and casting The last material to be mentioned was too properties — for a price. From an engin-
porosity problems. Since most racers will advanced to be discussed in the first edition, eering standpoint, a preliminary step would
merely buy components already produced but it now deserves its own section. The be to optimize the structural design with
in magnesium, it doesn’t matter that it is ex- failure point of any material is usually existing materials, as described in Chapter
tremely difficult to form or weld. And while determined by inescapable impurities in the 15. But for the average constructor, who uses
it isn’t easy to ignite, once temperatures get composition. But if the material is literally the “‘build it, break it, build it stronger”
high enough it will burn intensely — grown, molecule by molecule, in an inert technique, there are still definite advantages
especially the fine particles from machining atmosphere, then absolutely pure fibers or to the new materials. And there are a whole
Operations. whiskers can be created. These pure lot of problems!
Titanium is the strongest metal available filaments of carbon, graphite, or boron have The subject can be broken down logically

93
 would have a tendency to fracture. In this
case, Kevlar might be a better bet. A chassis
is a different story, where torsional stiffness
is the primary consideration. The graphite
McLaren MP4B is a good example of a
chassis that is torsionally stiff enough to let
the suspension do its job — while coming
nowhere near its ultimate strength limit
(even in crashes, which are far above
ordinary racing stresses).
In a serious engineering design, you get
to the point where the optimum material can
vary with location in the same given part.
In a simple beam, you might want upper
layers of carbon to have the maximum com-
pression resistance, and lower layers of
Kevlar for its tensile and impact properties.
In other cases, it might be best to alternate
layers of carbon and Kevlar to take advan-
tage of both. In fact, most composite
weavers offer fabrics with any combination
of carbon, Kevlar, and even glass in the
same cloth. It isn’t unusual to see a race car
tub with distinct gold/black weave patterns
Sandwich composites: flexible aluminum honeycomb core which can be bonded illustrating the mix of Kevlar and carbon.
in compound curves (left), and various combinations of aluminum/Nomex/car-
bon/Kevlar (right). The Weave
Historically the pattern, or weave, of the
into two sections, “the materials” and ‘‘the of glass. Although carbon is closely related fabric has not been a consideration in race
construction processes.’ However, it is im- to diamonds, the cost is not due to the raw car parts. In fiberglass, you use what you
possible to provide a working knowledge in material, but the processing — in relatively can get, which is usually 50/50, or equal
such a short space, even to those who are limited production. Most carbon fibers are, numbers of strands in each direction. When
experienced in fiberglass layups. Because of in fact, produced by heating rayon fibers to engineering a part to the optimum, however,
the aerospace heritage, there is a vast a temperature at which they “carburize.” the directionality of primary stresses must
accumulation of knowledge which is being So as applications and volume increase, the be determined in each area. Ina wing, for
obsoleted and updated daily. In addition, the price will continue to decline drastically. example, the greatest stresses are lateral
required tools can range from special 50 The price was about three hundred dollars tension in the top surface, and lateral com-
dollar scissors to hundred-thousand dollar a pound just fifteen years ago! pression in the bottom surface. If you use
autoclaves (pressurized ovens). So the com- Although carbon fibers have gotten most a 50/50 cloth here, you essentially waste
panies with the tools and technology are not of the publicity, DuPont has a relatively new most of the 50 percent of the strands that run
terribly interested in an application that does aramid fiber called Kevlar, which is a in the longitudinal direction. On the other
not promise a market of perhaps a quarter reasonable compromise between glass and hand, in the sides of a chassis tub which is
million dollars. Still, anyone can start carbon. (Nomex is a related aramid fiber.) torsionally stressed, a 45 degree orientation
experimenting on a small scale, and work Interestingly enough, Kevlar fibers actually of fibers is the optimum layout. Also, here
their way up. have a slightly greater strength-to-weight the peak loads will be symmetrical, in left
ratio than most graphite fibers. On the other and right turns, so a 50/50 weave would be
The Fibers hand, while Kevlar has twice the stiffness appropriate.
Carbon/graphite fibers, at first glance, per pound of fiberglass or metals, it has only Among the dozens of weaves available
look like nothing more than black fiberglass. half the stiffness of carbon. And Kevlar (even with the same filament material)
Technically, there is a distinction between currently costs less than half as much as another important consideration is
carbon filaments and graphite filaments, the carbon. So there are some very important “drapability” or the ability of the weave to
latter being a slightly more refined product, tradeoff considerations to be made in the conform to tight compound curves. Given
but for our purposes they are essentially the selection of materials. the limited availability to racers, however,
same. Apart from the shape, however, there At this point, we should analyze the this selection criteria can be overlooked
is no comparison with fiberglass. Pound for distinction between material strength and slightly by using a little more “‘cut and lap”
pound, graphite fibers are two to three times stiffness, and the proper application of each in the layup.
as strong as glass fibers and three to five in a race car. Which panels and parts do you The extremes of weaving ratios are
times as stiff, depending on the grade. (Here want to be strong, in terms of ultimate reached at “unidirectional” cloths, in which
we are talking about the fibers alone. When failure, and which ones do you want to resist perhaps 95 percent of the fibers are in one
bonded into a “composite,” the comparison bending? A good example might be a front direction. The other 5 percent are merely used
is not quite so amazing, as will be shown spoiler. Making it out of carbon fibers to hold it all together until layup. In this case,
later.) The cost of these fibers is also would give it the minimum air pressure there is essentially no strength in the cross-
amazing, starting at about $30 per pound in deflection per pound. However, when it ply direction, except for what is provided by
large quantities, or about 10 times the cost impacted the pavement at high speed, it the resin, so successive plys must be carefully

94
oriented with respect to each other.
Beyond (or before) weaving is the case of
pure roving, or “rows” of roving in “‘tape”’
form. Roving is used primarily for filament
winding of cylindrical or spherical pressure
containers and is of small consideration to
the race car builder. The tapes, however, are
rows of filaments from 1 to 24 inches wide,
held loosely together by a viscous semicured
resin pre-impregnated in the fibers
(“prepreg”). These tapes can be relatively
easily applied in complicated molds —
given a carefully planned number and orien-
tation of plys. This, of course, gets us into
the glues.
Resins
Polyester resins are so common as to be
almost beneath mention. You can buy a
quart can and hardener at any hardware
store, and simply follow the directions on
the label. With a little more sophistication,
you can distinguish between laminating, This relatively simple carbon composite tub is made up of sub-assemblies, the top _
sanding, finishing, jell coat, and tooling and bottom halves being joined at the beltline. (Renault)
polyesters for the appropriate application.
In aerospace work, there is even a vinyl
ester resin which is good for corrosion times” and “five times’? mentioned cesses which might be used by race car
resistance, higher temperatures, better wet- previously. In the first place, all the fibers fabricators or smaller aerospace contractors
out, and more impact flexibility. However, cannot be working all the time — unless you who are locally available.
even the best of them are not as strong as have a rope in pure tension. Varying loads The simplest form of composite structure
the epoxies. For race cars, the only advan- on a structure require some redundancy of is the familiar wet layup of woven fiberglass
tage of polyesters is their cost, which is materials which will not be carrying loads cloth with polyester resin. While anyone can
about half that of epoxy — and familiarity. in any given situation. Secondly, the fibers do this in their own garage, it takes years
Epoxy resins have a poor reputation out- are just a part of the finished structure. of experience to develop the skills for
side of the aerospace industry, where they Anywhere from 30 to 60 percent of the making good molds and strong, light,
have become the standard for strength in weight will be resin, which contibutes attractive parts. One of the best freelance
composites. At one time the epoxies were relatively little strength. (Imagine a chassis bodybuilders is Wayne Hartman, who
not only harder to work with, but more toxic made out of just a thin film of epoxy resin.) makes everything from aftermarket spoilers
than polyester. Now, however, the situation If there is 50 percent resin content in a to CanAm and Indy car bodies. Recently
has reversed, with safer and more diver- laminate, then the strength to weight ratio he has picked up a lot of experience in wet
sified epoxy resins. You still can’t drink of the structure will be about half that of layup of graphite, Kevlar, and epoxy. So a
them, but at least they won’t knock you the raw fibers — at best. little of his practical knowledge will be
down if you open the can in a closed room. A real world example of first-stage gains described here.
Epoxies come in two types: Catalyst-cure were the first G.T. racer body skins pro- Compared to fiberglass, the greatest dif-
and thermal-cure. The catalytic resins are duced in carbon filament. Using conven- ferences in the use of carbon and Kevlar are
available in familiar over-the-counter for- tional wet-layup polyester techniques, and in cutting, wetting, and draping. Carbon cuts
mulations ranging from one-ounce tubes of just substituting 50/50 dry woven fabric, and trims neatly because each fiber is
“5-minute”’ epoxy to commercial 55 gallon fabricators were able to reduce the weight relatively brittle and tends to snap off. But
drums of 2-part resins, with cure times of of composite body parts by an average of Kevlar is so incredibly tough that special
from 30 minutes to 2 hours. The thermal- about 30 percent. While this is worth killing edge-hardened shears are recommended,
cure resins are not commonly used outside for in most cases, it is nothing compared and trimming finished parts can leave a
of the aerospace industry, because of the to what is available if aerospace processes mess of frayed edges. One tip is to trim with
fabrication problems of even, controlled are used in combination with these a sharp blade while the resin is still slightly
heating, and the effects of temperature on aerospace materials. tacky. Wetting out, or fully saturating the
ordinary mold materials. Recently, however, fibers, is harder than with fiberglass for two
epoxies have been developed with cure Aerospace Composites—How to... reasons. First, with carbon you cannot see
temperatures as low as 200 degrees The first McLaren tubs were built by an air pockets through the material, as you can
Fahrenheit, with a small loss in cured aerospace materials company called with fiberglass. You have to get right in there
strength properties. Hercules, using facilities that are very ex- and feel for bubbles with your fingertips.
Up to this point, we have been looking pensive and relatively rare, and some tech- Secondly, some of these fibers are designed
at the relative strengths of fibers, and the niques that are closely guarded secrets. primarily for pre-pregging, and are not
resins used in “wet layup.” When you put While the more exotic processes will be coated for ease of manual wetting. It is a
these together in a finished composite struc- explained for their academic interest, we good idea to get a lot of advice on resin/fiber
ture, the gains will not quite be the “‘two should first start out with more basic pro- compatibility from your supplier, and to

95
 of resin is reduced — although it is usually
insignificant compared to fiber costs.
Once you get the plys in place, they must
be heated. To really optimize the resin pro-
perties, the heating cycle should be carefully
controlled per supplier recommendations,
including rise time, cure time, and post
cure. For less critical applications, some
racers have been known to use heat lamps
or heat guns, when a large enough oven is
not available. Naturally the mold has to be
fairly heat-resistant, to avoid warpage — if
not total collapse — under these conditions.
Pressure Curing
To get really intimate contact between all
the plys and the resin, some sort of pressure
application is mandatory. In mass produc-
tion, matched metal dies are most practical,
but for just a few parts, or very large areas,
vacuum bags are used. These are carefully
laid out, multiple-layered bags which are
sucked right down on the surface of the
prepreg plies. Of course, this means that the
mold has to be even stronger, to take
pressures of up to 2,000 pounds per square
Chevrolet Engineering test facilities which can rapidly stress-cycle a bare foot. An alternative, where possible, is to
chassis and driveline for durability—in this case, a lightweight Stingray. use a bag skin on both sides of every
(GM Photograph) surface, so that the mold is merely com-
pressed, and therefore requires no extra
start with the slowest-setting resins. The will not even mention. Unfortunately, structural strength.
drapability problem ought to be obvious. because there is almost an infinite number This pressurization of the curing plys not
You are not going to get a very stiff carbon of combinations of fibers, weaves, and only guarantees a good bond, but it can be
fiber to bend into as tight a radius as resins, there is no standard product that you used to squeeze out a little excess resin. This
fiberglass. And even if you do manage to can walk in and buy off the shelf. Also, is more complicated than it sounds, and has
force it, the fibers can develop micro-cracks while temperatures of from 200 to 500 become a proprietary process that not many
(in corners of less than about 1/8 inch degrees are necessary to cure the resins, companies want to talk about. Essentially
radius) and severely weaken the finished they will slowly set if left at room what it requires is multiple layers of
part in critical locations. temperature for some time. This is not an specialized materials between the bag and
Even the best of wet layups have two alternate method of curing, but a limitation the part. First is a release agent that can take
major drawbacks which oppose each other. on shelf life. And because of the automation the temperatures. Then the edges must be
Either the finished part has too great a resin of the impregnation process, few companies sealed with “side dams.” Then a teflon-
content for optimum strength-to-weight will start up a production run for less than coated porous fiberglass release bleeder is
ratio, or you run the risk of having air a thousand-dollar order. But because of this applied, and covered with more layers of
pockets or voids between plys. In aerospace production process and shelf-life problem, bleeder cloth (each layer will bleed off about
construction, more than 1/2 percent of air sometimes roll-ends or remnants are 2 percent of resin). These bleeder cloths
content in the cured part is unacceptable, available if you make the right connections. also insure that pockets of air do not get
while 2 percent void content is common in The advantages of prepregs are both in trapped in cavities, but are vented to the
the best wet layups. To work the air out, the fabrication (given the right facilities) and vacuum line. Over this goes another release
general practice is to pour in an excess of final strength-to weight ratio. Many of the film to protect the nylon vacuum bag. If you
resin, work it in, and squeegee it from the nuisance problems of dry fabric are avoid- want to go to that much trouble to get the
center out to the edges. Still, it is hard to ed by starting with a slightly tacky sheet of last 5 to 10 percent weight reduction, you
get the resin content below 50 percent by material. A backing paper prevents the should probably contract out to someone
weight using this method. All of these layers from sticking on the roll, and it can with the tools and experience.
problems can be reduced however, by the be peeled off as each ply is laid. Because a vacuum bag is limited to
advanced technologies of prepregging, heat But the real advantage is in low resin con- atmospheric pressure of about 14.7 psi, the
curing, vacuum bagging, and autoclaves. tent. Prepregs are delivered fully saturated next step up is a pressurized oven, or
with only 30 to 40 percent resin, and using autoclave. Autoclaves are no small invest-
Prepregs the right fabrication techniques, that can be ment, since they have to be built to take both
Fabrics and tapes which have been pre- improved to perhaps 25 percent at the ab- the temperature, and pressures up to 100 psi.
impregnated with a heat-curing resin are the solute minimum. Obviously, for a given Hercules has an autoclave for curing rocket
standard material in aerospace. Any kind of strength in the fabric, a 25 percent weight motor cases that is 10 feet in diameter and
fabric is available with any kind of resin: reduction over wet layup is worth a little ef- 35 feet long, but many other companies have
epoxy, polyester, polyamine, and others I fort. Not only that, but in volume, the cost smaller versions.

96
 Anyone as ingenious as a racer can come
up with alternatives to this hardware. Paul
MacCready, of Gossamer Albatross fame,
built his own oven to make 20-foot long
graphite spars. All it consisted of was a
20-foot length of steel tubing a couple of
inches in diameter. To apply the pressure,
he wound his prepreg tubes with heat-shrink
mylar tape, which contracted under curing
temperature to apply the required pressure.
When you get up to 300 degrees and 100
psi, the mold takes quite a beating. Resin
can be forced through microscopic
porosities in the best of mold surfaces. At
this extreme, carbon fiber/epoxy molds, or
metal molds, or metal-sprayed epoxy molds
are required. It is possible to make an
extremely tough mold by metal-spraying
onto a pattern (even clay), and supporting
this metal film with a metal-filled epoxy. An
added advantage to this process is that
heating and cooling channels can be cast
right into the mold, saving the cost and
wasted heating in an autoclave. A good sedan roll cage may have enough tubes, cross-bracing, and
It is worth keeping in mind all through triangulation to be an adequate space-frame in itself.
this explanation of design, fibers, weaves,
resins, and processes, that everything is
interchangeable. A computer design in then rigid enough for the final cure without torsional stiffness per pound that only
fiberglass cloth can be cured in an autoclave, molds or bags. The design, tooling, and tubular space frames and monocoque con-
just as unidirectional carbon fibers can be first prototype tub probably costs less than struction can be considered for race cars.
used in a polyester wet layup. The explana- $100,000, and subsequent tubs could be While popular choice has fluctuated be-
tion has merely proceeded in a logical made for a fraction of that, with material tween the two for years, it is probable that
sequence, from the simplest, most familiar costs only about a couple thousand each. the best solution is a combination of the two.
— and structurally least efficient — to the More recently, the technology has spread Monocoque, or stressed skin construction,
most exotic in each category. Perhaps it out to all Formula Car teams. Many of is most efficient where large areas carry a
seems like a waste of time to describe the them, for example, rely on a British fabri- distributed load, such as aerodynamic sur-
upper limits of a technology which is cator called Advanced Composites Com- faces and fuel containers, or where the skin
unavailable to all but a few aerospace pany. Basically, the same techniques are serves a double purpose as structure and
fabricators. But in fact, this upper limit is used, except that the tubs are often as- body. On the other hand, a tubular space
where many professional racing teams are sembled from a number of smaller pieces. frame is most efficient where there are a few
today, and where everyone else will be a few One external shell is formed from left and specific highly-loaded points to be
years later as the costs of winning escalate right halves, while another uses a top- connected, such as engine mounts and
and the technology filters down. bottom split. In many cases, longitudinal suspension brackets. Other important con-
Even though teams try to keep a lid on Kevlar reinforcement is bonded in at the siderations include space limitations, access,
their advantages as long as possible, some belt line, primarily for impact resistance, ease of construction and repair, and cost.
educated guesses can be made about their and all bulkheads are added separately. So the choice between monocoque and
chassis design and_ construction tech- Most competitive Formula One cars are space frames is not an “‘either-or” based on
niques, based on examination of the bare already at the minimum weight, so the stiffness versus weight. The selection should
tubs. Looking at the computer-generated advantage of these tubs is often only in be based on particular requirements in each
structural drawings, it is apparent that increased chassis torsional stiffness and particular area. The choice is further con-
they use breakaway or dissolvable pat- the availability of more ballast. First fused or assisted by the popular use of the
terns. In other words, the walls of cavities designs in this new method are conserva- engine and transaxle as a structural member.
like the fuel cell and footbox areas are tive, naturally, with one of the early Ad- The primary requirements for a race car
formed around plaster or soluble plastic vanced Composites tubs quoted at 84 frame are ultimate strength, or resistance to
materials such as styrene foam. Layers of pounds. Even the original McLaren -- structural failures, and torsional rigidity.
unidirectional prepreg graphite are then perhaps one of the best of them -- was no Other problems common to production cars,
laid up over these cores, a honeycomb core featherweight, as I once had the opportu- such as beaming rigidity and vibration fre-
placed over that, and outer skins applied. nity to lift a totally stripped tub, and it had quencies, are seldom a problem. By far the
The whole thing is then wrapped in vacuum to weigh at least 70 pounds. The cost is most important principle in structural design
bag layers and popped in the autoclave. At never conservative, however, due to the is that stiffness increases exponentially with
a relatively low temperature, the resin labor-intensive fabrication methods. increased cross-sectional area. So for given
kicks and becomes semi-cured. Then the strength requirements and minimum weight,
cores are broken out, melted out, or dis- TYPES OF FRAMES the taller and wider the frame, the better.
solved out with a solvent. The structure is The simple ladder frame is so poor in But since few racers build their own frames

97
 representative frame and suspension can
have racing loads cycled into it until the
weakest link breaks. Then it can be repaired
and cycled again until something else breaks
— and on and on until its strength and
durability are well known quantities. Then
an all-new chassis can rise from the ashes.
The cost of such a test program may seem
to be out of the question, but it will in-
evitably save weight and avoid failures, and
perhaps even help avoid a serious track
accident.
In addition to major structural changes,
degrees
in
twist
— ft-ibs
3000
torque
at it will frequently be necessary to add or
alter brackets and other miscellaneous
60 80 100 attachments. A failure here may not affect
station — inches from rear axle the safety of the vehicle as a whole, but it
Figure 42. Chassis Torque Deflection Curve can still cause the loss of a race in many
cases. It is important to remember that all
supports must not only carry some static
component weight, but that there will be
anyway, this section will concentrate on im- degree of twist is a minimum. Knowing how frequent high g loadings from bumps, and
provements and other modifications. many degrees of twist there is, you can there will be continual high frequency vibra-
The first step is in knowing how to form decide for yourself how important that much tions caused by the engine. Exhaust systems
and join metals, particularly riveting in the adverse camber is. However, if there is an and supports are particularly prone to failure
case of monocoque and welding in the case easily-met minimum weight, more rigidity from high temperatures and vibration
of tubular frames. Durability and safety are is better. If the handling doesn’t respond to fatigue. The best insurance is to make all
so important that only a fabricator familiar bar changes, or the front wheels seem to brackets as stiff as possible with adequate
with aircraft construction techniques, or “patter” excessively over rough sections of triangulation, and to feed their loads into
perhaps a certified welder, should be track, regardless of shock absorber settings, major chassis components.
allowed to build or modify any critical com- then the chassis may need more stiffness to The best race car chassis may be designed
ponents. In the search for absolute weight resist torsional vibrations. very close to its ultimate limit of failure. As
efficiency, the chassis is only as strong as When the chassis is bolted to the floor or a chassis gets older and gets handed down
its weakest link — which could be one bad a bedplate, and weights or other large, and raced over a number of years, many
weld. For those fabricators with some measurable forces are available (such as changes take place. Metal fatigue may begin
experience who must work on their own hydraulic or pneumatic cylinders), it is also to affect chassis components that are not
race cars, Carroll Smith gives many impor- quite educational to stress the suspension. easily removed and inspected, and a suspen-
tant construction tips on riveting and Rough calculations can be used to estimate sion bracket is as important as the suspen-
welding in his books. Otherwise, the best the input forces due to tire traction in ac- sion itself. Then too, racers tend to update
bet is a quick course in aircraft fabrication celeration, cornering, and braking, and the older cars with larger tires and engines. So
or at least a handbook on the subject, such loads can be simulated by forces acting on as the chassis gets older and weaker, it is
as Ron Fournier’s Race and Custom Car torque arms bolted to the hubs. With enough also subjected to greater forces from in-
Fabricator’s Handbook. hardware, an accurate simulation can be creased weight and cornering loads. A car’s
made of combinations such as increased weight should never increase merely by the
STRENGTH vertical load, and braking and cornering added weight of a heavier component; there
Before a chassis can be modified, there forces on the outside front suspension. The should be a proportional increase in the
must be some means of determining how corresponding deflections or deformations strength (and weight) of supporting
strong it is already. Early in the original are likely to be frightening enough to structure.
construction stage or at any time thereafter, motivate stronger reinforcements. It isn’t Large production cars modified for racing
the bare chassis and suspension components unusual to see the suspension and/or frame usually don’t have critical strength
can be loaded and twisted while deflections deflect by some degrees under realistic problems. Still, they aren’t designed for the
are measured. The maximum twist load that loads. kind of loads that can be generated by in-
a race car is likely to be subjected to will A great advantage to such static testing creased tire capabilities, and they are apt to
be the entire front weight (including aero is that the weakest locations can be pin- have an undesirable amount of flex. The
loads) carried by one front wheel, which pointed (see Fig. 42) rather than having to most common problem is how to remove
may then be accelerated upward at 2 to 3 make intuitive or massive corrections. By unnecessary structural weight and yet add
g’s due to bump loads. That is multiplied using a rigid base surface and a number of back the minimum amount needed to in-
times the chassis track width to get the dial indicators, it is possible to localize the crease torsional rigidity.
torque in pounds-feet. With a torsional test major sources of deflection and to evaluate So much can be done to a passenger car
setup as described in Chapter 13, the chassis the corresponding reinforcements. body and frame that it may be fastest to start
can be stressed and measured, and if the Once this setup and development is com- with a new unleaded and unpainted shell.
angular deflection is excessive, the frame or pleted, it isn’t hard to go to the next step The first step, when allowed by the rules,
suspension should be reinforced. For light — durability testing. If there is some means may be to acid-dip or cut away all non-
road-racing cars, 3000 pounds-feet per of continuously varying the loads, a structural components — and even all struc-

98
tural components that will later be re-
inforced by a mandatory roll cage. At one
time there were a couple of aerospace com-
panies in the Los Angeles area that had the
capacity to dip complete full-size
automobile bodies. The operation takes
some care in planning for the amount of
metal that can be safely removed without
weakening large panels too much, and it
may require extensive masking for those
areas where no metal removal is desired. It
is also somewhat expensive, but it is a very
rapid way to cut the weight by many
hundreds of pounds. Even when the
minimum race weight is specified by the
rules and is easy to meet, acid-dipping is
still a worthwhile process. By removing the
maximum amount from the front and top
of the body, it becomes possible to replace
the weight in ballast where it is more useful.
It may be possible to move the center of
gravity down and rearward by inches by the
careful placement of hundreds of pounds of This balsa wood model of a sedan subframe can be qualitatively evaluated
ballast. Unfortunately, it seems that Federal for rigidity per pound with various combinations of bracing.
OSHA safety rules and EPA hazardous
waste regulations have curtailed acid dip-
ping, so it may be impossible to have it done structural purpose. As far as the rules go, which may be in the way. Also, cutting and
in the U.S. any more. all it has to be is a protective box around rewelding all that sheet metal can have a
Whether dipped or not, all spot-welded the driver. But by fully triangulating the box disastrous effect on body stiffness, which
body seams, and all intermittent welding on for stiffness, and connecting it to the front can be corrected with properly located roll
frame rails, should be completely seam- and rear spring mounts, it can become a cage tubes. The same is true for suspension
welded into one rigid unit. The factory fully adequate space frame in itself. Some travel clearances, where lowering the chassis
doesn’t do it for cost reasons, but racing racing sedans would probably be just as creates some interference with the existing
demands it for the added strength and strong with none of the production body frame. There can never be too much suspen-
durability. In the case of previously used components or frame members at all. The sion clearance — until the frame or body
components, sandblasting is not only body tends to be no more than an easily hits the ground.
necessary to clean the metal for welding, but replaced external skin. ‘Finally, all that roll cage material can
is an alternate means of weight reduction Where some attempt is made to maintain protect a lot of other valuable things besides
for non-visible panels. the original body and frame — because the the driver. For safety, the fuel cell should
Cast-iron engine blocks and heads have rules demand it or cost is a major problem be well surrounded, and for insurance it
been acid-dipped with some success, but the — then the cage will have to fit whatever helps to have the engine room somewhat
lost weight is hardly worth the effort in- exists. The major points to be connected by enclosed also. Any serious accident will
volved. Usually a finished block must have the roll cage tubing are: front spring and probably tear off the wheels and suspension
all its machined surfaces masked off before shock mounts, engine mounts, front sub- and a lot of sheet metal, but everything else
dipping, including bores, bearing surfaces, frame mounts, and all rear leaf spring and should be salvageable.
and all flat mounting faces and threaded shock mounts. In the case of unit- If it isn’t obvious, any rectangular
holes. An alternate procedure is to start with construction bodies, it will be neccessary arrangement of reinforcing structure is very
an unmachined raw casting and have it com- to have reinforcement plates to spread the inefficient in terms of stiffness per pound.
pletely machined after dipping. In either load on the sheet metal wherever a tube is There are very few places in a race car
case, there is always the unknown question to be welded in. Otherwise, in the case of where full triangulation of all members is
of durability from reduced wall thicknesses. separate frames, the cage must be welded impossible. It is not too practical to have a
The first and major reinforcement to a in to the frame itself. If the body normally diagonal tube crossing the windshield, and
production chassis is the roll cage. Most rests on rubber insulators, it will add to it will be necessary to leave a space up front
race sanctioning bodies specify minimum chassis stiffness if they are removed and the for engine removal — although bolt-in
requirements in terms of tubing size and body is welded or bolted rigidly to the diagonals are sometimes used. If there is
locations and amount of triangulation. frame. If the engine is not used as a struc- any question about optimum tube locations,
NASCAR probably has the safest and most tural component, it is probably better to it is fairly easy to make small test models
elaborate roll cage specifications, which are keep it on rubber mounts to absorb some out of balsa wood. These models can be
good to follow in any instance where the of the vibration. loaded or twisted by hand and measured for
weight is no great penalty. But for lighter The best time to open up the wheel wells deflection with a dial gauge. Making a
cars on slower tracks, the SCCA specifica- and fenders for greater tire clearance is number of configurations and comparing
tions are relatively safe.. during construction. Tire sizes will often their stiffness per pound will at least give
As long as the roll cage has to be there become larger than anticipated and it can a relative value for different tube
for safety’s sake, it might as well serve some be very difficult to relocate roll cage tubes arrangements.

99
 should also be considered structurally. As
long as a body has to exist for aerodynamic
reasons or rule requirements, it might as
well serve another purpose as a structural
member whenever possible. Assuming that
the final external shape is resolved in
original design and won't have to be ex-
perimented with or changed, the only
restrictions are due to access. Unfortunately,
practically every enclosed space on a race
car will require access at some time or
other, so it has to be a matter of com-
promises. At any rate, locations requiring
only occasional access, such as over fuel
tank bladders or the pedal areas, can be
taken care of by using structural cover
plates.
At the very least, bodies should use their
skins efficiently enough so a minimum of
frame structure is necessary to support
Stress, deflection, and vibration data on a chassis can be accurately scaled them. Downward air loads are hopefully
up from models in PVC plastic, such as this 1/4 scale 1967 Camaro. great enough to require a lot of strength also,
(GM Photograph) and the lifting forces due to negative air
pressures shouldn’t be ignored either. With
On the subject of models, it is possible industry commonly does this with complex some forethought, a fiberglass body skin
to get very accurate data from small models, unit-body designs, and it is a simple opera- can be designed with integral stiffeners and
which can be scaled up to the final chassis. tion for much more straightforward race car a continuous shape that is totally self-
Properties of metal components in full scale chassis. It takes some complex mathematics supporting. And finally, even if the body is
can be duplicated by the properties of to make accurate scale factors, but the strong enough in itself, it had better be
certain plastics in smaller-scale thicknesses. process is still valuable even for qualitative attached with some care. It is nice to be able
Steel, for example, can be simulated by the evaluations. SAE Paper No. 710262 goes to get a panel or body off in a hurry to make
use of clear, rigid polyvinyl-chloride-acetate into more detail. The effort spent in building repairs, but not if an aerodynamic surface
(PVC-A), which can be easily fabricated by models of a proposed chassis could save a can come loose at high speed. The loss of
thermoforming and welding. It is possible lot of expensive rebuilding or reinforcing in bodywork has had fatal consequences in
to fabricate a perfect scale model of a the final product. racing — not to mention all the lost races
proposed chassis, perform small-scale tests Race car bodies are primarily an aero- it has caused. With the magnitude of
for strength, deflections, and vibration fre- dynamic consideration, and better covered downforces that are being generated, the
quencies, and then accurately scale the data in that chapter. But they tend to make up strength of body panels can be every bit as
up to the actual chassis. The automobile such a large part of a race car that they critical as suspension components.

100
 11
SAFETY
Race car safety is a responsibility that 12-point bolt heads indicate extremely high- threads and even half-heights for shear-only
tends to be left up to the sanctioning bodies. strength bolts that will outlast just about applications. Most can be reused a few
Every racer’s first consideration is simply anything they could be attached to. Then times before the nut can be freely turned by
to win, and the thought of something going there are also titanium bolts, which are just hand. For temperatures over 250 degrees it
wrong, or of an accident, is usually so as strong and yet save some weight, but will be necessary to use all-metal lock nuts,
remote that few are prepared for it. Even which are seldom worth it for their tendency which may not be reusable, or castellated
when it is considered, not many racers are to gall and corrode — and their expense. nuts with cotter pins. For absolute secur-
anxious to sacrifice weight — much less Racers who know (accurately) what loads ity, and a constant visual indication that
money — for safety equipment that gives are to be carried by specific bolts can also everything is tight, it is hard to beat safety-
no improvement in performance and will save weight by not using unnecessarily large wiring. It may even be the only answer in
only be put to use one time. But safety diameters. Ultimate tensile strengths vary the case of bolts in tapped holes. After the
involves more than protection in case of an from 125,000 psi for AN bolts to 160,000 wire is threaded and twisted, it should be
accident. An equally important aspect is psi for NAS bolts. It is a simple matter to wound slightly clockwise around the.
avoidance of the accident in the first place. divide the load by the strength to get bolthead or nut before anchoring.
Accident avoidance is something that means optimum bolt cross-section area, even if it Tension bolts must be carefully pre-
money in the bank, by protecting the car in- isn’t easy to anticipate exactly what the real stressed with a torque wrench. For really
vestment and insuring a finish. Racing is an loads are going to be, due to accelerations, critical applications, such as connecting rod
inherently risky business, but the risks can shocks, and vibrations. It takes a com- bolts, it is far more accurate to gauge the
be greatly reduced by taking certain precau- plicated analysis with a lot of other tension by measuring bolt elongation with
tions and knowing things that are easy to considerations, so a large margin of safety a micrometer, but the torque wrench is
check. should be allowed. otherwise more convenient. The problem is
The strongest bolt is worthless if the nut that bolt torque is strongly affected by fric-
ACCIDENT AVOIDANCE
can’t be kept on. Racing vibrations and fluc- tion at the threads and bearing surfaces.
FASTENERS tuations in stress will inevitably loosen any This can be minimized by using washers
The first and most obvious means of nut that isn’t locked on in some way. Lock under each bolt head and nut and by
accident avoidance is to make sure nothing washers are relatively undependable and lubricating the surfaces with an anti-sieze
becomes disconnected or falls off. This have a tendency to tear up the light alloy compound or moly grease. Proper bolt
requires a good knowledge of fasteners and materials in race cars. The most familiar torques for each type and size and proper
how to fasten them. Most auto mechanics lock nuts are the type with integral plastic torquing techniques are usually provided
are familiar with the SAE standards for rings which are now available even in hard- with torque wrenches.
bolts, which are coded on the bolt head. ware stores. They come in fine and coarse When all else fails, there is always
There is no reason to ever use anything less
than a Grade 8 bolt on anything that could
cause the loss of a race or loss of the car
if it failed. But in addition, better bolt
standards are set up by the government: AN
specifications for aircraft; and NAS
specifications for aerospace. Such bolts are
not only of a much higher quality in strength
and dimensions but far more expensive and
difficult to purchase. However, due to
government inefficiency, they are often
available at surplus stores at prices usually
lower than new SAE bolts. All a mechanic
has to do is search through bins with a ruler
and a list of the sizes needed. It helps to
know the code of designations, but the im-
portant considerations of diameter and
length are obvious — keeping in mind that
AN/NAS bolts have a shorter thread length
so that the threads never carry shear loads.
Bolts' designed primarily for shear have
thinner heads, and in many cases the heads
come pre-drilled for safety wiring. NAS

101
 however, fast and sensitive crack-detecting
systems are available at reasonable prices.
But first, the surfaces of all critical com-
ponents should be carefully prepared and
cared for. In initial fabrication, large radii
and careful polishing (in the direction of
stresses) reduces the concentration of stress.
In addition, shot-peening increases the
ultimate strength by pre-stressing the surface
material. From then on, the components
should be carefully protected from corro-
sion, since stress and corrosion contribute
to each other.
If there are no obvious surface flaws,
there are still three methods of finding any
microscopic or invisible imperfections: X-
ray, Magnetic particle, and dye penetrant
inspection. X-ray is relatively expensive and
necessary only once, to show up any sub-
surface flaws, voids, or casting porosity. If
it is going to be done, it is best to do it
before the raw forging or casting is
machined, to save that much expense. But
if there is a history of component failures,
and the manufacturer doesn’t X-ray, it may
still be worthwhile to do it after purchase.
Magnetic particle inspection, or
Magnaflux, is the most popular technique
for finding external flaws, but of course it
is only possible with ferrous parts. The
component must be removed and cleaned,
magnetized in a solution of metal particles
—~
and fluorescent dye, and inspected under
AHact REQUIRED
BET On
HNC
NO
$ USE AS eR
TG VAD® black light. Any surface cracks or flaws will
show up as obvious red lines. All non-
ferrous materials such as aluminum and
The very thorough NASCAR inspection form, which attempts to insure both safe- magnesium must be inspected by a dye-
ty and equality in rules enforcement (and sponsor credit). penetrant technique such as Zyglo, which
is similar except for the magnetization.
Portable and inexpensive kits are available
Loctite, an essential series of liquid resins Because they are hollow, shear strength is which can be used at home. However, pro-
used to lock any two tightly-fitting fasteners a little less than a bolt, although they are fessional service is available at any airport
together. It requires careful pre-cleaning and usually made from high-grade steels. Flush large enough to have a good engine
application but is self-hardening and quick-release fasteners such as the Dzuz rebuilding shop, and some race sanctioning
stronger than most lock nuts. In‘ addition, fastener are inconspicuous and neat but have bodies require official inspection certificates
it prevents fluid leakage and corrosion. less holding strength and can be difficult to for critical components.
There are many grades and types for ap- operate in a real emergency. They should The big question is in trying to determine
plications from locking to sealing to be used only for non-critical, non-load- which parts are critical and how often they
bonding, and for other cases of excess bearing body surfaces. For the same should be inspected. There are a few life-
clearance such as bearing fits. However, reasons, all body panels should have edges or-death components on every car, such as
since it takes some time to apply correctly, protected so that the airstream can’t get axles, hubs, spindles, hub carriers, and
other locking methods are sometimes more under them and rip them off. If overlapping steering arms, which could cause a serious
convenient. edges can’t be arranged correctly, it may be accident. But there are other highly-stressed
Other fasteners may have to be released necessary to seal the gaps with racer’s tape. parts, especially in the engine, transmission
during a race, for access or to replace a and driveline, which would merely cost a
damaged body component. A number of INSPECTION lost race and some internal damage. The
quick-release pins or fasteners are available, Since the nature of race cars is to have cost of inspection is insignificant compared
but the most common are the familiar hood all components as lightweight and struc- to normal teardown time and the potential
pins with a wire locking clip. Where turally efficient as possible, frequent loss from failure. Where possible, of course,
removable shear pins are required, as in inspection for potential failure is mandatory. the best practice is to anticipate the poten-
hinges or on highly-stressed components, Given a high-power microscope and enough tial for failure, and minimize the damage by
ball-lock pins, or pip-pins, are the best time, a mechanic could go over every com- providing a fail-safe design. The best
answer. These are hollow pins with button- ponent inch by inch to find beginning stress example of this is dual isolated braking
operated detent balls to hold them in place. cracks. Due to similar aircraft requirements, systems. The more a driver knows about the

102
condition of his car, and the care the
mechanics are putting into it, the more he
can concentrate on the other risks of racing.

DRIVER PROTECTION
When the most careful efforts toward
accident avoidance fail and the inevitable
accident occurs, the driver is going to be
praying for every protection device
available. There are so many possible types
and combinations of accidents that it is im-
possible to be prepared for everything. But
the driver, engineer, and mechanics should
be aware of some of the more likely ones,
so they aren’t left to the sometimes arbitrary
standards set by sanctioning bodies. Their
safety rules are the minimums they will
allow, and more is always better. In the long
run, more protection will even be worth
what it may cost in lost performance due
to increased weight. In fact, one of the
biggest problems in providing accident
protection is minimum weight regulations
that are too hard to meet. If the minimum
weights are high enough, everyone can add
more protective (and stiffening) structure
and better fire prevention equipment with
no one at a disadvantage for it.
To understand a total safety system, it is
necessary to break an accident down into
four common ways a driver can be injured:
excessive decelerations, impact with the
interior of the car, collapse of the car, and
post-crash hazards such as fire or suffoca-
tion. Any one of them alone can be fatal,
so each one must be anticipated and pro-
tective measures taken.
DECELERATION
Healthy humans can survive decelera-
tions of over 50-80 g’s for brief instants
under ideal conditions. Using the equa-
tion for deceleration versus distance and
speed from Chapter 5:
(.0334) (mph)?
(distance)

will show that it is theoretically possible to Cause or effect? Sometimes it isn’t easy to resolve which came first, the impact
stay within a 50-g limit at 75 miles per or the failure, but here are a couple of reasons for caution.
hour if the driver is evenly decelerated in
about 4 feet. In other words, the driver
could walk away from a collision with a owners and race drivers debate the relative the driver doesn’t contact any rigid part of
concrete abutment at 70 mph if the front of value of rigid versus absorbent barriers, and the vehicle interior, which could cause bone
the car could be designed tocollapse evenly whether the errant race car should carom breakage or tissue damage. This is the pur-
for those 4 feet. Of course, it isn’t easy to off or be swallowed by the barrier. Chain pose of the seat, seatbelts, helmet, and other
provide such ideal collapse, but then it link fences have proved to be able to stop padded surfaces. The ideal seat might be
isn’t common for drivers to ram walls light race cars at about the correct rate — something like an astronaut’s couch, de-
head-on either. The point is that more with minimal damage to the car, much less signed to distribute high g forces evenly over
available crush distance is a tremendous the driver. But, given the necessity of rigid the maximum body area and to enclose the
advantage to survival. walls around some race tracks, the chassis driver as much as possible to prevent
Even at 200 miles per hour, 27 feet of designers must provide as much carefully undesired motion in any direction. No
constant deceleration is theoretically sur- thought-out crush space as is reasonably padding is necessary if the seat contacts only
vivable, but that is obviously beyond any possible. soft fleshy areas, so a perfectly rigid shell
potential in chassis collapse. Race track A survivable limit of 50 g’s assumes that could be adequate. Of course, the mounting

103
 most valuable and irreplaceable component
in any race car is the driver’s head. Protec-
tion is a bargain at any cost.
RETENTION
Seatbelts are fairly well defined by most
racing groups. In any kind of accident in any
kind of car, they have proved to increase sur-
vivability by at least a factor of ten. The
desirability of being thrown free is a total
myth from the standpoint of tolerable g’s and
required deceleration space. The absolute
minimum belt standard is a four-point
system, and one or two extra crotch belts
to prevent submarining may also be
necessary. There are even systems that
include arm and helmet-restricting straps.
Belt anchorages are critical, in both ultimate
strength and the angle of pull on the driver’s
body. A final precaution is to protect the
buckle from accidental snagging, which
could release it at the most unfortunate time.
ES
: ae”
SONS
Re
e The impact-absorbing rigid foam in
helmets is a good example for protection in
This racing seat provides lateral support for left turns. Note the six- other locations. Soft, flexible foam rubber
point safety harness, and padded head rest with pit radio behind it. is totally useless in a race car, since it will
bottom from the slightest impact. Foam
helmet liners are designed to crush at the
optimum rate at high decelerations, and to
brackets must be able to handle the forces resistance, and spreads the impact over a have no rebound, since any rebound would
from a 40-g impact in any direction. It is large area. The 300-g impacts are absorbed produce further undesirable forces.
especially important to have a headrest by about 1 1/4 inches of polystyrene foam. Therefore, these foams can only be used
designed for such impact forces from the Oddly enough, more than a decade of once. In fact, they may even deteriorate with
rear, and perhaps even from the sides, aerospace research has shown that ordinary time and should be replaced or updated
except for the visibility problems. inexpensive closed-cell styrofoam is one of every few years. Ideally, the same sort of
Few people are aware of the science that the best impact-absorbers possible. rigid foam would be used to line every inch
goes into a crash helmet, much less the There may even be extra fire resistance of a race car interior. But these potential
standards which distinguish a bicyclist’s built into some helmets. In addition to all driver impacts are less understood, so other
helmet from a pro-racing helmet. The lining materials being made from a fire- rigid urethane or styrene foams may be ade-
Snell Memorial Foundation traditionally retardant cloth, there may be an extra flap quate. It might even be worthwhile to
sets the most stringent criteria for auto to cover the chin strap. Also, the Lexan face surround the entire seat with foam, although
racing applications. To guarantee a rea- shield is not only essentially bullet-proof, the familiar and convenient foam-in-place
sonable degree of protection, there are five it will stand up to 2 1/2 minutes in a fire polyurethanes give off poisonous gases
tests that must be passed for Snell approval. without melting through. In Formula One when burned.
A simulated head should receive an instan- racing, there is a requirement for an At the very least all rigid metal com-
taneous peak of no more that 300 g’s when emergency life-support air bottle, which is ponents should be as far from the driver as
impacting a steel anvil in a free fall from plumbed into a steel tube in the side of the possible. Under extreme conditions,
14.5 feet. Penetration resistance is tested helmet. This provides over a minutes worth seatbelts stretch many inches, allowing the
by dropping a 6.6-pound pointed striker on of unheated air to a driver trapped in flames. driver to contact ordinarily unreachable
the helmet from almost 10 feet. The chin Finally, some helmet manufacturers will surfaces. Similarly, the steering column
strap must hold an 84-pound weight dropped custom-fit the eye opening to suit a driver’s should be collapsible, or very short, or fitted
4.7 inches and any chin bar must resist an individual preference. For example, at Indy with universal joints, and it is a good idea
11-pound weight dropped 23.6 inches. In many drivers will accept the tunnel vision to have a rigid foam pad in the center of the
addition, all materials must self-extinguish of a mere | 1/2-inch high eye slit as further wheel.
after being exposed to fire for 30 seconds. flame protection.
Satisfying these requirements is a trade- Crash helmets may be closely regulated, COLLAPSE
off against size and weight, especially con- but there is still some choice as to style. The If the driver is not to strike any interior
sidering the normal cornering g-loads that full-coverage type is superior in every safety surface of the car, his compartment has to
race driver’s heads are often subjected to. consideration, although it is slightly hotter be rigid enough that it doesn’t collapse
The best helmets today require only 3.5 and heavier. That little extra area of around him. He should also be protected
pounds of material to easily exceed all fiberglass and padding can save a lot of face from striking any external surface such as
Snell standards. The external shell may be from impact or flame. If it seems like a lot the ground, walls, trees, or other immovable
made up of 4 to 8 layers of custom-weave of weight to carry in constant high-g objects. And after the car has stopped com-
fiberglass cloth impregnated in a vinyl-ester corners, then stronger neck muscles or a pletely, it should be possible to remove the
resin. This merely provides penetration lateral restraint strap will be necessary. The driver without special tools or large forces.

104
This is no great problem in large sedans,
where roll cages can be designed to with-
stand incredible impacts and repeated
pounding from rollovers and flips. All sanc-
tioning groups have elaborate specifications
for materials that can be used, size of tubing
with respect to car weight, fabrication
methods, and layout. In addition, restraining
nets are required or recommended for the
driver’s window, to prevent his head and
arms from flailing out against something.
It has to be an extremely violent or freak
impact to critically injure a driver in such
a cage.
Open sports cars have traditionally been
required to have only a rollover bar or hoop
behind the driver. This is relatively inexpen-
sive and simple to construct, but it ignores
the possibility of frontal impacts. Recently,
wiser drivers have been equipping open
sports cars with full roll cages forward over
the windscreen. Any disadvantage in weight
can be more than made up for in reduced
chassis structure required, if the cage is also
used for structural stiffness. A good example of safety fasteners, including the Dzuz fastener (top
For some obscure or commercial reasons,
right), ball-lock pin (bottom right), lock nut, and safety wire.
race sanctioning groups don’t seem to care
as much for the lives of drivers in
lightweight race cars. They still have good
standards for rollover hoops to protect the alikely possibility. But aerospace technol- crushproof, leakproof, and explosion proof,
driver’s head, but they ignore the mass ogy and lots of experience have produced they can still be cut or punctured by a sharp
damage possible from any other sort of some very effective avoidance and protec- object, so they need to be enclosed by a
impact. A frontal or diagonal collision, or tion systems. metal container. Most regulations require a
an end-over flip at any speed, almost The fuel cell has probably saved more minimum of 20-gauge steel or .060-inch
guarantees severe injury. There have been lives than any other safety device in a race aluminum for the container, with some sort
instances of drivers taking it upon car except seat belts. Only its cost prevents of double-wall honeycomb or sandwich
themselves to install cages or at least long it from being a valuable item in passenger metal being even more efficient. As long as
diagonal braces forward of the cockpit — cars. The flexible bladders themselves are that material is required, it ought to be
as is common in dirt-track sprint cars. But so well specified and standardized by sanc- possible to use it as stiffening structure,
they haven’t become popular yet, perhaps tioning groups that all a racer has to do is although it isn’t wise to use it as an external
because of the stigma of overcaution or the select one to fit his capacity, filling, and fuel skin which is highly vulnerable to impact.
imagined disadvantages in weight or a requirements. The bladders are made from In Indy-car racing, there must also be an
higher center of gravity. Without such 2 to 3 layers of rubber-impregnated nylon inner liner of 3/8-inch thick non-woven
equipment it is possible for the entire front fabric and vulcanized in an oven. Some nylon ballistics felt to provide even more
chassis of a race car to be ripped off in a Formula One bladders are even constructed puncture resistance. As a regular test,
collision, exposing most of the lower half with rubberized Kevlar fabric, which is the Goodyear drops their complete system —
of a driver’s body. The only alternative to same material used in bulletproof vests. filled — from a 65-foot tower onto concrete.
better protective structures being required If a standard volume or shape won’t work, In spite of all this technology, a fuel cell
is for someone to win a few races with a they can easily be custom-fitted to any may still be punctured or ruptured in a 200
roll cage in his formula car — or to walk application — at a much higher price. Most mph impact. However, Goodyear proudly
away from a violent accident in one. bladders are filled with blocks of a porous claims that since their fuel cell was
Sometimes the best that can be said for the foam, which reduces the interior volume by mandated in Indy cars in 1974, there have
safety of sophisticated monocoque formula about 5 percent and should be considered been no fatal fires nor even a major post-
car chassis is that they may distribute the when capacity is marginal. The foam has crash fuel leak. Occasionally a small flare-
load well as they collapse about a driver’s practically no effect on fuel movement (as up may be seen, but this is usually due to
body, and they are relatively easy to tear demonstrated by the fact that a 22-gallon ruptured reserve or surge tanks or broken
apart to remove the remains. bladder can be filled in 4 seconds), so it will fuel pumps and hoses, and not the 40-gallon
have little effect on slosh or leakage if the main system.
FIRE tank is torn open. The primary purpose of Even with the best fuel bladder and con-
When an accident occurs, there is still the foam is to prevent explosion in the tainer, it is still possible to lose the fuel
avoidance and protection to consider in the bladder by slowing down the flame front, through other broken lines or openings. For
case of fire. With the severity of impacts, but it also helps hold the bladder in shape all race cars that don’t ordinarily make a
the amount of fuel carried, and the lightness even when it is empty. pitstop for fuel, a flush positive-locking or
of race car structures, a post-crash fire is Although all bladders are practically screw-type filler cap is best. Where long

105
 eid Ine scmensh esol sam 2 Game
azainst 2) ip 3) sali of See Tel EE
fe onhoeet ite saawzsti: ao fe
Deer S DSS Soe msn Beg —
Ss we lee. Lee sess ee SS Ga

Gower hee the etwas of 2 sie oval

— @ ion & hos aw s ee
But tie Goer Gf an en Ce Ce
ins Sawaal
Ge am scons.
Atinonst chemical ine someoshes
Tsgunsd an all mace cers. they we asus
Tt bey) SUE WD DE OE a See
Taam Gee. Ther @ fe aoe
TET SISETS BBliahie. oweser, wath
:
a
ssnsons. These aoe somewhat betes —
The typical Gry-break facie system for seiam and sports cars mcinties hamiiy Ioninrent. 2 2ach es 1S own Ga
the sprimsinaded receptacle on the car (jop). and maichme nozzies on
Whe 3 sage te Gime.Tres
the 11-celion fie cams with vented handles (belpw).
can De at Ge tank Inca & eee aa
a >
V 1 DE Of fe fe bees £ es, oe
CONG: § COT oe eo a Ges
The! & SDSSRE0 In. Of 2 coma oF oh

5 fe WED oO Wet De i 2
© ese. Wii 2 Ss Gee ace et

3 alwaessthe possiniiny that © wall => a=

= the wom Ge or thet Ge Semel wort
te Stn cath DD = Ea a all Efe
ISS & essed whi fe cr & ell
TANS 2 SOE Stet Speed. fe chemcas
wall Se Gali set ap or Down aver Iw Te
tome dhe car coies ip 2 step. Te Ss
® te GE Te ave, Os Te 2
conibmaion of all of dhe abowe_, woth some

cam dhe cock. in any over, tte mam

2 sguited Gor fast ine] sips. mechanical Tael pumps. however, pothans SnOLED ip fie diover sp f worl dsappear
2 Ons-way Check waive or Tapper wae with the season that whee fie coon Ti pans of fe chess a io avey oo ee
Taandatory, 3p case fhe neck Sos opps ot Diowe oF Dies. fe DUM Sons amomacalls, accieS
Most mace salies Sven Segue Deak i may be posit @ as fale ine I ae Ge mae ca Ge see
cmnecon which prevent spilias= fom fines Winch “breast eet Gey ait Chemis were commonly CD. and “ars
either the filler neck of dhe ineime cans commonly avaliable for mace cars. em" which & sy 2 oe we
Fuel tank wens can de 2 Digger DroDIeT take sede. These woe davis
They have io aliow ar > ote 2s es) & SHS Sse. 2 Tease COC Deeks ay
consumed, ie ar ont m2 oory 2s the tank heb pve 2 Sand Tame. Ths swan feemge Se of fe CO> on Goan am
i deme Glied. ae dump apy overiow fe) cus all decmicay irom the bates ano 3 [ comes Oo a abot coms 99) dees
2 Ganesrons locauon. and awe 2 check Tas and located io cosy ace Faire). and dhe dry Chem) ses
save 1D Drover ceamase © fhe ca BE OP comms wotkers. i ss oven hanty ip hae 2 WHC Tai 2 es of fe co. pie. and
sie Gown. The best solumon& 2 pombe. secon switch om fe ime wit seach of fe sapccidliy
the Insc of an one Ones fhe
Tanspais Overiow caich cap maaich ee
anoine: dyDredk commecnon on fhe went Gaion as come wp be asd det
The fuel pomops and Ges ae gest 2s pacecatt spats —asi a EE the tiemsty. Wot any dass ¢ hae moe ot
portalip sates. The ies should be sone batiery Should be enclosed mm an ali piastic These irawbacks.
Du on 2 perpound Das
1D the Tos Deced iecaons and perhaps 2 2a 3 Goes ewe Scie than fie
even be Shicided Dy heavy met nibme. in snouid be well pretecied or shee Sier chemidis. Tie ony smal ds a

fhe most widies cases, 2 ie wall When very avontance measur fails and advan: & fe cos of stil, 2 SI
awindbiy be pulled apati. so i =a pood aiee the cat Giecs pp wp mm Tames, © oo lee
to have the pumps 26 Close ip the tanks 2 wt ie mor pwEcoD equpneT iene Drowns that ther Gaion SE = a
possidic, 19 Bleck die Glow when the Slee Sometimes 2 tace track fier omuck cant o= suey meric, epeculy sme as
tmcity gocs of NASCAR segnmes at accent ie wer 2 coe. and 2 et Dona applica & mm ama

16
 Femeet cof = 2 tenes
ava aves Tatt canis 2st ae

WH cat Je oe BM oo SS ees

Gita TSE 2 ee Ise Gere

Iams Se eesitie
a fer et well oe

Gest af = feunt = fit © pme= =

diez fir 2 (eas GP aoa “fe ir

he Tames or Se is ete we Sel et

Oy Gee steed? ar ar Tames IF
SE Way Cat te Tei ap Sete es Feat
= et
meal T= sitio = & oe

Ir went 3e Gatea Gs =
af Giese pein weet Ge aon & decracs Se of « Spl Se comin. want (ef Ge Giadder
ont moms Catineous
ft Set ee ee
oat wT alow sone acfow ir oir sally won WW draws drtvers win ay
71: amas fe coo 3] SSS eee ee: Se
Moss sas te similar Et hes te
sme qu lve of lets Tam foal Soe DSoeer ant aa 2 oo
[essen tee sect = Nome reste oe EEE Sie mm ie jms Aiee a fears © Semi Ge owe ome or =
URS Wi? Tas adie 2 we oe md f& cao = @ ff aie
fan 2 Ge socom oes cme fe SS aeior alow eaters DT ee
if sme wes Semmaly) Tr ep Se he Tie est Greomer suf Ges Shei siowes,
SD Gesher on 2 ie Speerseet Keo @ei Ses SE Ge Gemes mesma =
ail Siang is Gome watt 2 sonia her i] samcioming snuies awe iiss of her
yes. medians. at gmesdores
DSU 5 SRS eer SH. SS IT SE iat Servers oust Gilow. Ant een hex
imine: hea Sr ass ose ear
Aaittest lave=s oe aoe asst Pe oo Tas. Goves aut Ses Be BST mae GSS Gest Mey as Sugetines wr fetter =
Giese he Tere rewst Gime & 2 Sone Tt cite Gress ofWome. wat lewis
aiiet sae. So some “sot of aS ae caer ies emit om. St ome 1S ame ast pe wertit Emsam of fe
cme. lewiter Home wes acct Ser ins Mies Gf Gmsime mess Aderhe moss
SSanc= Far ww would > want ar cece WSs Univ Siemiv Deter
far wine 2 SS SS SCs. ay sail be
Bessie oP comme tf ie me Re car

fattne suct & Sow seat of : See Qiaek Gre mee Gert Snes Gmesrt
4a Ge al Geet of fe sie aver Ge ce eet Ges 2 Sre roet nent DOSES
Gf ie ca Sac as ite cai et
Gmamcemscc of ewe of er gmes Rewer Swuid Je wort ume of. ib aio. =
@e Gees. Rese anf ser Ges ges af 2 Lower fee sheti — oe ne Rex ert amass oy smal] mmacs ov
Gemctes lee fea Geet bo Se Goset oe — wail genes he dmvers See Siding over suet sumaess. Tires will
Semmes ems Ge er eet whet eos Geer dame ie 2 Ser me setaniv ay Jan. <r © Gowid 2 esse
When we 2 tal peter fe oie 2] © he nis on wiasser is et Fite or
Whee if comes © nuitaver
ants, mere SS Gt Greet comer wit flame Ser ie ier damawest sv Gasily thar the ofc ae
Sar Dees NE made Pe ee [SSeS or “ewan” Wares x5 awe ifQue, esc is always 2 Clasc= © Sve
Briers “tummzet Qlyer sais we Sewn Gat 2 Sa@ieige: sat wil Samstinns Geer he a=
12
THE DRIVER
Race car drivers have to perform such performance considerably. With the kind fitted to each driver. It isn’t necessary to
complex and strenuous tasks that it is im- of accelerations possible today, lateral sup- have the steering wheel at arm’s length,
portant to make their job as comfortable as port is even more important than vertical because the ideally set-up car should never
possible. Every means of simplifying the support. To prevent the driver from hav- need more than a quarter-turn of travel at
driver’s job should be explored: fitting, con- ing to hold himself upright by the steering the wheel. But it is a good idea to have it
venience, comfort, sensory feedbacks, and wheel, well-fitted lateral support is re- far enough away so the driver doesn’t elbow
mental programming. This chapter may quired at the thighs, hips, rib cage, and himself in the abdomen when an emergency
seem to be somewhat dehumanizing in especially the shoulders. If these contact correction is necessary. The brake and
considering the driver as simply a part of points distribute the load over large, fleshy clutch pedals should be far enough away so
the guidance and control system to be areas of the body, there is no need for the driver’s leg is almost straight at full
optimized. But in fact, the ideal driver is padding. In fact, padding adds practically travel since that is where it develops the
one who acts very much like a machine or nothing to impact safety, and is likely to maximum force. It is also a good idea to
a computer. His function is to determine the obstruct cooling air flow to the driver’s have adjustable pedal stops to prevent a
ultimate limit of the car’s performance at backside. A series of holes drilled in a strong and overenthusiastic driver from
every point on the track, and to guide the rigid seat shell will improve circulation, bending the linkage. The brake and
car to that point repeatedly and precisely. allow perspiration to escape, and reduce accelerator pedals must be positioned
A sophisticated electronic guidance system weight. Mounting an unpadded seat shell perfectly to the driver’s satisfaction so that
might be built to do the job even better, rigidly to the chassis also improves the he can heel-and-toe (or toe-and-heel) to his
except for the necessity of constant sensing driver’s ability to sense every minute force preference during braking and downshifting.
and adaptation to other cars and drivers. and motion inthe car. A seat I designed for The optimum position is probably to toe the
Unfortunately, race driving is no place for myself and Mark Donohue (shown below) brake pedal and heel the accelerator, since
any human characteristics such as emotions, is a good example, although it only fits the brakes require more sensitive control at
errors, or imagination. For the fastest and drivers in the 140- to 180-pound range. that point. At any rate, the steering wheel,
most consistent laps, the driver must be able Any seat should be fully adjustable in pedals, and shift lever should all be fully
to concentrate totally on keeping the car at travel, height, and tilt, until the driver is and independently adjustable, until the
its critical limit. satisfied that he has exactly what he wants. driver is satisfied and has a written record
Possible considerations are: arm exten- of all the necessary locations carefully
FITTING AND COMFORT sion, visibility, leg travel, and leg interfer- measured.
The driver has such intimate contact with ence with the steering wheel. Enough drivers have blamed a stuck
the seat that a proper fit can improve his All the controls should be individually throttle for loss of control, for the entire
accelerator linkage to be carefully designed
and installed to eliminate it as a possible
cause. Since the engine is apt to move
around due to accelerations and torque reac-
tions, any rigid linkage of rods and ball ends
will probably be out of alignment at some
time. The best solution is a tension cable
in a flexible steel housing. Motorcycle
control cables are adequate, but heavier
industrial control cables and housings are
available which can also take some com-
pression forces. The reason for compression
action is in case the throttles stick or the
return springs fail. A toe-hook on the pedal
can be used to push the throttles closed. Not
less than two of the highest quality throttle
springs should be used. And, needless to
say, all throttle butterflys and shafts should
be absolutely free from misalignment or
binding. The accelerator linkage, as well as
the brake and clutch linkage, should be
designed with its inherently variable force
This form-fitting racing seat shell provides full lateral support for the and movement ratio in mind. Ideally, the
legs, chest, and shoulders, plus an integral head rest. ratios should vary so that there is quick

108
action at the start of travel to take up slack
in a hurry, and slow action at the end to
allow precise modulation.
All that is required of gauges is that they
be accurate, easy to read, and reliable. The
necessary gauges, in approximate order of
importance, are: tachometer, oil pressure
gauge, water temperature gauge, oil
temperature gauge, and fuel pressure gauge.
If any other component in the car is
marginal in performance or durability, a
gauge may help to keep an eye on it,
whether it is the charging system, differen-
tial temperature, brake vacuum level, or
whatever. But the driver is usually so busy
that more than four gauges are almost
beyond comprehension. At that, it can be Se.

a good idea to back up the really critical All instruments in this sedan are arranged for the needles to be vertical
gauges with big, bright warning lights. Not at peak operating conditions. Note the pit signal reminder list.
only will they alert the driver when he can’t
easily check the gauges — say in a series
of turns — but they provide a quick double- The value of vehicle feedbacks to the possible within the limits of performance,
check if an instrument seems to have failed. driver was mentioned in Chapter 7. The and the driver should be protected from
Failure in an oil or fuel pressure gauge with better a driver can perceive these signals, excessive heat, fumes, buffeting, and vision -
a remote line can be disastrous, so it is a the more precisely he will be able to con- interruptions.
good idea to provide shutoff valves in their trol the vehicle. The feedbacks can be Sealing the firewalls ahead of or behind
lines to prevent loss of fluids during a race. categorized by the driver’s senses: vision, the driver is not only an obvious safety
Every gauge should be assumed to be for the perception of the location of the car precaution but keeps heat and fumes from
inaccurate until it is tested and calibrated. (and other cars) on the track; feel, for the leaking into the cockpit. Gasoline and ex-
A new tachometer should be taken to a perception of forces, accelerations, jerks, haust fumes can have serious effects on a
reliable test shop, and the person doing the and vibrations — both linear and rotational; driver before he is even aware of a problem.
testing should recognize the importance of sound, for the perception of road speed and For the same reason, most regulations re-
even a 100-rpm error at 8000 rpm. The zero component failures; and perhaps even smell, quire tailpipes to be rearward of any cockpit
or idle calibration can be off, as long as the for the perception of overheated com- openings. Even when weight is a problem,
needle is accurate at the top end. Temper- ponents. So far, it hasn’t been possible to a little bit of fiberglass or asbestos insula-
ature and pressure gauges and switches can measure the relative value of these sensory tion and sealant may be worth its weight in
be tested at home using boiling water and inputs in optimizing control. It would make driver endurance. In front-engine cars, a
welding gas regulators for reference. In an interesting research project to try and tremendous amount of heat is radiated from
addition to a tachometer, it might be good study them by establishing a simulated the firewall, transmission tunnel, and the
insurance to have an electronic rpm limiter. driving task, while totally eliminating each floor above the exhaust pipes. A driver will
The best choice is one that doesn’t suddenly input in sequence and measuring the dropoff probably know quite well when more cool
kill the engine but gradually breaks up the in performance. It is interesting that even air ducting is required, and he deserves all
sparks instead, and one which is nearly Formula car drivers are often moved into he wants. The effect of cooling ducts on
failure-proof. Still, the driver should be a more vertical sitting position — which external air drag is hardly worth con-
provided with a bypass switch in case of can’t help but improve visibility and the sidering. But it won’t help much if the
failure, or in case he absolutely has to have perception of horizontal forces and acceler- radiator is ahead of the driver and is pouring
an extra few hundred rpm to draft past a ations. Anything that can be done to im- hot air either into the cockpit or into the
relatively equal car. prove the driver’s intimate contact with the driver’s cool air ducts.
To minimize a driver’s visual transition vehicle and its motions will help — at least For longer races the driver may require
from the road to the instruments, they until the contact reaches the point of a drinking water supply. Water might be
should be as close to the driver’s line of sight discomfort. provided during pitstops, except for the
and as far away as reasonable. They may Aside from all the factors involved in the obstruction of full-coverage helmets and
be arranged in any grouping the driver is driver’s control system, there is still the very interference with other more necessary pit
most familiar with, but rapid compre- important and often neglected consideration chores. The water container will have to be
hension is the primary consideration. Each of comfort. There are many areas where insulated and mounted near the driver.
gauge should be well identified with large small changes can noticeably improve the Location of the container and drinking tube
labels, and the danger zone should be driver’s maximum effort and endurance — can be a problem, to avoid natural siphon-
marked in red. It helps to rotate each dial although they may be so subtle that the ing action under hard horizontal accelera-
face until the optimum needle position is driver is not consciously aware of them. In tions. If the driver should stop sucking and
perfectly vertical, so an undesirable condi- the midst of a race the driver probably won’t take a breath at such time, a drastically tilted
tion is immediately apparent. In addition, notice that vibrations are affecting his vision water level could maintain the flow enough
all switches must be well labeled and within or that engine noise has deafened him to the to choke him. The container should be more
arm’s length — and yet not close enough to sound of grinding brake pads. Vibrations than a 45-degree angle below the driver’s
be hit accidentally. and noise should be reduced as much as mouth.

109
AVRIL ROBARTS LR@
 The windscreen on an open car is usu- strip and skid pad. The long straightaway minutes than by hours of practice driving.
ally a compromise between visibility and gives the driver a chance to practice up- and If a faster, more experienced driver is
wind forces or buffeting. The optical downshifting and braking at the limit of available, he can set a performance baseline
problem is that any transparent material lockup, and gives him a feel for high-speed to be matched. Or two fairly good drivers
becomes very difficult to see through at a straight line stability. The skidpad will give can compare their recordings and learn from
large angle, and pits, dirt or oil will make him a chance to get used to the maximum each other, since each will be better than
vision practically impossible. The proper lateral capabilities without the risk of sliding the other in some distinct areas. But even
angle, however, with perhaps an added off a dangerous turn. And it is not a a single driver can learn by studying the
curved lip, can deflect the airflow high coward’s way out. Some of the best race output charts and seeing where his perfor-
enough so the driver can be positioned to drivers in the world insist on driving a new mance is not at the limit. In any case, it is
see completely over the windscreen. But this car on the skidpad before taking it to a race more convenient to use a distance-base in
still allows the driver’s goggles or face shield track. a recorder instead of a time-base so that all
to collect some trash, so spare lenses or Another practical suggestion is to get as laps come out exactly the same length on
stripable covers should be provided. Flat many driving miles on a particular race car the chart paper and can be directly overlaid
black paint on all visible surfaces will cut setup as possible. Not all drivers can step for comparisons. For example, two drivers
down on glare that might temporarily blind from a passenger car into a race car and may have about the same laptimes in the
the driver. make the instantaneous transition required same car but learn from comparing record-
The final, if not most important con- in responses and control movements. The ings that one is faster in the corners, while
sideration, is for the driver’s mental more familiarity a driver has with his car, the other brakes later and harder. Both may
comfort. If the driver can be sure that the the better he can control it and the less be able to learn to drive faster by knowing
owner, engineer, and mechanics all know chance there is for him to make a mistake. exactly where someone else has already
their business and that they care a lot for If the race car is a converted production done it. The recordings will also graphically
the driver’s health — and for finishing races sports car or sedan, it might even be illustrate just how smooth a driver is, and
— then he can be confident that they have possible to add a few street items such as where and how he can improve in this
done everything possible to keep the car mufflers and lights, and get a lot of ordinary Tespect.
together. A driver who has faith in his car highway miles on it. Such training instrumentation isn’t just
and the team can afford to concentrate on Of course, it is insane to try to practice for beginners. For the experienced profes-
the business at hand and will probably feel race driving on public streets and highways sional driver, it may be the only way to find
more free to take other risks that only he in any kind of car. The conditions are so areas for improvement when there are no
is responsible for. far removed from actual racing that it is a observers or instructors as skilled as he is.
total waste of time, energy, the car, and the It isn’t easy to sense the point at which the
LEARNING tires. The only way to practice realistic race car is at its absolute limit without taking
There is still the matter of programming driving is at the absolute limit of traction the chance of going over that limit. It is
the computer/driver/machine to take full and at both edges of the road. If it were much more precise--and safer--to find the
advantage of the race car’s potential. This possible to do that on public streets, it would limits by electronic instrumentation. In
doesn’t sound quite so cold and calculating be homicidal. The only time a real racer addition, the speed and acceleration curves
if it is called “‘learning” and the driver is drives fast off the track is when he’s in a are invaluable records for setting the car
an actively willing participant. Of course, hurry — not because he’s practicing. up for that track the next time.
everyone knows how to drive and the racing However, there is one kind of practice that For example, an article I wrote in Sports
driver knows his business better than anyone is safe in public — smoothness. When a car Car Graphic, way back in 1971, showed a
standing in the pits. But he doesn't is balanced precariously at the limit of trac- comparison between Mark Donohue and
necessarily know the peculiar characteristics tion, any sudden movement of the controls Milt Minter in a Trans-Am car at River-
of that particular car or setup, and he can’t can send it off the track or lose some tenths side. Since that story would be hard to
out-sense electronic instrumentation. of seconds. Every driver, race or otherwise, find, however, here is a condensation:
For the driver who is just starting out in would find it productive to practice smooth First let’s look at speed and lateral g
road racing — and for some drivers with driving at all times. recordings of Mark Donohue’s best single
a lot of unsuccessful experience — the first A real race track can’t be beat for realistic lap, with corresponding scales on the left
step is to learn the friction circle concept. racing practice. Unfortunately, the rental and right sides. (See Fig. 43) The curve
This was explained more logically and cost for exclusive practice time is usually begins as he crosses the start/finish line at
completely in Chapter 2, but in a sentence, out of the question, and race weekends are about 120 mph and immediately starts into
it means that the fastest way around a track so crowded that very little practice time is the left hand Turn 1 (refer to the course
is to keep the tires at their maximum fric- available to any individual driver. Amateur map). This is a flat out sweeper, and
tion capability in any direction — laterally Taces may provide inexpensive practice or therefore he only reaches peak g’s of about
or longitudinally, and especially during the development time for professional drivers, 0.80 (A). Then by the time cornering g’s
transitions. This means that braking, but by the time a driver has reached that have dropped off he has reached a top speed
cornering, and accelerating will usually stage, he can usually adapt to a track of 140 mph (B), and starts braking for Turn
overlap each other. As easy as it is to say, relatively rapidly. 2 as the g’s rise in the opposite direction,
it takes a lot of practice and sometimes a The best investment in cost per hour of to an average peak of about 1.2 g. After he
lot of relearning. learning time is probably an electronic slows to 110 mph in Turn 2, he accelerates
If a driver is stepping into a race car for performance recording system. If the race flat out all the way through the esses, being
the first time, or into a radically different car is wired to record both speed and lateral slowed in places by the cornering drag at
kind of race car, the experience probably acceleration constantly (or the net accelera- the tires. As can be seen from the g-curve
ought to be taken in small steps. One of the tion in any horizontal direction) as described “spikes,” the car is being whipped violent-
best places to start is at a combination drag in Chapter 13, the driver can learn more in ly from side to side, going from 1.2 g nght

110
 epproximate: ly 5 seconds —
3i T T T | 1.25
70

4 weOy

TURN
RIGHT

° G

TURN
LEFT

75

MARK’S VELOCITY.
MARK’S g’s eeee

$1.25

Figure 43. A speed and g recording of Mark Donohue compared to a speed recording of Milt Minter (dotted line) in the same
Trans-Am car at Riverside. This project was done for Sports Car Graphic Magazine by the author in 1970.

to 1.2 g left and back in less than 3 seconds

at speeds up to 130 mph! Then comes the
hardest braking point (C), where the car is
slowed from 130 to 50 for Turn 6, even
while the car is negotiating Turn 5 at a
nearly 1.25 g spike. There is a dip in Turn
Tl SF
6 where the car becomes light, which can
be seen in the g curve as it drops from 1.25 RIVERSIDE INTERNATIONAL RACEWAY
g to 0.80 g and back (D), and then g’s fall
off gradually as Mark accelerates toward
Turn 7. This time he only has to brake from
15 to 50, though looking at the course, it’s compared, even with the same laptime, esses, he is never able to make up the lost
hard to imagine that as the minimum speed. there are variations of over 2 mph just due momentum. The condition is compounded
Even though there is no straight between 7 to driver inconsistencies, even for the very by letting up in Turn 3 (T). But in braking
and 7a, the transition from left to right turn best drivers. So it takes a lot of practice to for Turn 6, Milt’s speed curve rises above
allows a short burst of acceleration to 68 interpret the results where minor vehicle Mark’s (U) where he gets off the brakes and
mph (E). Then g’s again drop off out of 7a changes are made, but it’s still far better than accelerates for a split second through Turn
as he accelerates down the back straight. merely taking the average of all the improve- 5. Nobody knows ail the answers. Given
There is another flat out left sweeper which ments, “dis-improvements,’ and incon- enough time, Mark might have gone back
is only good for a 0.65 g peak, before sistencies over a full lap — which is what and picked up a few tenths here himself. But
braking for Turn 9 at again a top speed of laptime comparisons do. back to Milt, who then loses ground in the
140 mph. Mark’s minimum speed in 9 is Now let’s look at a comparison lap by last half of Turn 6. By slowing more in the
95 mph, but due to the banking, and by Milt Minter in the same car. In less than dropaway portion of the turn (V), and ap-
changing the radius of his path, he reaches 20 laps, he had familiarized himself well parently taking too early an apex, he loses
and holds a maximum lateral acceleration enough to gain 0.6 seconds over his best a little, though he manages to reach the
slightly above the 1.25 g scale limit (F). time in his own car, though still 1.2 seconds same top speed between 6 and 7. In 7, he
Then he accelerates out of 9 and returns to slower than Mark. With two very good shows he knows the way as well as Mark,
the start/finish line. Quite a round trip for drivers in the same car and lap times so but in 7a he drops down about 3 mph slower
only 88.9 seconds. close, you would expect the traces to be (W). All down the straight, he never quite
However impressive that may be, its almost indistinguishable. But look at the dif- makes up his lost turn-exit speed, and that
greatest value is in comparison to another ference in Turn 2 and through the esses (S). combined with ever-so-slightly early
chassis or engine configuration, or to Milt brakes early, losing 4 mph, and to an braking means a lost 2 mph at the peak (X).
another driver. At that, even when two 8 mph slower apex speed. Then, though he Through Turn 9, Milt shows an excellent
consecutive laps by the same driver are accelerates at the same rate through the. copy of Mark’s trace, as most of his per-

111
 percent of all college students. And they are
likewise defined as tough-minded, self-
reliant, and no-nonsense when it comes to
the job of winning. The motivation for this
drive doesn’t seem to be important. It can
be either to attain recognition in front of
large audiences or simply a personal obses-
sion to beat the other guy.
Similarily, the most successful driver is
usually highly assertive and has a high need
for dominance. His assertiveness shows up
as independence, aggressiveness, and stub-
bornness. He is a self-motivated, self-
directed person. He will be quite willing to
argue for his own point of view, to be a
leader of whatever groups he must be
associated with, and to be able to persuade
and influence others. However, since the
need for this trait is often strongest off the
race track, some relatively quiet and humble
drivers have done very well with the
assistance of team managers who always
handled those particular conflicts.
The author comparing the second-by-second track performance recordings of two To balance these strong drives and traits,
attentive drivers to those of Mark Donohue (right).
racing drivers must also have good control
over their emotions and behavior. They are
perhaps compulsively self-disciplined. But
formance was, but still a few mph slower. While there aren’t any distinguishing at the same time, they are not neurotic nor
Such comparisons are becoming com- physical characteristics, there do seem to be unusually sensitive to emotional problems
monplace today, considering the avail- a number of personality characteristics that in themselves or others. In most situations,
ability of electronic instrumentation and frequently set the great driver apart from especially during a race, they are more
two-car teams. Some drivers actively study most other people — even those who are casual, relaxed, and composed than the
the perfomance data on their own, and successful in other competitive activities. A average population. They are relatively
some even consider it valuable to hire an psychologist, Dr. Keith Johnsgard, has made decisive and imperturbable, and don’t waste
experienced coach to study the data and studies of over five hundred race drivers, a lot of effort in self-criticism. They are
make recommendations. from beginners to World Champions, and emotionally mature in their personal respon-
But for actual beginning race-driving found definite distinctions between them and sibility. Another balancing factor is that
training and experience, it's hard to beat the average population. The discussion of most of them are able to listen to criticism
the professional race driver schools, where these characteristics might help a car owner from others objectively, and they welcome
they teach such things as the proper posi- or team manager to select the optimum constructive suggestions from qualified
tion and control motions, paths to take on driver component for his car or help him persons.
the track, and how to deal with other cars. to program that component for better per- There are a number of other character-
Second best, but far less expensive, are the formance. Although that may sound coldly istics which fall outside the range for the
SCCA driver’s schools, where they teach exploitive, the fact is that the driver is average population. Racers generally are
about the same thing with perhaps less usually getting paid for the ultimate more intelligent and more capable of
experienced instructors. And finally, it performance, and he ordinarily wants to abstract thinking. They have a good percep-
helps to read and think about the right way improve himself. tion of reality. The brighter a driver is, the
to do it. There are many good books on Probably the most significant trait in more successful he tends to be. He also has
driving technique, but Carroll Smith's is successful drivers is their overwhelming a high capacity for endurance in the face of
the most technical. motivation to win at any cost — the killer frustration and physical hardships, and he
instinct. Many times this isn’t apparent in is willing to adventure, to take chances in
EMOTIONAL PROGRAMMING a driver’s outward appearance, but it can pursuit of his goals.
Great racing drivers are made, not born, show up in different forms. Some drivers But what do all these studies and analyses
and no one can make a driver great but demonstrate it on the track by driving like mean to the race driver himself? If a driver
himself. The difference between a very maniacs when they get behind — although is relatively unaware of his own attitudes and
good driver and one who becomes a suc- those types don’t usually last very long. The feelings, it is possible for him to take some
cessful professional is all in their heads — opposite extreme is the driver who spends psychological tests and discover his strong
not their physical makeups. There seems to every waking hour building and developing and weak areas. He might learn how he
be no correlation between any physical his car, and honing his own skills to perfec- compares with other more successful drivers
characteristics such as reflexes, eyesight, or tion. But in either case, there is an unusually and try to learn to compensate or develop
build and a driver’s accomplishments. Of high need to be best, to be successful, to his personality to overcome the slightest
course, most drivers are physically strong accomplish tasks requiring skill and effort, emotional handicap. It is important to keep
and quite healthy, but that is probably an to be a recognized authority. Their average in mind, however, that these statistical
effect, not a cause. need for achievement is higher than 90 studies are all averages of a large number

112
 wil
iit

‘cnnnianquioni

|
og
UT
¥

of drivers. Some World Champions are low petitive sports in that an extremely complex an immediate automatic response to ease up
in some desirable characteristics, and some machine is involved. A ballplayer or a track on the brake pedal — as though that tire
losers are high in all of them. and field man is born with most of the hard- were simply another part of the driver’s
There is another personality characteristic ware he needs to compete, but a driver must body.
which is not recognized by standard learn to use complex extensions to his body. It is theoretically possible to build an
psychological tests, and that is an empathy And the characteristics of those extensions ultimately perfect electronic driver. Enough
for machines. There are two diverse ap- can change totally from car to car, and even is known about the psychology of driving,
proaches to the matching of the man and the from moment to moment in a race, re- race strategy, cybernetics, computers, servo
machine. At one extreme is the highly quiring tremendous adaptation skills. It systems, and inertial guidance to do the job
logical and systematic driver who learns usually takes years of practice for the racing — given enough time and money. It is even
everything there is to know about the driver to learn accurate perceptions of all possible right now to mathematically select
machine and who adapts that machine to fit his feedbacks and the proper control the perfect line for a car to take in each
his own precise personal preferences. At the responses required. It must be as though the corner, although it takes a precise map of
opposite extreme is the driver with strong steered wheels were the driver’s hands and the course and a lot of computer time. Such
instincts for the right responses, who can the driven wheels were his feet. He doesn’t an electronic driving machine would be able
take any vehicle and adapt to its particular really feel the steering wheel and pedals as to drive any race car at the absolute limit
characteristics. The former method seems much as he feels the track at the tire contact at all times, always select the optimum path,
to be more reliable and consistent, while the patches. There must be a direct, instan- never fatigue, and never make a mistake.
latter usually takes many thousands of miles taneous connection between the ground and But then, which is more interesting, com-
of experience. But in either case, there is his mind — without conscious comprehen- petition between machines or competition
a strong feel, or empathy, for the car. sion of all the hardware in between. When between men in equal machines? It is an im-
Race driving is unlike most other com- a tire locks up in braking, there has to be portant consideration in racing even today.

113
13
TESTING

Automobile testing is an incredibly broad The second point is that all tests should — the results. Therefore, there are certain
and complex subject worth a book in itself, be baselined. There is no way to know characteristics that are essential to a test or
but the more limited discussion of race car whether a change is positive or negative development driver — above and beyond the
testing can be covered in a chapter. All too unless there is some well-known, fixed basis skills of a mere race driver. Of course,
often race car mechanics, drivers, and even of reference. If a change in suspension whether racing or testing, consistency is of
engineers assume that all you have to do is geometry seems to make the car faster, it primary importance. One superfast lap out
put the right pieces in the right places and must be possible to go back to the original of ten scattered laptimes is meaningless in
you can go out and win races. But in fact, setting, to make sure the change wasn’t either case.
testing and development are far more im- simply aue to driver improvement. Since the driver is the most variable input,
portant than basic design. It is almost a Sometimes it is even more important to be nothing can be determined until he can con-
truism that a well tested and developed old able to go back to the original condition tinually repeat good laptimes within a few
design will be more successful than the most when a change had negative effects. tenths of a second. This requires the driver
advanced new design. Most of the other Third, never make more than one change to be familiar with the car and the track and
chapters in this book mention specific tests at a time. If two things are changed at once to have a lot of practice before vehicle testing
in their areas, but this chapter will go into and the car becomes slower, there is no way begins. It also helps if the driver has had
more detail about the necessary test hard- of knowing which change had what effect. some experience with a wide variety of
ware and equipment and the proper test It is quite possible that one of them actually handling characteristics, so he will know
procedures. added to the performance. On the other what feels good or bad — and how to cope
hand, the car may be made faster through with the bad. Still, consistency is far more —
PHILOSOPHY mass changes but no one will ever know valuable than ultimately low laptimes or
Before getting into the specifics, there are exactly how or why. years of experience.
a number of basic points to be made con- It may be a good idea to make changes There are other personal traits that mean
cerning testing of any kind. First, it is im- large enough so the results will be obvious. a lot to the mechanics and engineers and
portant to realize that there will always be This makes it possible to bracket the speed up development immensely. The
a large number of uncontrollable variables optimum result, avoiding the need for driver must know the language — as ex-
that can have great or small influences on eternal and indeterminate small im- pressed in this book, for example — so he
the area to be studied. In some cases it may provements. The exception is any place can communicate what is happening to his
be possible to compensate for them, but where a great change may make the car crew. He also has to be physically sensitive
even if not, they should be known and con- dangerously uncontrollable or liable to to all the various feedbacks he gets, such
sidered before any results are accepted as critical failure. Track testing can be far more as steering wheel forces and movements,
fact. At the very least, all vehicle and dangerous than race driving, even if there vibrations, noises, smells, and so on, and
environmental conditions should be re- are no other cars on the track. Many com- to be able to distinguish the most subtle
corded so it is possible to analyse incon- ponents are usually being altered and vehicle changes. In addition, he has to have the
sistencies later. A partial check listof factors characteristics change a great deal between patience to wait through interminable delays
is given in Table 8. runs. In addition, few teams can afford the and interruptions and to watch for the most
Plus, of course, notes concerning the corner workers and safety personnel that are minor improvements. Finally, the test or
purpose of the test and any driver com- always present during a race. At the very development driver has to be absolutely
ments. As memorable as all those may seem minimum, an ambulance and para-medic honest with himself and the crew, to avoid
at the time, it won’t be that easy to recall should always be present. the search for problems that were actually
the details, and there is always the possibility The most valuable test instrument is the the result of driver error.
that the person who remembers won’t be driver. He is the one who operates the test,
around forever. and he is the one to recognize — or suffer HARDWARE—THE INSTRUMENTS
Before getting into the description of test
instruments, three measurement terms must
Location, date Suspension Tire size Spoiler heights be understood: precision, accuracy, and
Driver Spring rates Compound Wing angles repeatability. “‘Precision’” is determined by
Mechanic Ride height Rim width Temperatures the smallest increment on the measuring
Engine Shock settings Pressures Barometer device. On a yardstick, precision may be 1/4
Transmission ratios Caster, camber, toe Rolling radius Wind inch; on a micrometer caliper, probably .0O1
inch. “Accuracy” is the relationship to
Weight Anti-roll bars Condition Track condition
absolute standards as established by the
Table 8. Check list factors for engineering development National Bureau of Standards and depends
on the accumulation of errors in successive

114
 < SSS
axvug NOlLVOllddvy
SLNIOd

O22 OYUVWVD
SNITGNVH 3SYNOD
HLON]I
- Olt” 1334

as ae ee "Lavy HSINTA= ROWLVIOT o é

21n3Iq
“ppYW jauUeYD-9
eyep Axjauraja}
uraysdspasnAq joforAay_)
18 24} 696T O1BUIED sNpo.nUI
uN paonpoid
sty} de]-au0~4099.1
SumpJo ay) S,.10yjNe SUIALIp*9]Ays
yy sea sty} Jooford
JVY} ISAT paye.ysuOU
ay} ap UOTJIB.197UT
JO [e197e]
puke [eurpny|Suo‘$90.10J-3
]
Yorpay 0} 94} UONILY,, 39.119
, Jdaou0d
JO aoe. “BuIALIp

115
9}0NBY) [18.19 SuTye.AQ
Se [esaxe]
$3 asta SUISOFUY SUN} °Z ‘¢ pur"p
 dianapolis cars, a skidpad isn’t apt to have
a high enough speed capability to raise tire
temperatures to a simulation of high-banked
turns. However, the smaller asphalt oval
tracks themselves serve roughly the same
purpose. In a pinch, any long (180 degrees
or sO) constant-radius turn is better than
nothing, if a precise path can be marked on
it and followed by the driver.
Going to a race track is the last stage —
or as a last resort can serve as the first stage.
At any rate, consistency and safety will be
the greatest problems. A road-race course
is designed to be a challenge to the driver
and the total capabilities of a vehicle, not
to isolate test factors. It could help to put
down very distinct markers (braking and
cornering) to follow, so as to reduce driver
inconsistencies. There usually isn’t much
The interior of a Chevrolet R & D instrumentation van, with (from left) choice among available road-racing test
automatic-wind light table, 10-channel strip chart recorder, telemetry tracks, but if a track has some particularly
receiver, and portable calibration reference box. (GM Photograph) unsafe areas, the car can be timed through
limited sections, since overall laptimes
probably aren’t that important. A final con-
measuring devices. For example, when a costs, so you should investigate their most sideration on any surface is its cleanliness.
tachometer is calibrated, what is used as the up-to-date information before committing Obvious dust, dirt, or gravel that can blow
reference? “Repeatability” is a function of to one. This chapter will focus just on the or wash across a track surface can be a real
the human element. How close can a per- sensors as they apply to specific tests. For problem in any minimum-traction sections.
son “eyeball” the readout, and was there more on computerized data recorders, see But in addition, oil from broken engines or
any change in the object between Chapter 15. oil heat-seeping out of the track can hurt.
measurements? On the other hand, a lot of tire rubber can
Test gear can range anywhere from a FACILITIES help, unless rains have washed the surface
stopwatch to hundred-thousand-dollar The first thing necessary is a reliable, clean.
rolling laboratories of telemetry and re- repeatable, and safe surface on which to test. A stopwatch is an absolute necessity. The
cording instruments. The major auto and A great deal of the basic vehicle test work older dial watches are adequate, though they
tire manufacturers can afford to be the most can be done on simple drag strips, most of can’t match the capabilities of modern elec-
advanced in this area, with lightweight which are longer and wider than typical race tronic, digital-readout watches with
systems that can record anything they could track straightaways. Drag strips are usually memories and printers. Any watch is good
possibly want to know about a car or driver inexpensive to rent during the week, are enough for test work as long as it has a
at any time or in any place. It is possible smooth and level, and are far safer than a precision of hundredths of a second and has
to record simultaneously and continuously racetrack or deserted country road. If they split action for sequential timing. The elec-
dozens of channels of information on speed, can be rented exclusively, two-way runs can tronic variety, however, offers the option of
acceleration, displacements, locations, be made, which improves accuracy and cuts remote control and can be triggered by ex-
temperatures, forces, pressures, strains, and down on test time. This may be a good first ternal switches or photocell timing traps. If
events (See Fig. 44). exposure for any new car and/or driver, as automatic timing traps are not used, the
As of this writing, the most advanced it allows familiarization of the shifting, greatest inaccuracy with any watch will be
data acquisition systems are being used by braking, and high-speed stability without the the operator. At best the operator will find
Formula One teams with factory support. complication of cornering problems. it difficult to get his own two thumbs to
Traditionally, the major racing engine Ideally, the dragstrip should have an agree within a few hundredths of a second,
builders have led the way because of their adjoining skidpad on which to test steady- much less his perception of when a car
digital engine sensing and control needs. state cornering. There are few criteria for passed a certain point. Automatic timing
The key players are Bosch (German en- a skidpad, other than it be level, as smooth traps have become inexpensive enough for
gines), Pi (affiliated with Ford), Delphi as possible, and have a consistent and high- the amateur, but the time and difficulty in
(GM), Magneti-Marelli (Italian engines), coefficient surface. It is also necessary to setup seldom justify the increase in timing
TAG Electronics, EFI (Racing engines in have a guide circle painted on the surface. accuracy. In any case, some common sense
the US and Japan), CDS (American racing The accepted minimum radius seems to be is necessary in the analysis of elapsed times
aftermarket), Motec (Australian racing about 100 feet, while the maximum with any watch on any course. It is not
supplier), Racepak (primarily drag rac- desirable size is limited only by the available simply a matter of taking the best time
ing), Stack (derivatives of their instru- surface area. In small or isolated cities it (which may be timer error), or the average
ment panels), and smaller-scale data re- may be possible to rent or borrow a large, of many times (which may include a driver
corder suppliers such as Kwik-Data, Corsa, unobstructed parking lot, although it takes error). The most reasonable figure is
and Tach-Mate. All of these companies a strong underbase to resist the pounding probably an average of the three best times
sell competitive data acquisition systems, and cornering forces from large race cars. after any unusually low time has been
but with a wide range of capabilities and In the case of NASCAR sedans or In- thrown out.

116
 EST COST COMMENTS
Steering wheel angle Rotary potentiometer $20-100 For driver training, understeer angle, frequency response
Steering torque Strain gauge $20-200 Pure research, suspension and power steering development
Throttle position Rotary potentiometer $20-100 Driver development and comparisons on race tracks
Brake application Brake light switch $10-20 Driver development and comparisons on race tracks

Distance, longitudinal Wheel pulse counter $100-300 (or event mark) Important as a track reference
lateral Down-looking camera $50-500 Driver development (tracking ability), deviation from path
Speed, longitudinal Tach-generator/ pulse counter $50-300 Key factor, always used, precision considerations
lateral Differential of lateral location Useless

Acceleration, longitudina Accelerometer/slope of speed $100-500 Important for coastdowns, braking, power, and traction tests
lateral Accelerometer $100-500 For driver and tire development, must correct for roll
vertical Accelerometer $100-500 Only for ride quality, no standard analysis criteria

Angle in roll Photo analysis, or For cornering suspension analysis

pitch Suspension deflections, or For braking suspension analysis
yaw Humphry position gyros Most expensive gyro package, seldom justified

Angular rate in roll Differential of angle Relatively unimportant

pitch from suspension pots, or $50-100 Unimportant
yaw Rate gyros $200-500 Frequently used as a “response” measure

Individual wheel-speeds(x4) Pulse-train or DC generators $50-200 For brake lockup or differential testing
Individual susp. defl. (x4)} String-rewind or linear pots $50-200 For aero downforce or roll/pitch angle
Brake temperatures (Xn) Temp-plate or thermocouples $20-600 Slow changes, electronics seldom justified
Brake line pressures (X2) Pressure transducers $50-200 Mostly for research, transducers can “soften” brakes

Air speed Prop-generator/ press transd $200-400 Transducer location is critical, using zero wind day is better
Wind angle Wind vane potentiometer $50-100 Same as above
Body pressures (Xn) Pressure gage/multi-manometer $50-400 Also requires pressure rakes or body pressure taps
Aerodynamic drag See speed (coastdown) Take coasting curve slope manually or electronically
Aerodynamic downforce See suspension deflection Multiply chassis deflection times ride rate in Ibs/inch

Tire slip angles Sth wheel arm $100-1000 For tire research only
Driveline torque strain gauge and slip rings $500-2000 Unimportant, can also get from accel curve (-) intertia and drag
Event marker Manual/automatic trigger $10-200 Automatic photocell or trip switch is best
Voice record Tape $50-100 Available on some data cassette recorders

lEngine rpm Tach gen/ pulse counter $50-200 Often more useful than speed record on race tracks
temps/ pressure See brakes Slow changes, analog gauges usually adequate
other sensors Use existing ECM transducers Add similar sensors where necessary

Table 9. A summary of most possible vehicle data recording measures and transducers, with approximate cost and special notes.

ELECTRONIC INSTRUMENTS (1) Using electronic test equipment If the test instrumentation is made up of
primarily because it sounds important, and interchangeable modular components, it can
not because it is rationally justified. be a general-purpose tool for many types
When it comes to instrumentation selec- (2) Using specific transducers just of testing, and its cost can be spread out over
tion, it sometimes isn’t easy to reach a because they are available, and not because years of vehicle development. Therefore it
compromise between a vehicle engineer they are mandatory. should be considered as a system, and ten
who knows what he wants but not how to specific questions should be resolved before
get it, and a test engineer who is more (3) Taking far more data than will ever
assembling that system:
interested in sophisticated electronic be analysed.
gadgetry. Although the value of good test (4) Spending more time and money on
instrumentation is without question, there development of new instrumentation tech- (1) What is the ultimate purpose: general
are four common mistakes that must be niques than on the problem it was meant to driver or vehicle evaluation, or a specific
avoided: solve. problem to be solved, such as specific com-

117
 and distance is proportional to the sum of
the pulses. In addition, a math channel or
electronic circuits can be designed to dif-
ferentiate the velocity change to give a
signal proportional to acceleration. A
more precise digital pickup is the optical
sensor, which can be bought off the shelf
with over a thousand pulses per revolu-
tion.
For recording straight-line or slow cor-
nering speeds to a greater precision, say
for the first critical inches in a dragstrip
launch, a fifth wheel may used, since it has
minimal diameter change effects. But for
most race car use, a front wheel is best,
preferably the outside wheel, as it has the
least effects from lifting or slipping.
For recording laptimes on the data, or
recording a reference to the car’s location
on the track, infra-red triggers are com-
monly used. In fact, they are often used
with a simple onboard timer and readout,
Advanced sensors available from Pi, including left to right: tire pressure transmit- without the need for an expensive multi-
ters, laser ride height sensors, and driveshaft torque sensors and transmitters. channel datalogger. These receivers work
with a trackside infra-red beacon trans-
ponent durability, more aerodynamic able. An onboard strip-chart recorder is mitter, commonly mounted at the start-
downforce, or turbo lag reduction? about as basic and easy to comprehend as finish line. More sophisticated systems
(2) What parameters absolutely must anything, with immediate mechanical feed- use many coded beacons located around
be recorded, and what precision is re- back of both calibration and performance. the track to give segment times.
quired to distinguish meaningful results: However, compared to modern computer Track location recording will become
10 percent? 1 percent? 0.1 percent? systems, they have some major drawbacks, more common and much more precise as
(3) What is the maximum number of such as size and weight, sensitivity to g- better GPS systems are demilitarized. At
channels absolutely necessary at one time? loadings, and frequency response. Onboard the time of this writing, systems were
Could you get by with eight?...sixteen? tape recorders with download to a strip being demonstrated that had astounding
(4) What data sampling rate is most chart were once popular, but are as compli- resolution of small fractions of an inch.
appropriate: ten per second, or as much as cated and expensive as computers, which This will make it possible to evaluate the
500 per second? Does data change slowly, keep getting cheaper and more powerful. different lines being used through corners,
or fast and continuously, like shock travel,? Whether the data recording is analog or and if sensors are mounted at the front and
(5) What is the most practical data digital, most of the measurement sensors rear of the chassis, even the pitch and yaw
logger, considering: channels, precision, willbe somewhat similar. The transducers angles.
sample rate, total capacity, cost, power, are where the real engineering comes in. It Accelerations can be measured in many
space, weight, and frequency response? is always necessary to convert whatever ways. Forward or coasting accelerations
(6) What is the best output data format? quantity is to be measured into electrical are usually so small that taking the slope of
Is it enough to see plotted curves, or is a signals that are compatible with the re- the speed curve (differentiation) is most
numerical printout needed for later auto- corder. The most common measurements accurate. For transient lateral accelera-
mated analysis? desired in race car development are usually tion (non-skidpad) or braking, however,
(7) Is data transmission necessary -- speeds, accelerations, displacements, the most direct method is the electronic
either car to pits or to office? Does an observer forces, and temperatures, and occasionally potentiometer accelerometer. These can
want constant trackside monitoring? pressures and stresses, although many more occasionally be found through aerospace
(8) What transducers must be bought or are possible. surplus stores in the desired 2.0-g range
rented? (See Table 9) Speed in miles per hour or rpm can be and are easily wired in with an accurate
(9) What data analysis procedures are sensed by two methods, depending on the fixed reference voltage. A bubble level is
necessary, such as damping, averaging, or use of the data. Sometimes a tachometer also necessary to insure a perfect zero free
scaling? Is raw data adequate, or a few generator may be used, which when from gravity effects.
simple calculations, or is computer reduc- mounted to a wheel or the engine, gener- Displacement sensors are most com-
tion required. ates a precise DC voltage signal propor- monly used to record suspension travel, or
(10) Is it cost effective, or would com- tional to rpm. When it is also necessary to steering or throttle movements. In most
mon manual test equipment be adequate? know the total distance or number of revo- cases a linear potentiometer may be avail-
Would you settle for 90 percent accuracy lutions, it is more practical to use a digital able with enough travel to fit the total
for only 10 percent of the cost? sensor. A number of magnetic slugs can be displacement. But a more flexible device,
Up to the last edition of this book, glued evenly around a rotating circumfer- in both mounting and total travel, is a
analog data recorders were a reasonable ence, and a magnetic pickup is mounted rotary potentiometer attached to a spring-
option, and still might be considered from near enoughto produce digital pulses. Speed loaded cable reel. This can easily be
a cost-saving standpoint if already avail- is then proportional to the pulse frequency, attached to any fixed surface, with the wire

118
cable connected to the moving component.
Ordinarily, the output inches of travel can
be equal to inches of vehicle component
movement. However, aerodynamic
downforce has become so critical, and so
sensitive to ground clearance, that it must
be measured more accurately than spring
deflection, to account for tire deflection.
To deal with the moving ground problem,
very precise laser displacement sensors
have been developed for racing applica-
tions, although they are about the most
expensive sensors available. Four are often
used, although two at one end and one at the
other will give both pitch and roll angles, as
well as average ground clearance.
Forces are difficult to measure if no
movement can be tolerated, in which case
strain gauges are necessary. Otherwise,
the easiest technique is to allow movement
against a known-rate spring and again
record the displacement. Wheel loads or
vertical aerodynamic forces are most often
measured in this way, by recording the
average suspension deflection at a specific
speed. Strain gauges have the capability of
measuring forces very precisely with no
displacement, but they are inconvenient
for all but the most precise stress work.
They require professional attachment and
careful calibration, and they tend to record Measuring suspension deflection with a rotary potentiometer and cable reel strap-
every little unimportant force or vibration. ped to the fender, with the cable leading to a swivel at the center of the hubcap
Forces from fluid pressures can also be (left), and a linear potentiometer hose-clamped to a shock absorber (right).
measured with aerospace pressure trans-
ducers, to record fuel, oil, or air pressure.
Torques at the driveshaft or axles are
obviously very important information, and
&7
just a further application of strain gauges.
However, they require either slip rings
(which are impractical under race track
conditions) ora short-range data transmit-
ter from the rotating shaft. Although not
uncommon to passenger car development,
only recently have they become available as
a part of racing data acquisition systems.
It may also be desirable to record tem-
peratures in the moving race car, other
than by the fluid temperature gauges avail-
able to the driver. Thermocouples can be
mounted on the brakes, exhaust, or air
inlets, and wired into the recorder for a
continuous record. However, in most cases
the peak temperature is all that is wanted.
If so, it is far easier, faster, and cheaper to
use a simple heat-sensitive positive indica-
tor. The most familiar device is called
Temp-Plate, which consists of a row of
colored dots onan adhesive strip. The dots
blacken successively with precise increases
in temperature and are available in incre-
ments of 10 to 50 degrees, anywhere from be nae Oe

100 to 1100 degrees Fahrenheit. A rotary potentiometer (left) which be coupled to the steering shaft via a negator
Tire temperatures are critical to setup, spring, and a tach-generator (right), being mounted to the hub of a front wheel,
as explained in Chapter 1, especially real- with wires taped to the fender.

119
 can be checked with an electronic stop-
watch, or a constant accurate time reference
can be provided with a standard frequency
oscillator wired to the recorder. It can also
be useful to have an accurate digital
voltmeter when setting up a recorder and
potentiometers. Finally, it helps to have all
calibration gear easily accessible and to have
the recorder controls within easy reach of
the driver. It is often necessary to record just
one small segment in a series of laps or for
the driver to stop and check calibration out
on the track.

PRESSURE INSTRUMENTS
For aerodynamic work beyond random
experimental changes in shape and the
measurement of lift and drag, some pressure
instrumentation is desirable. Very little is
available off-the-shelf to suit race develop-
ment work, but it is relatively easy to build.
A commercially available pitot-static tube
is useful to measure both free-stream
dynamic pressure and free-stream static
pressure to use as a reference. Other
necessary commercial products are fluids
and hoses. Water is perfectly adequate, but
a colored manometer fluid will have the
same density and be easier to read. For the
measurement of multiple pressures at one
time, a flexible plastic hose called Strip-a-
Tube is available with up to 10 separate
tubes formed in one flat strip. Otherwise,
a hundred feet of model airplane hose will
work as well, though somewhat messily. A
manometer, pressure rake, and surface
pressure strip will probably have to be built
from scratch.
A multi-tube manometer will require at
least 12 tubes; 10 for measurements and one
J on each end for reference static pressure.
A capacity of 12 inches of water will handle
most pressures at reasonably fast air speeds.
Dual pot-type accelerometers (top), provide a bi-axial recording of g-level, while The easiest material to use is acrylic plate
the digital thermocouple readout (bottom) can also provide a recordable output. and tubing, which can be sawed, drilled,
time on-track data, as opposed to waiting to monly recorded in the development of high and glued to shape.
get pitlane readings with a contact pyrom- downforce bodies. However, the demand A pressure rake is simply a row of short
eter. Non-contact infra-red temperature for more data points, and the recognition of lengths of brass tubing soldered to a flexi-
sensors have been available for decades, slow data sampling requirements, has led ble metal strip. The strip can be taped above
and so inexpensive that it’s surprising they to multipexing. Scanivalve technology, a body surface or in a duct, and connected
aren't used more often. Some teams go so where many pressure taps are fed into a by flex hose to the manometer.
far as to have 3 sensors reading each of all single electronic sensor, has been adapted The surface pressure strip is designed to
four tires during development (of such from wind tunnel applications, andas many be taped flush to a body of any compound
variables as camber angle), then during the as 64 separate surface pressures may be shape, to read the air pressures acting
race, just monitor one average reading per recorded on a single analog channel. perpendicular to the surface with a
tire (when allowed by the rules). When anything is measured indirectly, minimum of interference to the airflow. The
Tire pressure recording gets back to the as with transducers and a recorder, accu- best material is probably flexible vinyl
torque problem of transmitting data froma rate calibration is critical. Forces can be sheet, which can easily be cut and glued into
rotating object. However, since this has calibrated by loading with known weights, a ten-channel strip and yet will be flexible
become relatively common in passenger and displacements calibrated by measur- enough to follow all but the sharpest body
car applications, it should become less ing with a tape measure. With accelerom- contours. To reduce airflow disturbance, the
expensive and more practical in racing, eters, 1.0 g can be checked by using the pressure pickup holes should be in the
especially from a safety justification. earth’s gravitation, while anything greater center and smoothly drilled, the strip should
Aerodynamic surface pressures are com- will require a centrifuge for calibration. Time be of minimum thickness, and it should be

120
e
taped to the body as smoothly as possible.
It is very important to make every hose
and tube air tight, or observed pressures
will be meaningless. It is even necessary to
test for pressure leakage between tubes or
channels, especially in the surface pressure
strip. This can be done by applying pressure
to one channel at a time and watching for
pressure readings in neighboring tubes.
Leaks can be found by watching for bubbles
when the entire setup is immersed in water.

WIND TUNNELS
In the first edition (1976), I dismissed
wind tunnels with one paragraph, saying,
“For a race car that is already builtand run-
ning, it is far more practical to do
aerodynamic development by measuring
changes on the existing bodywork. For a
radical new design or elaborate changes,
however, a wind tunnel can save time and This strip of Temp-Plate being attached to a transmission case will
money. . . . But it is such a complex opera- blacken its indicator dots up to the maximum temperature encountered.
tion that an experienced test engineer has
to set up and run the tests and analyse the
output data.” Recently, however, a number instrumentation is simply an electronic is scale. To get reliable results, the product
of big-buck Formula One, CART, and recorder capable of storing time/distance of speed times vehicle length should be con-
factory-sponsored race car teams have been and chassis downforce or spring deflection. stant between test conditions and operating
spending fortunes in wind tunnel rental — as described later. Although it may seem conditions. For example, if a quarter-scale
and even in building their own. So it seems that road testing takes a lot longer, allowing model is to be used, then tunnel airspeed
the subject should now be covered in more for individual run times, consider the time should theoretically be four times the real
detail. it can take to get a wind tunnel up to speed vehicle speed — and a 600mph tunnel
I have done race car aerodynamics and stabilized after a model change, and presents other problems! Of course, most
research both in wind tunnels and full scale the time required to average out normal Tace car testers violate this rule, but not
on the track, and in my opinion, for any tunnel fluctuations (G.M. uses 30 seconds without careful consideration of the errors
given amount of my own money, I would per run). it can create in local surface flow variations.
probably get better results per dollar real- For wind tunnel testing, whether rented The size of the model is also important
life, on-the-road. Certainly, those who have or built, the most important consideration in avoiding “blockage” in the test section.
made investments in tunnels can justify
them strongly. For example, Tony Cicale
(chief engineer for VDS, Newman/Haas; e i;4 AA ,°
Andretti, and Patrick Racmmg, among
others), spends over 200 hours per year in
wind tunnels, including G.M/s and some LULL
rolling road tunnels in England. He feels
that it is an efficient way to quickly weed
out 70 to 90 percent of those ideas that give
no improvement. The remaiming ideas are
then tried on the track, and even then only
about half of them will show up on a stop-
watch. But that still puts him ahead of those
who do no testing at all.
In comparison, road testing of
aerodynamics only requires a typical vehi-
cle, a thoughtfully selected site, and some
general-purpose instrumentation. The test
vehicle doesn’t have to be a final, priceless
race car, but simply an aero “mule” with
@ minimal rolling platform and a stock or
well-used engine. The site can be any long,
surveyed flat straightaway which has a
reasonable number of still, dry days.
(Granted, I'm lucky to be in Souther
California, with usually 2-4 hours of pre- The manometer is connected by tubing to a pressure pickup strip taped
dictable dead calm per day.) And the to the body, with pressure holes numbered from 1 to 10 from the nose.

121
 a full scale tunnel such as Lockheed’s. In
addition, there can be extra setup and
cleanup charges of a day each — which is
non- productive time. It inspires a person
to consider building his own.
Building your own tunnel sounds like a
fascinating project. A number of ac-
quaintences have done it, and even I was
asked to do so at Chevrolet in 1966. But
even if you could justify the considerable
capital investment in limited usage or
rentals, there are are a number of areas
where state-of-the-art engineering cannot be
overlooked, if test results are to be at all
meaningful. Design considerations include:
scale and size, fan and drive, open versus
closed circuit, nozzles and diffusers, flow
quality, balance precision, and moving
ground planes. Then of course, there are the
ultimate problems of final calibration and
correlation with the real vehicle.
The previously determined test section
size and airspeed define the fan and motor
requirements. The magnitude of the prob-
lem is illustrated by the ideal blockage ratio
of 1:20. If it takes 100 horsepower to push
Pressure measurement equipment being fitted to a duct includes: multitube a car through the air at 100 mph, then it will
manometer, Strip-a-tube, pitot-static tube, and pressure rake. take more than 2000 horsepower to blow 20
times the air at the same speed over a sta-
tionary vehicle. In fact, GM’s new tunnel
For relatively blunt objects like open-wheel tunnels with up to 120 mph through an 80 has a motor rated at 3000 continuous
race cars, the wind tunnel test section area by 120 foot section -- but which are un- horsepower or 4000 horsepower for 15
should be at least 20 times the frontal area available to all but government projects. minute bursts. Every time one of these
of the model. Otherwise, the tunnel walls The tunnels most frequently used for auto- monsters is switched on, the lights dim at
can create unrealistic and indeterminate ef- motive research in America are at: General Hoover Dam. Of course, for smaller scales,
fects on air flow. Obviously, then, a quarter- Motors, Warren, Michigan; Ford, power requirements are much less. To make
scale model can require an eight-foot square Dearborn, Michigan; Lockheed, Marietta, a good prediction of any design parameters,
test section. Georgia; Chance-Vought, Dallas, Texas; use engineering data from such references
Model-building sounds like fun. But get- Northrop, Los Angeles, California; Na- as Low Speed Wind Tunnel Testing by Rae
ting the necessary detail in a model can be tional Research Council, Ottawa; Califor- and Pope, or Wind Tunnel Technique by
almost as expensive and time-consuming as nia Institute of Technology; University of Pankhurst and Holder.
building the full-size body. To get truly Maryland; University of Kansas; Univer- Most automotive tunnels in this country
representative airflow, not only must all sity of Michigan; University of Wichita; are of the “closed-loop, closed-section’”’
ducts and radiator cores be accurately Cornell University; and the latest entry, type, which means that the air recirculates
simulated, but most seams, gaps, fasteners, Swift in San Clemente, California. through a continuous tunnel, and the test
and surface imperfections should also. And General Motors’ latest tunnel has to be section has closed walls. Although it may
on top of that, on most race cars the wheels one of the best, but don’t even think of cost more to build, it saves a great deal in
have a dominating effect on airflow, and to trying to get in there without an unsolicited operating power requirements, and it pro-
run a model without rotating them at typical invitation from the chairman of the board. duces much better airflow quality. Some
speeds may be pointless. Finally, the forces For normal production vehicles, it's al- European tunnels use open inlet and exhaust
to be measured can be a real problem, con- ready in use two shifts, seven days a week. bells venting to and from the atmosphere,
sidering that for the same airspeed, quarter- So it turns out that access or scheduling and they may also have open-walled test
scale model forces are about one-sixteenth problems are roughly equivalent to weather sections to allow freer airflow around the
of the true forces. So to show incremental problems in road testing. For details on vehicles.
changes, it may require a readout precision how they built it, see SAE paper number Flow quality can make or break a tunnel,
of grams. 820371. Most auto manufacturers use and must be considered from day one. The
Renting a wind tunnel is a fairly common smaller ‘‘climactic’’ tunnels for cooling spinning propellor and any airflow path
practice for infrequent users. Once the scale and ventilation tests, but these aren't so- changes can generate intolerable turbulence,
is determined, the search begins for the phisticated enough for accurate lift and which must be filtered out with vanes and
most appropriate facility, in terms of size, drag measurements, but only for A-B tests screens. There are precise means of measur-
speed, precision, availability, and of course, on the same model. ing this turbulence, for rating tunnels, but
cost. They range from small-scale univer- Commercially available tunnel rental it can be vividly illustrated by the ever-
sity tunnels, which may take on a limited rates range from about $100 per operating popular smoke stream, which eventually
project for a student thesis, to giant NASA hour in a small one, to over $1000/hour in breaks up and diffuses depending on the

122
level of random turbulence. Flow quality
also includes boundary layers, which can
build up on the walls and especially the
ground plane, which is supposed to repre-
sent a moving surface.
For absolutely true simulation of race car
airflow, the floor of the tunnel should be
speeding along at the velocity of the air.
Government and industry tests have shown
the effect to be almost negligible for pro-
duction sedans — but not for race cars. Of
course, there are some examples of rolling
road wind tunnels in existence, but they
haven’t been cheap, and they generate some
formidible problems of their own. First, it
isn’t easy to get a belt to run smoothly at
the speeds required. Second, race car
aerodynamic downforces have a very strong
tendency to lift the belt itself, until it con-
tacts the model and upsets force readings.
And finally, the belt makes it necessary to
support the model from the top, rear, or
sides, with struts which can also upset
airflow quality.
Aerodynamic forces are usually measured
at the four wheel contact patches, to avoid
disrupting the airflow. A strain-gauged or
mechanical beam balance is located below
the tunnel floor, and is designed to provide
the appropriate sensitivity in three axes.
This is often complicated by the need to shift
the model position to simulate movements
and crosswinds. A common alternative for
models is an internal strain-gauged force
balance, which must be adapted to each
model, but at least makes it easier to use
a moving ground plane. To analyse the
reams of output data, most modern tunnels
have computerized data reduction for instant
feedback.
Calibration of the airflow and balance in
a new tunnel can take weeks or months,
depending on original planning and the
precision desired. First the longitudinal gra-
dient or pressure change due to boundary
layer buildup is measured, and corrected or
allowed for in drag measurements. Then the
lateral survey or velocity profile is taken as
a measure of gross airflow variations and
boundary layers. This usually requires
changes in wall shapes, and fan and corner
turning vanes. And finally, the turbulence
factor is determined by comparing the drag
of a known object such as a sphere with its
drag in other tunnels. This can be improved
by the addition of screens or honeycomb
sheets, with a corresponding loss in
airspeed.
Calibration of the balance is a serious ex-
ercise in precision and patience, even though
only drag and front/rear downforce are im-
portant in race cars. Force balance factors
such as sensitivity, linearity, repeatability, Wind tunnel extremes, from the 1/25 scale demonstration tunnel (top), to NASA’s
hysteresis, drift, and travel must be Ames behemoth (note truck), and the full scale G.M. research tunnel (bottom).

123
 of the question for anything more than A-B
tests, because horsepower measurements
are limited by tire slip on the rollers, mak-
ing these dynamometers less accurate.
The most practical dynos may be the
portables, whichcan be bolted to the engine
bellhousing on an engine stand or to the
output shaft on a transaxle. Of course, all
engine development work should be done
by the builder in a dyno cell before the
engine is even installed. But the racer with
a few less dollars or an off-brand engine he
has to develop on his own can learn a lot
faster on a dyno than by cutting and trying
and measuring laptimes. Although straight-
away acceleration runs are simple and pro-
vide real-world information, a dyno can
save time and improve precision.

TEST PROCEDURES
Even with the best equipment, a test may
A half-scale champ car on the Swift wind tunnel moving ground plane. The model
be meaningless if some small step is
is supported from the overhead strut, while the wheels are attached to side struts.° overlooked. An accurate test requires careful
preparation, precise execution, and often a
lot of practice. After the driver has become
documented in each direction, and then or a can opener, is just a tool. Whether it familiar with the car and track, sometimes
checked in interaction with other force is used is less important than how it is it takes still more practice to become
inputs. Then the effects of external vibra- used. ..which is solely a function of the familiar with a particular test procedure or
tions (fan, airflow, and model mount reso- wisdom, skill, and creativity of the user. So the instruments involved. If there is room
nances) must be checked under running the decision whether to test on the road or for a passenger, and if he doesn’t suffer from
conditions, to indicate whether any damp- in a tunnel is a situational one, each having motion sickness while reading instruments,
ing or averaging is necessary. Off-the-shelf its Own appropriate time and place. While some of the responsibility and concentra-
commercial strain-gauged balances can be playing with scale model race cars in the tion can be taken off the driver. But for
a bargain at several thousand dollars. comfort of a high tech facility like a wind single-seaters, there isn’t much choice — the
Correlation of wind tunnel data with tunnel may sound on the surface to be a driver will sometimes have to be a test
final vehicle performance is often taken for hobbiest’s delight, maybe you can see why engineer also.
granted. After all, if facilities are available I prefer being on the road most of the time.
to measure full scale aerodynamics, why go ACCELERATION
to the tunnel at all? So engineers usually OTHER EQUIPMENT Acceleration tests are probably the most
simply assume that the best wind tunnel Tire or suspension development will in- common. Although an engine may be
model configuration will produce the best evitably require the use of a tire pyrometer, perfectly developed and tuned to the last
vehicle. Manufacturers’ aero research, or electronic thermometer. Tire company horsepower, there are other factors that must
however, has to be proven -- to the cus- engineers are usually available to take be resolved on the track. The dynamometer
tomer and tothe EPA for mileage tests. But temperatures at professional races, but for can’t quite simulate the response require-
very carefully-controlled comparisons have personal use, a reasonably quick, reliable, ments of a driver accelerating out of a turn,
shown tunnel-to-tunnel variations of 5 to 8 and accurate pyrometer can be bought for or the inertia, air drag, and air pressure
percent in drag coefficients -- on the same a few hundred dollars. Tire temperatures effects on the system. When gear ratios are
models. And no one can say which is the change so fast that convenience and speed changeable, they too must often be selected
most accurate answer, because there is no are essential. For truly accurate measure- by trial and error.
standard reference. In similar comparisons ment, some companies have developed For most racing situations — except drag
with real vehicle aerodynamic coastdown infrared sensors which can read and record racing — acceleration from a standstill is
tests, various wind tunnels have given drag surface temperatures on a tire at high speed. irrelevant, and acceleration with wheelspin
figures from -6 percent to +11 percent They are so expensive, however, that stop- is more of a tire test than anything else.
different, depending on the vehicle. These ping and contacting the tire is usually fast Therefore, a good acceleration test will be
variations have not yet been explained by enough for most engineers. from the end of wheelspin capability up to
the experts, so who are we to say what our Engine development has reached such a the maximum possible speed. If the gear
accuracy will be? The amount of money level of sophistication that it is beyond all ratios are already determined, the engine
available in professional racing, and the but the largest professional teams. But even output can be analyzed in just one gear —
critical importance of underbody aerody- for them, the engine is often considered a from some reasonable rpm to the upper rpm
namics, leads to continual escalation in simple “lump” which is purchased and limit. Before the test, certain conditions
tunnel capacity, so that we may soon see installed as is. Still, the cost of must be standardized, such as tire pressure
full scale models and speeds, running on dynamometers has come down to the point and temperature, engine and gear lube
full-speed rolling belts. where one can be bought for the price of temperature, exact total vehicle weight, and
Still, even the best tunnel, like a computer a V-8 engine rebuild. A chassis dyno is out precise shift rpm. If there is any possibility

124
of a road grade or wind, all runs must be
made in both directions. Assuming that i
acceleration is measured by the time interval
between velocities rather than with an ac-
celerometer, the velocity signal must be
accurately calibrated, and there must be a
precise “‘event’”” mark at each given speed.
If the time is measured with a stopwatch
against the tachometer, a greater time span
will improve accuracy. If the time interval
between speeds is taken from a continuous
speed recording, it is much easier to iden-
tify precise speed points but still necessary
to have an accurate time base on the
recorder paper.
For a simple A-B test between two con-
figurations, it is a simple matter to compare
elapsed time between two speed or rpm
points. However, it is possible to convert the
figures into acceleration in g’s, or thrust Getting out of shape in a brake balance test. Note wires leading to
force in pounds, or horsepower at any given tachometer-generators mounted on the right front and rear wheels.
point. In addition to the change in miles per ~
hour (V-v) and elapsed time (t), it is also
necessary to include the exact test weight No wonder few drivers are willing to use aerodynamic drag and downforce contribute
(W), times a factor to include rotational their brakes to the limit — even if they are to stopping power. To measure the actual
inertia. The equivalent weight due to rota- aware of what the brakes are capable of. The rate of deceleration, it is necessary to have
tional inertia may vary from 3 percent for ideal test location is at the center of a long an electronic accelerometer wired into a
a large sedan, to 6 percent for a light race straightaway such as a drag strip, so that strip-chart recorder. But aside from pure
car, but is roughly equal to the weight two-way runs can be made and so that there research, about all that will show is whether
percentage of the wheels and tires. is plenty of safe escape road. A race car the driver is using all available brake
Therefore, the equations are approximately: should be tested with a full fuel load for capability at all speeds.
_ (.045) (1.06 W) (V-v) maximum stress, but since the tanks are Braking stability tests are probably more
force unlikely to be exactly at the center of gravity meaningful with respect to vehicle develop-
(t) and fuel slosh can have a considerable ment. The question of whether and when
power = (.0027) (force) (average mph) effect, the car should also be tested for the front or rear brakes lock first was
balance while half-full and empty. It is also discussed in detail in Chapter 5. If the
Both force and horsepower figures are ap- necessary to warm racing brakes up to some braking g’s are much lower than the tire
plicable only at the specific average mph typical operating temperature and break in capabilities seem to be, or if the rears lock
point, or (V + v)/2. By taking data at a any new pads before testing begins. This is first, some front/rear ratio change will be
number of points, it is possible to plot a a good time to evaluate the modulation, or required. In addition to the previously
curve of the net acceleration force or the ability of the driver to hold the brakes explained procedure of an outside observer
horsepower above that required to overcome right at the verge of lockup without flat- determining which locks first, an onboard
friction and aerodynamic drag. An article spotting a tire. It should be just barely pos- recorder can be used. Having a velocity
by Ronald Brown in Road & Track, sible to get short, instantaneous bits of sensor on both a front and rear wheel can
November 1970, goes into a more detailed sliding without full lockup, with the per- even measure the relative slip rates in
explanation, with examples and illustrations. formance smooth and consistent throughout braking. If a two-channel recorder is used,
Of course, simple knowledge of the each and every application. A good test of both wheel speeds can be measured, but
horsepower curve doesn’t make a race car consistency is to make repeated complete even a single channel recorder can be used
any faster. It is possible to plot the thrust, stops from a specific speed while measur- if a circuit is designed to compare two speed
or power, or g curve in each gear, however, ing each distance. Declutching at the same signals constantly and read out the differ-
to locate the optimum shift rpm for each time will prevent engine interference or ence. Finally, even if the brakes could be
gear ratio. When many ratio selections are stalling. The driver should be able to begin perfectly balanced for continual variations
available, this can be used to help select the the braking within a few feet of a given fixed in weight distribution, aerodynamic
optimum gears — in addition to the tech- marker and keep each stopping distance downforce changes, and engine drag at the
nique described in Chapter 9. Finally, within a 5 to 10 percent variation: This is rear wheels, there would still be another
acceleration tests will give an elapsed time also a good test for maximum average change from straightaway tests to a road
evaluation and subjective feel for engine capability in g’s, knowing the initial speed course. Since the ideal situation is to brake
response to sudden throttle applications, in miles per hour (V) and the stopping deeply into a turn and there is more lateral
especially at low rpm or during a fast shift. distance in feet (d). load transfer at the front, it will probably
(.0334) (V)? be necessary to reduce the front bias even
BRAKING stopping g = more to avoid inside front wheel lockup at
d the limit.
Brake tests are all too often left until the
end of the straightaway on a road racing The actual instantaneous deceleration will Brake durability tests are fairly difficult
course, where any error can mean disaster. be greatest at higher speeds, where to perform, and there is the basic problem

125
 can be recorded in a number of ways. If the
car has a very precise and accurate
speedometer, a stopwatch can be used be-
tween any two speeds. Otherwise it will be
necessary to use an electronic speed signal
and record it with a precise time or distance
base. Another alternative that is sometimes
used is electronic speed traps in the road.
OP OLE Krai” However, at least three triggers are required,
and precision, accuracy, and reliability tend
to be worse than with continual speed
recording.
The test procedure requires considerable
care in execution. To minimize friction drag,
or at least to make it relatively constant, the
tires are overinflated and tires, brakes,
lubricants and bearings warmed up. (Tire
drag will vary noticeably unless they are
warmed up for at least 20 miles.) The driver
accelerates to the highest speed that will
allow some coasting distance, shifts into
neutral, and coasts with a minimum of
steering fluctuations. The precision of this
test is so important to the automotive
industry, that an elaborate standard was
created for it (SAE J1236), which includes
estimation figures for rotational inertia. But
still, to get good repeatability, nothing may
be changed from run to run: accessory
power, vents, precise coasting location,
turnaround technique, or perhaps even
minor weight changes due to fuel
consumption.
The air drag in pounds at each speed is
calculated the same as acceleration force
was, where the difference in two speeds is
(V-v), the elapsed time is (t), and the test
weight of the vehicle includes some rota-
tional inertia factor (1.06 W). Therefore:
The ever-popular smoke flow and ink droplet tests can give a rapid indication of
approximately which direction and how smoothly the air is flowing. (G.M.
(.045) (1.06W) (V-v)
force =
Photograph) 9)
That will be the drag force at the average
of accurately simulating real requirements. tests it is a good idea to keep a record of speed (V + v)/2. It is generally assumed
Tests can be made by driving back and forth pad thickness versus miles to get an indica- that air drag increases with the square of
and braking on a long straightaway if all the tion of durability, if that isn’t known from velocity, but plotting a number of points at
necessary maximum and mimimum speeds previous racing experience. Enough testing widely varying speeds will indicate the
are known, but a race track is an unques- may even show the variation in wear rate actual force curve with mechanical friction
tionable real situation. The most valuable at different peak temperatures, which may drag included. Alternately, it is possible to
information is the temperatures at the disc, be very important if pad life is marginal in design an electronic circuit to differentiate
the pad material, and the fluid in the caliper. a race. the speed curve continuously and produce
Even without a recorder and thermocouples, a signal proportional to g’s.
positive-indicating Temp-Plate will give the AERODYNAMICS Aerodynamic downforces are best
peak figures if the brakes are truly forced Aerodynamic drag and downforce tests measured in pounds at the front and rear
to their limits. (However, continuous recor- are most easily performed on a long axle centerlines and at a given fixed speed.
dings show that “heat sink” after the car straightaway, preferably one that is perfectly As explained in Chapter 6, the forces should
comes to a stop can raise some component flat (or at least surveyed so that the precise be measured at a constant speed, to include
temperatures far above any seen on the grade — and its variations — is known) and rearward load transfer due to air drag. The
track, perhaps giving a false reading.) The free from any winds. The only way to know test setup includes some sort of suspension
most severe conditions are; full fuel load, that grade or wind has a minimal effect is deflection sensors mounted at the front and
high ambient temperatures, maximum trac- to run all tests at the highest possible speed rear, and either observation or a continuous
tion coefficient, and a comparatively low and to repeat each test in the opposite direc- recording of the deflections. For a solid rear
average-speed course, to minimize high- tion immediately. The most convenient test axle, the mounting point should be the
speed cooling. During these temperature for air drag is the coast-down run, which center of the housing to eliminate any asym-

126
metry or roll effects. Most independent
suspensions have an anti-roll bar between
the two sides, so the best mounting location
is a lever arm fastened to the center of the
bar. The best calibration, or scale factor, is
probably to have 1 inch of chassis deflec-
tion equal to 1 inch on the chart recorder.
After the instrumentation is set up, the
actual suspension ride rate in pounds per
inch must be determined. Assuming that
downforce is going to be recorded — rather
than lift — the best method is to load the
chassis with known weights and note the
corresponding deflections in inches. For
close approximations and relatively linear
suspension spring rates, it can be simplest
to have a known-weight person climb on the
Measuring cornering power and handling on a skidpad. The car is held
chassis, and then note that single
at maximum speed on a precise circle—indicated by the white line.
weight/deflection point on the output chart.
Otherwise, for greater precision, or non-
linear springs or anticipated high loads, it tunnel work he did on downforce. Just camera to record the water levels.
will be necessary to use a number of large. imagine, if the FIA ever wants to truly The test procedure is to drive at a specific
test weights. It is also important to load the regulate downforce, all it has to do is install constant air speed — repeated carefully run-
chassis directly at each axle centerline, and such a plate in every race track straightaway, to-run — while the pressures are recorded.
to carefully bounce out any friction or Air pressure measurements or air velocity It is a good idea to approach the speed
hysteresis in the suspension before each profiles on a body are fast and easy to record slowly, because extreme pressures can run
reading. Knowing the pounds-per-inch rate with a multi-channel recorder, hundreds of the fluid out the top or bottom of some
and the deflection in inches due to pickup points, pressure transducers, and manometer tubes and it can be difficult to
downforce, it is a simple matter to calculate multiple-scanning valves. However, with the work air bubbles out again. It is probably
pounds of force. less elaborate and less expensive equipment fastest to mark the plastic face of the
The test itself takes some care to get any described previously, taking pressure manometer with a wax crayon and copy the
accuracy or repeatability. Headwinds or measurements over an area of any size is various inches of water on paper during a
crosswinds can have such odd effects that a long and laborious process. Ten points or pickup relocation stop. For accurate figures
averaging two-way runs under such condi- readings are about all that can be con- in inches of water, it is also important to
tions is only slightly better than nothing. It veniently taken at one time, so it takes a lot keep the manometer tubes perfectly vertical
is also necessary to have the test speed be of equipment-relocating and test reruns to while recording. Later, the figures can be
as constant as possible during a run, and cover a large surface. converted to average air pressure perpen-
also from run-to-run. Suspension friction Part of the test setup is to find an accurate dicular to the surface, or velocity of the
can be great enough to hold the chassis as and constant source of reference static airstream at a given speed, by the
much as half an inch away from the true pressure. A race car will have varying relationships:
force-balanced height, so it should be pressures everywhere in and around it as the
carefully shaken free at the test speed. If speed changes. Therefore it will be best to 1 inch of water = (5.2) psf
road roughness isn’t enough to keep it free, use the static pickup holes in a pitot-static 1 inch of water = (.00049) (mph)?
it will be necessary to blip the throttle and tube, and the tube will have to be mounted
rock the steering wheel slightly to get it some distance away from the body to avoid Of course, it is also possible to measure true
balanced out from previous acceleration or interference. Four to 5 feet above a body, air speed of the vehicle by comparing the
cornering deflections. While the suspension or 8 to 10 feet ahead, should be good static pressure reference to total frontal
deflection instrumentation is set up, it can enough, but a check can be made by using pressure from the pitot tube.
be worthwhile to record maximum pitch a very sensitive absolute pressure gauge at
angle during braking and acceleration. The rest and at high speed. Commercial HANDLING
figures may come in handy when working mechanical absolute pressure gauges may Skidpad tests can be the most productive
on suspension travel, anti-dive or anti-squat also be used for body pressure readings, and exciting tests for a road racing car, but
effects, or in correcting accelerometer except that their cost and bulk doesn’t justify they are also highly demanding on driver
readings. any added precision. precision and consistency. It may take a lot
John Barnard, a brilliant designer for The pressure pickup strip and total of practice laps before the driver can main-
McLaren, Chaparral, Parnelli, and Lola, pressure rake are taped to the body or in tain a nearly constant throttle and steer angle
has used an even more clever aerodynamic a duct and connected by flex tubing to the and get lap times to repeat within a few
test setup. Michelin has an electronic load- multi-tube manometer. The two outside hundredths of a second. But a lot of valuable
cell plate built flush into the surface of a manometer tubes are connected to the static data can be obtained at low cost and low
3-km long straightaway. The response of this pressure reference, and the manometer is risk on a skidpad, so it is well worth the
system is so quick that it can measure the filled about half full with colored water. It effort. The test setup will vary depending
load on both front and rear tires passing over is practically impossible for a driver to read on the data needed, but driving technique
it at 150 mph. So Barnard simply used it or mark the fluid levels at speed, so a is simply to keep the car on a precise
to verify the the months of scale-model wind passenger will be necessary or perhaps a circular path at the highest possible constant

127
 120 eye and thumb, or an automatic trap switch
to start and stop the watch. A timer’s ac-
curacy can be improved by having a distinct
A EXTREME U mark on the circle and noting the instant the
8 UNDER- ] car’s front tire hits it. It is also a good idea
ge) 00 STEER for the timer or someone else to watch the
=
&
LIMIT car and make a note whenever it moves
noticeably off the marked path. As with any
=
wo other data-taking, any particularly low
rc) figure will have to be backed-up before it
2 60 is credible. Of course, tires should always
=
= be scrubbed in and warmed up until lap-
times reach a consistent and low level, and
the turn direction should be reversed
Sued TYPICAL frequently. The relationship between lateral
8 RANGE / acceleration, radius, speed, and elapsed
5 time is given by:
</ (.067) (mph)? CE22)i(n)
0 NEUTRAL § % OVER-STEER ©) ae OG:
0 0.2 0.4 0.6 0.8 1.0 where the radius (r) is to the center of gravi-
lateral acceleration (q’s)
ty of the race car, or the inside tire path plus
0 10 15 20 25 30 35 one-half the tread width, plus some slip
speed on 100 foot radius (mph) angle displacement.
Steady-state stability can be evaluated at
120 the same time as maximum lateral accelera-
tion. This is the tendency of the race car to
maintain a stable non-oversteering condition
at top speed on the skidpad, as discussed
in Chapter 7. As various factors, such as
© 90
Lo?)
springs, anti-roll bars, or tires are being
S changed, the steering angle necessary to
2 / balance the car will change. It should be
= ,O) possible to see the increase in laptimes as
BS 60 DISCRETE DATA POINTS 4 the car becomes unstable and uncontrollable
£ FOR TEST CAR 4 with oversteer. But it is also possible to
measure the oversteer or understeer by
2 of
recording the steer angle. At a very low
E2 30 ~~ -0-
-- speed — say 5 miles per hour, or the
minimum possible — a given steer angle
= ~_-O- -o STEER ANGLE AT (Ackermann steering) is necessary to keep
O° —————- 7 RO MPH ON RADIUS the car on the circle. As the speed is in-
creased, the necessary steer angle will either
increase (understeer), remain constant
0 (neutral steer), or decrease and require
0 0.2 0.4 0.6 0.8 1.0
quick and violent corrections (oversteer).
lateral acceleration (g’s) Examples of understeer/oversteer plots are
shown in Fig. 45. The vertical scale is the
Figure 45. Measures of understeer taken on a skidpad, in general (top), and taken steer angle required to drive the car on a
at six discrete speeds on a specific vehicle (bottom). 100 foot radius circle. The horizontal scale
is lateral acceleration in g’s, which increases
as the square of increasing speed. (Speed
speed. To a driver who is used to clipping tage is that it is hard to take the average is shown just for comparison on the lower
corner apexes with the inside front tire, it speed, and the recorded speed is affected scale. The speed scale would change with
is probably easiest to try to keep the inside by lateral tire slippage. A more direct different radius skidpads, while the g scale
front tire on the painted circular stripe. That method is to use an electronic accelerometer remains relatively constant.)
tire will also be least affected by a change mounted laterally in the car. But again, it In the lower sublimit area of the curve —
in friction coefficient due to the painted is hard to take an average of all the fluctua- usually up to perhaps 0.3 g — the curve is
stripe. tions, and the reading has to be corrected relatively flat, and can be presented in a
Maximum lateral capability can be for roll angle. The easiest, fastest, and prob- single number of degrees per g. But up
measured by one of three methods, or all ably most accurate method is to take the lap- closer to the limit of lateral acceleration is
three together, if accuracy demands it. The time for each complete revolution on the where the interesting things happen. With
first is to use an electronic speedometer and circle. This merely requires a split-action a normal understeering car, the curve will
record the speed constantly. The disadvan- stopwatch and someone with a very precise continue to climb, usually at an increasing

128
rate. In other words, it requires more and
more steer angle, until no amount of steer-
ing will keep the car on the path. An
oversteering car, however, will reach a point
where the curve flattens out and makes a
sudden break downward. This is where
“countersteering” is necessary to stay on the
path — a very unstable situation. The rare
neutral-steer car will have a very flat curve
from zero and not break up or down. It is
generally accepted that the straightest curve
is best, and with enough rise that it never
breaks downward under any transient
conditions.
It is important to mention that many static
vehicle parameters can affect these curves,
and so they are sometimes factored out. The
Ackerman steer angle is the steering
necessary to stay on the radius at just above
zero speed or g’s, and is a function of
wheelbase and steering gear ratio. Therefore
it is usually subtracted out, defining zero
steer at just above zero speed. Steering gear
ratio will also directly affect the slope of the
curve, doubling the slope with a doubling
of the gear ratio. However, because the
change in steering wheel angle is what the
driver actually perceives, gear ratio is not Tire engineers using an electric pyrometer to measure the temperature — and its
normalized out except for engineering variations across the tread — of a race tire in the pits.
analysis of front/rear tire slip angles.
To get these curves, it is usually neces-
sary to use an electronic data acquisition and starve the pickup, it is better for it to For real-life conditions such as bumps, dips,
system. Vehicle dynamics engineers ordi- happen on a skidpad where there is a greater and combinations including acceleration and
narily use recorders which take a minimum opportunity to shut the engine down im- braking, only a test on an actual race track
of 8 channels of data simultaneously. mediately. The same is true of fuel pickup will suffice.
Here, just 2 channels would be adequate: systems, especially in trying to determine
steer angle and speed. Another less precise how completely they can drain the tank TIRE TESTS
alternative is to increase speed in steps, and without picking up air and leaning-out the Tire tests are one of the most valuable
make a visual note of the steer angle in engine. The greatest problem with the test uses for a skidpad, once the chassis has been
degrees at each speed. is that in many cases the worst possible fairly well developed. There is no other way
Transient tests can only be positively situation is compounded by the addition of to accurately determine the optimum tire
made on a skidpad that allows entry and exit braking or acceleration forces. compounds, temperatures, camber angles,
from the steady-state circle, or on a race Suspension deflections can be important or pressures. Results obtained from race
track, of course. However, if there is even to know, especially in development of track testing would have to be far more
a little extra room on the skidpad, it is possi- geometry and ride rates. Deflection significant to. eliminate driver inconsis-
ble to get a feel for transient response. There measurements on the skidpad will merely tencies. The test setup simply requires a race
should be enough understeer so the car can show the maximum roll angle, how close car and driver that can run all day long at
be accelerated relatively rapidly without the components are to bottoming, and how low speeds and high lateral acceleration,
losing traction at the rear. In other words, great the jacking effect is. The test setup is with no fatigue, overheating, or variation in
there shouldn’t be any drastic throttle about the same for aerodynamic downforce, performance. It is also important to
oversteer. Conversely, there shouldn’t be a except that the deflection sensors are remember that only a pair of front or rear
sudden change in stability due to a complete mounted as close to the wheels as possible, tires can be tested at once, since the car will
throttle-off at the limit. It can also be an instead of in the center of the car. Ideally, be limited by either front or rear cornering
educational experience to brake hard from there would be a continual recording at each capability. The best practice is probably to
maximum lateral acceleration while trying wheel, but if only one channel is available develop front tire cornering performance
to maintain the path radius. rather than four, the test must be carefully first (since the car should be understeering),
When it comes to checking fuel and oil repeated four times. The best location to and as it becomes better, to keep increas-
pressures, the skidpad is far safer than a race record suspension deflection is probably at ing the front anti-roll bar rate as necessary
track, if not quite a perfect simulation. the centerline of the spring, although it may to avoid oversteer. When no more front
Naturally the driver should always be aware also be useful to know the wheel deflection. cormering power is available or the front roll
of oil pressure at high lateral accelerations, In that case, a swivel anchor can be mounted rate is so great that it lifts the inside front
but the skidpad allows him to watch it more to the center of the wheel, and the sensor tire off the ground, then it is time to work
closely over a longer period. If the oil level can be mounted outside and above the wheel on increasing the rear tires’ capabilities. The
is ever going to settle to one side of the pan — as on a fender lip or extended bracket. rear anti-roll bar rate may then be increased

129
 action watch) as it isto finda spot where the
car can be seen at many locations around
the track. At some tracks, it is possible to
see a car most of the way around the track
RIVERSIDE from the roof ofthe timing tower. A sample
RACEWAY of segment times is shown in Fig. 46.
This is also the best way to find out how
a competitor’s car really compares--as op-
posed to average laptimes. If the other
& driver is sandbagging, it will probably show
OBSERVATION POINT up in a particular segment. But if the other
interval times in seconds
car is quicker, it is helpful to know exactly
AREA:
where, to know where there are some capa-
CAR cn
bilities tobe gained. In the example shown,
8.60 | 12.85 | 6. car B is apparently better in acceleration,
|
DIFFERENCE
| +.10 [+.05 |—.10 }+.05 |[—1. oeI — Oo which means that car A is probably ad-
equate in handling and braking but should
Figure 46. Comparative Interval Times for Two Cars at Riverside have more power or lower air drag. Some
people have also used radar guns or elec-
tronic gunsight tracking devices to record
comparative speeds around an arc. There
to create oversteer, as a rear tire limiting work input under a given ambient are other more sophisticated spying de-
condition to overcome. temperature. vices to analyze competitor’s cars, which
Tire test procedure is simply to record When tire compound and temperature can are moreaccurate and more complete, such
average laptime or lateral acceleration for be held constant, then optimum tire pressure as computer analysis of exhaust sounds to
each configuration change. However, it will and camber angles can be determined. resolve an rpm curve. But, needless to say,
be necessary to monitor tire temperature Proper camber angle will show up in skid- they are much more expensive.
constantly, since its effect is great enough pad lap times, tire wear profiles, or Probably the best race driver teaching
to cancel out other test conditions. The first temperature differences across the tread, but device known to man is a continual re-
test with any tire should be a comparison the last method is quickest for tire develop- cording of speed and horizontal accelera-
of g’s versus temperature, to determine the ment work. The pyrometer must be used tions around a race course. It won’t say
optimum and the dropoff on either side of rapidly to get three readings (inside edge, much about the car unless there are other
the optimum. Since it is difficult and ex- center, outside edge) before the natural heat recordings of the same car in another
pensive to record temperatures continuously, conduction in the tread evens out the configuration, or other recordings of other
it will be necessary to stop the test at inter- temperatures. It isn’t reasonable to expect cars, to compare with. But such recordings
vals and check the temperatures as rapidly them to be exactly equal, however, since the will tell a great deal about the driver’s ability
as possible with a tire pyrometer. The best car will be at maximum lateral acceleration to take advantage of the car’s capabilities.
technique is to run in two or three lap in- camber angle only for very short periods. The best test setup is to have a two-channel
crements, with one person timing, and When everything else has been developed recorder with speed and acceleration inputs.
another taking temperatures as fast as the to the optimum on a skidpad, it can be a The speed can come from a front wheel
car can be stopped from its high lateral con- good place to teach the driver what extreme pickup, preferably the outside wheel, to
dition. It shouldn’t be necessary to take over variations in handling feel like. If the car avoid lift or lockup problems in cornering
half a lap to stop, and the tire technician is ever going to lose a shock absorber, or and braking. The g sensor can be either a
should be right there at the stopping point. break a front anti-roll bar and oversteer, or single lateral accelerometer or a combina-
Within 10 or 20 laps the tire should be at have a tire go soft, or otherwise become un- tion of two arranged at right angles. An
its maximum temperature, or past its peak manageable, the skidpad is the safest place electronic circuit can be designed to
cornering capability. It can also be valuable to learn the feel and the corrections re- calculate the net horizontal acceleration in
to know just how fast the tire cools off, to quired. Just knowing the feel of ordinary any direction and produce a signal propor-
get an idea of what the true temperature is changes from oversteer to neutral steer to tional to the percentage of traction used
while the tire is working. (See Chapter 2) understeer is an invaluable aid in later versus traction available. A more thorough
This can be estimated by watching the analysis of a race car on a race track. explanation is given in Chapter 12.
temperature fall in a given location over a Other tests that can be performed on a
matter of seconds and projecting the result. TRACK TESTING race track were previously explained under
Since the test accuracy is poor, it is a good An actual race course is the last place straightaway or skidpad testing. However, in
idea to repeat it after, or as, the tires cool where any serious or accurate development general they tend to be tests of the track con-
off, and in both directions around the pad. work can be done. Only after the car has figuration rather than of the vehicle.
Of course, the outside front or rear tire been otherwise ideally set up will race track Unusual suspension deflections are mostly
temperature is of greatest importance. laps be meaningful, and then primarily with dependent on surface condition or grade
From then on, all tests with that particular respect to the driver’s performance. For changes or bankings. Different tire or brake
tire should be run at that optimum vehicle evaluations, it will still be necessary temperatures are a function of track co-
temperature or at least corrected for any to break the track down into braking, cor- efficient and distribution of time spent on
dropoff. This data will also come in handy nering, and acceleration segments, as comering, acceleration, and braking.
at the race track, to determine whether a tire opposed to over-all laptimes. The timing Vehicle transient response characteristics
compound is too hard or too soft for the isn’t as difficult (with an electronic split- will change with respect to the types of

130 Hi
corners on a given race track. If there is
enough time, a race car can be set up to the
Po Aw
(W) — (X)
Wr Se
optimum for each particular track’s
predominant characteristics, but it is likely L
to take at least a few days of track rental and
exclusive running. It helps a great deal to
have a lot of experience at a particular track
and at a wide variety of tracks. But for those C.G./"
with no experience or for a new track, a
careful analysis of speed and g recordings h
can work almost as well. 4 > Ey

Of course, in the absence of any recording

instrumentation at all, stopwatch times in
various track segments will probably indi-
cy
cate the worth of any vehicle changes. W (25

Handling or tire improvements should show

up in the low-speed cornering segment
times, aerodynamic downforce should show SCALES (AW)
up in high-speed cornering segments, engine
power should show up in the straightaways Figure 47. Locating Center of Gravity
and so on. The overall improvement will
probably show up in total laptimes, but less ©
significantly.
This leads to the question of the relative braking capability reduction with lower STATIC TESTING
value of various vehicle improvements. temperatures. However, it may be somewhat
Racers tend to concentrate on making im- difficult to quantify the exact value of the There are also a number of static tests or
provements in the areas they understand or change in the factor. But once the relative measurements that may be necessary or
enjoy the most rather than those areas where positive effect of various changes can be desirable for further development work. In
the potential gains may be greatest. It isn’t estimated, the proper concentration of the first place, it will frequently be
that hard to determine just what these efforts can be allocated. Of course, even the necessary to know the exact rolling tire
relative values are, however. A computer most scientific approach must be adapted diameter. Since there is some static deflec-
can be — and has been — used to put a to fit the capriciousness of racing regulations tion due to weight, the most accurate
numerical value on various race car im- and the availability of time and dollars. method is to measure the rolling distance
provements, but any racer can find the Most racers feel that durability tests are and divide by 3.14. In fact, it is more
numbers for his own car on any particular also beyond their limits in available accurate to roll the loaded tire for 10 full,
track by the handicap method. If it isn’t easy resources, so they tend to use experience, exact revolutions and divide by 314.
to improve a car’s performance, it is all too intuition, and luck instead. The only way ~ Center of gravity height is of more than
easy to diminish it. All that’s necessary is to really know whether a car can be raced just academic interest, since it is required
to know the amount by which the effect is hard for 24 hours or 500 miles is to race for many performance calculations. There
reduced and the increase in laptime which it hard for that time or distance. If it is done are many ways to locate it by suspending the
results. The effect will be linear enough over in a test session and nothing breaks, then car from a single point or by mounting it
a reasonable range to project from a the car can be totally rebuilt as new for the on a teeter-board, but the following is easiest
decrease in performance to an increase in real race — and still fail due to some for most racers’ limited facilities. (See Fig.
performance due to a positive change. random faulty new component. Even if ex- 47) First, it is necessary to realize that any
The most obvious example is in deter- perience is the best judge of durability, that fluid movement during a test can falsify the
mining the effect of reduced weight on lap- sort of experience can be bought. A brand results. It helps to drain all fluids, perhaps
times. All that is necessary is to plot new design may be faster but it definitely even oil in the pan and water in the cooling
laptimes versus fuel consumption in pounds. doesn’t have a history of reliability. The best system. At least two precise platform scales
Say that a 2000-pound race car consumes insurance is in knowing that a particular will be necessary, plus a hoist or forklift to
200 pounds of fuel during a race, while its design or component has been around for raise the front of the car. With the rear
average laptime decreases from 90.0 seconds a long time without unusual failures. At the wheels resting on the scales, the front is
to 88.2 seconds. Then, assuming that all very least, it is a good idea to know a car raised while the increase in rear wheel load
other factors remained equal during the or component’s history, and its average life is noted. Obviously, when the center of
race, a person could project that a 10 percent expectancy before inspection, rebuilding, or gravity is directly over the rear axle
reduction in vehicle weight would produce replacing is necessary. On the other hand, centerline, the weight on the rear will be
a 2 percent reduction in laptimes — at that if durability testing is feasible, the biggest total vehicle weight. Well before that point
track. (This is true for race cars with mistake is to try and make the car survive is reached, the center of gravity height can
centrally mounted fuel tanks. For extremely under those conditions. Instead, the idea is be calculated by:
rearward tank locations, weight distribution to try and break the car under reasonably (AW) (L) VL? — x?
changes are an uncontrollable variable.) severe simulated usage, rather than Hes
Other factors may be as easy to degrade as pussyfooting it around. However, it is wise (W) (x)
weight: power reduction with throttle-stops, to test on a track where a failure doesn’t have Where: (h) is height of the center of grav-
air drag increase with a flat plate, tire consequences as critical as they could be in ity above the rear axle centerline in feet, (W)
traction reduction with lower pressures, and a race. is total vehicle weight, (AW) is the increase

131
LIVERPOOL JOHN MOORES UNIVERSITY
LEARNING SERVICES
 brake rotor and hub with the wheel and
tire. Examples of equivalent added mass
(per corner) are: passenger car steel wheel
with 235/75R15 tire = 20 lbs.; Corvette
wheel/tire/rotor/hub = 31 Ibs.; pickup truck
steel wheel with LT245/75R16 tire = 38
Ibs. If vehicle yaw polar moment is nec-
essary, the same setup may be used, where
(r) is the radius of the three cable anchor
points in inches, and the general equation
is: I (ft-Ib-sec2) = (r2)(t?)(w)/473(L).
Addenda to vehicle measures and
sensors (p. 117). At the time of the latest
book revision in 1992, two more impor-
tant measures, real-time tire surface tem-
peratures, and aerodynamic ground clear-
ance, had become more easily obtainable.
Relatively inexpensive infra-red tempera-
ture sensors with voltage or digital readout
are available from mail-order catalogs
such as Omega (203-359-1660) and Wahl
Torsion test of a roll cage under construction, using a known body weight, an an- (800-421-2853). These can be aimed at
chor at the right front, and measuring deflection at the left front wheel. the running tire surface and the data re-
corded or provided as visual feedback to
the driver. Aerodynamic ground clear-
ance is so critical that it must be measured
in rear scale weight, (L) is the wheelbase of locations. The race car chassis is mounted more accurately than suspension deflec-
length in feet, and (x) is the raised height with the rear hubs anchored firmly to the tion sensors can, because of tire deflec-
of the front tires in feet. The higher the front floor, the front center of the chassis resting tion. Laser displacement sensors are avail-
tires can be raised, the greater the increase on a knife edge, and the weight beam con- able from Keyence (310-595-5646), al-
in rear scale weight and the more accurate necting the two front hubs. As the weights thougha little pricey, atabout $1500 each.
the calculations. The rear wheels may be are loaded on one end of the arm, torsional
blocked on the scales, but the front wheels deflection can be measured with dial gauges
must be free-rolling on a perfectly hori- at two points on the arm. The deflection in
zontal surface to avoid any weight effects at degrees for each torsional load is equal to
the scales. This technique is compared to the difference in readings, divided by the
others used by GM, Ford, and Chrysler in distance between the gauges, times 57.3. If
SAE paper number 920050, which shows the stiffness falls much below 3000 pounds-
differences of an inch or two in calculated feet per degree, then it may help to take
c.g. height. However, a discussion of deflection readings at various locations
compliances in the test rig, the tires, and along the chassis length to find any par-
chassis components, shows how the error ticularly weak sections. Then it may be
can be minimized. possible to reinforce the chassis with panel
Chassis deflection tests can ‘also be stiffeners or tubes and recheck the over-all
important for a race car, especially rigidity.
lightweight, open cockpit chassis. Beam Polar moment of inertia of the entire
strength is seldom a problem if the torsional vehicle is relatively unimportant, except
stiffness is adequate, since nothing is as in sophisticated vehicle dynamics simula-
critical in the longitudinal plane as front/rear tions. However, the polar moment of
torsion resistance and suspension alignment. rotating driveline components should be
There are no accepted minimum standards, measured and added to the effective ve-
but 3000 pounds-feet per degree of twist is hicle mass in acceleration, braking, and
acceptable for most lightweight road-racing coastdown calculations (see p. 126). The
cars, with a figure many times that being simplest measurement method is to use
common for a sedan with an integral roll therotating oscillation pendulum and equa-
cage, or a Formula car with incredible tion shown in Fig. 47B, in which (t) is time
aerodynamic downforces to deal with. for one complete rotation cycle in sec- a ee
The test procedure requires some heavy onds, (w) is the weight of the rotating rotary oscillation
equipment, including anchors in a concrete object in pounds, and (L) is the length of
floor; hundreds of pounds of weights; a the cables in inches. For greater accuracy,
long, stiff beam arm to hang them from; and make the cables about six feet long, care- equivalent added weight = (9.8)(t?)(w)/(L)
solid links to replace all four suspension fully locate the lower pivot points, use a
springs. It is most informative to measure stopwatch to time at least 20 complete
the deflection with dial gauges in a number cycles and divide by 20, and include the Figure 47B. Measuring Polar Moment

132
 14
THE PITS
Even if a race car is perfectly set up and polish. Polish will never make the car faster right time. This includes everything from
the driver is fast enough to put it on the pole, but it is a great psychological advantage over a packing list, to schedules and reservations,
the race can still be lost in the pits. In fact, competitors when they realize that to equipment in the pits. The assembly and
a great many races are won by slower cars everything else has already been done to pre-race check lists can be started by using
because their pit crews were quicker or they your car — at least as long as your car is a commercial race car inspection form, and
used better strategy, or simply because they fast and doesn’t break. additions can be made by going through this
prepared and maintained the car better. book and including more details. The
There are a number of important considera- RECORDS logistics list is a combination of all
tions beyond the engineering and develop- Check lists are highly underrated and necessary procedures and all necessary
ment of the machine. Ideally, the entire underused, considering the complexity of a pieces of hardware that might be needed.
team, including car, driver, mechanics, pit | race car. And they shouldn't be regarded as The best way to reach a compromise be-
crew, timer/scorer, signalman, and equip- an insult to the intelligence of a mechanic, tween taking everything and leaving out that
ment will all function perfectly as one because no one can remember every little last important item is with careful hindsight
well-oiled computer. critical detail. When more than one person and a reasonable amount of forethought.
The attitudes of the individuals who make is involved, each will assume the other has Experience will indicate what parts are most
up the team and the way they relate to each done everything. There are even many dif- likely to fail or wear out and the tools
other is a primary factor. The team ferent types of check lists. There can be a necessary to do the job. The time will come
members seldom get the recognition that a detailed assembly check list for separate when an experienced mechanic will tire of
driver, owner, or sponsor will, so they have areas on the car such as engine installation, the check list and graduate from it. And the
to be rewarded in other ways, whether by suspension assembly, or brake rebuilding. time will invariably follow when some tiny
money, association, travel, or simply love Then there is the pre-race check list, to detail will be ignored or forgotten and it will
of the action. In any case, it takes a tremen- make sure every last-minute preparation cost the race.
dously wise team manager to hold them all detail has been checked, such as all fluid There are other necessary records besides
together. Supervision and management skills levels, tire pressures, fasteners, and ad- those kept on the engineering, testing, and
may even be more important than ultimate justments. And finally there is the logistics development of a race car. Once the car is
technical knowledge, and it couldn’t hurt check list, to make sure everything and at the track, it is essential to know just what
to study ordinary business management everyone is always in the right place at the kind of mileages can be expected from all
techniques. Personal factors such as com-
mitment, responsibility, teamwork, and
personal pride are as valuable as knowledge
or experience. When all those qualities are
found in a mechanic or manager, he ought
to be recognized and rewarded, because
such people are hard to come by at any
price.
The team manager is also responsible for
a great deal of planning and paperwork —
the dull, but critical clerical work. This
includes records, entry forms, do-lists,
check lists, schedules, reservations, pur-
chasing, and bird-dogging late or lost items.
There are no end of projects to make a race
car faster or more reliable, but some sort
of priority must be established. At short in-
tervals — from weeks to days — the do-list
should be revised to suit the remaining time
schedule. It may even be necessary to break
all jobs down into three categories: Must do,
Important, and Also. The first category
would include work on safety and durabil-
ity, plus any necessary repairs. Second
would be any changes that might make the A modern computer-intensive pit booth may have monitors dedicated to timing and
car faster or more comfortable. And last scoring, fueling calculations, pit stop video, and data download or telemetry.
would be cosmetic jobs such as paint and

133
 Number Driver
11/zis |¢|s /¢ /7 |? 19 Vj2i3 is |e |7 3) ¢€ ‘4 ‘iz + Gl\7 '
(|$|Panweces SUS US |ISNWS (elt@|te\"G@| ic] | (| (el
16|1 [NO] IG] [6] 16]1G |1G} (Cle) 4g CATA
) OLLNE 1G (SI IGIG t HTEI2 48198 4848143 |€2148148 48 48 48/48/48 1 |1 [1GIG] 6 1G|G|G/6/
G|G|G| cic
4/8 |\SuRnev €2/424242/62 |42 42 |42142 [4242 [42 TiS isis C| ce] C1 © [166 |6] 15] 8 eShis las |ssl sltS) isles
6|DeNonve 4 GG GG G/C|G| | G| Gl G| G| Gl isia2 (| 1 | 6| G] GI GI (ol @ 15/43 48'4 24, elas
2| SAVAGE (AS 4 Cha} eustists lististis [ass |is| C(O) 1 | 1} ele Wististisuish Cleleteyulesaiel
cde
@|REvson zlzlq ieee ovireeri «tse le felt |tl unit |i ele |r| re leecher
ie) 2| 2 larlerleriaziqziazleriazazed <2] <:
I7\Pese 9 \77 Z\2 [FNS |WG7e rig elses WE] IG|7G/ 47/4 2Zaz 2] faze Zizlz|zlezize ZiZ2/2lz/ zI] z
z\|Euto Irvina|2 |0 |0 [Me2 eeios2) zi z| zi 212 |Z| feslas|ayjus|~? |oZ |e? or |4z|4z| 76/76 Wo!
TS 7G 7G) 76/76)7: 7O| 7%
I|\LESCcIE GBIG|PGIS [1S |S | V| Cs CP CejcPlaFie GP] 62|19 72 [79 29|72|79| TI. 7PY 26/7] REE TT
|78/79 79 AAA:
MINTER 110 | 6YGUCR 79/79) 79 77|77|72 TZIILVA7I179.| Z| 2% 1212/2 |e fe jz 7179/79 179 GPicziC q 77| 79 |79
e|ADAHOwICce FG]7G) 79| TY 79| CF [CF [OY 2-9] G2! C9) 67/29 631 EF 6919 |F194 NSE SUIS PSUS ESS | TV |F1F 69168 169) 6918
HesGas VW6UC1 lgaleaialealealee BUFF I) BAL A 199199199199) 9 169 169161169 (6 $6 ]4 7940) C1 |<
Woop DTH IG! & | | | 1S | He] | | +199) 77] V7) HO) 4667169 169/46) 4 H4O14 4G 7017268 RGR ESIC ETICI AGs cies +
9la|Van Vack |79¢9/ti¢e TLR A] A) FAT 199179 SPOT 14C]4 (eld 72] 7 217PITE|7E OF CF TEAGU Gt |GUGt Uce| 7] 75] 78
Ss & | & 78) 78178 78) FA|77) 17) 77717 17) 4C] 4G 40 |77) 7717 721721G1 (IF 19.9| 92 IPT, Gl /7P 79) 78 |cz| G2 <2
Sit Frecvs S |%z TVITTAA| A 79/7740] <0 40/40 /4e)78 |7¢| 72) 77172 |72161 | IAM ITGCAGCUGI |G Gl TWHTS |GUGU7, Gt C2] 2] 19] 18
son 82 4179 BUG UIT
TT Vi78 179/78 |TP ITELC ENCE NG | [Gl A" CDG2| 62 [| 75|75] 7S]? @ |7S|TSS) F2| sz\ez| 1¢| 221 79
(LL BYURG FB|F26Z/B27 7/62/60 60/97] Gor |ot |} 97127/77 77191197] 6.1275 |75 |75|62|@2|62|O26 z |82| F282) 08 1s |0s |771 977 2
UP| SCALLON F978 [08]F V7/%1/APc 1197/4971 97/9 G2] 62/62/6262 |orK18175|5 (SI7| 22 |22|S2S2|S2jS 2S [250 |2S eS)z5| zs|es|eo
6|t| BAKER F/V7GU7S| OU)19 Gt S262 crjaricz|cruerzie| 19/18 |/F |12 [18 7 |s0s)
2254 DzseS 2SeS]z 1g |relestee 77127|7P|
SO| St
UVimEcvirE GZ} 0BiG0) F [BUNS 0S |eF)03] -¢| eg 08] 1g |75 |75 || 75] 7S |7S |selzs|P2| 35/22/75] 19 108] 08 [8 08 9? 89 |S0} so} so| so]34/74
Vl) westr rasasiasi7 mleseeapp hie TE |3S|3S|TS(RS |25
|25/34 12519 7(34)
1 7|$0|$O|S0|S0|S50|
SO| SO|SO| SO $4) 34| 34| 75| 75175
Z|S|HAMLTON s|7s|75135|7 VS|IPS|
7S 7S]35] <5/35] s5|3S]zs |25] 25 |ZzS] $O|S0|34|22 197/24 |z5 [SCL
34 Se 34/34/3434) 75/75 |75|24| 24/24
S| FoRDHAM G1 6rl4 (3S s\3s\3sissigz €2zzs|zs\z so] sdso|saza 34/2 2/35/2475]19 2¢\2al2 1G. 24|z7¢ z¢|z4|z¢|22 zz|zz
(iHAVES 3 35/s0l2 S\US 25 |2S|2S|z5| 8z|82/2134 |34}74/34) €2| F213 19) Slie Solgar Qaezlzzj2z22/2z| zz| zz zz|2z| 22/44/4444

Actual lap chart data from an early Trans-Am race. Different colored pencils are used every ten laps (of the lead car) to he
keep order, and circled numbers indicate pit stops, or car retirements in a shorter race.

consumables. Total fuel and oil mileage how long it takes to close up on the paced tions such as spectator’s questions, rain, pit
should be obvious. But it is also necessary pack, how much time is required in the pits, stops, calls of nature, and cars blowing up
to know at what exact point the car’s per- and how many positions will be lost to those must be completely ignored. It will prob-
formance suffers from fuel or oil starvation, who don’t come in. This is a case where ably also be necessary to be in place and
because it is quite possible that a poor fuel even the best records will only provide a working whenever the cars are on the track
pickup will cause the engine to lean out with better than average liklihood of making the — even during practice — to make checks
a few gallons left in the tanks. It helps to right decision. on competitors’ cars, and to become familiar
carry the absolute minimum weight in with all the car numbers and colors. The
fluids, but it may be more valuable to have TIMING AND SCORING pit position of these people is particularly
a safe margin to allow for changing condi- Timing and scoring are the most difficult important. They should have a clear view
tions. In long-distance races, tire wear and kinds of records to keep during a race. of a reference point on the track (and some
brake pad wear must also be known with Simple lap timing can be done by almost space ahead to anticipate in) that will not
some accuracy so that pit stops'ican be anyone, but the added complications of be obstructed by spectators or other cars.
precisely scheduled. Careful pre-planning counting laps, taking intervals, and scoring If a timer can see far enough ahead, it may
in scheduling replacements can save a lot all cars on a lap chart makes it a challenge be possible to present the driver with infor-
of time. Brake pads should be changed at for two or three experienced people. The mation on his current lap — instead of on
the same time as tires — and the car should subject is too involved for the space allowed the previous lap. Timers should be easily
also be fueled if enough crew members are here, so a potential timer/scorer should accessible to the manager — and no one
allowed. But in some cases it is only prac- enroll in one of the training schools held else. The primary purpose of a timer/scorer
tical to have two wheels off the ground at regularly by sanctioning groups, and get is not to check the accuracy of the officials
once and since the outside tires wear faster, some practice with them at amateur races. but to provide continual running informa-
it may be quicker to change only one side But for those with a little experience already, tion for the manager and the driver.
per pit stop. Careful record-keeping will a few tips may help. (For the future of The relative value of the different types
make it possible to keep the number of computerized timing and scoring, see of information to be gathered varies com-
pitstops and the time lost to a minimum. Chapter 15.) pletely with the nature of each race. For the
With the increased use of the pace car First, a timer/scorer must have a very average short amateur race, all the driver
during a race (for safety reasons or perhaps strong and stable personality, to face the really needs is his time interval from the
sometimes just to spice things up a little), inherent pressures and to maintain the next car ahead or behind, and perhaps the
that becomes an important strategy con- necessary concentration. There will be 20 laps remaining. It may not even be
sideration for pitstops. To avoid losing a lap to 60 cars going by — multiplied by perhaps necessary to keep a lap chart to identify
by being passed by the pace car and pack hundreds of laps — and all a scorer has to which cars are immediately competitive and
while in the pits, it is necessary to know do is miss one car to disorganize a lap chart which ones are some laps down. Over-all
such things as how long a paced lap takes, completely for the rest of the race. Distrac- consecutive lap times are insignificant

134
 =
to be most valuable to the driver. The driver
12/3 /¢ 6/7 Qe ({2/3)/4/S16/ 7/8/19 ( + 7
may also arrange certain signals to the pits,
(6\ (G6) (GK(E) (G| (E| 1G] (6 th
to request particular data such as intervals
‘S|& |@js7]42]47 |4ele2|4y i IS |1S tS [1S [0S SSIS| (GHICUGHNGNGGVG NIG
or remaining laps to a pit stop.
+7]& G|G@ IS ISS @|\G|G 6(6|G|G|C|G\e |o|G |e Some sort of communication system with
Mite] elt TC histis
|Glala@ ACCOR tClelele ji dt WMiazlazlaziaziezie the driver is always needed. Two-way radios
CL FE PT eT Phe PE CT t (148 43148 48 144 48 FP
are becoming less expensive and more
2/47/42 AZ 4242/42 2 42} 427(42 Z 12 12 12 [Z
popular, although they can complicate
76¢|7G Zil2z2Ziz\e 2lBlz |zjz Z |G8|E8|C8 |e
matters and are not yet foolproof. Inter-
Ctl eat ea |e 76,76 CEI GS OS |6% CP IGP IGP
ference, background noise, batteries,
CEIOUIO? IG: GCL LF |7F [TA
antennas, overlapping channels, and
£9159 116' 7% [TE 77 [GI IGl
unwelcome monitoring may have to be dealt
4614, (4 CIGUGITE
with. However, they do avoid the problems
*e|72 78 (79|7¢kez of hand signals and allow much more in-
G(|G\ 2|G2|62| 8z!
formation to be transmitted. Otherwise a
129 |79 BeUse 82 &2|§2 2 RQ—+—
driver is limited to a small number of pre-
TTI
arranged signals for anticipated situations.
2s|zs |2S]
He may have signals to request certain in-
7s er formation or to warn the crew that he needs
iva
fuel, oil, water, a tire, brake pads, or even
a driver change. But if it is something else,
such as a vibration or a miss, or he needs
a door taped shut or a wing adjusted, it will
take some explaining. If the crew isn’t
prepared when the car comes in, it may save
time for the driver to go back out for another
lap while they get the parts and tools set up.
Signaling the driver from the pits is more
unless the driver is trying to improve his a race, but at best the information will prob- common in amateur racing. The methods
consistency or finds it necessary to main- ably be a few laps old, even if electronic can range from a simple blackboard, to
tain a certain predetermined average speed. computer scored systems are used. If there commercial number flip-boards, to metal
In a long-distance race, perfect strategy are enough stopwatches to go around, it may boards with magnetic letters and numbers.
would determine the ideal average speed to be best for each member of the team to have Regardless of the board, there are certain
have the car just slightly in the lead at the a specific timing assignment and perhaps sign conventions that are ordinarily used:
finish without overstressing it to the point even keep track of laps on a couple of the P-3 indicates third position; L-29 indicates
of failure. Total number of laps completed nearest competitors. Before the race, the either 29 laps down or 29 to go (the driver
may be extremely important — particularly driver and manager should come to an ought to know the difference) with a final
when fuel mileage is marginal — but that agreement as to what information is going IN sign for the last lap before a pit stop; and
is simple enough so that more than one
member of the crew could keep a running
tally. When races are run on total length
(versus overall time), it is important to
distinguish race laps remaining to the driver
so he can plan his Lee accordingly. vot

scheduled fuel stops, ‘itwill be necessary to

keep a lap chart. Otherwise, no one will be
quite sure just who is immediately ahead or
behind, much less what position the car is
in. A driver may find it possible to pick up
a position or two with a little extra effort,
if he knows where he stands. Conversely,
he could risk destroying the car trying to i
pass another car that was actually a lap
down. Keeping accurate lap charts, however,
requires careful training and a lot of ex-
perience. This is especially true where
mixed classes and pit stops are involved, and
, on short, high-speed tracks where thirty-
three cars may be coming around once every
minute. These, and endurance races which
run from 4 to 24 hours, are a real challenge
even for professionals. It may be possible Signalers at the pit rail, with various types of pit boards. Some of them
to get car positions from the officials during are also doing their own lap and interval timing.

135
 PIT STOPS
There are two kinds of pit stops to be
prepared for — the ordinary scheduled pit
stop and the unexpected occurrence. The
first kind of stop requires a great deal of
planning, special hardware, and practice.
The hardware ordinarily consists of every
conceivable device that will speed the
addition of fluids or the changing of com-
ponents. Quick-fill fueling systems are an
art in themselves, but since overhead tanks
have been generally replaced by eleven-
gallon cans with dry-break couplings,
fueling has become much more standardized
— and safer. But the internal design of cans,
tanks, and spouts can often be revised to
pick up a few seconds. The air vent or
overflow tube may also be reworked to get
the air out faster and insure complete filling.
Oil and water, when necessary, can be
added by using dry-break connectors in the
engine compartment. Because there is less
danger of fire, a predetermined amount can
be injected very rapidly from pressurized
containers.
With the method of tire changing being
generally open under most rules, there have
been a lot of clever developments to speed
up the process. Knock-off hubs are not
allowed in most cases, so it is a matter of
getting the lug nuts off and on again in the
absolute minimum time. Ordinary
pneumatic impact wrenches are used, but
they are souped-up, or modified by drilling
out the air passages, and the pressure
regulators are boosted way up. Even the
sockets are custom-ground to slip on more
easily. All of the wheel studs and lug nuts
are modified to guarantee that they will go
on right and fast with no chance of cross-
threading. The threads are usually machined
Os off the first 3/4-inch of the stud to guide the
nut on straight. And all nuts are carefully
NASCAR tricks of gluing greased lugs to the wheel with trim cement (top), and lubricated before the race.
lightening pit jacks with aluminum components and skid plates (bottom). Special lightweight hydraulic jacks are
available, and some crews even add more
exotic materials to lighten them further. Jack
wheels are seldom necessary, and they have
+3.5 or —3.5 indicates that the driver is tinguish individuals at the pit wall. The a tendency to sink into hot asphalt, so a skid
either 3.5 seconds ahead of or behind the board should be well marked and its visi- plate is usually substituted. The jacking
next nearest competitor. Laptimes can bility checked at changing sun angles. The point on the car is well marked, as is the
usually be given in two-digit numbers, such one person who is allowed out to the pit wall proper inserting distance on the jack. The
as 6.5 to indicate a lap of one minute and for signaling will be in the best position to jack is pre-set to about the right clearance,
56.5 seconds, since the numbers shouldn’t watch for any obvious problems on the and the jacker should know just how many
vary by more than a few seconds. An addi- outside of the car — a loose panel, body pumps are necessary. It is also a good idea
tional blackboard offers the chance to make damage, leaks, or whatever. It might be a to try getting the jack under the car with
other odd comments to the driver, such as good idea to have another person on the one or both tires flat on that side. The old
“faster” or “conserve fuel.’ But they had other side of the track to watch the other “over-center” type jack once common in
better be simple and obvious to avoid con- side of the car. After a little shoving out on Indy Car pits has been replaced by onboard
fusing or worrying the driver. Finally, the the track, a driver may be very interested air jacks. It seems that elimination of the
signal board will have to be well located and in knowing whether a bent fender seems to jacker, and the time he took, has more than
easily identifiable, since the driver is very be rubbing on a tire. made up for the slight increase in weight.
busy out there and it can be hard to dis- The wheel can be mounted much faster

136
if there is a tapered centering cone on the
hub and the stud holes in the wheel are
chamfered slightly. There are a number of
ingenious methods for getting all five nuts
started rapidly. The most popular is to have
new nuts pre-attached to their proper
bearing surface on the wheel, by gluing
them in place with trim cement. The first
few threads of the studs are removed to
provide a guide for the nuts. They slip on
the studs as the wheel slips on the hub, and
the impact wrench will still turn them up
tight. The old hot nuts simply lie where
they fall. They are generally expendable
and not re-used anyhow, since they are
softer than the studs, to save the stud ifthey
get cross-threaded. And it's a good idea for
the changer to carry an extra nut.
Brake pad changes are least frequent and
most difficult. Obviously they should be
scheduled for the same time as a tire change
when the car is up and the wheels are off
already. Then the replacement of pads Crew prepared for a tire change pit stop, with tires, jack, and pneumatic wrench
shouldn’t take more than another minute at ready. Lightweight tools, radios and protective wear make stops faster and safer.
the worst. The greatest obstacle is that the
pads and disc will probably be literally red For metal-bodied cars, the most com- in a few minutes. If the old glass is merely
hot, so the old ones will have to be removed mon problems are fenders collapsed onto cracked and not shattered, it might even be
with pliers or asbestos gloves, and the new the tire and broken windshields. Heavy left under the new windshield. Other than
ones must be installed with great care. Re- sledges and long, forked pry bars are com- external problems, about all a crew can do
tainer pins must be fail-safe and yet easily mon in NASCAR pits to bend the metal is to have as many tools and spare parts as
removed and reinstalled. The problem of away from the tire. A pneumatic chisel possible — at least suspension, brake, and
squeezing the caliper pistons back to make might even be used to snip the wheel well engine components — in the back of the pit.
room for thicker linings may be handled open. The problem of windshield replace- Still, there may come a point where the time
with a vacuum retractor system, as de- ment is eased by merely resting the glass in required to get the car going again will put
scribed in Chapter 5. Another method, arubber-lined frame, and holding itin with it too far back to be worth the panic and
which is cheaper but slower, is to build a the required safety clips. It ought to be effort, unless a few points may still be
screw-operated spreader device to push the possible to get the old windshield out and a gained for the season championship.
pistons back in. It is worth another men- new one in within acaution period, or with- Since the pit crew is what makes or
tion here that all new pads should be pre-
burnished, or broken-in before race time.
That fairly well covers the scheduled pit
stop, but there must also be some anticipa-
tion of and preparation for the emergency
stop. Experience and pessimism are the best
helps here. If what was going to fail were
known, it should be improved or made fail-
safe instead of making it rapidly repairable.
The only exception is collision damage that
prevents the car from continuing or slows
it down for a greater net time loss than a
pit stop would cost. Ordinarily, about all that
can be repaired fast enough to keep the car
competitive is bodywork. Racing cars usu-
ally have lightweight fiberglass body com-
ponents that can be replaced rapidly with
quick-release fasteners — providing there
are spare panels immediately accessible.
Otherwise, a lot of silver racer’s tape,
bungee cords, and wire may be used to hold
the remaining fiberglass together. Some
thought should be put to the changed
aerodynamic characteristics also, as the loss
A NASCAR fuel stop, where in addition to the fueler, a man is dedicated to
of certain surfaces may make the car
dangerously unstable.
catching any overflow. Note the mandatory use of fire-resistant uniforms. (Chevrolet)

137
 watching, plotting, and practicing.
Safety is still a problem in the pits, even
with safer fueling rigs. It is wise for fuelers
to wear flameproof coveralls or old driver’s
suits, even if it looks kind of peculiar.
Goggles and fireproof hoods ought to be re-
quired protection. And the crew has to
recognize the fact that another driver in the
pits may care more about a fast pit stop than
their personal safety. The pits are just about
the easiest place there is to get hit by a
speeding car. The faster a pit stop is, the
less exposure time there is for the crew.
If the driver has first choice of pits, there
are some very rational selection criteria.
Usually, it is best to have the pit closest to
the slower end of the track. If there is a slow
turn just before the entrance to the pits, that
end is the best location, because the car is
already going slowly and that gives the
The beginning of a fast fuel and tire stop. Two men fuel, two remove length of the pit straight in which to ac-
and replace the tires, one jacks the car, and one carries the spares. celerate again. Other less significant con-
siderations are: to be close to facilities, to
be close to the official timers, or to have the
breaks a fast pit stop, their attitude and are far greater than any possible gains in high ground if it rains. In practice, the driver
skill is more important than the hardware split seconds. And yet, it can make a should also practice full-blast pit stops to
involved. First, it is critical that everyone difference when races are won or lost by a learn just where the last possible shutoff or
know the rules of the game. If too many car length. About all that can be expected braking point is. If he overshoots his pit, he
people are involved, if they move too soon, of the crew during an emergency stop is that will probably have to go around the track
or if they do something they shouldn‘t, the they be fast thinkers and even faster workers. again — whether he has enough fuel or not.
car could be penalized. Second, they must It is worth a note here that any time the car If all goes according to the script, the ac-
know whatis happening at all times and be comes in unexpectedly, time should be taken tual racing pit stop itself will be anticlimatic.
prepared for immediate action. They may to fuel it up unless it has enough to finish The car will be in and out in a matter of
get no more than about 10 seconds’ notice the race. That precaution could save another seconds with little apparent concern or
before the car comes toa stop in the pits, or fuel stop later. drama. If things don’t go so well, the longer
they may have to take the car into the The actual steps and movements in a car sits, the sicker it gets. The engine may
garage area for major work--if that is legal. fueling or tire changing ought to be planned start to overheat if it has to idle much longer
But most important, the crew must have a and choreographed as well as any ballet. than an ordinary tire change, and if it is shut
lot of practice at doing the jobs they know The crew members themselves can probably off, there is the possibility it will heat lock
they will have to do: fueling and tire come up with the optimum arrangement and not restart. It may be possible to spray
changes. There is no such thing as too after a few rehearsals. The overflow vent the radiator with water, but it makes the pit
much practice--just as a driver can’t prac- man might steady the first fuel can while a mess for fast starts. Also, the brakes will
tice his job too much. But as a crew the fueler returns for the second can. The be blazing hot from the last stop, and they
member does his job, he ought to have his jacker might clean the windshield while the can boil the fluid or warp and crack the
priorities in order. First and foremost, the tires are being changed, or pass a new tire discs if the wheels aren’t kept rotating. And
job has to be done right. There usually at the precise moment the old one comes finally, when the pit stop is over, one
isn’t any second chance--a wheel may fall off. The instant the jack drops might be the assigned crew member must act as a traffic
off, or the pits may go up in flames. Doing signal for the driver to drop the clutch. cop and keep the driver from pulling out
it quickly is only the second priority. The There are a thousand possible arrangements in front of another pitting car. Otherwise,
possibilities for lost time due to a mistake and tricks available to the crew that keeps the race may end right there in the pit.

138
 15
COMPUTERS IN RACING
Computers are the most valuable tool sion data, braking predictions, and driver 360, with a maximum capacity of two
in racing these days, from initial design to performance models. As larger computers megabytes. To fully simulate one minute of
analysing results. Mention that you've became available, these programs were all driving took two hours of computer time, at
used one in the development of a new race pieced together into two gigantic and al- over $500 per hour. Today the same pro-
car, and you're assumed to be a potential most unworkable versions. gram could be run comfortably at home on
winner. You will intimidate everyone but One, called DK4 (simply a sequential a PC. Butit still takes just as much time to
those who used computers before you, code) was an incredibly detailed simula- gather the necessary information -- even
especially those who have found they are tion of every moving part in a vehicle though much of it turns out to have minor
not a substitute for experience and com- chassis, from shock absorber bushings to effects on the total analysis.
mon sense. Computers definitely have an tires. In fact, its size was its greatest Though I occasionally used DK4 to
important place in racing -- although they _ handicap. It could take weeks to gather or study generic behavior (What effect does a
are sometimes used in the wrong place. approximate over a hundred real vehicle locked differential have on understeer?)
Always keep in mind that: a smart man measurements necessary to even get it my main interest was not the behavior of
knows about high tech; an educated man started. (See Fig. 48) Tires had to be tested single components on simple maneuvers,
knows how to use high tech; but a wise for slip angle, camber angle, and thrust but optimizing laptimes around a race
man knows when not to use high tech -- effects. Characteristics of shock absorbers, course. Over time we had combined straight
when it may not be time/cost effective. springs, rubber suspension pivots, rotating line acceleration models with straight line
Believe it or not, the use of computers inertias, and wheel geometries had to be braking models, and with actual measured
in racing does have a history, actually known. Then there were whole-body pa- cornering speeds (like the Indy Ford model).
going back to the early sixties. That was rameters such as aerodynamics, center of But constant-radius turn calculations were
the period when Ford was actively going gravity locations, polar moments, and too simplistic for real world driver behavior.
after Indy and LeMans, and Chevrolet was stiffnesses. It was discovered that in many Computer simulations stagnated there
quietly but strongly supporting Chaparral cases, it was almost as fast to build and test, a few years for a lack of knowledge of what
and Penske in road racing. The factories as to estimate and simulate -- and the drivers did in transition stages. The theo-
had plenty of young engineers and the results were often more convincing. reticians wanted a driver with "eyes", a
latest computers to apply to a field which Another problem in that era was the "feel" for lateral acceleration, "foresight",
had been traditionally empirical. (In rac- limitations of computers. These simula- and "learning", to steer around acorner on
ing, lots of experience and a little science tions were created on G.M.'s state-of-the- the optimum line at the maximum speed.
usually beats lots of science and a little art computer, which in 1968 was an IBM/ But as the only experienced race driver/
experience.)
Ford published their technology first, a
computer simulation for predicting the
performance of their V-8 in the first mid-
engine chassis at Indy in 1963. Although
it was a crude model, using driver-re-
ported turn speeds and estimated vehicle
drag figures, the acceleration figures were
based on real dyno runs, and were quite
encouraging -- and accurate. When Ford
went to LeMans, there was a major engi-
neering effort in computer simulations of
engines, brakes, and gearboxes, which
were described in a series of SAE papers
published in 1968.
But at roughly the same time, Chevrolet
was working with Cornell Aeronautical
Laboratories on much more ambitious gen-
eral-purpose vehicle simulations for pro-
duction chassis development. And this
work could also be transferred to race cars.
These programs started out as individual
studies in suspension analysis, accelera-
tion prediction from engine and transmis-

139
(1) INPUT OPTIONS Method of measurement (+) for asymmetry No's Tables
Masses Sprung weight scales 1
Unsprung weights -- F/R scales 2
Fuel weight and location scales 4
Polar moment of inertia in yaw, pitch, roll swing the entire vehicle from a torsional pendulum 3
Chassis Dimensions -- wheelbase/tread tape measure 2
Sprung center of gravity location -- 3 dims scales, and a means of hoisting one end of the car 3
Unsprung cg locations -- F/R tape measure 6
Static ride height log book D2
Chassis torsional rigidity support on 3 wheels, measure to reference plane 1
Suspension net characteristics: (or derive from links coordinates)
Roll center -- F/R -- and movement curve tape measure or plot from drawings 2 2
Camber change curves -- F/R dial indicator readings through vertical travel 2
Toe change curves -- F/R bump steer rig 2
Anti-squat -- Anti-dive tape measure and plot from drawings e)
Ackerman geometry tape measure and plot from drawings 1
Static camber, toe, caster -- F/R -- [L/R] log 7+4
Camber compliance -- F/R lateral load stress and dial indicator 2
Steer/toe compliance lateral load stress and dial indicator 1
Spring/shock Spring rate or curve -- F/R -- [L/R] manufacturer data or spring tester 2 +2 2
Bump rubbers F/R manufacturer data or spring tester Z 2
Variable bellcrank geometry F/R measure and plot 2 2
Shock curves -- F/R manufacturer data or shock dyno 2
Third spring/damper F/R manufacturer data or shock dyno 2
Anti-roll bar F/R log, calculations 2
Diagonal comer load bias log, from scales 1
Tires Diameter as a function of pressure manufacturer data, or test 1
Basic traction ellipse ratio manufacturer data 1
Basic Cf skidpad tests 1
Lateral force = f(slip angle, load, camber) manufacturer data 343
Lateral force = f(drive/brake, temperature) manufacturer data 2-2
Spring rate, damping as a function of pressure manufacturer data 2
Driveline Engine torque curve, ambient effects dyno 1
Engine braking drag dyno, or free spin coastdown : 1
Transmission ratios, shift time calculations, data acquisition system 6
Differential bias as a function of torque manufacturer data, or dyno 2 1
Trans and differential efficiency dyno 2
Driveline, wheel, tire, and brake inertias polar moment pendulum 4
Brake bias and curve calculate from brake hardware, valve manufacturer 1 ]
Aero map Drag as a function of height wind tunnel or DAS ]
Downforce as a function of height -- F/R wind tunnel or DAS a)
Ambient effects weather station 1
Lateral data [for stability studies] wind tunnel 1
Track map Derive radii and straights from data, or derive from DAS 1
Survey X,Y coordinates track owner, survey, or GPS 1
Widths, camber, grades, Cf, every 50-100 feet |DAS, or walk and measure (every 10 feet if traction limited) 3

(2) PROGRAM OR VEHICLE CONTROL POSSIBILITIES: Control function:

Evaluate: (a) Parameter tradeoffs (used in initial design or trackside development) Optimum track lap software
(b) Steady-state stability (for finding basic setup parameters) Constant lateral g simulation (turn circle)
(c) Response to controls (check the effect of trackside balance changes) Transient response software
(d) Reaction to disturbances (like bumps, and crosswind stability) Also needs shock and lateral aero data

(3) OUTPUT POSSIBILITIES:

(a) Vehicle performance plots, identical to plots from familiar data acquisition software, or
(b) Tables of comparisons, with optimum chassis setup numbers identified by making repeated iterations.

Figure 48. HOW TO EVALUATE RACE CAR COMPUTER SIMULATIONS. To get accurate results, these are the most
relevant factors they should have, or variables you might want to be able to evaluate. The sheer volume of input data
required looks formidable. However, compared to the costs of running repeated track tests, it can make simulations very cost
effective. If A-B comparisons are adequate, rather than absolute accuracy, then estimates or default figures can be good
enough. And once the basic parameters have been established for the first runs, then subsequent runs are much easier.

140
 3-D FORMULA 1 SPEEDS

Figure 49. Computer-predicted speeds of a Formula One car simulated at Watkins Glen in the 70's, overlaid on a 3-dimensional
course map, produced from typical vehicle and accurate track data by Calspan engineers. (Ca/span)

engineers in Chevrolet R&D, Don Cox PERFORMANCE SIMULATIONS track? Or do you want to design an original
and I sat down and brainstormed out the vehicle on paper and predict how it will
simple truth. A racer doesn't drive from a Race car computer simulations may compare to an existing design? And how
straight, through a iurn, to a straight. sound like fun, especially to new computer accurate do you want your results to be? Is
Instead, he drives from an apex, througha hobbiests, but are they worth the consider- an A-B comparison good enough or do you
straight, toanapex. Coming out ofan apex able investment in terminal time? Useful need close correlation with the real ve-
is relatively easy -- it's merely like balanc- results depend on the amount of detail and hicle? Only after resolving these questions
ing onatightrope. But going into the apex track validation, regardless of user expertise. will it be possible to decide which of a
is a little worse -- it's like leaping onto the A good starting point may be the U.S. hundred factors can be safely ignored.
tightrope blindfolded. On a computer, DOT. As part of their research in wutomo- One of the drawbacks to using a com-
however, time can be run in reverse, so it tive dynamics, they have sponsored their plex computer program is that you seldom
was possible to work from both apexes own computer simulations based on roughly learn ‘“‘why’’ it came up with a particular
toward the middle, selecting the absolute the same original concepts. Currently, “‘what’’. Alright, so two-wheel drive gives
optimum braking point. Based on this they are being simplified by Systems Tech- better laptimes at Mid Ohio than four-
theory, and the other known tire and ve- nology to run on PC’s (See SAE paper wheel drive. Why does it? A human may
hicle conditions, it was relatively straight- #2000-01-1620). These particular pro- not be able to comprehend the underlying
forward to complete the driver control grams are intended for passenger car work. logic well enough to ever know. The
simulation, although the mathematical Although the basic framework can be used intermediate steps and sub-conclusions may
equations took months to formulate. to produce realistic race car simulations, it be buried in a stack of printout sheets and
Eventually, others discovered the tech- is necessary to get typical racecar parameters plots. Sometimes it is more effective to use
nique and assembled their own versions. and formulate typical racing maneuvers. a simpler, more comprehensible analysis.
Acouple ofengineers at Calspan (Cornell) Because simulations are an interesting The following considerations will define
produced a version whichconcentrated on challenge, and there are already many on the type of simulation necessary to get
the nearby Watkins Glen track (See Fig. the market, the first question is how to significant results:
49), and it was also used by Goodyear to select the most appropriate one. Suspension simulations are those in
predict tire performance at new tracks. The first step is to define precisely which only the mechanical, geometric de-
Milliken Associates then inherited the pro- what you want to know -- what you want the flections of the hardware are considered.
gram, adapted it to modern computers, program to concentrate on. Do you want to At the first level, just one wheel location
and from DOS to Windows, as part of a anticipate the effect of various changes on may be modeled, including the wheel,
DAS package. And many others have your existing race car? Do you want to knuckleor upright, all locating links, spring,
since followed in these footsteps. study the optimum driving style at each shock, and anti-roll bar. The model ends at

141
 simulations, but are a quantum leap ahead,
in that tire models and overall inertias must
be incorporated. The ordinary rubber tire
is about the hardest thing there is on a car
to mathematically model. Good real-world
representation requires a tire to be tested in
a laboratory in all combinations of load,
camber, slip angle, cornering force, and
thrust -- while controlling temperature,
pressure, and wear. Then these data points
must be represented by tables or general
equations of relationships. After that, the
problem gets complicated. The vehicle
inertias -- vertical, longitudinal, lateral,
pitch, roll, and yaw (plus the oscillating
inertias of the unsprung suspension com-
ponents) act against the ground through
these four flexible tire contact patches.
Four points creates a “‘statically indetermi-
nate plane,’’ which usually requires
“‘iteration’’ or repeated approximations to
approach a true force balance.
@ . Finally, how is the model used when it
Figure 50. Data input format for a simple two-dimensional analysis (Mitchell) is finished? That is, what kind of maneu-
vers or inputs are used to make it act like a
real vehicle? Some simple vehicle test
C:\Program Files\Mitchell\Simulate\Demo_ Isw.dyn
maneuvers such as the “‘step steer’’ or
sinusoidal steer inputs can be used for
comparison with real tests, but they are not
representative of very many actual driver
behaviors, on the street or on the track.
And what kind of results are to be pre-
First Setup {2726
fe
sented? Is a single number adequate, or are
complete plots of paths and angles neces-
sary? ;
Race course simulations are some-
thing else. Here we are talking about only
the ultimate limits of performance (hope-
fully), representing a driver with very
well-known and repeatable behavior (on
the track). None of those sub-limit maneu-
ver niceties such as tire data maps and yaw
responses are necessary here. You define
the boundaries of the race course, select a
*‘line’’ just as a driver might (See Fig. 52),
RACE CAA
. 334 Ibs L__ SIMULATOR and then instruct the computer to travel on
that line as fast as it can, restrained only by
the theoretical upper limits of the engine
and tires.
But of course it’s not as easy as it
Figure 51. Aclever computer representation of vehicle dynamics by auto writer Bill
sounds. First a path must be defined in
Mitchell, including tire loads, forces, and camber, and spring and bar loads.
computer terms to go from a tight radius at
the chassis, which is considered a fixed, tances to determine true lateral chassis the apex to an infinite radius in the straight-
solidanchor. This program can give single motions. (See Fig. 51 ) The next level of away (constant arcs, spirals, parabolic, and
wheel angular changes with travel, and sophistication is to include engine and hyperbolic paths have beenused). Then the
force paths due to wheel loadings. (See Fig. brake torque inputs for their effect on lon- computer must be told how to balance
50 ) The next step is to include the oppo- gitudinal suspension motions, as in squat cornering versus acceleration to get the
site side, to see how one complete end of the and dive. And the highest level is a consid- best laptime, while considering the effects
vehicle might respond with respect to the eration of the actual compliance of links, of weight transfer in determining whether
ground -- in roll angle and vertical deflec- bushings, and pivots as they are stressed, the limit is the front tires, rear tires, or
tion. Then the other end of the chassis and and not simply as assumed infinitely rigid enginetorque. And in the real world of race
suspension can be added in the same way -- objects. tracks, there may be considerable changes
which will require a comparison of front/ Vehicle dynamics simulations gener- in grade, camber, and surface coefficient
rear weights, roll centers, and roll resis- ally require the above suspension from corner to corner. Another major and

142
unique race car consideration is aerody-
namics, which can change immensely
between 50 and 200 mph. And finally,
these days it can be very useful to be able to
vary the model between rear wheel, front
wheel, and four-wheel drive.
Depth of analysis. Regardless of the
type of simulation, you will have to study a
candidate program’s input variable cat-
egories, and the detail within each one.
Otherwisethe approximations and assump-
tions will affect results. For example, is
tire data represented by a simple friction
ellipse, or is it in tabular or coefficient
form, and does it go into all the affecting
variables of interest, such as temperature or
wear sensitivity? Do the required aerody- \ DIST
namic coefficients vary as a function of ride
height and ambient conditions? Do the
suspension models incorporate bushing
compliances? Are shock absorber charac- PATH OF CAR |

teristics required, perhaps varying as a

function of speed, travel, or even tempera-
ture? And does the program actually use
these parameters, responding with appro-
priate results? Also, the file sizes and the
precision of tabulated input such as engine
torque curves and tire data should be con- 1/2 CAR WIDTH
sidered.
Vehicle control input. There are three Figure 52. Mathematical/geometric representation of a racing line through a
main types: Most common isa vehicle path typical turn at Watkins Glen in the Calspan race car prediction model. (Calspan)
controlled by duplicating the output from
anactual track test session. This isa simple panies should warn you not to expect abso- quote examples of exactly how much they
way of studying the effect of configuration lute correlation in speeds and laptimes to alone have improved laptimes. Perhaps
changes and correlating them with testing. be too close, unless they reach the point the ultimate step would be to provide real-
More scientific is a path defined by fitting where every important parameter is accu- time inputs, such as computer joy-stick
curves to a surveyors track map (including rately represented. However, you should steering, throttle, and brakes, and observe
track widths). This allows the prediction of expect results to at least indicate the proper corresponding vehicle behavior represented
performance and setup change effects ona direction of results, ifnot also the approxi- as images generated on the screen. Very
track that has never been run, orhas changed mate relative value of results. Possibly a simple versions of this can be seen in any
since the last race. Its accuracy will depend demo version of the software would allow computer-game arcade. But like any other
on the inclusion of hard-to-define local you to input a change in which you know speed equipment, you pay your money and
surface variations such as camber, grade, the likely results from your own experi- you take your chances. And more and
roughness, and grip. And in a separate ence. Until you have personally verified more of the technically sophisticated pro
category is a path resulting from driver the predictive value by correlation on your racing teams are taking the chance.
control inputs, either duplicated from test own car, it can be dangerous to have too You can expect race car simulations to
data or standardized *‘open-loop’’ inputs much faith in it. become more commonly used in the future.
such as the step-steer or sine wave. These All of these race car computer simula- There will be more available to choose
runs are mostly used to study transient tions are extremely challenging and from, and at much lower prices, for the
stability as a response to a precise and therefore as fascinating as the real ve- following reasons: (a) Continual advance-
repeatable simulated driver. hicles. They provide the opportunity to ments in computer memory and speeds
Output presentation. Most common research minor and major changes without makes it easier to develop and run them.
is plotting results in the same format as a the limitations of actual test drivers who (b) There will be more computer program-
familiar data acquisition system - perhaps sometimes don’t repeat too well, may not mers available with an interest in racing.
a host DAS, as they become affiliated with be sensitive to the immediate question, and (c) The basic mechanical and dynamics
SIMs. Output should also be available as can get hurt if someone makes a mistake. modeling has already been done, and made
data files such as ASCII or comma delim- The simulation will repeat its maneuver publicly available by the US Department of
ited files that can be analyzed in other precisely to the nearest fraction of a foot, Transportation, and it has been well vali-
software such as spreadsheets. Output and determine whether there was a differ- dated. (d) There will be a larger market
from asimulated run mightalso include the ence of a millisecond between two from more knowledgeable users. (e) The
limiting conditions in every location, such configurations. necessary common input parameters, such
as engine power limit or oversteer/under- Such programs have given me as much as vehicle, track, shock, and tire data, will
steer. insight about the behavior of race cars as beamortized and become more easily avail-
Accuracy/correlation. Most SIM com- actually driving them. Granted, I can’t able from outside sources.

143
 optimize the strength-to-weight ratio so
that no area is either unnecessarily weak or
unnecessarily heavy. The process is inde-
pendent of the type of material or the type
of structure. It could be used on a balsa
monocoque or aspaghetti-tube frame. With
that introduction, let’s consider the actual
i NeneNigha procedure, using the parent program
IT SAAS ““NASTRAN”’ (NASA Structural Analy-
sis) aS an example:
First you have to define the necessary
shape, and certain locating points such as
the suspension and engine mounts. You
have to provide space for the driver and
Figure 53. A dramatic extreme example of computerized finite element analysis fuel and a hundred other bits. This is as far
as most people get in their consideration of
commonly done by auto manufacturers—in this case, a Mazda 323. (Mazda)
chassis design. But for computer analysis,
all of this information must be represented
COMPUTER AIDED DESIGN tines which can identify weak areas, by numbers, and not drawings. Every
reconfigure the design, and rerun the analy- point must be defined by three-dimen-
The use of computers in the production sis, the same as an engineer would. If sional co-ordinates, which in some cases
of race car hardware is commonplace airflow is involved, say it’s an intake mani- means locating thousands of three-digit
among the major designers. However, fold, CFD may be available to study differ- dimensions.
there may be some confusion in the proper ent flow path designs. If it’s a moving For each point, the computer needs to
terminology, between the acronyms of mechanism such as steering components, a know what material properties apply. In
CAD, CAM, CAE, CNC, CFD, and FEM. combination of parts may be joined and the case of a sheet metal monocoque, this
Computer Aided Design (CAD) implies then animated, to see what interferences is relatively simple, because the material
using aterminal to work out the geometries may exist. If high temperatures are in- will haveuniform thickness and symmetri-
and relationships of individual compo- volved, as in header pipes, a thermal analy- cal properties. With fiberglass or graphite
nents, as a draftsman would with pencil sis can show how that heat is transferred composites, however, both the thickness
and paper. CAM can mean either Com- and dissipated. Then, as the design is and the direction of maximum stiffness are
puter Aided Manufacturing (production finalized, cumulative tolerance checks can infinitely variable (as discussed in chapter
components), or Computer Aided Machin- be run automatically before having compo- 10). This is a structural advantage but an
ing, which takes these digitally-defined nents fabricated. The final digital model analytical nightmare.
dimensions and uses them to automatically can then be sent directly toa CNC machine
form the components. More recently, this such as an automatic mill or lathe -- or even The Input Process
is called CNC, for Computerized Numeri- the latest rapid prototyping methods of This is the tedious part. If manpower is
cal Control. But since few readers of this stereolithographic production, or the laser- cheap, you can sit down at any computer
book have such facilities, this section will generation of models layer-by-layer out of keyboard and type in those thousands of
concentrate on the more useful application fluid materials. numbers -- and just hope you didn’t make
of strength analysis, using Finite Element an error. The use of “‘interactive graph-
Modeling or Analysis (FEM or FEA), and The Theory of FEA ics’ however, makes the job faster and
Computerized Flow Dynamics (CFD) Finite element analysis is simply a math- more fun -- and less expensive in computer
which is a developing substitute for the ematical technique of doing strength analy- time. Using a video terminal, you can
wind tunnel or flow bench. Actually, all sis with a computer. Before computers, it begin by just locating a few overall critical
the above techniques, as well as vehicle could take engineering man-years to per- points, and then use stored computer logic
dynamics modeling from the previous sec- form such an analysis on aircraft or auto- to provide more detail. With simple com-
tion, may be lumped under Computer Aided mobiles -- and the results were very often mands, and/or a digitizing tablet, you can
Engineering (CAE). So much for the questionable. Now software packages are identify a point or line; redraw a hand
alphabet soup. available which can reduce the process to a sketch; duplicate, enlarge, or rotate any
As engineering software has evolved matter of days. Even relatively inexperi- element; and call up stored shapes. You
into comprehensive packages, the root func- enced engineers can learn such systems, can define a surface by four points and then
tion seems to be “solid modeling.”’ That although a basic course 1n structural analy- ““mesh it,’ or break it down into any
is, before any kind of development can be sis is highly desirable first. number of sub-elements. And if your
done, the first item of business is to get a Even without a computer, a good struc- chassis is symmetrical, you can just create
numerical 3-D model of the object -- tural engineer can estimate the location of one-half and then double it. The finished
whether a bracket or a complete chassis -- the worst stresses and easily calculate the image is then automatically digitized, or
intothe computer. Once that isdone, many worst loads. From there he can work down converted into the necessary data points.
analysis operations can be performed onit. through less and less critical areas. But a Alternatively, if a real model or prototype
FEM canbe used to determine its strength, computer can just as easily go to extremes has already been constructed, a 3-D digi-
once boundary conditions are defined, such of refinement, breaking a structure into tizer can scan the necessary surface points,
as space and geometry limitations, mate- hundreds of thousands of ‘‘elements,’’ as using either a laser or an electronically
rial properties, and all load inputs. In fact, shown in the computer-generated drawing located mechanical arm.
there are even automatic optimization rou- of Fig 53. In this way it is possible to The most comprehensive FEA pro-

144
grams can run any conceivable stress test
on the structure: static or dynamic, linear
or non-linear, impact or fatigue. Dynamic
tests are primarily for vibration frequency
analysis; non-linear tests are for investiga- AIA TS
tion of structures as they yield; fatigue is
ces
A KH bs > ‘&

not usually a problem in the short life of a

OBEN
TA L A
LOL A
KA 1] , Ly
HAN
racecar; and thermal oraerodynamic stud- ¥ ral

ies don’t apply to the tub. So for our

purposes (race cars) the simplest test is
adequate--and expensive enough. A simple
Static torsion test is not only the most
appropriate for a race car, but it is the
easiest to verify in the finished product--
and to compare to previous empirical de-
signs.
At some point it’s necessary to make
an educated guess as to what loads will be
encountered and what stiffness is required.
It is commonly assumed that a chassis
which is torsionally stiff enough will be
adequate in beaming. It also used to be
assumed that 3,000 foot-pounds per de-
gree of stiffness provided a rigid enough
chassis for the suspension springs and
anti-roll bars to balance the handling.
With the incredible aerodynamic loads
now seen, it’s hard to say what is adequate,
but the best Formula One cars are probably
over 40,000 foot-pounds per degree. The
scientific way to determine the loads would
be to instrument a car for vertical, longitu-
dinal, and lateral g loads, and get a record-
ing of the conditions at all tracks in the
series. This would give not only stiffness
requirements, but peak bump loads and
amplitude/frequency plots for fatigue
analysis.

Understanding the Output

There are just as many options for the
form in which the computed answers are
presented. The simplest would be just one
figure for stiffness in foot-pounds per de- a
gree, or perhapsa plot of stiffness down the ‘ey

y £ sh a Wwe
| lcci
. %
.

length of the chassis. Another option is a ae F

computer-generated drawing of the struc- Figure 54. The Hercules Aerospace finite element model of the McLaren MP4B,
ture in its unloaded state, over-laid with an which introduced the FEM concept to racing (top), and the final tub (below).
exaggerated drawing of its shape in a
stress-deformed configuration, showing case. In fact, there are unanticipated situ- In the most efficient design and develop-
exactly what is bending where. Finally, ations which must be allowed for, like an ment operation, all possible approaches
stress contour maps may be drawn in color off-course excursion, or combinations of are compared in terms of cost/benefit ratio.
over all the surfaces, giving a vivid visual load such as hitting the brakes and a bump Rumor has it, that in the aerospace busi-
representation of high stress concentra- at the same time. When you get right down ness, weight reduction is effective at
tions. These are almost equivalent to to it, it’s hard to even make a subjective $100,000 per pound. Even in mass-pro-
layout drawings for locating extra layers of decision: What kind of safety factor does duction automobiles the figure might be
reinforcement material. one select? 1.5?...2.5?...ten? But given a about the same, spreading the cost over
Ideally, a designer could keep modify- complete historical survey of what the ve- 100,000 cars. In fact, the auto industry has
ing the structure, changing the location hicle should encounter, and leaving the found finite element analysis to be more
and thickness of panels until all stresses consequences of deviation up to the driver, cost effective than exotic materials for
were equal at the limit. This would give continual redesign and rerunning is cer- weight savings so far. After they get the
the ultimate optimum strength-to-weight tainly better than experienced guessing. design optimized, then they will start us-
tatio. However, this assumes that the real ing more aluminum and composites. For
world loading conditions are accurately Cost just afew Formula cars, however, the cost/
known, and that there is only one worst Even in Formula One, there is a limit. benefit ratio may be more like ten thou-

145
 of course that information is never gener-
ally available to privateer racers or con-
structors. However, a strong racing-ori-
ented engineering program under Harry
Law at Clemson University produced a
number of valuable publicly available SAE
papers (See numbers 983051, -53, and -
54), that demonstrated what FEA could
contribute. Graduate level students used it,
with carefully measured dimensions from
some familiar chassis, to establish that the
baseline torsional rigidity was about 10,000
foot-pounds per degree. They then ran
dozens of cases in which they added or
strengthened structure to see what the rela-
tive effects were.
The analysis easily showed that
the weakest areas were at the cowl and the
un-triangulated engine bay. Eventually,
A finite element model of an actual NASCAR chassis, as it was being optimized for repeated computer runs brought the rigid-
torsional rigidity at Clemson University (Prof. Lonny Thompson) ity up to about 23,000 foot-pounds per
degree, a number that was felt adequate in
sand dollars per pound. proof. Those who have done FEA a few maintaining the desired suspension dy-
The actual costs of running a finite times recommend experienced specialists namics. Being realistic, they also analyzed
element analysis are a little hard to predict, who can save a lot of time, money, and the engineering tradeoffs in terms of added
given all the unknowns and the potential mistakes. While the designer can’t keep up weight and CG height. The final configu-
complexity. Most teams probably don’t with the latest in computer technology, ration was plus 40 pounds and a 1/2-inch
even know what it costs to analyze their neither can the programmer be aware of all higher CG, numbers which could be re-
cars, but we can use one early model as an the race car design requirements. It takes lated to laptimes through the use of a
extreme example for estimation of a pro- a combination of talents with expertise in racetrack simulation. Finally, to verify the
fessional analysis. The structure was bro- many areas to make it all come together. FEA, other students constructed a precise
ken down into about 800 elements. Based torsional test rig -- also to resolve the
on computer time and man-hour costs, just History of FEA precision and repeatability of testing. After
the input or setup time would take about To my knowledge, the first computer- running a number of different chassis, their
two weeks and cost roughly $3,000. From ized structural analysis of a complete race figures indicated that some common chas-
then on, each computer run on each con- car chassis was the McLaren MP4B For- sis were less than 60 percent as stiff as the
figuration change might be about $500, the mula One chassis, which was the most original baseline chassis in this study.
number of runs depending on the bril- sophisticated race car chassis of its time.
liance of the first layout and the final The first hint was a computer-generated How To
refinement desired. So the total computer structural drawing which appeared in a Computerized structural analysis of race
development time could have cost less than press kit at the 1982 Long Beach Grand car chassis and components is not totally
$10,000. The weight saving due to graph- Prix (See Fig. 54). Since McLaren’s suc- out of reach of non-factory teams. Software
ite fibers was said tobe 35 percent less than cess rate that season was largely due to the companies do not ignore any market for
an aluminum tub, but how much can be team and their drivers, the basic tub never their products, scaling down their more
credited to the computer is unknown. did get appropriate recognition. omnipotent programs to fit lesser budgets
However, if we estimate the tub to be McLaren had asked Hercules Aerospace and computers. Granted, you still need a
around 50 pounds, and the finite element of Salt Lake City, one of the largest graphite basic background (at least a first engineer-
analysis reduced the weight by 10%, that is fiber producers in America, to produce the ing course in structural analysis) to be able
5 pounds at $2,000 per pound--which is first one-piece true monocoque structures to understand the simplest programs. And
certainly within reason. ever seen in racing. So naturally when if you don’t spend more than a few weeks
On the other hand, assuming you al- Hercules took the job, they decided to use per year in structural design, the familiar-
ready have a decent PC, a slightly less the standard aerospace technique of finite ization time required will never pay off. So
ambitious analysis can be done in your element analysis to optimize the design. instead of duplicating standard textbook
spare time for a fraction of the cost. Sim- Eventually Hercules admitted that the origi- explanationsof FEM programs, Iwill briefly
plified versions of NASTRAN currently nal hand-calculated design by McLaren’s mention how to select and use them with
sell for a few thousand dollars, and with John Barnard was intuitively ‘pretty close,”’ respect to race cars.
them, you can make as many runs as your but this was still a dramatic first applica- Cost and compatibility with your ex-
spare time and deadlines allow. But for a tion. isting hardware may be of first importance
few applications, the best value might be At the time of this writing, FEA had to the non-professional non-commercial
hiring out the analysis, or merely making even reached NASCAR, the last bastion of user, although an engineering professor
a few simplified calculations on a pocket intuitive race car design. Factory assisted once told me the most effective strategy for
computer. Or, as is usually the case, rely- efforts from Chevrolet and Ford were being a department of five or more engineers
ing on intuition. quietly used to improve the torsional would be to fire one engineer and buy the
In any case, the process is not bullet- strength-to-weight ratio of stock cars -- but best software package. At the time of this

146
writing, apparently the most popular and
most powerful package was I-DEAS (Inte-
grated Design Engineering Analysis Soft-
ware), which with all added modules could
be in the $50,000 range. Other packages
that should be investigated include CATIA
ot al« Ace
bas
Loa
(more design and surfaces oriented), z ari weeyooh! y dshoul have been
Unigraphics (more manufacturing ori-
ented), ProEngineer, and Autocad. The
FEA techniques they use may be based on
NASTRAN, ANSYS, ABAQUS (prima-
rily non-linear), and STRUDL (civil struc-
i near
ae

tures). Although an adequate full-bore voce: End of steering 2.7oe San

HO Bee

program could be in the $5-10,000 range,

simplistic training programs are as little as
a few hundred. This may be adequate for Paid Se a ~ should
have been
many racers, when selection is based on the
something like this
following capability considerations:
Types of analysis include statics and
Less Throttle
dynamics, 2D or 3D, and heat transfer, = Opposite Steering ~ i : ee Understeer
which is hardly critical in race car struc-. And Throtile on : ; i Front Tires Overused
= OFS correction
tures.
Ala ooows Throttle on
Variety of components that can be (138
modeled includes plates (panels), beams
(including tapered beams), and shafts. One Typical analysis process using one simple screen of MoTeC data, as presented by
small drawback in some scaled-down pro- Claude Rouelle in his Race Car Dynamics Seminar. This represents only four
grams is that they only accept point loads. channels of data (lateral g's, steering, speed, and throttle) in a 100-mph turn. The
That is, aerodynamic or gravity loads can- notations with arrows are the comments added in class.
not be distributed over a surface, but must
be reduced to point loads at nodes or analy- with thousands of nodes (the original DATA ACQUISITION SYSTEMS
sis intersections. McLaren example was about 600 nodes).
Graphics capabilities can include in- It’s simply a matter of learning to live with Anyone with enough time can design a
put visualization, zoom/rotation/pan of the limitations, by first analyzing witha ‘ “broad race car on their computer, or simulate one
image, overlaying of the strain-deformed brush’’ and then focusing in on critical on their computer. But the greatest excite-
image, and animation of the image in areas with increased element densities. ment comes from actually coupling the
dynamic analysis. Other features not usu- As problems are simplified to accom- computer up with a real race car to take
ally available in lower cost programs are modate smaller computers, accuracy of real data on a real track. Race car devel-
automatic generation of meshes, or panel results may begin to suffer somewhat. Con- opment usually requires flexible general-
breakdowns, and automatic input data er- sidering that real-world race car loads are purpose equipment, inthe number of chan-
ror checking. only approximated anyhow, errorsare prob- nels, types of transducers, sampling inter-
Output variations beyond graphics in- ably not significant. But this points out that vals, and total capacity. Electronic race
clude transfer capability to larger comput- FEA’s greater value is in showing relative car data acquisition started in the late 60’s,
ers for more detailed analysis -- a strong effects rather than absolute values. That is, although it was limited to analog tape
point for a NASTRAN-compatible pro- it is better to run a computer analysis on a recording systems, as described in my first
gram. And of course there may be CNC- known, existing chassis (perhaps compar- book, ““Chevrolet -- Racing?”’
compatible output, if such machinery is ing output to real test results), and then Computerized data recording is com-
available, even on a contract basis. make simulated changes in the chassis, mon in engineering. There are dozens of
Ease of use may be a key question for such as thinner material sections, to study such systems available; dedicated multi-
occasional users. Procedures and com- the change in performance. Inother words, user, multi-tasking systems that can even
mands are easily forgotten without interac- what is the predicted tradeoff in reduced use the analysed data to automatically
tive input, or automatic prompting. To weight versus reduced stiffness? control subsequent test procedures. Ona
reduce memory requirements, some pack- In the near future, expect to be able to smaller scale, anyone witha PC can gradu-
ages have minimal prompting. For initial lease software just as needed from applica- ally build their own small system out of
familiarization, a good interactive package tion service providers, and to be able to run commercial modular components such as
might take a few days to learn, while a poor analyses at home via the Internet. Mas- a plug-in board and general purpose off-
non-interactive program might take few sively parallel processing will soon make the-shelf data analysis software. How-
weeks. But then as procedures become analyses a thousand time faster, and a ever, the best of these are seldom adequate
familiar, the repetitive prompting isn’t there hundred times more efficient. But in the for the in-the-field race car requirements
to slow down the process. fast-changing world of software, no book of vibration, noise, and heat resistance,
Speed and capacity might be the last can be up to date. The only way to get the and cost, portability, and size restrictions,
consideration. At worst, a maximum ca- latest information on FEA products is from the reason for development of dedicated
pacity analysis on a PC might take over- monthlies such as Mechanical Engineer- race systems.
night -- which is human downtime any- ing, Machine Design, SAE Journal, and Origins of digital DAS
how. But that may still allow a structure your own specific PC magazine. BMW may have been the first with

147
 reality, not that many applications outside
ofengine development and “‘maybe’’ over-
steer/understeer balance demand instanta-
neous real-time observation of data while
the car is on the track.
Over the last couple of decades, com-
puterized DAS has evolved along with
advancing computer technology, driven by
many overlapping commercial interests:
(a) Factory engine development engineers
need faster and more precise control sys-
tems, especially in telemetry, and they can
afford the best. (b) Chassis development
engineers can count on the DAS manufac-
turers to continually provide them with just
a little more capabilities that they can
apply. (c) The dash gauge and instrument
manufacturers increase their product so-
phistication by incorporating memory that
can also be downloaded for later analysis.

Hardware Selection considerations.

Specialized data loggers are necessary,
since most computers are not compatible
A demo sample of data acquisition hardware, utilizing CAN bus, or multiplex tech- with the race car environment. Actually, a
nology, in which all sensors can be chained together on one cable of only five wires. logger is acomputer, but one that is “‘hard-
ened’’ and dedicated to primarily collect-
digital race data-taking, in their engine intervals, analyzing them on an Apple, and ing, storing, and transmitting data, rather
development for Brabham in 1983. Work- worked up to a portable computer, analyz- than analyzing it. You will find a lot of
ing with the Bosch computer-controlled ing at least eight channels simultaneously. similarity between many systems, since
injection system, they already had digi- Although drag racing seems like a low- they’ve evolved from similar roots, and it’s
tized engine signals such as rpm, tempera- tech sport which contributes little to circuit easy to copy features, and it has become a
tures, and pressures. So it required little racing, it’s interesting to note that their ““mature technology.”
extra electronics to temporarily store a data recording requirements may be even It’s not appropriate to focus on specific
sequence of such figures in an onboard data more demanding, where things happen so DAS products in a book that faces years in
buffer. When I first saw the system, they fast -- sub-5-second runs over 300 mph -- print, while the products change monthly.
had a separate circuit board in the shielded that only a computer can handle them. As The key is in knowing the most important
engine control module, with a 16K RAM early as the 1983 season, the Kenny general selection factors, or how to make
that could store up to 1000 data points from Bernstein team was developing a system an intelligent choice for your specific needs.
each of 16 channels - which seems amus- that could take 32 channels of data at 100 Even when software demo discs are avail-
ingly primitive today. Then in the pits they samples per second. Eventually Bernstein’s able, unless you try all of them, you may be
would download the data for analysis. As engineers came up witha simpler and more so impressed with one that you miss some-
it was primarily an engine project, how- marketable system called Racepak. thing important that’s available from an-
ever, they probably didn’t use the system In 1985, Harry Gant’s NASCAR other company.
for much chassis development. Chevrolet had an extremely sophisticated Initial cost will probably be your first
About the same time, Honda racing telemetry system. The onboard data acqui- consideration, if you aren’t lucky enough
engineers asked me to come in and discuss sition system was a 40-pound box that could to be with a top series pro team where
systems that might be useful in their new continuously monitor 24 different sensors budget is no problem. However, you should
F] activity. After a two hour monolog on ata rate of 55 samples per second. The data really think in terms of cost/effectiveness.
my part, they asked some questions that was then transmitted to a receiving and You could probably learn everything you
indicated they knew a lot more than they analysis console via a 4-watt transmitter, need to know with a four-channel system
were telling. As it turned out, they were and a 30-foot receiving antenna. However, for about $2000, but it would take a lot
about to sign with the Williams Team, at the end ofthat season, NASCAR officials more time and laps. As the cost of track
where Frank Dirney was already using a 4- became wary of such systems actually being time and car wear escalates, amortizing a
channel tape recorder on his cars to gather used to monitor a car during a race, and more expensive system over years of saved
data for later computer analysis. outlawed them. test days makes a lot of sense. Costs have
Eventually, Williams/Honda made a Incidentally, data telemetry will not be tended to hold steady in roughly a $500-
deal to share computer technology with an covered in great detail here, even though it 1000 per channel range, depending on
American CART team headed by Ian Reed, sounds exotic and is commonly used in software bells and whistles -- plus the host
and he brought in a hardware expert, Kurt Formula One and CART teams. My own computer cost, which can be justified for
Borman, who spent a couple of years devel- experience with it convinced me that it was many other applications. On the other
oping a data-taking and analysis system, seldom worth its development and mainte- hand, an experienced electronics techni-
later marketed as the Motech. It started nance problems unless there was someone cian could probably put together his own
with 4 channels, taking data at 1/2 second on it full-time. When you get down to system for half that, using inexpensive

148
modules and general purpose analysis soft-
ware like Labview, DaDisp, MatLab, or Excel.
Precision or resolution may be a key 0.00 s xf (21
issue in logger selection, depending on Session 210
ateral Accel Vs Distance (No Filter) [1]

your specific test requirements. Common Outing 020

Driver Luke
Z 2 aa 4 ce eo

Channel Elapsed Time (Section Change)

8-bit precision allows you to make each Lap 1 Time FL Diff Otm Diff
measurement to one part in 256 (two to the 028 24.10
2a93|
1:25.84
125.77
cos]
| 0.0000
——
=
eighth power) over your full-scale range. 030 24142 1:47.78 022.07 ——

That is, if you have a 200 mph range, each

Theoretical Fastest Lap
233 1 25,54 000.40

step increment will be 0.8 mph. If changes Fastest Rolling Lop

in the measured quantity are fairly slow

and smooth, your program can mathemati-
cally interpolate to smaller fractions. Often,
12-bit precision is necessary, at one part in
4096 (two to the twelfth power), especially
when operating in a small part of a trans-
ducer's total range, or differentiating a
signal. Note the distinction between data
precision and operating system precision,
which might have 32-bit bytes.
Channel capacity required in the log-
ger will depend on the complexity of the
tests and the operator’s experience. Start-
ing out, eight channels can satisfy most
needs, and any morecan overwhelmall but A Pi data analysis screen, just one of an almost unlimited number of possible
a full-time engineer. Expandability in configurations, which requires a steep learning curve and experience to fully exploit.
terms of added plug-in channel modules,
may be an important consideration if you a magnifying glass and expensive tools to training program, at least classroom, if not
expect to be growing with the system. Over assemble them. On the other hand, you’ll one-on-one. Ideally they will have enough
32 channels may be available on a logger probably find that the most common sys- customers to justify a track support engi-
-- but then you have to consider the added tem failure will be in the connectors, so neer at major events, equipped with spare
cost and maintenance of 32 transducers. don’t expect inexpensive Radio Shack hard- hardware and minor wiring repair facili-
Channel types are as important as ware to stand up. ties. This is expensive service, and should
number, since requirements are different Total memory capacity of loggers is be factored into the original price consid-
for analog sensors such as temperature or reaching a level where they should easily eration.
pressure, and digital channels for rpm, meet any data acquisition demands be- Familiarity to users. In some series,
speed, distance, and events suchas laptime tween pit stops. Already, many megabytes an overwhelming consideration may be
or segment timing from external trackside of data canbe stored on aremovable memory the established “‘default standard,’ abrand
beacon inputs. card the size of a stick of gum. Compacting that has established a near monopoly. This
Maximum sample rate you’ll need it more just makes it easy to lose. means that if you want convenient support,
should be at least 3-4 times the maximum Size and weight are hardly important and the ability to pirate top crew from other
oscillation rate of your data. A 20-Hertz considerations any more, even in formula teams, you have to followthe herd. Atleast
system would be adequate for overall ve- cars, since memory and processors have this is true until another manufacturer
hicle dynamics, aero loads, and driver shrunk to the point where their container comes along with some breakthrough in
evaluations. Atthe other extreme, analysis size is limited by space required for the data acquisition that provides a great com-
of engine vibrations, injector pulses, or harness plugs, switches, and readouts. petitive advantage.
torque fluctuations on each revolution could Demonstrated hardware durability
require 1000-Hertz to get meaningful data. Source -- supplier company and debugged software. A newly-an-
Faster sampling gives greater precision -- Who you buy your data acquisition sys- nounced apparent innovation in any tech-
but it can also produce excessive amounts tem from might not seem as importantas its nology is only as good as its development
of data. The best restraining force is to specifications -- at first. But you will find program. The world’s greatest new thing
frequently look at the mountains of data that it’s a major investment in familiariza- will drive youcrazy ifit hasn’t been through
that are ignored and/or thrown away, and tion time, as well as money, and you be- a “‘beta test’ to identify and fix the inevi-
reduce expectations accordingly. It’s hard come committed to one manufacturer even table problems that only a million user
tocomprehend --even with computer analy- if another eventually becomes more appro- trials can uncover -- unless you recognize
sis -- a megabyte of data (24 channels X priate. The day will come when you'll wish the importance of the innovation, and your
500 samples/sec X 80 seconds) on every lap. you had considered the following factors: role indebugging, and are willing to put up
Harness connector size, durability, and [Or manufacturers will use this as their with it (perhaps fixing the bugs yourself to
availability should be considered, espe- checklist for long-term survivability. ] keep an advantage).
cially if you expect to be modifying the Training, track support, spares. In Company stability and growth. His-
system, as opposed to having a permanent high-tech companies, size matters. They torically, there have already been many
pre-wired harness. Some of the more should be big enough to have people dedi- good DAS products from companies that
sophisticated aerospace connectors are ex- cated to personal service. Unless it is a very disappeared for various reasons, leaving
tremely expensive, and so small you need simple system, there should be some sort of “‘orphan’’ systems behind: A successful

149
 the track timing system, or isolating laps
for easier overlay.
Color-coded data curves, along with a
color reference table, are mandatory for
fast identification of multiple traces. It’s
important to be able to assign selected
colors for consistent reference and group-
ing, instead of random assignment.
File storage and access convenience
will become more and more important as
your data library grows. It will be useful to
search and identify data by many catego-
ries, such as date, or track, or car, or driver,
or test type.

Additional comparison features

Transfer/storage media. Hard drive
capacity will probably always keep ahead
of your data storage needs. But the rough
racing environment demands a lot of
backup for lost or damaged data. You
might archive on tape or removable disc by
years, but consider that someday you might
want to recover and compare the last five
A very basic Racepak data screen familiar to drag racers: engine rpm, wheel speed year’s data to study the effect of setup
and acceleration in g's, illustrating wheelspin, shift points, and driver irregularities. changes at one given track.
Download method/time. For really
technology company has to put just the Software Features -- obvious or expected large data sets, or for professional series
right resources into future development -- Overlays are one of the most important where you might want to download during
enough to keep ahead, but not enough to applications of DAS. It’s hard to get any- a fuel stop or tire change, the time and
bankrupt them. Or you could buy an thing meaningful out of data without direct method can be critical. Physical contact
inexpensive system and expect it to be dis- graphically-visual comparisons. Only when possibilities are cable connection or plug-
posable. you superimpose one lap speed or throttle in memory modules or cards of various
Upgrade/expansion. Advancements trace over another do the improvements types and sizes. Non-contact options might
come so fast in this business, that you stand out. It may be used to compare setup include infra-red or telemetry.
should expect new hardware to be compat- changes, driver technique changes, and iden- Screen arrangement options. More
ible, and software upgrades to be inexpen- tify why thereisan improvement in laptimes. sophisticated analyses may require mul-
sive and convenient to install. In fact, new It's also important to be able to recover and tiple windows to be visible simultaneously,
software should be available via the Internet. compare data from different sessions. such as speed trace, gauges, and track map.
Hardware upgrades are most commonly Selection of graphed channels allows Every user can decide where and how large
more memory, more channels, or faster you to tailor the output to your immediate the windows are located on the screen, for
processing. Trade-in might be possible, needs. Traditionally you will record far faster comprehension.
making upgrades less expensive, and pro- more channels of data than you expect to Sample frequency variation on dif-
viding a ‘‘trickle-down’’ affordability to use, as insurance, then only display what ferent channels. Instead of sampling all
newcomers. The transducers usually have you need in any analysis. And you should channels at the same rate, it can save a lot
common signal electronics, if not connec- be able to easily call them up in pre-selected of memory space by only taking data at its
tors, andcan be carried over for many years categories, such as the handling group, the relevant rate. Engine torque data may
through many systems. engine group, or the aero group. require 100 samples/second or more, but
Component cost/quality/availability. Scale/calibration/zero adjustment is inlet temperature probably won’t change
Take a good look at the transducer catalog. necessary to correct for unintentional drift over many seconds.
Don’t buy an inexpensive data logger that or mechanical shift in the transducer sig- Noise filtering options. Filtering is
requires expensive heavily marked-up nal. It also allows you to isolate curves by another subject worth an entire chapter,
transducers, just because they have special shifting them vertically with an offset, and whether the noise is physical, such as
connectors installed on commonly avail- to scalethem up to fill the screen, ortoshow pavement roughness, suspension or lateral
able components. The experienced instru- more detail. vibration, or aerodynamic turbulence, or
mentation technician might save a bundle Zoom for time or distance range is a way electronic noise from ignition or radio
by buying military surplus transducers, or of focusing in on smaller and smaller incre- interference. Instead of simply slowing the
buying them independently from the source, ments of time, starting with the “‘big pic- sample rate, or averaging data points, it
and wiring them himself. For example, ture’ of all laps in a session, down to can be useful to have other mathematical
some fairly accurate transducers may be visualization of each individual lap, or as filtering options, to avoid hiding real spikes,
common in most production cars, such as small as one individual sample time-step. points of interest, or sudden changes that
rotary throttle pots, rpm triggers, mass Re-zero of the time base is a way of filtering can smooth over.
airflow sensors, wheel speed sensors for resetting the time base to zero at the start- Time-base versus distance-base op-
ABS, and yaw gyros. finish, for correlation with stopwatches or tion. It’s natural and obvious to plot data

150
as afunction oftime. But when overlaying a Urack Master - [WC with Air-Fuels.AVW]

curves, the car’s instantaneous location

will shift slightly due to any variations in ks|Wheel |Serateh Pad| |c
speed. This can be cured by automatic AF 5: 13.8 AF AF 1.134 AF AIF AVG_AF_L: 0 Winston Cup Test, Run 12, Lap 3 s
yAF
conversion to a distance-base, so that the AF 6.13.2 A/F AF_2) 14.6 AVF :
horizontal axis is in feet instead of seconds. AF 7: 13.4 AF AF 3: 12.6 AVE Session Notes
It also allows data to be located more — — raised front float level
accurately when matching track distur- AF 4 126 AF AF 8: 13.2 AVF another O60

bance features that show up in the data. AVG AF L 13.38 A/F IAVG AF R: 13.29 AF||Right side got richer, picked
up .2sec
Select best. It may take some time for
Signal Minimum Maxirnurn Average Range
the human mind to scan a table of numbers AVG_AF 13.28 16.09 1379 2.806
AVG_AF_L 13.05 16 82 13.71 2776
and identify the best lap, so that should be AVG_AF_LR 13.04 1663 13.86 3.586
a simple and convenient software function, ENGINE_R- 030.357 Winston Cup Test, Run 12, Lap 3
whether for best lap, best segment, or top
aN

speed. rt ee 7683 ram - i

Histograms. Bar charts that summa- (Gl) ==
*
Nee EASE
rize data by categories of time increments
give a quick visual reference for such fac-
tors as time spent at different rpms, differ-
ent throttles, braking, or left/right lateral g’s.
Min/max/average over a defined pe- .
riod. This is another statistical manipula- 0.0)
] [
tion that is difficult for humans, but a 000 6000

meaningful quick reference for straight-

away max or cornering min speeds. CDS Track Master screen layout possibilities for analysing engine air/fuel ratio
Math channels for data manipulation. performance related to track location on a superspeedway.
This is a way of combining different chan-
nels of data to produce another artificial send it to a database for report writing. ing, perhaps also using GPS, which will
one. For example, a plot of the point-by- Annotation. It can be useful to be able become continually more accurate and less
point differences in speed between two to add operator or driver comments or expensive. However, knowing the precise
laps, or plotting the vector sum of lateral engineering setup specs to the data on the track geometry doesn’t always identify the
and longitudinal acceleration to give a screen, to help explain odd data without driver’s path on it.
“‘friction circle,’’ or converting suspen- having to refer to a separate logbook. In the Superimposed data on track map.
sion deflections to pitch or roll angle, or near future it should be possible to add This is another visualization trick that
combining steer angle, speed, and lateral comments in audio. plots one or two channels of data such as
g’s to indicate understeer, or differentiat- Animations of gauges, controls, or sus- speed, braking, or lateral g’s, perpendicu-
ing speed to get acceleration, or accelera- pension deflection. This helps to visualize lar to the track curve, or vertically in a 3-D
tion in g’s can be compared to engine rpm, rapidly changing or complicated data, such plot. This is low resolution data, but it may
driveline ratios, mass, and drag, to read out as gauge readings or chassis motions, espe- provide easier location identification than
an engine torque curve. cially with telemetry in real time. It also distance-based plots, especially for less
Curve fit. If noisy data should theoreti- helps the driver recall and understand what technical drivers. Sometimes digital data
cally fit a smooth exponential curve, then he was doing, in slow motion. may be available in text boxes on the map.
itcan help to use an automatic curve fitting Track map creation, with correction Hot link between graphs/windows.
routine. Then, generation of the equation for grade and camber. Although it’s com- When a number of data windows are open
for that best curve fitcan reduce a lot of data mon to generatea track map from speed and simultaneously, it can be convenient to
to simple coefficients, such as reducing an lateral g’s, these maps are more for rough have them linked so that a change in data
aerodynamic coastdown curve to a coeffi- location visualization, using the cursor to selection or calibration made to one, is
cient of drag. identify data location, than for accurate automatically made to the others.
3-D Graphs. Visualization of three vehicle location. When ‘‘closing’’ lap Fast Fourier Transform or FFT. This
variables simultaneously on three axes, positions don’t match up, ‘“fudge factors”’ is a way of plotting data not as a function of
such as isometric carpet or waterfall plots, are called on. Without full correction for time or distance, but with respect to data
or by using variations in line color to the roll angle contribution to lateral g’s, frequencies. For example, suspension de-
represent the third variable, as in superim- and correction for camber and elevation flection for an entire lap can be reduced to
posing another variable on a speed-time changes, the map will be just an approxi- how often it oscillated at different frequen-
trace. mation. Other sources for map error in- cies. You would expect to identify two
File transfer, to a simulation, spread- clude inaccurate identification of a curve’s major peaks on the graph, one at the wheel
sheet, or database. If your DAS software start and finish, from normal steering noise, frequency around 10-15 cycles/sec, and
package doesn’t have some desired fea- and variations in surface traction coeffi- one at the ride frequency of about 2-5
tures, your files should be in a format that cient. cycles/sec. It can also be used to correlate
can be transferred to another that does. For Track map surveys. For really accu- a suspicious vibration with likely sources
example, you might want to use the track rate data matching and simulation, the map known to be rotating or contacting at the
map and speed trace data in a lap simula- might be obtained from engineering sur- same frequency, like gear tooth contact.
tion, or send the data to a spreadsheet for veys (sometimes available from the track And it can be used for spying: to discreetly
advanced graphing and curve fitting, or owners), or from low speed position record- record and plot a competitor’s engine rpm

131
 —— : Satin iennEEEnnn EEE ———————
| video of the driver’s actions, suspension
| Circle=1G) |
4 | deflections, or even impacts, in perfect
sync with the data.
Control capabilities. This is the fea-
ture that actually preceeded digital data
acquisition, where engine companies were
| using computers to control spark and fuel,
8 which required electronic sensors and
a || onboard memory. Control requires special
1
0.0 | hardware that doesn’t just record, but can
7 || also output signals in real time. Currently,
| |
just a few higher level DAS companies
|| || provide this feature, because it is primarily
4 | useful to engine developers. It can also
|
iT
|
|
|
allow trackside manipulation of 3-D fuel
and spark maps.
=} i \

Computerized chassis control. The

— = - —————— = = Je = SS
engine control capabilities are easily
adapted to control anything else on the
Brake | | |
chassis that can be driven by a servomotor:
20| Left | | | Right ride height, anti-lock brakes, anti-
I I T T T T I i} T T T T l wheelspin, differential bias, stability con-
20 0.0 -2.0
trol, aerodynamic devices, etc. Unfortu-
nately, most race sanctioning bodies have
Combined lateral and longitudinal g's over a full lap of left/right turns, showing the become extremely restrictive in this sort of
effects of high speed downforce and engine torque. (from Buddy Fey's "Data Power") technology, for various justifications.

history on each lap, just from it’s sound. removing that responsibility from the In Development:
Data alignment, using a manual driver, or for higher-speed data sampling. GPS position monitoring. Even the
““nudge’’ or calculations. When overlay- For example, a crash recorder reacting toa most sophisticated GPS can’t locate a car
ing comparison traces, it can be hard to fit high g trigger, which can fill an entire data fast enough and precisely enough to re-
them precisely. A manual nudge function logger in a matter of seconds to record an place wheel distance tracking and inertial
may be used to shift a curve by as little as impact, but over-writing old data if it’s a reference -- yet. However, the basic con-
one pixel. Mitchell Software was the first false alarm. cept can be used with locating beacons
to present a technique of aligning traces CAN data bus and smart gauges. Mul- which are located more closely than an
mathematically, using road roughness re- tiplexing many data streams (both to and orbiting satellite. It may not be long before
peatability when plotting in distance-base. from the computer), on simpler few-wire a system is available with transponders in
Multi-user screen tailoring. Teams harnesses, is becoming more popular. It precise locations in the front and rear of the
with more than one data analyst may find requires processing power at remote loca- race car, so that not only can the instanta-
it useful to have software that responds tions, either at a few distant nodes, or at neous track position be recorded, but also
with screen arrangements based on the user every individual sensor. It requires more the pitch and yaw angles.
identification. sophisticated hardware and software, but Onboard in-wheel telemetry. Tire
Pit strategy. In series which allow may be less expensive in the long run. It pressure monitors are already commer-
onboard computers for real-time race data allows ‘‘smart’’ sensors and gauges whose cially available, which transmit four sig-
collection, this may be used to help keep functions can be modified in the field. nals to the vehicle computer. So we will
track of fuel and tire consumption, as a Onboard data manipulation. The data probably soon see multiple tire tempera-
driver readout of fuel remaining, or with logger can be programmed with its own tures transmitted real time also. And one
telemetry to the pit crew. math channels, to combine information of tire company is developing street tires with
In-car display (dash). Compatibility immediate value to the driver, such as the magnetic materialsimbedded inthe sidewall
between the DAS and dash, may avoid previously mentioned understeer calcula- in such a way that their stress-induced
duplication of many functions, from the tion, or instantaneous speed improvement deflection can be monitored as ameasure of
black box to sensors. It also provides more over a baseline lap. load and traction.
possibilities for feedback to the driver. Multiplexed slow data. A combina- Wearable computers with head-
Optional displays may be selected for test- tion of CAN bus and selective sample rates mounted monocle display. These are not
ing, racing, yellow flag conditions, or pit- can be especially useful for taking lots of for drivers, but for the pit crew. The
ting. Programmable warning condition lights temperature or pressure readings for aero- military already has a voice-controlled,
and shift lights may also be available. dynamic development, andcramming them wireless, net-connected, wearable computer
Threshold triggers. These can be pre- all on one channel. However, this may for aircraft and vehicle maintenance. And
programmed or driver adjustable, to intel- require more careful coding or identifica- I have already tried out a fingernail-size
ligently focus on critical conditions, in- tion of each signal. eyeglasses-mounted screen display and free-
stead of on simple single peak values -- Merging of TV images. A picture may space mouse that could make trackside
such as noting an over-temperature limit, be worth a thousand data points. Small, computing hands-free.
but only if speed is over a certain level. Or lightweight, inexpensive digital onboard Data mining. One of the greatest data
it can be used as a recording auto-start, cameras may be used to incorporate in-car acquisition problems is the mass of data

152
 Track Master - [Road Atlanta WSC.AVW]
that defies fast analysis. Eventually, expert
crew chiefs will be studied in great detail,
and their mental processes programmed
into software that duplicates their analyses Min. REAR_RIDE
REAR_RIDE -0.628 in
and decision-making. This should make -0.288 in
expertise more widely available, leveling
Min. FRNT_RIDE
the playing field somewhat. 0.773 in

Fiber optics. Transmitting data in a

race car can be much cleaner with optics, FRNT_RIDE )
-0.712in /
which aren’t affected by radiated electrical 9
noises as wiring might be. The main
disadvantage is greater difficulty in mak-
ing good connections. So it may be limited
to major runs or pre-assembled lengths.
HEAVE
-0 49984 in
Application REAR ROLL -0.1025 deg

Making a DAS that can survive in the

race car environment of heat, vibrations, OOF ti iy ; ar
Wat Wk Tate tT | asd ME Veta
and tremendous electrical noise (electro- lo dha i mh eV Seal avery fh
hate pra MA RONG Th ca otacsall Raker

magnetic interference -- EMI, or radio

1Miy heyHy
frequency interference -- RFI) can be a real 0 Oh Nt
ya ay +}7 ralij
= f \ |
problem. EMI is a concern in ordinary FRNT_ROLL -0.139 deg 05 -{---45 { “hy
passenger cars, given the output of dis-
tributors and the sensitivity of computer-
controlled ignitions. But far worse is the The CDS technique of illustrating wheel and chassis movements and ride heights as
case of race cars with no steel body enclo- a function of track location at Road Atlanta. The upper planeis a static reference plane.
sures. Most exposed wires, whether for
power supply or transducer signals, act just tion curve on an accelerometer may show it Analysis
like antennas, and can easily give you an to have 5,000 ohms resistance at 1.0 g’s. So The typical race team manager, if pre-
engine rpm signal on every channel -- they build one circuit to convert that to 1.0 sented with a typical computer output of
whether you want it or not. And if the volt per g’s, another to convert the voltage graphs and printouts crawling with num-
vehicle battery is used for the logger or to binary-coded-digital (BCD), and then bers will say, ‘‘What the hell does all this
transducer power supply, there may be all calibrate the output to plot 1.0 g’s as exactly mean? Isthecar better or worse?" To avoid
sorts of sudden surprises on the line. The 2.00 inches. the familiar problem of too much informa-
best approach is to expect the worst. Make A more straightforward calibration tion and not enough knowledge, the DAG
sure that every module and transducer is method might be called ‘‘mechanical had better have a very good idea of what the
totally enclosed by metal shielding, all through-the-system’’. You stand the ac- desired answer is, and how to weed it out
connecting cables are externally shielded, celerometer up on one end, using the earth’s fast. And if a lot of data runs are taken, it
and all shields are grounded to each other gravity to provide a 1.0 g input, and make will be very important to have good identi-
-- and not the chassis. Components should a short, steady recording. Then, when fication of each, before transferring
them to
probably use their own isolated battery playing the data back, simply adjust the mass storage.
power, unless integral with the ECU. gain until it shows 2.00 inches on the One of the best features of computer
Heat and vibration might not be a prob- screen. Some other calibration reference analysis is its ability to ‘‘cleanup’’ the data
lem if the logger can be located at the inputs are: photo optic sensor - 60 Hz or smooth the curves. Most raw data will
engineers’ preference. But all too often ““blink’’ from incandescent lights; suspen- have interference or “‘noise’’ superimposed
space and weight and physical interference sion deflection - tape measure; steer angle on it, whether because of vibrations caus-
problems (and driver prejudice) prevent or throttle deflection - protractor; pressure ing the transducer to oscillate above and
such freedom. Electronic components are sensors - column of water or mercury; below the true reading, or an extra signal
fairly resistant to shock and vibration, but angular rate - motor driven turntable or (say from ignition interference) added to it.
when you mount it on a circuit board timed skidpad circle; engine or vehicle In either case, various methods of averag-
(which can vibrate like a diaphragm on its speed - hold a constant tach reading; dis- ing readings over longer intervals will
standoffs) and connect loose wires, and tance - drive stop-to-stop over a measured smooth out the curves. Or if the generic
hard mount it all to an engine shaking at mile or 1/4 mile. shape of the curve is known (such as a
15,000 rpm, failure is guaranteed. And Since data recording components will coastdown curve) then a ‘‘least-squares’’
with all the necessary electrical shielding, inevitably fail at the track, you should curve may be fitted to the data. There are
internal heat buildup alone might be worse frequently scan the output to make damn a number of increasingly sophisticated fil-
than external sources. sure something is being recorded from tering techniques, going so far as to deter-
When it comes to actually taking data every transducer, or you may go home and mine how far each incoming data point is
on the car, key DAS issues are calibration, sit down to analyze what isn’t there. Be- from a predicted curve based on previous
confidence, and tracing (or trouble-shoot- cause it can be very difficult to trace a data. For more detailed information on
ing a lost signal). Instrumentation engi- digital signal at the track, the safest bet is data taking and analysis, with lots of ex-
neers using interchangeable general-pur- to use modular ‘‘black box’’ construction ample graphs, you need to get Buddy Fey’s
pose equipment traditionally rely on elec- and have backup components that you can book, ““Data Power: Using Racecar Data
tronic calibration. For example, a calibra- switch around to localize the problem. Acquisition.”’

153
 constant advancements, thischapter would
soon become obsolete if specific products
were described. So it will instead focus on
general capabilites and selection criteria.
= {2} SAMPLE.TAB
Gg Fuel Injection
For information on making the general
{aa} Iniection
{5q) Inj, Phase
manufacturer, software, and hardware de-
2g) Inj f (Battery)
{0} Cylinder Trins
cisions, refer to the DAS section.
a) Fuel Constants
& 3 Spark Advance
There are two primary computer con-
(30) Spark Map
on) Dwell f(Battery)
trol functions available: spark timing as a
feo) Cylinder Trims
{cc} Spark Constants
tion > Injection function of rpm and throttle angle, and
(5) Gy Lambda Control
gq) Lambda setpoint
injector timing as a function of rpm and
fon) Lambda Channel 1
fo) Lambda Channel 2
throttle -- or airflow. Functionally, these
{o)
&:) Lambda Constants
& (@ Sensor Calibration
aren’t much different from the old meth-
EQ Correction Tables
{20} Ini. F(4arbox)
ods of mechanical distributor and carbure-
lon) Ini. F (Water)
fon) Ini Fin)
tor controls, which used centrifugal
Inj. F (Fuel P)
Spark f (Water)
weights, vacuum actuators, or pressure
foo) Spark
& GQ Accel
t(Air)
/Decel
differentials. However, the computer can
} (Qj Reserved
(4 (2g System Constants
be faster and much more precise, by using
3 [@) Control
g) IRL 1.02 3-D ‘‘maps’’ which are plots based on
dyno runs and stored in memory for rapid
lookup. In addition, these maps may be
modified in real time, depending on chang-
ing conditions such as temperature or pres-
sures or other variables, actually making
EFI Technology software showing 3-D spark and fuel timing maps as a function of them maps of 4-D or 5-D or more. Also,
rpm and throttle angle. Note the variable spacing of the data points. other functions besides spark and fuel flow
can be controlled by the computer, such as
ELECTRONIC ENGINE CONTROL quires signals from a number of engine a turbo wastegate, or the length of variable
sensors which are accessible. For example, inlet bells (or possibly exhaust pipes), or
Computer-control of both engine spark the mass airflow sensor may produce a DC perhaps in the future, a variable cam-
timing and induction will eventually be- signal that is proportional to the airflow advance mechanism.
come almost universal -- even in racing, rate. Say the signal is 5.0 volts at maximum In selecting a system, the first consid-
probably even in NASCAR, as they find it air flow. If a leaner mixture were needed, erations mightbe hardware, or the number
more meaningful to the manufacturers another potentiometer might be added, and and types of input sensors that.can be used
and as easy to regulate as carburetors. adjusted so that at maximum flow the to calculate the output controls. Pressure
However, it will only be covered briefly reading was only 4.8 volts. In other words, sensors are obvious, especially barometric
here. As mentioned in Chapter 8, racing the computer would think there was less pressure, fuel pressure, and possibly low
engines are most often bought as a pack- air, and therefore provide less fuel at that oil pressure as a safety factor. Tempera-
age from specialist builders with a fortune point. ture sensors might be used for the coolant,
in dyno equipment. So engine electronics Another modification possibility is con- the ambient air, and the fuel, which can
alone deserves its own book. Telling you stant ignition advance by moving the dis- become surprisingly hot in return-flow
how to reprogram your engine’s computer tributor or crank sensor, or changing the systems. Also, the exhaust gas tempera-
is almost like telling you how to design mixture only at wide open throttle by mov- tures of each cylinder can be valuable, as
your own tires -- at least it used tobe. ing the throttle potentiometer. Although commonly used in dyno development.
Today, if you want more performance tricking sensor signals can be used to make ‘“‘Lambda’’ sensors, or post-combustion
out of a production engine, there are broad across-the-board mixture or timing oxygen sensors, might be installed in one
aftermarket chips, and plug-in devices to changes, this is a crude last-ditch measure. or more tailpipes for “‘self-mapping closed-
rewrite the spark and fuel maps. To satisfy Dyno tests would more likely show a need loop’’ control systems. A knock sensor
governmentemission and mileage require- for complex changes in the advance or could be used, or even one per cylinder,
ments, and to control all the other elec- mixture curve, which would justify using a just as they have become more common in
tronic devices in a production vehicle, the user-programmable aftermarket engine passenger cars. And where mileage is
microprocessor has to handle hundreds of control unit. important in longer races, a fuel-flow meter
functions, from cold start to the air condi- For pure racing engines, the rapid ad- could be important, if fuel is not measured
tioning compressor. And these computers vancement of inexpensive computer power, by recording the injector flow. Finally, a
have to be produced inexpensively by the and the availability of more computer- driveline torque sensor may be used for a
millions for mass production automobiles. literate racing engineers, is revolutioniz- more direct measure of performance. So
And they have to be bulletproof. There- ing electronic engine controls. Competi- check a system's specs for the number and
fore, the development cost of the hardware tion components and software become types of input channels available, and check
is minimized in racing applications, if it cheaper and easier to work with as more their catalog for the sensors you might need.
can simply be adapted from existing com- and more companies enter the market. Besides the obvious control map func-
ponents. Often both the engine control system and tions above, there may be other software
Even when aftermarket performance data acquisition are available from the “‘bells and whistles’’ to influence your
devices aren’t available, it’s still possible same companies because of the overlap in decision. Field adjustment of the engine
to trick a system, since the computer re- sensors and processing. Because of these control parameters should be easy, using a

154
 Paim OS Emulator . Palm OS’ Emulator

laptop to edit the maps. There should be Keyin corner temps Loa Polookrl 195-200-205 214-230-234
195-200-205 214-230- ~1.2psi less press
an option for alternate reference maps in 187-184-193 PYROM 208-212-216 7-5 camber
out 1 camber in
memory, preferably selectable by the O82122.16
13deg
driver. Sensor input signals should have PUSH
easy calibration adjustment, including
non-linear calibration curves, and 187-184-193 208-212-216
e 1.2psi more press
adjustable width “break points” on both -.3 camber out A camber in
throttle and rpm axes. And expect
complete and independent control of spark
and fuel for each individual cylinder. BACK] COMMENTS [PLOT ho: §
COMMON SENSE WARNING
Whatever the application of computers
to racing — simulations/predictions, CAD-
CAM-CAE, data recording, setup advice, or
Palm OS’ Emulator
engine controls — there should be one
overriding rule: never violate the laws of MOSTLY- or it VARIES turn-to-turn | CAMBEROFF? ] = fae
common sense! Or in other words, beware LOOSE Inears PUSH
of the following pitfalls:
turn entry
|+?- FRONT PRESSURE?) 40.
; LOOSE In mid turn
Experience beats theory. Before using 13
computers for anything, you need a lot of LOOSE In turn exit PUSH
hands-on experience with the real object that Initial brake PUSH
is to be modeled, tested, or controlled. MORE REAR WEIGHT?
Otherwise some very real problems will LOOSE initial throttle PUSH §
leap out of the final product and bite you. | OTHER CORNERING PROBLEMS
|MORE FRT AERO DOWN |
There are a large number of factors we all BACK TO TIRETEMPS] DEFINITIONS 30 4 BACK 42,000

have to assume or take for granted to

simplify any complex analysis. But the
computer takes nothing for granted.
Don’t overlook theory. It will help you TRACK SETUP “AdVisor”
avoid reinventing the wheel. A little Good trackside decisions require: good aerodynamics, and driver technique.
literature research will tell you if anyone has data (analysed data), plus experience, and Finally, it gives an explanation why it
already attacked the problem or written the total recall of all factors -- and quickly. To should work (and why maybe not), plus
program, and what surprises may be in resolve this problem, I’ve created a new tradeoffs, and refers to more information
store. Or it may help you understand the category of computer applications for by page number in books like this.
inevitable questions which arise from the racing, the computerized race track setup How it works -- The logic is based on
very best research. advisor, or an “expert diagnostic system,” an in-depth analysis of which adjustments
Don’t over-trust your computer geek. If which runs on a handheld computer so are most effective for each combination of
someone promises to assemble a computer that it can be used with one hand while problems in each part of each turn. The
system or write a program, ask for the most actually working on the car. The intent advice is based on typical responses from
pessimistic cost/time estimate, and then was to “level the playing field” for racers highly experienced, pro racing setup
double it. It is unknown for any computer who were less inclined to read a textbook engineers -- except that it's faster, and it
expert, infinite in wisdom and logic they like this, but who need a “problem- never overlooks any important factors.
may be, to ever accomplish a goal within oriented” advisor, and can’t find or afford Since the basic physics are the same for
their own estimate. Even if it looks like it is a good human version. It’s what I call a most cars, only two software versions
approaching success — that is the point Pocket AdVisor and Data Analyst, or were required, either for sy~metrical cars
where adventuresome new ideas are tried, PADA™ (an elaboration on a PDA). on asymmetrical courses (road racing), or
setting the project back by weeks. These What it does - (ater basic shop setup asymmetrical cars on symmetrical courses
people tend to be oblivious to drop- to factory suggestions). Just like an (paved ovals).
deadlines like race day. expert track engineer, it first accepts tire No computer experience or manual is
Don’t over-commit to computer outputs. temperatures. Then it analyses them, and required, and there are hundreds of advice
They may make predictions or analyze data makes specific recommendations about screens. It only runs on specific low-cost
to one part in a thousand, and yet be off by a what chassis setup factors to change -- PDA’s, some of which are also
factor of two, because of some oversight or and how much, in the case of tire compatible with special snap-on modules
assumption in the program. Output should pressures, camber, and crossweight. appropriate for racing, like an integral
be considered a guide, or possibility, which Then it takes driver comments such as pyrometer, a 4-channel analog DAS, a
is to be evaluated by common sense, and "tight or loose", "in or out" of the turn, “g-9” friction circle analyst, and an
then ‘gingerly tested for credibility. If you and compares them to tire temperatures, accelerometer-based horsepower analyst.
put any more faith in a computer than an and makes more recommendations about The software starts at just a few hundred
idiot-savant off the street, then you may be shocks, anti-roll bar, brake balance, dollars, which is a lot cheaper than any
gambling your life or other serious invest- weight distribution, stagger, track bar, expert’s time. For the latest information,
ment on a 50/50 one-bit electronic flip-flop. spring rates, toe-in/toe-out, geometry, contact [email protected].
155
16

FRONT WHEEL DRIVE RACING

At the time of this writing, it looked like The first and most basic question is: will of the weight is on the rear. So far, the
front wheel drive was the hottest growth FWD ever be competitive with RWD in tendency has been for FWD racers to main-
area in high performance vehicles. Front racing? No. . .and yes. . .depending on the tain conventional appearances with equal
wheel drive cars were projected to even- car and the course. To compare the two, first tire sizes, and try to reduce front static
tually account for 90 percent of the market, let us assume all else is equal under the weight from perhaps 65 percent to 55
and the manufacturers saw racing as their rules, including power to weight ratio, percent to avoid extreme steady state
best chance to prove that the concept was, weight to tire size ratio, aerodynamics, etc. understeer.
in fact, practical for high performance. So Under these conditions there is no way on Second, consider straight line accelera-
to write this chapter, I researched the hell earth that a high-powered FWD car can tion and braking, with two hypothetical race
out of the problem, especially in the IMSA- accelerate and brake as well as a RWD car. cars, a RWD with 65 percent rear weight,
RS series for FWD sedans. As it turned out, As we will see, it simply violates basic laws and a FWD with 65 percent front weight,
the technology was in its infancy. Although of physics. Therefore, a FWD race car both with a 90 inch wheelbase, 13.5 inch
a number of talented people had a lot of ex- doesn’t stand a chance on a traditional road' high center of gravity, and 1.0g tires. Given
perience in the field, no one I spoke to was course, although perhaps it would on a high enough engine torque, the wheeispin limit
willing to confess to knowing any more than speed oval like Daytona or Indy. for the RWD car will be 0.76g acceleration
I did about the engineering theory or the First let us consider steady state turns. while the FWD car will only pull 0.56 g’s.
development needed. What seemed to be The fastest way around a skidpad or a high- Even if you figure a relatively low-powered
missing was the sort of basic research we speed oval 180 degree turn, would be for amateur sedan that can’t spin its tires above
did on RWD race cars at Chevrolet in the four equal tires to share equal loads and 20 mph, the same wheelspin handicap ap-
sixties. equal driving torque. Excess load or excess plies to the inside tires coming out of a turn.
However, no one seemed to be interested torque on any single tire will diminish its If you try to alleviate the wheelspin problem
in sponsoring such research (at least pub- cornering capabilities. In the case of Indy by adding more weight up front, braking can
licly). Eventually, I may find time to use one cars, with 60-65 percent rear weight and become nearly impossible. As it is, with 65
of my vehicle dynamics programs to isolate rear drive, the obvious solution is bigger percent front weight, and a 1.0 g lockup
and study the different factors — without rear tires. In NASCAR, however, the rules limit, only 20 percent of the weight remains
driver non-repeatability, and then validate require front engines and equal tires front on the rear tires. It can be very difficult to
the conclusions using my electronic in- and rear, so the front weight and rear drive control a car when the rear wheels lock up
strumentation on a real vehicle. As it stands, tend to balance out. So it would seem that — or come completely off the ground.
the book couldn’t wait, making this chapter FWD, with both front weight and front drive These are non-negotiable laws of physics.
rather theoretical and speculative, or full of is a double loser — unless much larger tires Occasionally someone will demonstrate
more questions than answers. are used up front or more than 50 percent that a FWD car can outpull a RWD car on

(Steven Schnabel)

The drivers of FWD and RWD cars may use considerably different techniques in getting through the same corner.

156
ice — which seems like a more extreme
traction problem than a Formula One car
on asphalt. But the dice are usually loaded.
Your typical RWD sedan might have less
than 50 percent static weight on the driving
wheels, while the FWD has up to 65 per-
cent. Add to that the fact that you can’t pull
over about 0.1 g on ice without studs or
chains (meaning very little weight transfer
is available to either hurt the FWD nor help
the RWD), and of course FWD appears
better.
Finally, what about the transient combina-
tion — which in racing is usually accelera-
tion out of a turn. As any new FWD race
driver will tell you, this is the first shock-
ing realization that things are not the same.
At the steady-state cornering limit in a turn,
the car understeers — perhaps a little more A demonstration of the limitations of front wheel drive acceleration. (Jeffrey Zwart)
than with RWD. But when you start to get
-hard on the throttle, it understeers a lot
more. This is vastly unlike the normal RWD by the same four tenths in acceleration (both different vehicle paths or “‘lines,” it appears
response of either a more neutral power-out were within a few tenths in overall lap- that there isn’t as much as a few tenths dif-
or power-induced oversteer. So the driver’s times). This indicates that there are some ference between most familiar driving styles
first puzzle is “How do I get the car to rotate vastly different but equal ways to drive a — for identical vehicles. But that may not
at the apex?” The front tires are already race track — perhaps depending on chassis hold true when both driving style and
saturated, and any acceleration will transfer setup. If we can understand why this is true, chassis setup are changed together. And at
load off of them, further reducing traction. then it may be possible to take the best of that, some of the quicker computer times
You only have one choice — decelerate both. I was told the biggest difference be- were on lines not commonly accepted as the
briefly to give the front a bit more corner- tween the two cars was that the better braker best.
ing power, which is devastating to laptimes. had a locked differential and the better 2) Differential Selections
(Or some rally drivers use the “handbrake accelerator had a limited slip — which Current popular thinking among FWD
turn” to cause the rear tires to lose traction sounds backwards. race drivers is that: A) Open differentials
for an instant.) Then the next problem is In preliminary computer studies of are unacceptable because they allow inside
how to deal with excess understeer when
powering out. Again, just one choice: First
85
gradually reduce steer angle and then apply
more power. You have to drive out of a turn
instead of drifting out.
But so much for the academic question 80 BASELINE: 67% FRONT WEIGHT
of FWD versus RWD, now let’s consider
how to make one FWD faster than another 2 INCH LOWER C.G.
FWD: 75
1) Driving style, or line
50% FRONT WEIGHT
Everyone knows that FWD race cars must
be driven differently than RWD race cars, 70
but there isn’t perfect agreement as to how.
This was made vividly apparent in a careful feet
examination of times and lines by top IMSA mph
in
speed
65
drivers through a 180 degree hairpin. If you
take total turn times, sometimes differences
aren't too obvious. In comparison between
two of the fastest Mazdas, total turn times 60
for one were 15.60, 15.72, 15.74, 15.94, (15.75
average) and 15.54, 15.62, 15.66, 15.76, (15.65
average) for the other. But the real surprise 55
was in comparison of turn entry and turn
exit. Timing segments were chosen to get
about 8 seconds worth of straight braking
and turn-in to the apex, and about 8 seconds
worth of accelerating out into a straightaway.
One of the Mazdas was consistently faster
by over four tenths of a second in the Figure 58. Computer-predicted transient turn speeds are useful for analysing dif-
braking segment, and subsequently slower ferent configurations of a front-wheel drive race car.

157
 mention other suspension considerations
such as roll-rate distribution and suspension
geometry.
A second consideration can be the “spec
tire.’ In some cases it appears that heavier
cars can saturate “specified” tires when too
much load is carried on the front. That is,
on hot days on hot tracks, they simply
overheat and blister tires, and the only cure
is to slow down or carry more weight on
the rear.
Finally, steady-state turn speed may be the
setup criteria. It appears that on a skidpad,
the highest g level can be reached with
Firestone
slightly more weight on the rear (with equal
tires front and rear). Front cornering power
which is diminished by thrust effects, would
then be balanced by rear cornering power
which is diminished by greater load — with
final trim setup by roll resistance distribu-
tion. But what is the tradeoff between higher
Inside rear wheel lockup when a lightly-loaded rear end is trail braked deep into minimum turn speed, and the potential abil-
a corner. However, low weight minimizes flat-spotting. (Geoffery Hewitt) ity to accelerate out earlier and faster?
4) Brake Balance
FWD has created some unique problems
front wheelspin. B) Locked differentials Accelerating out of a turn with an open to solve in front/rear brake balance. Most
allow faster exit speed but won’t “turn-in” differential, the inside front may spin, obvious is the extremely light rear weight
under braking and create a lot of steering putting a serious limit on acceleration. in braking, which may approach 15-20 per-
wheel fight. C) Limited slip differentials are However, this depends on minimum turn cent at the limit. Not only is this a stability
a good compromise, but are unpredictable exit speed, front/rear weight distribution, problem in straight line braking, it almost
in handling and weak in durability. traction, and torque availability. It should be guarantees inside rear wheel lift and lockup
Let’s look at what ideal requirements established on which cars and which tracks when braking deeply into a turn. (However,
might be, based on an analysis of moments this is not a consideration. Then again, a it may be that such lockup is not an impor-
about the center of gravity in transient study of different driving styles may show tant concern, because the light or zero load
maneuvers. that more distinct separation of cornering creates minimal flatspots). While electronic
Braking into a turn, a totally open front and acceleration produces equal or better anti-lock systems may be seen as an answer,
differential has the advantage of providing laptimes, raising the critical front wheelspin it remains to be determined whether such
equal left/right braking balance. That is, limit to a straight-line consideration. devices have other dynamic problems in
there is no oversteer or understeer moment, Accelerating out with a locked differen- brake-in-turn. Without cornering g sensing
and the car should respond freely to steer- tial is claimed to be faster. Theoretically, and compensation, it seems that they would
ing under braking. On the other hand, it more torque can be used, and the greater allow much greater retarding force on both
may allow the lightly-loaded inside front to pull on the outside wheel provides an outside tires — understeering or pulling the
lock early, although this is perhaps less of “oversteer” moment to counteract the car out of the turn just like a locked axle.
a problem than inside rear lockup. ' dreaded limit understeer. On the other hand, Second, the question of engine drag and
Braking with a totally locked front there is an unknown tradeoff balance. More downshifting must be re-evaluated. At high
differential creates an understeer problem outside front thrust also reduces front cor- rpm’s, engine drag aids the brakes (causing
as you attempt to turn-in. The more heavily- nering power — which is an understeer con- earlier rear lockup with RWD, or earlier
loaded outside front will be able to provide tribution. That is, you can’t take advantage front lockup with FWD) while at low rpm‘s,
more braking force, and will try to pull the of the greater moment to reduce understeer, engine inertia reduces the tendency to
car out of the turn (understeer). This has unless more cornering power is available at lockup at the drive wheels. If the driver is
not been a serious problem with locked the front, i.e. less weight or bigger tires than not to declutch entirely during braking, then
axles on RWD race cars, since the rear axle with an open differential. This definitely it should be resolved under what speed or
carries a much smaller percentage of the deserves a computer simulation of wheel conditions to set the mechanical brake
braking force. loads and moments about the center of balance, to avoid low speed rear lock.
Perhaps the best compromise in braking gravity. The wheel fight question will be 5) Lift/Drag, Track width tradeoffs
is a limited slip with only a fixed pre-load addressed under “Suspension Geometry.” It has been proven in theory and in prac-
breakaway (clutch pack or reverse ineffi- 3) Front/Rear Weight Balance tice that for high powered race cars,
ciency) and no response to pinion gear At first glance, it seems that FWD road downforce is about ten times as important
torque inputs. Or the Weismann may be racing cars require a much higher percen- as drag in reducing laptimes on road racing
more appropriate, as it may or may not tage of weight on the front to avoid courses. Yet in sedans, racers generally put
allow one-wheel lockup depending on wheelspin. However, this is strongly de- equal effort into improving both, going to
whether the engine is providing a retarding pendent on driving style and differential great lengths to reduce air drag. Perhaps a
force on the ring gear. (This deserves type, as well as the previously mentioned few simulations or controlled tests should
further analysis). turn speed, traction, and torque, not to be run on these cars also. It might be found

158
that within body shape limitations,
downforce is all important here also.
Probably very few racers even know what
the tradeoff is in lift versus drag on their
own cars. That is, what is the drag penalty
in pounds due to a downforce gain in
pounds, and vice-versa.
A related question: What is the tradeoff
between air drag and cornering power (car
width)? Many racers intuitively widen their
cars to the limit to get wider tires and a
wider track, while others do just the
opposite. They try to keep the tires inside
flush body sides for the lower drag it pro-
vides. Which is better in overall laptimes?
Or is it possible that either one is correct,
depending on the class or power-to-weight
ratio? A side profile indicates that the more successful front wheel drive race cars are
6) Front/Rear downforce distribution tending to use more aerodynamic downforce at the front.
All RWD racers work toward an extreme
rear bias in aero downforce. The rear is
where the traction loss is, and that is how wheel camber and straightaway accelera- understeer by overloading a front tire, but
you reduce the possibility of dreaded high- tion/braking camber. What are the values conversely it decreases understeer because
speed oversteer. Theory indicates that the of this tradeoff? Is front camber unimpor- the more heavily loaded tire is now driving
net aero force vector (total downforce plus tant in straight line acceleration where small instead of dragging. The same outside tire
air drag) should pass through the center of sedans are not limited by wheelspin and drag effect also adds to understeer at the
gravity, or that the downforce ratio should have less front lift due to weight transfer? rear, although with less effect due to less
be about equal to the static weight distri- And how much does excess negative camber weight there. The current practice of adding
bution. But again, do any sedan racers even hurt front braking traction with a lot of nose rear roll resistance until the inside rear just
know what their current downforce distribu- dive — or can that be reduced with anti- barely comes off the ground in the worst
tion is? dive geometry.? case — braking into a turn — may be a good
Complaints have been heard that it is dif- Front/rear roll resistance distribution, rule of thumb. But this could also be a func-
ficult to balance FWD oversteer/understeer which is all-important to RWD race car final tion of the first three questions: Driving
stability in low and high speed corners. balance, may not be quite as useful a tool style, differential type, and weight distribu-
Perhaps it has been overlooked that in FWD. In RWD, a bigger anti-roll bar up tion. It may be that other chassis changes
mechanical fixes should be used for front provides more understeer for two are more practical for final oversteer/
mechanical stability problems (no change reasons: it reduces front cornering power understeer balance.
with increasing speed) and aero fixes should by overloading one tire, and it increases drag Steering wheel feel, or feedback, is
be used for aero problems (both change with on the outside tire, trying to steer the car generally more important in street-driven
speed squared). out of the turn. But in FWD a big front bar passengers cars, although it has been men-
The first thought is that perhaps FWD has conflicting effects: it still increases tioned as a significant problem in race cars.
should be the opposite of RWD, in having
a higher downforce on the front — in the
same ratio as the weight distribution. This
corrects for the drive thrust traction loss at
the front and should maintain a fairly neutral
balance, although it ignores the conse-
quences of high speed aero oversteer.
7) Suspension Geometry — for
cormering power, balance,
and feel
All FWD cars, race or otherwise, are
generally limited by front wheel traction,
making rear suspension analysis relatively
unimportant. And since all current FWD
race cars are derived from production
sedans, MacPherson struts are the major
problem. Certain considerations carry over
from RWD race car practice, such as harder
bushings to reduce compliance and modified
camber change curves to keep the tires more
vertical in cornering. However, this is
sometimes limited by the rules and the cor- Front wheel static camber can appear extreme when it has to compensate for the
responding tradeoff problems with inside geometric disadvantages of the MacPherson strut.

159
 transient response. If the theory turns out
to be true, then the effect should be
evaluated with respect to FWD in
competition.
IN RWD race cats, long straightaways
allow the front tires to cool off more than
the rears, which are more shielded and
under driving thrust. Therefore, unless there
is severe braking before the next corner, the
car will tend toward more understeer at first,
evening out toward neutral as the front tires
heat up. But in FWD, it is the rears that
should cool off more, creating initial
oversteer. The situation is even more
extreme at the end of a long straightaway,
where in addition, the front tires may be
absorbing more than 80 percent of the
braking energy. So at the start of the turns,
The upper limit of rear roll stiffness or front weight bias, which seems to be the the front tires may already be temperature-
accepted setup standard in front wheel drive race cars. (Dennis Ashlock) saturated, or over their cornering capability
peak. When good data and simulations are
available on RWD racers, it should soon be
Sometimes, especially with locked differen- and rear suspensions in alignment with applied to FWD as well. This analysis leads
tials and a large front wheel scrub radius respect to each other, and to give anti-roll to the possibility that the smooth “blend”
due to wider tires, even race drivers have bars a stiff base to act against. But if FWD style of braking and cornering may not be
complained of wheel fight. Solutions to this rear suspension is just along for the ride, appropriate for FWD. It is possible that
would seem to be obvious, if difficult with and anti-roll bars are no longer terribly totally separate braking and cornering
the geometric limits of MacPherson struts. effective, then stock chassis stiffnesses could phases may be as fast.
Simply reduce the scrub radius and caster be adequate. It could be argued that the roll In retrospect, the engineering analysis of
as much as possible. But it is surprising to cage has to be in there anyhow, and it might RWD race cars eventually proved that it was
find that one important aspect seems to have as well be put to work. But if overall'weight hard to make a mistake with them in basic
been ignored. Static geometry analysis is far is critical, why make the cage more than chassis setup. Most factors were obvious or
from appropriate, especially in a race car what is necessary to protect the driver’s self-correcting in their inherent dynamics —
under high g loading, The centroid of the area? at least once everyone got over their basic
tire contact patch surely moves a matter of Center of gravity is a factor here also. We mistrust of rear-bias, rear-engine racers in
inches under severe fore/aft and lateral can jump to the conclusion that a low center the early sixties. Proper static weight,
loads, when it is most critical. It may be of gravity is more important than ever in dynamic weight transfer, aerodynamics, and
that this effect can be used to advantage, or FWD race cars. While a high center of driving styles turned out to usually be in-
that it can be exploited through the use of gravity can help in RWD acceleration and tuitively correct. Now, although FWD
more innovative geometries such as the braking, it is a constant loser in FWD. So seems to be better in many ways for naive
BMW two-lower-link “virtual center” Mac- if there is no negative tradeoff to lowering street drivers, it is going to be extremely dif-
Pherson strut system. the center of gravity in FWD, perhaps the ficult to make these cars right for the track.
8) Torsional Stiffness and extreme coverage in current roll cages is one Ultimately 4WD may become com-
CG. Height of the easiest places to reduce top-heaviness. monplace enough to be a production road-
Is FWD chassis torsional stiffness as im- 9) Tire Heat Transients race possibility (especially if tires are
portant as it is in RWD race cars? Its Research is under way to determine the limited). If so, we may find it impossible
greatest value has been in keeping the front effect of varying in-use tire temperatures on to screw up that layout.

160
 17

RACE CAR OF THE FUTURE

If you happen to think that race cars are be more sophisticated than on Formula cars. surfaces could move at the driver’s com-
the least bit exciting, the race car of the Various aero prototypes from G.M. and mand, or as an automatic function of car
future will vaporize your cortex. Rules Ford already have surfaces (or in fact whole speed or cornering g’s, to decrease air drag
willing. body movements) which can be altered to in the straights and increase downforce in
Over the years I’ve done quite a bit of improve aerodynamic drag or downforce at the corners. We could forget about the
research and writing about the future of speed. And just look at open-wheel cars in tradeoff problems in lift/drag. We would see
technology in transportation, and if there’s general. Most of their designers would love moveable flaps, spoilers, ducts, and perhaps
anyone who doesn’t want you to know about to cover those massive spinning rubber even air drag brakes again, which Jaguar
their future plans, it’s race car engineers. cylinders with aerodynamic bodywork — pioneered at LeMans in 1957. Such devices
To them, however, the future is seldom but the financial interests feel it would are neither too expensive nor too dangerous
further away than next season. If they have destroy a certain image. for the cheapest homebuilt aircraft.
an innovation in mind, there’s no reason to Speaking of rotating cylinders, one which But even within the letter of the rules,
hold anything back (except perhaps a time was carefully located at the rear edge of a there is a lot of unplowed ground. A surface
or budget limit). And when you consider body might generate enough airflow on its doesn’t have to be moveable to have varying
the volatility of rules, who can plan much surface to draw a large amount of free- aerodynamic characteristics. Boundary layer
more than a season ahead? Whatever man’s stream air around a corner. The cylinder vacuum or booster devices can control
mind can imagine, man can build — but not might take up a great deal of space and airflow by moving air through strategically
if there are artificial restraints in the name weight, however, and must have a surface placed holes or slots in the body. Slow-
of safety, by sometimes capricious velocity many times the velocity of the moving air can be sucked off the surface or
rulesmakers. Once upon a time there was vehicle. Driving such a cylinder would high-velocity air can be blown out and
a series called the Can-Am, with unlimited require some complicated mechanism and downstream to counter skin-friction drag.
“knife fight’ rules, which spawned would consume a small amount of power. Of course, it is difficult to determine just
moveable aerodynamic surfaces, vacuum Another disadvantage is that it is probably where and what needs to be done, and it
traction, 8.7-liter V-8’s, and 1200-hp an illegal “moving aerodynamic device” does require some power to move the extra
turbomotors. But now the upper limit to under current international racing air. The power source might be the inlet to
racing technology is often determined in regulations. a supercharger, or a separate suction/blower
committee meetings and courtrooms. If the arbitrary and outdated international air pump might produce a net advantage in
rule against “‘moveable aerodynamic critical situations such as on an airfoil. And
AIR FLOW surfaces” were forgotten, we would if it’s not acceptable to utilize existing air
Aerodynamics is a perfect example. How immediately see some phenominal power sources such as turbo inlets and
ironic that we are moving into an era where technological advances (or at least spin-offs exhaust outlets, then “fluidic devices” may
passenger car aerodynamics are allowed to from the aircraft industry). Large body be the answer. In this case, a venturi tap

(Ford Motor Company)

Ford’s Probe IV has a front air dam which automatically lowers with increasing speeds. Legal on the street; illegal on the
track. It also has moveable front wheel fairings which would probably also be unacceptable on a race car under current rules.
 ground effects, exposed spinning wheels,
open cockpits, and external wings, the
optimum solution may take a little longer
to reach. There is a strong indication that
lateral airflow considerations, or large
controlled-vortex flow, or a toroidal airflow
path may be productive new approaches.
For example, wherever the upper body
pressure starts to go negative, deflect it
around the sides. In other words, the airflow
might be impacted on a sloping nose to
create downforce pressure, and then be
“screwed” around to the sides with perpen-
dicular wing stubs or a cascade of vanes,
where it would create balanced Jateral lift-
ing pressures instead of vertical lifting
pressure on the body top.
RETURN
TO THE TUBE FRAME
Race car chassis’ have had three great
revolutions: from the platform or ladder to
the space frame, from space frame. to
monocoque, and now to aerospace analysis
and composites. Each year’s chassis appears
to be the ultimate — until the next innova-
tion comes out. How can we imagine what
the next great revolution will be?
Somehow it seems that monocoque may
have overstayed its welcome. In the be-
ginning, these structures saved weight by
doubling as the body, or external surfaces.
But now, with aerodynamics decisions
changing frequently, redundant surface skins
are necessary, and are often added and/or
modified from race to race — as the
monocoque definitely cannot.
But monocoque may not even be the most
efficient structure, depending on the specific
loads. True, in aircraft it is almost universal,
where loads are primarily aerodynamic and
evenly distributed. But in a race car you
have mostly “point” loads, at spring and
shock and suspension and bodywork and
engine mounts. These point loads must be
These smoke tunnel experiments illustrate the possibility of reducing drag on bluff
distributed over a large area in a mono-
bodies by sucking the core out of trailing vortices. (Lockheed)
coque. To optimize such a structure, the
panel areas which are not in a direct load
provides the vacuum source, and pneumatic At the moment, it hasn’t been demonstrated path between points must be reduced in
equivalents of amplifiers and switching gates that the reduction in drag is greater than the thickness — ultimately to where they do not
can be used to control airflow — with no power required to pump the necessary exist, and you eventually end up with an
mechanical moving parts. And the same volume of air. The extreme example, of open (tube) frame again. Someday, someone
devices can be controlled automatically by course, would be a gigantic pump to remove in Formula/Indy cars is going to say, “Hey,
fluidic logic, as they sense speed or lateral all air from the front of the body and push why do all those production-derived race
acceleration. it out the back — in other words, a jet. cars not use the existing production
A final device is another suction system And yet within the spirit of the rules, monocoque shell — but instead install
which reorganizes the turbulent separation there are plenty of untried fixed body elaborate tube space frames inside?” And
airflow at the rear of a body. As the air shapes. The ultimate goal is relatively when someone finally figures out a way to
separates from the body, it forms cyclically simple: to maximize air pressure above and efficiently join non-weldable composite
alternating vortices which spread out in the behind the body, while minimizing it to the tubes, as in some of the newest bicycle
wake and dissipate drag energy into the air. front and bottom. With free-air shapes (no frames, we will see composite space frames.
By sucking the heart out of the first vortex, ground effect) it is easy to reach the Lee Dykstra disagrees, however, citing the
and pumping it out longitudinally behind the optimum. In fact it can be done by com- use of machined alloy plates to spread
body, Cornish has eliminated successive puter iteration, using aerospace 3-dimen- point loads. So a better predictionis ‘‘the ee
e
a e

vortices and improved the trailing airflow. sional potential flow panel analysis. But with right construction in the right place.’’

162
WHEEL CONTROL
Suspensions have reached the point where
they don’t offer much potential growth in
race cars, considering they were essentially
eliminated for a period a few years back,
when aero downforce became too great. But
now it looks like there are two possibilities
in the near future.
Rising-rate suspensions — avoided
because of chassis balance. problems —
could turn out to be a good solution to the Figure 59. A single-spring rocker-arm suspension designed for a variable rate
downforce/travel problem after all, once spring. All roll resistance must be taken by an even stiffer anti-roll bar.
designers face the realization that ride rate
and roll resistance can be totally separated designed for ultimate ride in passenger cars carbon/Kevlar monocoques are almost to
— and perhaps should be. Where left/right gets borrowed and adapted to someone else’s that point now. Perhaps what would be ideal
pairs of spring/shock units are centrally race car. A related example is production is a stronger cocoon with a final brittle frac-
located — almost touching — why not use computer-controlled fuel injection systems ture limit — such as carbon composites —
just one? The chassis could be hung from now spilling over into race cars. And even- so that at the ultimate survivable impact, the
one spring at each end with any rising-rate tually, the artificially-intelligent look-ahead- shell would disintegrate like an eggshell,
desired (with appropriate linkage changes; sensor system will be used to anticipate setting the driver free. But it will take a lot
see Fig. 58), and only the anti-roll bars and corners, braking, and bumps, and pre-set of development to determine the proper
shocks would control roll and individual the suspension to cancel them as they are “destruct” point.
wheel damping. I hinted at this in the first encountered. So far there has been some resistance to
edition of Race Car Engineering, and full roll cages over the drivers of Formula
actually saw an example in Rich Sloma’s SUDDEN IMPACTS cars for the fear of trapping them. But it is
Trans-Am Corvette in 1984. However, in this When push comes to shove, impact interesting to compare current fuel dragster
case the purpose was to reduce total rear roll protection is as important as chassis chassis. It is not uncommon for these
resistance, as is done at the rear of Formula torsional stiffness in a race car structure. machines to go out of control or actually
V cars, rather than to allow rising rate. In You need a certain amount of yield, to explode at over 250 mph. Either the tum-
any case, Sloma reports that it is a great absorb the impact, but then you need a final bling or the engine destruction often breaks
luxury to be able to tune roll rate and ride limit that says ” go no farther” or it may the chassis in half, allowing the roll caged
rate independently. collapse upon the driver and crush or trap driver capsule and nose section to go its own
“Active suspensions” may still be im- him. So perhaps deformable metals will way. Granted there are seldom barriers or
portant in the future, even if they have always be more practical around the other race cars in the path, but then Formula
been outlawed inanumber of racing series periphery, with a very stiff cocoon simply cars aren’t crashing at 250 mph either.
at this writing, to eliminate their outra- encapsulating the driver. Open-wheel Nobody accuses dragster drivers of being
geous development costs. Again, we see
a situation similar to aerodynamic regula-
tions, where passenger cars will soon have
more sophisticated suspensions than all
but F1 race cars. However, there may be
difficulty in defining and enforcing these
restrictions, due to the previous accep-
tance of electronically-controlled shock
absorbers. Already, “fast-shocks” can
electronically adjust the damping dozens
of times over one bump input cycle. With
enough dollars, an “active” system could
probably be hidden within astandard shock
absorber envelope, using bump inputs to
store energy, and a micro-processor to
release it at the proper time and directions.
Or, fully electromagnetic dampers might
eventually become powerful enough to
actually drive the suspension, and be to-
tally self-contained except for the electric
power supply. As passenger car suspen-
sion electronic control becomes more com-
mon, and the problems of sensors, valves,
computers, power supplies, and logic are
worked out and amortized in production, An early Formula One Lotus hydraulic rear suspension, showing rams on the in-
it could be very reasonable that a version ner rockers, height sensors, and accumulator. Complex but effective.

163
 not cast or layed-up composite components,
but actual filament-wound parts. This is
already common practice for cylindrical
aerospace objects such as rocket cases and
production automotive driveshafts. Perhaps
piston cylinders aren’t that far removed. And
imagine a one-piece connecting rod, made
from one continuous filament wrapped
around both the pin and big end. It would
require a take-apart crank, but even that has
been done already. If thermal properties are
seen as a problem, consider the latest
breakthrough: metal-matrix composites.
These are carbon filaments imbedded in
lightweight metals such as aluminum,
instead of epoxy. . .the fabricated indestruc-
table cylinder.
For really high temperature parts such as
piston crowns, exhaust valves, and turbo-
chargers, ceramics are already approaching
production. But it seems as though carbon/
carbon composites (now used for aircraft
and race car brake discs) should be making
some inroads here. A motorcycle company
recently experimented with sintered carbon-
fiber pistons, but the aerospace industry has
fairly well wrapped up all research facilities,
Polimotor, the first composite plastic racing engine, which weighed only 160 pounds and the U.S. Department of Defense has
in its 2-liter, twin-cam configuration. (Amoco) classified all information. When these
materials do become more available, it will
be a close race between carbon/carbon and
sissies, and their cages haven’t dampened remains: how to anticipate the need and reinforced ceramics.
spectator appeal. It has to happen in open- trigger the system in time. Racing engines may also have truly dry
wheelers eventually. sumps someday. The use of ball-bearing
Fire control is almost good enough, with PLASTIC ELECTRONIC rods, camshafts and followers, and more
the current technology in fuel cells, dry- TURBOMOTORS wear-resistant piston-ring and cylinder
break fittings, and fireproof clothing. But Racing engines of the future may be com- materials may make engine oil superfluous.
a possible final step is chemical treatment pletely non-metallic. The filament- (If a take-apart crank is necessary for one-
of fuel. NASA has been experimenting for composite blocks made for Ford by piece rods, then roller bearings are also
decades with chemicals which will cause Polimotor, however, are just a tentative first possible for main bearings.) Most people
fuels to suddenly jell, or become more step. When the basic materials are finally don’t realize that oil is used for cooling as
viscous, so that they neither spill nor proven, then more design and tooling much as for lubrication. In fact, Porsche
vaporize rapidly. One small problem sophistication will be justified to produce racing engines perhaps should be called oil-
& if cooled rather than air-cooled. But if oil is
not needed for lubrication, why not just
open the crankcase up and air cool the
bottom end also? You will never see this
happen in production engines however, due
to the problems of piston blow-by emis-
sions and radiated engine noise.
Fuel and ignition control systems wili
require an engineering sophistication
available only from major manufacturers.
But we are already seeing this support from
Porsche, BMW, Renault, Honda, and
Buick. So racing advancements will
probably follow production component
availability. One idea likely to be used is the
replacement of electronic computer control
modules by optic and fluidic equivalents, to
overcome heat, vibration, and electro-
magnetic interference. Since this is already
Mono-filament wound carbon reinforced connecting rods, such as these under recognized as a serious problem in military
development by VW, are a racing possibility for the future. (Volkswagen AG) aircraft, a solution is surely not far away.

164
Future opto-fluidic computers will be
integral with the engine, and the only ex-
ternal connections will be a throttle control
and fuel line. At the same time, valve timing
will also fall under computer control — first
electronically, then via fluidic logic and
hydraulic control. Finally, as more fuel
consumption limitations are imposed, the
onboard computer logic will be given the
added responsibility of calculating and
predicting appropriate instantaneous power
limits — to insure finishing the race on the
last drop of fuel.
Turbochargers have gotten race cars a
long way into the future already, but they
are potentially so much more effective (and
promotable in passenger cars), that we will
see a lot more development yet. The lag
problem is more important on the street than
in an engine which seldom drops below
6000 rpm, but still there are some innovative
possible ways to keep turbo pressure up.
Some Formula One teams have developed
bypass loops which recirculate gases (to
reduce backpressure drag on the blades) Ceramics are a good possibility in very high temperature racing applications, such
when the throttles are closed. When they as these experimental NGK turbocharger blades. (NGK Insulators)
find some way to create a temporary vacuum
at both the turbine and compressor wheels,
then they might be able to maintain turbo compounding, or coupling the turbo shaft reciprocating part of the engine) which
rpms clear through a pitstop. Who knows, to the driveline, is already common in merely converts turbine pressure into
maybe there is some way to provide an on- aircraft. But in race cars, where there are relatively low rpms at the wheels.
board temporary reservoir, to accumulate transient problems in accelerating and
excess (blowoff) pressure, and meter it back decelerating, the necessary high ratio gear SHIFTLESS TRANSMISSIONS
just at the appropriate instant. train is more of a problem. Here we see a Racing transmissions haven’t changed a
If rules were no obstacle, the most prac- potentially better application for the great deal in principle for about the last
tical trick might be a supplementary nitrous infinitely-variable ratio driveline, in coupling century. Torque converters made a brief
oxide system. The engine control module turbine to engine rather than engine to axle. appearance in Chaparrals in the 60’s, and
could meter out a quick blast of nitrous Still, we are rapidly approaching the point transverse transmission transaxles were the
oxide just for that fraction of a second where where the turbine becomes such a hot setup for a while in “tunnel cars.” There
the boost lag hole is. It probably wouldn’t dominating factor, that it is considered the shouldn’t be many shocking surprises in the
take more than 10 pounds (100 seconds) of prime motivator — via a piston pump (the future either, since the most exciting
nitrous oxide per race, and be well worth ere

the added pressure bottle weight. Who

knows, maybe less ethical teams are already
smuggling such systems aboard.
Turbos are just the first step in recovering
some of the waste heat that is poured out
the radiators and exhaust. A hot topic in the
industry these days is adiabatic or low-heat-
rejection engines, which are actually
insulated to retain heat, and utilize ultra-high
temperature materials like ceramics to avoid
melt-down. Unfortunately, it looks like
diesels are the only practical application,
due to the heat intolerance or knock sen-
sitivity of gasoline.
Actually, as turbochargers become more
and more effective, we may see them
developed into turbine engines. When you
consider the fact that 1.5-liter turbocharged
Formula One engines are more powerful
than 3.0-liters without a turbo — you realize
that the turbo is already accounting for over Variable turbocharger inlet blades, indicated by these external pivots on a For-
50 percent of the power. Turbo- mula One Ferrari, improve response and broaden the useful torque band.

165
 million dollar development effort which was
recently abandoned by McLaren. Even
high-pressure steel disc friction devices are
being tried, although it’s hard to imagine
them capable of being scaled up very far.
As a sidelight, Formula One racers who
have gotten to field test protype PIV
transmissions have been shocked by fuel
consumption increases. Manual shifting
may give slower acceleration, but it con-
serves a surprising amount of fuel in engine
“downtime.” So better acceleration, via PIV
or whatever, may be ultimately limited by
fuel efficiency requirements.

TIRES, WHEELS, AND BRAKES

Racing tires may not get much wider —
even though we have thought that for
decades. But now most sanctioning groups
have put limits on wheel and tire dimen-
sions. So now the tire companies have to
concentrate on less obvious and more
difficult technologies, such as cord layouts
and rubber compounds. The potential of
dynamic tire temperature analysis, however
(see Chapter 2), leads to some other far-out
possibilities. If transient traction is in fact
limited by the time it takes a tire surface to
heat up in a corner, is it possible to take this
awareness and use it to “fill in” the missing
performance capability? Perceived necessity
is the father of invention. Let’s say it could
be desirable to artificially heat the tires more
rapidly than is done naturally by acceler-
ating, braking, or cornering, and cool them
faster than is done by natural cool air flow.
There must be dozens of ways to do it, even
Continuously-variable transmissions are a possibility in racing, ‘with the advent within the limitations of racing rules.
of more positive drives such as this metallic belt. (Fiat) The first thought is that heat is available
from radiator ducts and even, perhaps, ex-
haust pipes, if they could be varied, or
possibilities have been dreamed of for wisest decisions, so ratio selection will controlled. And, if not limited by regula-
decades. The automatic-shift manual trans- have to be computer-controlled. That is, tions about “moveable aerodynamic
mission is just arriving in F1, now that they the computer’s track memory will pro- devices,’ maybe cooling airflow could be
have onboard computers and servo sys- gram itself in practice to always be in the regulated in some way. One far out possi-
tems to take care of shift logic. right gear at the right time — with avoid- bility is electronic control and heating.
Another type of "quick-shift" transmis- ance of any great torque changes (up or Imbed resistance wires in the cord body, the
sion has been developed by Weismann, down shift) during cornering or braking. same as “‘steel belts,’ and provide them with
which has a roller clutch engagement con- The positive infinitely-variable (PIV) thermostatically-controlled current via slip
tained totally within each spur gear hub. In transmission and electro-static clutch fluids rings. (Someone has surely dreamed this up
addition to reducing size and weight by have also been “‘just around the corner” for already as an aid for winter driving.) Going
literally stacking all gears face-to-face, it quite a while. PIV seems to be made in one step further into outer space, it might
allows instant electronic selection of any heaven for racing, where it would be nice be possible to utilize all that wasted heat
ratio. Perhaps the ultimate spur-gear trans- to run an engine at one peak torque rpm all from nearby brake rotors. Just find a way
mission will be a series of separate 2- through its acceleration speeds, running to open and close a conductive. pathway
speeds that can be selected in all possible through a continuously-changing transmis- from the rotor to the rubber. They really
combinations. With two 2-speeds, there sion ratio with absolutely no power-robbing aren't that far apart. However, when I
would be four ratio combinations; eight slippage. Such transmissions are just now showed this section to Goodyear engineers,
ratio combinations with three 2-speeds; going into production for low-power they refused to comment, and instead re-
and so on. And with electronic shifting econoboxes, although with little more ferred me to their patent lawyers.
there would be no handicap to changing exotica than variable-sheave belt-drives for Race car wheels are always ripe for inno-
gears many times per second. snowmobiles. Non-slip mechanical versions vation, since they seem so straightforward
Given this over-abundance, the driver in the 300-500 horsepower range won't and easy to produce. Current modular
can’t be expected to always make the arrive soon, in spite of a rumored multi- designs appear to give anyone all the con-

166
figurations they could desire. But there is
still potential for improvement in terms of
aerodynamics and weight, especially con-
sidering that a pound off the wheel rim is
worth two pounds off the chassis (due to
rotational inertia effects). Filament-wound
composite wheels of carbon/Kevlar/epoxy
have to be the answer, when someone
figures out how to fabricate them. It will
take a computer-controlled filament-winding
machine with a shuttle perhaps free-flying ah
in 3-dimensions. Then, of course, it will 59 Le Tie emetic 3G 23.6 SEC

probably have to wind around a dissolvable + 17 Se

core or mandrel to produce a true solid-

surface monocoque wheel. The solid
surface wheel disc will also be necessary
to reduce air drag — unless someone comes
up with a far more efficient vane-pump
spoke design to help pull air out from under
the chassis.
Brake discs have related design problems.
The solid carbon/carbon composite disc is
rapidly taking over in Formula One, even A computerized ‘‘expert system”’ such as this one being developed by the author
though it is thermally not very efficient as may provide far more valuable and more immediate information for the crew.
a solid disc. Manufacturers are coming up
with ways of making them vented, as all
cast-iron racing rotors are, even though they far out schemes myself. would have accumulated statistics on all of
are extremely difficult to lay up and sinter The race car of the future will be well them, such as liklihood of breakdown,
in complex shapes, and extremely tough to supported by the computerized pit crew of driver consistency, and laptime and pitstop
machine or drill out after sintering. Due to the future. We will be seeing a lot more potential. During a race, then, it might
the poor thermal conductivity away from computers at trackside, not just for conven- automatically note, ” Car 27 is the nearest
rubbing surfaces, double discs at each wheel tional timing and scoring, but as expert potential threat.” And after a simple input
may be the ultimate solution. Another strategy advisors. Breakthroughs in artificial of ‘““Why?”, respond with, “Car 27 is 32
fascinating prospect for carbon/carbon is in intelligence have produced low-cost “Expert seconds back, gaining at 0.6 seconds per lap,
multi-disc racing clutches — in both drive System” frameworks which can be used on with 80 laps remaining, can potentially pit
and driven plates — due to its strength, low an ordinary IBM PC. In addition to taking 6 seconds faster, and is statistically very
density (low inertia), and high temperature over current calculator operations such as reliable.’ Or, being able to constantly
friction characteristics. when to fuel and how much to put in, they predict the location of every car on the track
will begin to make strategy recommen- at every instant, it could make lightning
ELECTRONICS AND EXPERT dations such as whether and when to come strategy recommendations such as, “There
COMPUTERS in on a yellow or to change to rain tires. I will be a clot of five slower cars ahead of
Although electronics may eventually be recently tried to use one first-generation our car in the esses on the next lap. Expend
proven inferior to fluidics and optics in program during a race, and found it merely every effort to pass before then.”
engine, transmission, and suspension logic allowed me to make mistakes with greater Such machines will probably never take
and control, it will become far more precision. But they will happen, and soon. over, because there are just too many
valuable in real-time race car test and feed- There are a number of valuable features variables to program, but they will definitely
back. The current use of computerized data- already available in Expert Systems. First, become very important support, offering
taking and storage for later analysis was they can accumulate the knowledge and continual surveillance and advice, and the
covered in Chapter 15, Computers in rationale of not just one, but many expert analysis of more information than is hu-
Racing. But there is a great deal that can team managers. And they may be fed this manly possible.
be done with such information in continual expertise not just by interviews with often
real-time onboard analysis, manipulation, inarticulate master mechanics, but by
and and feedback to the driver. Why distract observation of even a competitor’s opera- CONTROL
the driver with data? Well, the information tions. Then for any given situation, the Driver controls, such as throttle, steering,
doesn’t have to be presented digitally on a average or best possible decision may be brakes, and clutch, may eventually have
screen, any more than engine temperature derived from all of these simulated experts. their own competition between “more for
or rpms are. (Forget the faddism of digital But one of the most useful features in this the driver,’ and ‘“‘none for the driver.” The
instruments in passenger cars.) What would business, is that Expert Systems will operate first approach is to give the driver more and
the driver like to know, that would help him on imprecise “‘fuzzy data” and probabilities, better control over vehicle dynamics, such
win the race, and how can it be best pro- as opposed to unattainable absolute facts. as cockpit adjustable anti roll bars and
vided? I’m not going to sell out cheap here, In operation, such a system might be variable brake balance have been already.
because it’s relatively easy for hobbiests like programmed to continually monitor the laps The second approach, conceivable but
me to develop electronic hardware in our and laptimes and pitstops of every car on further away, is turning such controls over
spare time, and I’m going to try out some the track. Also, after a number of races it to the computer, just as spark and fuel

167
 Consider braking into a turn. If both out-
Main Switch Battery side tires are allowed to brake at their
maximum traction capability (with the
increased load), and both inside tires are
braking at their lesser maximum, then the
car probably cannot be steered into the turn,
but will straighten itself out. Perfect
(unwanted) stability. But if the driver wants
Control to steer and brake at the same time, why
not give him one control device to do both
Valve & — the “brake-steer pedal.”
rl Naturally, it’s probably beyond the
capability of even race drivers to take
advantage of such systems in real micro-
second response time. This is why com-
=, puters are now commonly used for engine .
ee as Auto & Crab
control. So again, it’s inevitable that
“stability computers’’ will be first developed
Select Switch for production vehicles (where average
driver capabilities are abysmally lower) and
then adapted for racing. The four individual
brake servos, and the two (or four) differ-
Four-wheel steering, under development for production cars, will make race car ential bias servos will be controlled by a
handling more consistent with aero, fuel, tire, and track changes. computer, based on signals from the steering
wheel and individual wheel speeds, not to
mention the possibility of a solid-state (non-
control have been. This is not only possible, hesion, the brakes and throttle are as gyro) yaw rate sensor — which will replace
and inevitable, but the first wave is already important as the steering wheel. This is due the old “seat-of-the-pants’’ sensor.
here. Various current production sedans to their effect on the yaw moment or rotation In a race car application, such a system
have computer-controlled anti-lock, anti- of the chassis due to changing lateral and might be capable of setting up and main-
wheelspin, transmission shifting,and power longitudinal forces at each tire, and not just taining the legendary perfect “‘four-wheel-
steering boost. And in some cases, like the steer angle of the front wheels. drift” without any steering at all. To go
braking, computer logic can already out Let’s suppose we could give the driver through a turn, the driver aims with the
perform most racing drivers. control over the drive torque at each rear steering wheel — which simply sends a
Given the race car driver’s obvious goal: wheel, with a variable differential bias. Say signal to the computer — and it decides how
accelerate and brake as rapidly as possible he could lean his throttle pedal to the right much inside braking to apply to start
while keeping on a well-defined path, his to send more torque to the left rear wheel, rotating the car, how much total braking is
skill and concentration are focused primar- thereby steering the car to the right — required to keep the car from losing ad-
ily on maintaining the limit of stability. perhaps without moving the steering wheel hesion on that radius of turn, and how much
Modern fighter aircraft have already pro- at all. Of course, too much pedal will do more outside rear wheel torque is necessary
gressed beyond this point, in increasing the the same thing — throttle oversteer — but to accelerate out on the required increasing
speed and response of aircraft to the point with a corresponding acceleration which radius.
where they absolutely cannot be flown by might not be desirable — say in a steady- Actually, all four wheels probably will be
mere humans — without the support of state turn. steered, imperceptably, or within a few
computer-controlled feedback systems. If all Perhaps theoretically even more useful degrees, by the computer, to keep each one’s
computers were to go out at once, the plane would be a 4-way brake valve, which pro- slip angle at the optimum for its load.
would immediately go out of control and portioned braking force to each wheel Such control systems will probably
crash. Whether the goal is to win a war or depending on where the driver positioned require a new breed of driver — the naive
to win a championship, if the computer can his foot. The driver could modulate the non-race driver who never learned to
do it better, it’s just a matter of time. force distribution so that a single lightly- manually control a car at the limit. And
Steering is assumed to be the race driver’s loaded tire (such as the inside front in a before you say “Over my dead body,’ ask
primary method of choosing his path. But turn) would not lock up early. Well, this is whether it is more undesirable to utilize
that oversimplification is obvious when he exactly what a 4-sensor, 4-modulator anti- such monitor and control technology — or
gets into a front-wheel-drive race car and lock braking system does — but it may not to have racing Czars say it shouldn’t be
discovers how much directional control is be what is wanted in terms of stability. (see done. And ruin the fun of all those
suddenly lost. In fact, at the limit of ad- Chapter 5) aerospace engineers.

168
 ' ABOUT THE AUTHOR
Paul Van Valkenburgh has been a draftsman, machinist, designer, modeler, fabricator,
engineer, aerodynamicist, test driver, race driver, programmer, author, columnist, illustrator,
photographer, editor, publisher, industrial film-maker, and proposalist;
.... working on the vehicle dynamics, tire dynamics, aerodymics, chassis, driveline,
brakes, ride and handling, acoustics, and ergonomics;
.... of cars, trucks, trailers, buses, motorcycles, snowmobiles, motorhomes, 3-wheelers,
bicycles, human-powered vehicles, electric vehicles, race cars from showroom stock to
Can-Am, and computerized instrumentation packages;
.... for McDonnell-Douglas, General Motors, Systems Technology (a vehicle dynamics
think-tank), and Petersen Publishing;
.... and consulting to various Japanese importers, aftermarket suppliers, small vehicle
manufacturers, racing teams, auto and bicycle tire companies, the U.S. Department of
Transportation, and the U.S. Olympic Committee bike and bobsled teams;
.... and writing for dozens of vehicle enthusiast magazines, and miscellaneous other
publications from SAE, Scientific American, and Encyclopaedia Britannica, and a regular
editorial column in Racecar Engineering magazine.
Previous automotive books include two classics, Chevrolet — Racing, Fourteen Years
of Raucous Silence, and Mark Donohue's biography, The Unfair Advantage, both re-
published in 2000. His latest book, about the future of automotive technology, is called
AUTOZ2010. (See page 176)
For all book updates and availabilities, contact [email protected]
BIBLIOGRAPHY: ORIGINAL SOURCES

TIRES AND WHEELS BRAKES AERODYNAMICS

Clark, Samuel K. (ed.) Mechanics of Bisimis, E. Influence of Antiskid Systems Abbott, I.H. & von Doenhoff, A.E. Theory
Pneumatic Tires. U.S. Government Printing on Vehicle Directional Dynamics. SAE of Wing Sections. New York: Dover
Office —0-308-639, 1981. Paper No. 790455. New York: Society of Publications, 1960.
Automotive Engineers, 1979.
Kummer, H.W. & Meyer, W.E. Unified Aerofoil Report. London: The Jim Clark
Theory of Tire and Rubber Friction. Douglas, JW. & Schafer, T.C. The Chrysler Foundation, 1969.
University Park, Pa.: Pennsylvania State “Sure Brake” — The First Production Four- Cornish, J.J. Some Consideration of
University, 1966. wheel Anti-skid System. SAE Paper No. Automobile Lift and Drag. SAE Paper No.
710248. New York: Society of Automotive
948B. New York: Society of Automotive
Engineers, 1971. Engineers, 1965.
SUSPENSION GEOMETRY AND
ALIGNMENT Leiber, H. & Czinczel, A. Four Years of
— Trapped Vortex Flow Control for
Experience with 4-Wheel Anti-skid Brake Automobiles. The Aerodynamics of Sports
Harvey, W.L. & Ressler, D.B. Competition
Systems. SAE Paper No. 830481. New
Setup — A Chassis Adjustment Analysis. and Competition Automobiles. Los
York: Society of Automotive Engineers,
Rochelle Park, N.J.: ADI Press, 1973. Angeles: American Institute of Aeronautics
1983.
and Astronautics, 1974.
Smith, C. Tune to Win: Fallbrook, Calif.:
Thnack. J.J. & Meek, J.F. Mark I GT Sports Fackrell, J.E. & Harvey J.K. The
Aero Publishers, 1978.
Car Disc Brake System. SAE Paper No.
Aerodynamics of an Isolated Road Wheel.
Winkelmann, O.J. Handling Requirements, 670070. New York: Society of Automotive Los Angeles: American Institute of
SAE Paper No. S220, New York: Society Engineers, 1967.
Aerodynamics and Astronautics, 1974.
of Automotive Engineers, 1959.
Madison, R.H. & Riordan, H.E. Evolution Fink, M.P. & Lastinger, J.L. Aerodynamic
of Sure-Irack Brake System. SAE Paper No. Characteristics of Low-Aspect-Ratio Wings
SPRINGS, ANTI-ROLL BARS, SHOCK 690213. New York: Society of Automotive in Close Proximity to the Ground. Technical
ABSORBERS Engineers, 1969. Note D-926. Washington: National
Kruse, D.F. & Edwards, R.C. Automotive Newcomb, T.P. & Spurr, R.T. Braking of Aeronautics and Space Administration, nd.
Suspension Bumpers. SAE Paper No. Road Vehicles. Cambridge, Mass.: Robert Hoerner, S.F. Fluid Dynamic Drag.
680471. New York: Society of Automotive Bentley Inc. 1969.
Midland Park, N.J.: Published by the
Engineers, 1968. Rusnak, R.M. & Schwartz, H.W. & Cole- author, 1958.
Speckhart, KH. & Harrison, E. The Design man, W.P. A Comparison by Thermal
Reilly, D.N. Aerodynamics of the Airfoil.
of a Shock Absorber to Improve Ride Com- Analysis of Rotor Alloys. SAE Paper No.
Road & Track Magazine, June 1969.
fort. SAE Paper No. 680472. New York: 700137. New York: Society of Automotive
Society of Automotive Engineers, 1968. Engineers, 1970. Romberg, G.F. & Chianese, F. & Lajoie,
R.G. Aerodynamics of Race Cars in Draft-
ing and Passing Situations. SAE Paper No.
710213. New York: Society of Automotive
Engineers, 1971.

BEST BOOKS AS OF 2000

Aerodynamics Race Car Aerodynamics, 1995, Joseph Katz Most of these books can be bought from
Chassis Fundamentals of Vehicle Dynamics, 1992, Tom Gillespie one of the following sources. At the very
Computers Data Power: Using Data Acquisition, 1993, Buddy Fey least, you should contact them for catalogs:
Driving Drive to Win, 1996, Carroll Smith
Engineering Race Car Vehicle Dynamics, 1995, Bill & Doug Milliken HP Books, 800-788-6262
Frames Chassis Engineering, 1993, Herb Adams Classic Motorbooks, 800-826-6600,
Oval Tracks Paved Track Stock Car Technology, 1997, Steve Smith Steve Smith Autosports, 714-639-768,
Success The Unfair Advantage, 2000, Mark Donohue Robert Bentley Publishers, 800-423-4595
Testing Race Car Dynamics [Seminar Notebook] 1999, Claude Rouelle SAE, 724-776-4970
Crewing Inside Racing Technology, 1995, Paul Haney Amazon.com

170
HANDLING FRAME
AND BODY THE PITS
Bergman, Walter, The Basic Nature of Vehi- Elliott, W.A. Plastic Models for Dynamic Stropus, Judy, Stropus Guide to Auto Tim-
cle Understeer-Oversteer. SAE Paper No. Structural Analysis. SAE Paper No. 710262. ing and Scoring. Norwalk, Conn.: Sports
957B. New York: Society of Automotive New York: Society of Automotive Car Press, 1975.
Engineers, 1965. Engineers, 1971.
Wyer, John, Motor Racing Management.
Ellis, J.R. Vehicle Dynamics. London: Fournier, Ron, Metal Fabricator’s Hand- London: St. Anne’s Press, 1956.
London Business Books Limited, 1969. book. Tucson, Arizona: H.P. Books, 1982.
Milliken, William F. Research ‘in Smith, Carroll, Prepare to Win. Fallbrook, GENERAL
Automobile Stability and Control. London: Calif.: Aero Publishing, 1975.
Campbell, Colin, Design of Racing Sports
The Institution of Mechanical Engineers, Cars. Cambridge, Mass.: Robert Bentley
1957. SAFETY incsw1973:
Henderson, Michael, Motor Racing Safety. Mezger, Hans, Engineering the Perfor-
ENGINE SUPPORT SYSTEMS London: Patrick Stephens Ltd., 1968. mance Car. SAE Paper No. 700678. New
Fisher, Bill & Waar, Bob, How to Hotrod York: Society of Automotive Engineers,
Small-Block Chevrolets. Tucson, Arizona: DRIVER 1970.
H.P. Books, 1973. Bondurant, Bob, On High Performance Smith, Carroll, Prepare to Win. Fallbrook,
Ganahl, Pat, Ford Performance. Santa Fe Driving. Motorbooks, Intl., 1982. Calif.: Aero Publishers, Inc., 1975.
Springs, Ca: S-A Publishing. Johnson, Alan, Driving in Competition. Steeds, W. Mechanics of Road Vehicles.
Ricardo, Sir Harry, The High-Speed Inter- Newport Beach, Calif.: Bond, Parkhurst London: Iliffe & Sons Ltd., 1960.
nal Combustion Engine. 4th ed. London: Publications, 1971.
Terry, Len, Race Car Design and Develop-
Blackie & Sons Ltd., 1967. ment. Cambridge, Mass.: Robert Bentley
Smith, Phillip H., The Design and Tuning TESTING Inc., 1973.
of Competition Engines. 5th ed. Cambridge, Brown, Ronald F. Your Car’s Horsepower.
Mass.: Robert Bentley Inc., 1971. Road & Track Magazine, November 1970.
Pankhurst, R.C. & Holder, DW. Wind Tun-
GEARING AND DIFFERENTIALS nel Technique. London: Sir Isaac Pitman &
Gregorich, H.L. & Jones, C.D. Mark II Sons Ltd., 1965.
GT. Transaxles. SAE Paper No. 670069. Rae, William H. & Pope, Alan, Low Speed
New York: Society of Automotive Wind Tunnel Testing. New York,: John
Engineers, 1967. Wiley & Sons, 1984.
Smith, Carroll, Prepare to Win. Fallbrook,
Calif.: Aero Publishing, 1975.

171
INDEX

Acceleration, 73, 80 Brakes, 43-54, 56, 72 Check valves for fuel, 79

downforce and, 62 anti-skid systems and, 53-54 for oil, 80
front wheel drive and, 156, 158 balance bar, 48 Circuit breakers, 82
measurement of, 119 break-in, 50 Clutch pedals, 108
testing, 124-125 development of, 48-50 Coastdown, 67, 126
torque and, 22 heat dissipation, 43, 52-53 Coefficient of friction, 7-Il, 14-15, 73-74
Accelerometers, 119, 120, 125, 128 inboard, 48 braking and, 44, 50
Accuracy, 1l4, 152, 154 parts of, 45-48 Coil spring, 26, 36-39
Acid-dipping, 98-99 pit stop and changing, 137 rate, 39
Ackerman steering, 23, 129 proper fitting of, 108 Communications, pit, 135
Acrylics (Plexiglas), 93 proportioning, 44-45, 47-49, 51, 168 Compliance (deflection), 23-24, 26, 142
Aerodynamic downforce, see Downforce suspension and, 21 Computers artificial intelligence and, 167
Aerodynamic theory of, 43-45 data recording and, 147-153
boundary layer, 62, 65 wear, 50-51 engine controls and, 153-154, 164
center of pressure, 62-63 Braking cornering and, 72, 157-158 simulations and, 7, 73, 139, 141-153, 148, 156
porpoising, 62 differential and, 90 structural analysis and, 143-146
separation, 55-56 kingpin offset and, 23 timing and scoring and, 154-155
skirts, 59-60 load transfer and, 61, 62 Constant-velocity universal joint, 91
stability, 72 testing, 125-126 Control (response), 70-72, 74-75, 167
turbulence, 66, 122 Bubble-balancing, 16 Control arm linkage suspension, 26
variable surfaces, 62, 69, 161 Bubble gauges, 29 Control arms, 21-22, 32
vortices, 55-56, 65, 162 Bump acceleration force, 19 Cooling, 56
wake 56 Bump rubbers, 37, 39 of brakes, 49-50, 56
Aerodynamics, 55-69 Bump/rebound, 20, 35, 38 differentials, 90
development, 66-69, 86, 144 Bumps, 27, 34 water-cooling system and, 81
front wheel drive and, 158 camber angle and, 21 Cornering, 9, 19
hardware, 63-66 shocks and, 35 differentials and, 90
testing, 126-127 tire absorption of, 21 handling and, 72-73, 76
theory of, 55-66 kingpin offset and, 23
wheel and tire, 60 CAD/CAM/CAE, 143 speed measurement, 118
Aeroquip flex hoses, 80 Calibration, 120, 123, 152 tire load and, 10
Air leakage, in tires, 16, 17 Calipers, 46, 50 Cracks in engine, 83
Aligning torque, 10, 19 mounting, 50 in wheels, 16
Alignment, 29-30 Camber angle, 10 Crash helmet, 104, 107
Anti-dive, 20, 34, 36 alignment and, 29-31
Anti-freeze solutions, 81 suspension and, 26, 27, 29, 32 Damping, 10, 35
Anti-roll bars, 33, 76, 129 tire development and, 14 stability and, 70, 72 See also Shock absorbers
development of, 40-41 torques and, 21 See also Tires Data logger (buffer), 118, 147, 149-151, 155
theory of, 34-35 Carbon/carbon (graphite) brakes, 51-53 Deceleration, 103-104, 125
types, 37-38 chassis, 93-96, 164 deDion axle suspension, 21, 28
Anti-skid brake system, 10, 45, 53-54 engine, 164 Deflection, suspension, 19-21
Anti-squat, 20, 22, 34 Carburetor, 75 air drag and, 67
Asymptotic stability, 70, 72 Caster angle, 29 balancing system, 42
Automatic transmissions, 91 adjusted, 30 measurement, 129
Axle solid, 21, 22, 27-28, 31 steering and, 23 testing, 132
swing, 21, 26-27 Catch cans, 80 Diagnostics, 154
Center of gravity, 10, 21 Dial gauges, 31-32
Balance, 16 handling and, 73-75, 160 Differentials, 73, 90-91, 157-158, 168
handling and, 72 locating, 131-132 Displacement ratio, 33
suspension and, 29-32 Ceramics, 164 Displacement sensors, 19
Ball-lock pins (pip-pins), 102 Chassis, 92-100 Distributor, 81
Battery, 106 anti-roll bar and, 35 Divergent instability, 70
Bearing, 37 deflection tests, 132 Double A-arm suspension, 26
suspension and spherical, 28 frame type and, 97-98 Downforce, 19, 39, 42, 64, 161
Belts, water pump, 81 materials for, 92-97 brakes and, 44, 48, 50
Black-light inspection 16 pit stop repairs, 137 force balancing and, 67-68
Bladder, 105 shocks and, 41 front wheel drive and, 159
Bolts, safety specifications for, 101 springs, 33-34 measurement of, 67
Bottoming, 39 stiffness, see Stiffness, chassis stability and, 60-63
Boundary layer, 62, 65 strength, 98-100 testing, 126-127
Brake discs, 46, 50, 51-53, 167 tube (space frame), 97-98, 162 theory of, 57-6
Brake fluid, 47, 50, 53 See also Aerodynamics trailing car, 69
Brake pads, 47, 50, 51, 137 Check lists, 133-134 See also Aerodynamics
Brake torque, 19 Drafting effect, 69

172
Drag Friction, 43 Ignition system, 81-82
gear ratio and, 84 differential, 90-91 Inboard brakes, 48
measurement of, 67 shock, 35 Independent suspension, 21, 31
racing and, 69 spring, 36-37 Induced drag, 55
testing, 126 Friction-circle theory, 11, 110, 153 Inertia, 71, 125, 142
theory of, 55-57 Front wheel drive (FWD) of wheels, 14, 43
Drag Coefficient (Cd), 56-57, 64 aerodynamics and, 158 Inspection, 102-103
Drag strips, 110, 116, 125, 147 brakes and, 158 Instantaneous center of rotation, 21, 27
Driveline, 91 chassis, 159-160 Instrumentation, 49, 71, 110, 156
Driver, 108-113, 135 differentials and, 157-158 Interference (EMI), 151-152
comfort for, 108-110 driving style, 157
emotional preparation, 112-113 tire temperature and, 160 Jacking effect, 20, 34
learning, 110-112 versus rear wheel drive, 156 Jerk (acceleration change), 35
protection, 103-107 Fuel, 79, 136, 138
testing and, 114 fire and. 105-106. 138 Kamm theory, 56
track testing and, 130 Fuel cell, 79, 105 Kevlar, 93-95, 97, 105
styles of, 76, 110-112, 157 Fuel pressure, 78-80 Kingpin, 10, 23
Dry-sump system, 80, 83, 87, 164 testing, 129
Ducts, 50, 68-69, 56, 60 Fuel pressure gauge, 79, 109
Laptimes, 131, 167
theory of, 66 Fuel system, 79-80
scoring, 134-136, 154-155
Durability, testing, 131 Fuses, 82
Lateral force, 19-20, 32, 34, 63
Dye penetrant inspection, 102
Gauges Lateral gust stability, 72
Dynamic balancing, 16
bubble, 29 Leaf spring, 36-37
Dynamics, suspension, 19
dial, 31-32 rate, 40
Dynamometers, 124, 153-154
driver comfort and, 109 Lexan plastic, 63, 93, 104
fuel pressure, 79, 109 Linkage, 38, 40, 89, 108-109
Electronic speedometer, 128
strain, 119 suspension, 22
Emergency pit stop, 137-138
Gearing, 12, 84-90, 125 Load levellers, 36, 42
Engine adiabatic, 165 Load transfer
cracks in, 83 maintenance, 89-90
ratio theory, 84-85 aerodynamic, 62
inertia, 45
transmission hardware and, 85 braking, 22, 44, 54, 125
measurement of, 124
Ground clearance, 39 computer simulations, 143
torque curves, 72, 84, 85
Ground effect, 57 handling and, 73-75
Engine support systems, 78-83
Guide vane, 66 suspensions and, 27, 34-35
exhaust, 83
Gurney flap, 65 tires and, 10, 15, 19
fuel, 79 Lock nuts, 101
ignition, 81-82 Gyroscopics, 14, 21
Locking differentials, 90-91
oil, 80 Loctite, 102
water, 81 Halon, 106 Long races
Exhaust system, 83 Handling, 70-77 timing and scoring, 134-136
control and, 72 tires and, 17
Fade, 43, 45, 49
development of, 75-77 Longitudinal force, 19, 20, 34, 46
Fasteners, 101-102
four-wheel drive and, 74 Longitudinal slip, 9
Feedback, 72, 109, 159, 167
hardware design for, 75 Lubricants, 89-90
Fiberglass, 93, 137
load transfer and, 73-75 Lug bolts, 13, 17
Fifth wheel, 119, 147
oversteer and understeer, 73-74 Lug nuts, 13, 16, 136
Filter cans, 80
stability and, 70-72
Finite element modeling, 143
test, 127-129 MacPherson strut, 24, 27, 159
Fins, 63
transient, 70, 72-73, 75 Magnesium, 93
Fire, 83
Header, 83 Magnetic particle inspection (Magnaflux), 102
driver safety and, 105-107
Heat Magnetic pulse distributor, 81
Fire extinguisher, 106, 164
driver comfort and, 109 Manometer, 120, 127
Fireproof clothing, 107
See also Temperature Master cylinders, 46, 47, 50
Fireproof hoods, 138
Firewall sealing, 109 Heat absorption, brakes and, 43 Materials, 52, 63, 88, 92-97, 100, 104, 144
Horsepower, 57, 78, 125 See also specific materials
Forces
transmissions and, 85, 89, 91 Mechanical force ratio (leverage), 33, 35
aerodynamic balancing ot, 42, 67-68
Hoses, 79-80 Metal fatigue, 98
measurement of, 119
for water-cooling system, 81 Models, 99-100, 121-122, 146
suspension, 19-21
Hubs Monocogque, 97, 105, 146, 162
Four-wheel drive, 74-75, 168
aerodynamic loads to, 68 Mounting
Frame, see Chassis
knock-off, 136 of tires, 16
Free height, 39
torques, 19 of wing, 66
Hydraulic jacks, 136

173
NACA/NASA, 55, 56, 58, 64, 66 Racing Shot-peening, 102
Neutral steer, 74, 76, 129-130 air drag and, 69 Signalling, 135
Noise, 79, 109 brakes and, 51 Skid, 72
gearing and, 89-90 Skidpad, 110, 129
See also Driver and specific listings handling and, 76-77
Offset, 13, 23 Racing radial, Il testing and, 116, 127-129, 156, 158
Oil, 80, 136 Radiator, 56, 66, 69, 81, 109 tire development and, 14-15
Oil gauges, 109 Skin friction drag, 55
downforce and, 60
Oil pressure, 129 Slingshot passing technique, 69
Rain tires, 13, 18
Oil starvation, 78 Rear steering, 24 Slip angle, 7, 9, 23, 74
Oiling system, 80 Slip rate, 9, 90
Record keeping, 133-134
100 percent anti-dive, 20-21, 34 Solid axle, 21, 22
timing and scoring, 134-136
Oscillatory stability, 70 alignment, 31
Recorder, data, 67, 130
Oscillograph, 118 suspension, 27-28
digital, 147-154
Overheating, 78 Solid stack height, 39
strip chart, 118, 147
Oversteer Spacers, 13
tape, 116, 147-148
aerodynamic, 57, 63, 68-69 Spark plugs, 82
Recordings, data, 8
handling and, 70, 73-74, 76-77 Speed, measurement of, 88, 118
driver performance, 110-111
suspension and 23-24, 34, 40-41 Speed traps, 126
speed, 88, 125
testing, 128-129 Spoilers, 63-64, 67
Relative value, 131
tires and, 9, 15, 129-130 Spring rate, 10, 33-34, 39
Repeatability, defined, 114
Reservoir, 80 Spring tension adjustment, 30
Panhard rod, 27 Resins, 95-96 Springs, 68
Penetrating dye, 16 Retainer pins, 137 development, 38-40
Pinion (input gear), 22 Ride, 70 theory of, 33-34
Pistons Ride height, 29, 37, 39 types of, 35-37
braking and, 50 Ride rate, 33-34, 39 Sprung weight, 13
caliper and, 46 Rim, 13 Stability
Pitot-static tube, 120, 127 tire mounting and, 16 airflow and, 60-63
Pits, 133-138 Ring gear (axle gear), 22 braking and, 90
communication, 135 Rocker arm suspension, 37 differential and, 90
record keeping, 133-134 Roll axis, 20 handling and, 70-72, 75, 128, 168
stops, 18, 136-138 Roll cage, 99, 105, 163 Stability effect (aerodynamic lift), 69
timing and scoring, 134-136, 167 Roll center, 20, 21, 27, 28 Stagger, 12
Plexiglass, 93 Roll rate (resistance), 21 Static alignment, 29, 30
Pneumatic chisel, 137 computer simulations, 142 Static balance, 16
Pneumatic impact wrenches, 136-137 front wheel drive and, 159 Static position, 19
Pneumatic impact wrenches, 136-137 suspensions and, 33-35, 40, 163 Static pressure, 127
Pneumatic trail, 11 Roll steer, 23, 26, 31-32 Static stability, 72
Polar moment, 71 Rotating cylinder, 161 Static testing 131-132
Potential flow theory, 65, 162 Rubber pivot bushing, 28 Steady-state stability, 72, 128
Potentiometers, 117-120, 153-154 Rubber spring, 36, 37 Steer angle, 23, 32
Precision, 114, 147, 149-151 Rubber suspension bushing, 26 Steering, 19, 108, 168
Prepregs, 96-97 Safety, 101-107 handling and, 72-74
Pressure accident avoidance, 98, 101-103 testing, 129
distribution of, 68-69 driver protection, 99, 103-107, 163 torques and, 22-23
measurement, 120, 127 pit, 138 Steering column, 104
tire, 10, 15, 17, 23 Safety factor, 19, 144 Steering kingpin, 10
Pressure caps, 81 Scales, for balance, 29-30 Steering wheel, 108
Pressure gauges, 127 Scorer, 134, 154-155 Step steer, 70, 142
Pressure pickup strip, 127 Seat Stiffness, chassis (torsional)
Pressure rake, 120 driver safety and, 103 aerodynamic loads and, 61
Pressure relief valve, 48-49 proper fitting, 108 computer simulations and, 144, 147
Profile (shape) drag, 55-56 Seatbelts, 104 frame types and, 97
Pump-up, 36 Servo-action brakes, 44 testing, 98, 132
Pumps Shaving, tread, 12 transmission effects on, 88
fuel, 79 Shifting, 84-86, 89 Stopping force, 44
oil, 80 Shock absorbers (dampers), 26 Stopwatch, 116, 125, 128
Pyrometer, 124 development of, 41-42 Strain gauges, 119, 123, 124
pump-up, 36 Stress, 26, 89, 144
stiffness, 35 Stress reversals, 78, 89
Quick-release fasteners, 102, 137
theory of, 35-36 Strip-Tube, 120
types of, 38 Structure (strength), 143-146

174
Surface pressure strip, 120 pressures, 10, 15, 17, 23 handling and, 70, 73-74, 76
Surge tanks, 79 steering and, 22-23 suspension and, 23-24, 34, 40-41
Suspension, 13, 68, 126-127 testing, 129-130 testing, 129-130
bottoming and, 39 thrust, 73-74, 84 tires and, 9, 14, 15, 129-130
hydraulic, 163 track and, 130 Universal joint, 91
measurement, 124 tread width, Il Unsprung weight, 13, 28, 33
simulations, 142 wear, 17 Urethane foams, 93, 104
Suspension geometry See also Camber angle, Temperature, Wheels
development of, 31-32 Titanium, 92-93 Vacuum power boosters, 44
handling control and, 70 Toe-in, 30-31 Valves, brake, 48
maintenance and service of, 29-31 Toe steer, 32 Variable rate springs, 37, 41, 33, 61, 163
theory of, 19-26 ; Torque, 21-26 Velocity profiles, 68
types, 26-29 aligning, 19 Vent hoses, 80
Swing axle, 21 anti-roll bar and, 34 Venturi, 57, 59
suspension, 27 oversteer/understeer, 73-74 Vertical force, 19
Torque arm, 22 Vertical strut suspension, 26
Torsion bars, 35-37 Vibration fatigue, 78, 144
Tachometer, 109, 118, 125
Team manager, 133 spring rate, 39-40
Water, pit stop and, 136
Telemetry, 116, 148 Total pressure rake, 127
Water container, 109-110
Temperature Track
ks eee 119-120 pit stop and, 133-136 wae Sohne ek etrt or
testing, 125
. testing, 130-131
Ke eliasFate
atts ge, e
of tires, 7, 14-15, 17, 129-130, 160 Track width, 75 - Wear strips, 39
Temp-Plate, 120, 126 Traction
oversteer/understeer and, 73
Weight,
Tension cable, 108 distribution, 22, 75, 156, 158
Testing, 114-132 roll and, 21
oem 98-100 stability and, 72 eaan A
f 109 Transducer (sensor), 54, 117-120, 147, 150-151 5 2
or gauges, Tras; Weight transfer, see Load transfer
instruments, 114-124 LAE - Wheel loads, 119
eee’
procedures, 124-130
front
anaes
wheelOCD
drive,7 156-157
95
De ee
Wheelbase, 75
static, 131-132 aaseen Sake Wheels
track, 130-131 E ohne? ic
Thermocouples, 120 testing, 129 BOR OOee
Thermoplastic materials, 93 2 tire, i, 15 construction, 13
Throttle, 74, 108-109 Fans hussion
stability and, 71-72 pe see ae
Timer, 134 development of, 85-89 mai neato for 93
Timing, 134-136, 154-155 dogs, 85-86 oh, Bae
segments, 130-131, 157 gear design, 87 it stop and mounting, 136
Tire chatter, 19
Tires. 7:18
gear ratios for, 85, 88-89
infinitely variable, 165-166
eee and weight-of, i 22
Re SR
break-in,
Braigas
17and.
49
synchronizers, 85
transverse, 86-88
sae ake Sone Tue
; :
ig and, Wind tunnel, 67, 121-124
bump absorption and, 21 Travel
Windscreen, 110
clearances, 12 dynamic displacement, 19, 35
Windshield, 137
coefficient of friction, 7-Il load levellers and, 42
Wings (or airfoils), 57-61, 63, 67-68
compounds, ll-15, 18 shocks and, 35-36
racing and, 69
computer simulations, 142 Treads
theory of, 64-66
contact patch, 12 design, 12-13
Wiring system, 82
design of, 1l-14 handling and width of, 75
development, 14-15 thickness, 7, 13 X-ray inspection, 102
downforce and, 60 Tubular space frame, 97
Turbochargers, 165 Yaw, 70-71
drag and, 56
driving and 17-18 Turbulence, 66, 122
Zyglo, 102
flat spots, 17
Understeer
leaks, 16-17
aerodynamic, 57, 61, 67
load transfer, 10, 15, 19, 34-35
braking and, 54
measurement, 124
differential and, 90
pit stop and changing,. 134, 136
front wheel drive and, 157
preparation of, 16-17

175
WHAT NEXT? MORE! The car book
FROM the future.
HOW TO FIND YOUR OWN
SECRET UNFAIR ADVANTAGES. 112 pages
Full of color
Now that you’ ve read this book, you’! finally be equal with 8 1/2" X 11"
everyone else who has also. Now, to get ahead of them, you’ll
have to do your own development and research into all new softcover

FEATURES
racing tricks. You won’t find those answers in print anywhere
— but there is a book that has clues on where to start.

AUTO2010 is a book totally dedicated to the future of high EDITORIAL INTRODUCTION

performance auto technology. It looks just like the car Writing in “Future Preseni” tense.
magazines you read today — except that it was written and CONGESTION SOLUTION I --
photographed in the future, using revolutionary computer Clippings from futures past.
imaging techniques. CONGESTION SOLUTION II --
Commuters, waiting for Prometheus.
AUTO2010 includes over 100 future automotive concepts CONGESTION SOLUTION III --
that have never been seen in print, such as: Finally, arrival of the Robocars.
Suspensions that memorize the track. LEGENDARY AUTO FUTURIST SYD MEAD
The 64-speed auto-shift transmission. A 50-year career of breaking away,
Scout-car and tow-car team strategy. RECORD GESTATION: THE 2010 LAUNCHER
In-wheel suspensions. 12 months from inspiration to market.
Independent 4-wheel slip angle control. FANTASY LIFE #34 IN A SERIES --
Automatic spin recovery. Auto design student Ira Gilford.
Filament-grown composite structures. ART CENTER COLLEGE OF AUTO DESIGN
Failure prediction without gauges. Mecca for committed carradicals.
Wind-vision goggles.
Selective filtering of racing sounds. ROAD TESTS
Real-time race participation simulators
WE’RE CLIMBING THE WALLS --
There are entire chapters on the technology expected in Indy Special test: the suction-traction Remora.
cars and Formula One by the year 2010, and a chapter on what ROBOTEST TECHNIQUES --
it’s like to actually DRIVE one of these cars. Other chapters Our new cyborg evaluator/writer.
describe the technology of production cars and methods of CHEVROLET IMPRESS --
reducing congestion and improving mileage (see complete For people who don’t like road tests.
contents atright). Perhaps you will be the firstto develop some PEGASUS 2010 --
of these concepts. If you order AUTO2010, your chances are Road/sky test of the latest fly-cycle .
a lot better.
SPECIAL TECH SECTION
AUTO2010 is not stocked in bookstores.
To order, just send a check for $16.95 (postpaid) to: EXTERIORS --
Shapes we never dreamed of.
Baul Van Yekenburgh POWER! --
Engines, motors, flywheels, and drivelines.
Seal Beach, CA 90740 CHASSIS -—-
Under the skin.
ata a Hj Cit et see
INTERIORS --
December 2010 Intimate contact.
a3tion Sulutiy
COMPETITION

2010 INDY 500 -—

“Training” for the win.
WANNA DRIVE AN INDY CAR ?
What it's really like out there.
SEN-SIM --
The world’s fastest race chair.
FORMULA ONE COMMENTARY --
A history lesson.
MOSCOW FINALE —
World car war 2010.
Fl TECHNOLOGY REVIEW --
Team secrets exposed.
176
 = : =
* :
ES
a a Fe
Ns i*. i
ae . m3

Ad) ac
P ¢ za a
f 2 d
*

me au pALIVER POOR Sten es

SOHN MOORES UNIVERSITY coal
oe eevee oeantsiad.
MTA. WIMEBARN STREET at ea
ys ee RRPOOL 22ER
O151 231 4020. |
 EVERYTHING YOU NEED TO WIN!
Ever wonder what top professional racing mechanics and engineers know that you don’t? What would you give to learn some
of their trade secrets—or get a job working with a team? With Paul Van Valkenburgh’s Race Car Engineering and Mechanics,
you'll get original information not available anywhere else. Over 200 photos and illustrations guide you through the complete
process of engineering a race car without a college degree. Whatever you want to know, it’s easy-to-find, and easy-to-read,
with inexpensive, practical suggestions for specific applications.

Aerodynamic drag and wings

Suspension geometry
Brake balancing and cooling
Track testing and shop testing
Making your tires work better
Gearing and differentials —
Driver education and pitstops
The use of computers and much more!

PHOTOS
MICHAE
BY
LUTFY

From dirt tracks to Indy, from slaloms to Daytona — winners use this book. Race Car Engineering and
Mechanics is the most nr “tool” for your toolbox — don’t go to another race without it!

Tony Cicale, winning Indy car engineer t

Lee Dykstra, winning professional sports car engineer. pa

UE
® =
Trevor Harris, winning IMSA GTP engineer. Pe:E

ISBN NO 82=
0768007 17%