.

A THEORETICAL AND EXPERIMENTAL STUDY OF THE

VIBRATION GALVANOMETER.

By Frank Wenner.

CONTENTS.
Page,
i . Introduction 347
2. Theory 349
3. Effect of Harmonics 356
4. Effect of Inductance 358
5 The use of a Transformer 358
6. The System of Units 359
7. Method of Determining Constants 360
8. Basis of the Equations 361
9. Method of Tuning 363
10. Galvanometers and their Constants , _ _ 364
1 1 Characteristics of the Instruments , 366
12. An Illustrative Problem 369
13. Design 375
14. Summary 377

1. INTRODUCTION.

The vibration galvanometer is an instrument for the detection

of alternating currents or electromotive forces. It differsfrom
other instruments for the same purpose mainly in having the
moving system tuned to the frequency of the current or electro-
motive force to be investigated. In one of the later designs the
construction differs from the ordinary D'Arsonval galvanometer
only in having a very narrow coil and having bifilar suspensions of
adjustable length and tension so as to give a period of from about
50 to about 750 vibrations per second. As the deflection of the
instrument changes sign as often as the current changes direction,
the image formed by the mirror of a line source of light appears
as a band, unless the frequency is very low. The width of the
band is a measure of the amplitude of the vibration.
347
348 Bulletin of the Bureau of Standards. woi. 6, No. 3

If the period or frequency of the current changed while the

is

free period of the instrument is kept constant, it is found that the

instrument has a high sensibility at and near the resonating
frequency, while at frequencies differing by more than a few per
cent from the resonating frequency the sensibility is very small.
The instrument, therefore, responds only slightly to the harmonics
in the current or electromotive force wave, so that usually no error
is introduced when measurements and calculations are made on

the assumption of a pure sine wave. This property makes the

instrument especially suitable for those measurements in which
frequency enters and in which null methods are used.
Max Wien pointed out some of the advantages of a tuned
*

instrument when he described his optical telephone, a vibration

galvanometer in which the essential part of the moving system
consists of a piece of soft iron on a very flexible diaphragm. The
soft iron armature is actuated by one or two bipolar telephone
magnets of the usual construction. The current to be investigated,
passing through the coils, causes a relative displacement between
the magnet and armature, and this is communicated to a mirror
in such a way as to give an angular displacement to the latter.
The displacement is read by the broadening of a line image. The
mirror is on a spring, and for a maximum sensibility the period of
this system, as well as the period of the current to be investigated,
is adjusted to the period of the diaphragm. Wien showed experi-
mentally that the deflection is proportional to the current.
Rubens next brought out an instrument in which the mirror
2

and soft iron piece are mounted together on a metal suspension

instead of a diaphragm. The bipolar telephone magnets are so
placed and their coils are so connected that the current causes an
angular displacement of the soft iron armature and mirror. The
frequency can be adjusted by either a change in length or tension
of the suspensions. This instrument is more flexible than the
optical telephone and a somewhat higher sensibility is claimed
for it.

Max Wien 3
developed an instrument having a very small
later
permanent magnet mounted on a brass wire suspension, the
1
Wied. Ann., 42, p. 593, and 44, p. 681 ; 1891.
2
Wied. Ann., 56, p. 27; 1895.
3
Ann. der Physik., 309, p. 425; 1901.
wenner.] The Vibration Galvanometer. 349

lengths and tension of which may be changed so as to give a

considerable range in the number of vibrations per second. The
coils are wound on a core of soft iron wire, and the pole faces of
this core are very close and parallel to the small magnet. The
sensibility claimed for this instrument is 16 times that of the
optical telephone.
Rosa and Grover 4
in their work on the measurement of induc-
tance pointed out certain advantages in the use of a vibration
galvanometer and give a curve showing the change in sensibility
with a change in tuning for an instrument which apparently had
two free periods.
Wells 5 showed experimentally that the deflection is proportional
to the electromotive force. He also plotted a curve showing the
change in the observed deflection as the free length of the sus-
pension was varied.
In 1907 Campbell 6 described a very convenient instrument of
the D'Arsonval type. This instrument is suitable for frequencies
of from about 50 to about 750 vibrations per second and is easily
adjustable over this range. In sensibility this instrument com-
pares favorably with Wien's when both are adjusted to the same
frequency. Concerning the theory of the instrument, Campbell
points out that the sensibility is independent of the moment of
inertia and inversely proportional to the frequency. He also
states that the damping is "both mechanical and electrical" and
should be kept small.
2. THEORY.

In any galvanometer, a current through the winding pro-

duces a force tending to cause a relative displacement between
the fixed and moving system. The resulting relative displacement
causes a part of the flux from the magnet to cut through the wind-
ing, and thus develops a counter electromotive force. The energy
available for producing a deflection is the time integral of the
product of the current and the counter electromotive force. As
galvanometers are ordinarily used, this energy is small, due to the
4
This Bulletin, 1, p. 291; 1905.
6
Phys. Rev., 23, p. 504; 1906.
6
Phil. Mag., 14, p. 494; 1907.
Electrician, 60, p. 60; 1907.
 —

350 Bulletin of the Bureau of Standards. \Voi. 6, No. 3.

fact that after the moving system has reached a displacement

equal to the final deflection the time integral of the product of the
counter electromotive force and current is either equal to or less
than zero.
If the motion is only slightly damped, a reversal of the current

at the end of the first deflection will cause a second deflection in

the opposite direction larger than the first. This is due to the
fact that the motion of the system again causes a counter electro-
motive force, and thus more electrical energy is converted into
mechanical energy. However, not all this energy is added to the
system, since some is lost by air friction, induced currents, etc.
By reversing the current at the end of each deflection the suc-
cessive deflections can be increased until the average rate at which
energy is lost by damping becomes equal to the average rate at
which electrical energy is converted.
In the work that follows we shall consider first the case where
the impressed electromotive force and current waves are both of a
sine form and later the effect of harmonics. With a sine wave cur-
rent and electromotive force we may assume a simple harmonic
vibration, and a sine wave back electromotive force.
In use the free period is adjusted to or near that of the current to
be investigated. If this adjustment is good the displacement of
the moving system leads the driving current by 90 °, and since
the back electromotive force depends upon the rate of change of
this displacement it is in phase with the current. If, however,

the tuning is not good the moving system leads the current by an
angle greater or less than 90 °, and consequently the phase angle
between the current and the back electromotive force is not zero.
As the power converted depends upon the cosine of this phase
angle, the tuning may have a marked effect upon the deflection.
As the amplitude of the vibration increases the back electro-
motive force increases and unless the resistance is very high it has
a marked effect on the value of the current. If the three vectors
the current, impressed electromotive force, and back electromotive
force — are in the same line, the power converted and consequently
the deflection is a maximum, if after a steady condition is reached
the back electromotive force is equal to one-half the impressed
electromotive force. Since the back electromotive force is pro-
wenner.] The Vibration Galvanometer. 351

portional to the field strength it may happen that an increase in the

field strength results in a decrease in the amplitude, just as in a
direct-current motor an increase of the field strength may result
in a decrease in the speed. For a given set of conditions the
deflection of the vibration galvanometer is a maximum for a par-
ticular field strength, just as for a given set of conditions the speed
of a direct -current motor is a maximum for a particular field
strength.
Since when a steady condition reached the power lost by
is

damping is equal to the electrical power converted, a decrease in

the damping must, if the other quantities are kept fixed, result in
an increase in the amplitude of the vibration. This, increase may
or may not be large, just as in the direct-current motor a decrease
in the load may or may not result in a material increase in the
speed.
The relations between the amplitude of the vibration, current,
electromotive force, frequency, etc., may
be expressed in terms
of the working constants, the more important of which are: (a)
The current sensibility, (b) the electromotive force sensibility, (c)
the power sensibility, (d) the ratio of the back to the impressed
electromotive force, (e) the time constant for current indication,
(/) the time constant for electromotive force indication, (g) the
resonance range for current indication, and (h) the resonance range
for electromotive force indication . By resonance range is meant the
fractional part by which the period of the current or the impressed
electromotive force changed to reduce the deflection by half.
is

These working constants depend upon the intrinsic constants

of the instrument which are: (a) The moment of inertia, (b) the
moment of damping (by which is meant the ratio of the retarding
torque to the angular speed when the coil is on open circuit), (c)
the moment of restoration (the ratio of the torque, tending to bring
the moving system to the null position, to the angular displace-
ment), (d) the moment of displacement (the ratio of the torque
produced by the current to the current) and (e) the resistance.
,

The intrinsic constants as here given and defined refer only to

an instrument in which the displacement is angular. The follow-
ing equations, however, are not limited to this kind of a displace-
ment but are equally applicable (except for the constant where an
35 2 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

arbitrary unit of displacement is chosen) to instruments of the

Einthoven or string type.
The equations giving the relation between the various intrinsic
constants, the deflection, time, and current or electromotive force
willbe stated and solved under certain limiting conditions. The
symbols used are as follows:
= angular displacement.
6
= time.
/

a = moment of inertia.
j3 = moment of damping.

7 = moment of restoration.
= moment of displacement.
t/t

p = resistance of galvanometer.
r = resistance of complete galvanometer circuit.
i = instantaneous value of current.
I' = maximum value of current.

/ = effective value of current.

w = power.
/
= frequency of current or impressed electromotive force.
f
= natural free frequency of moving system.
p = 27r/.
Po = 2tt/ .

= amplitude after a steady condition is reached.

<f>

e = instantaneous value of the back electromotive force.

E = effective value of impressed electromotive force.
E b = effective value of back electromotive force.
E = effective value of impedance electromotive force.
co = the angle by which the impressed electromotive force leads
the current.
7] = the angle by which the impedance electromotive force leads
the current.
a = the angle by which the back electromotive force leads the
7
current.
A = current sensibility.
V = electromotive force sensibility.
A = direct current sensibility.
f

7
When the electromotive force developed in the winding of the instrument is in
exact opposition to the current a is taken as zero as shown in Fig. 1.
Wenner.] The Vibration Galvanometer. 353

W = power sensibility.
D = ratio of back to impressed electromotive force.
Ti =time constant for current indications.
T = time
e constant for electromotive force indications.
R e
= resonance range for electromotive force indications.
R = resonance
t
range for current indications.
X = the step-up ratio of the transformer.
r = resistance of primary circuit.
1

Vx = electromotive force sensibility with transformer.

signifies that the quantity in question is referred to the set
of conditions given on page 359.
The motion of the moving system of a galvanometer is given 8
by
the equation
dO
^+%+^=^
d'-O
(I)

The free period

l/f = 2ir^a/y
or
r
il a =po 2
and if

i—V cos pt

Integrating (2) gives

W sin (pt+a) -PJ r

Ct -^[V*.'-W4-'i+cr]
t

°-Jp + J?P.--U-- (3)

where C and O are constants of integration and

PP
A short time after starting the current the second member of
equation (3) becomes zero, hence

»- ^f1 x (4)

where </> is the maximum value of 6 or the amplitude of the vibra-

tion.

8
Gray, Abs. Meas. in Elec. and Mag., Vol. II, Part II, pp. 392-3.
354 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

The current determined by the components of the

is electro-
motive forces in phase with it and the resistance or

E cos o) — Eh cos cr
(5)

The rate at which mechanical energy is supplied to the moving

system is equal to the rate at which electrical energy is converted.

Fig. 1. — Vector Diagram showing the Various Phase Angles and Relations between the Electro-
motive Forces.

The former is the torque [due to the current and equal to tyi\

times the angular speed [cpp cos (pt + o)] and the latter is the cur-
rent times the back electromotive force. Therefore

iyfr<f>p cos (pt + a) =--ie

or
E = ^4>pl^2.
h (6)

The substitution of the values of / from (5) and E b from (6) in (4)

gives
^/2-v^EcOS CO
= F=—+ (7)
<t>
T -ylp /3
2 2
ex
2
(p
2 -p 2
)
2
+ pyfr 2
COS <J

and the multiplication of (4) by (7) gives

2 = 2-\jr
2
I E COS 0)
(8)
-p (p o *-p
<t>
r[p 2 2
+a 2
(p o
2 2
)
2
]+T/p
2
/3
2
+a 2 2
)
2
pf 2
COS cr

Or if p=p , (4)1 (7)» and (8) become, since a is then zero,

_^2^rl ,
_ a/2tJtE COS » , ,
2 _ 2^ 2
/ E COS CO
<*> ' (9)
pp p(rP + $*) p>/3(r/3 + f*)
Wenner.] The Vibration Galvanometer. 355

Defining the current sensibility as the ratio of the amplitude of

the vibration to the effective value of the Current, the electromotive
force sensibility as the ratio of the amplitude to the effective value
of the electromotive force at the terminals of the instrument, and
the power sensibility as the ratio of the square of the amplitude to
the power supplied gives

V2 ^
2

A-'-y/
2
^ v- nndW- 2 ^ fieri

In the second part of (10) cos <o is taken as unity. This may
generally be done without introducing an appreciable error, since
in most instruments the impedance electromotive force, between
terminals, is small in comparison with and very nearly in phase
with the impressed electromotive force.
The time constant is given for current indications by equation
(3) or
T* 2a ( \

For electromotive force indications the back electromotive force

is zero when t is zero and increases directly with the amplitude of the

vibration. The current therefore starts with a certain value and

decreases as the amplitude increases, which causes the amplitude
to reach a certain fraction of its final value in less time than with a
constant current. If the instrument is tuned so that the back
electromotive force is in phase with the current it may be shown

that the retarding torque is — -r- instead of fi

-7- as given by

equation (1) for current indications. It therefore follows that

Te = (I2)
rJ+^- s

Since the resonance range is the change, in proportional parts,

in the impressed frequency from resonance, necessary to reduce
356 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

the deflection by half, it may be derived from equations (7) and (9)
by equating the second member of the first to half the second
member of the latter, or

^J2E^r COS ft) -y/2E\jr , v

V^ 2
+^ 2
(/>o
2
-^ 2
)
2
+^ 2
cos a 2/> (V/3 +^ 2
)

Then if we write p (i +R) we have, with a close approxima-

for p,
tion, 4p
4
R = (po — p
2 2 2
)
2
, which when substituted in (13), and writ-
ing p for p, where it makes practically no difference, gives

e> = Vt 2 cos <°( r fi +^ _^ 2

)
2
cos o-]
2
— 0*r 2
/ n

2<r^? r

Here, since there is a decided lack of tuning, cos a is small and in

comparison with 2 may be neglected while cos ft) has a value
between .8 and 1. Taking the latter gives

^^ V 3y8V +8/3rf 2 2
+4^ ( j
2arp

a value about 20 per cent too large if the resistance is small. If

the resistance is very large, the back electromotive force has no
effect on the value of the current, or only those terms of (15)
involving the resistance are of importance, so that

(l6)
*-2§r-
2ap
3. EFFECT OF HARMONICS.
In those measurements in which the frequency enters and null
methods are used an exact balance can not be obtained except
when the electromotive force and current waves are of a pure sine
form. In practice this condition is seldom if ever fulfilled, so the
vibrations caused by the harmonics may limit the precision attain-
able. One of the chief characteristics of the vibration galvanom-
eter is that the amplitude of the various harmonic vibrations is

small. because there is no resonance, and the turning

This is

moment changes direction at intervals short in comparison with

the half free period of the moving system. The amplitude of the
vibration produced by any particular harmonic in the electromo-
Wenner.] The Vibration Galvanometer. 357

tive force wave is given by equation (7) when the corresponding

value of p is substituted.
The substitution of 2>po for p gives

cos
= V2"fE 3
o>3

93
r^i&po +6 4 2
^ 4
+3pof 2
cos o-
3

Here, unless the reactance is large, (cos a>3 ) may be taken as

unity, while the terms in the denominator are all small in com-
parsion with 8rap 2
, so we may write

Denning the sensibility in the usual way

V - V2^
3
8 raps
Likewise

2/^rapo'

V,—fi±-
48 rap M
2
, (18)

V< V2 ^
SorapS

Corresponding substitutions in (4) give

A3 =^t
Sap **

a _ V2t
5
Map o ^
(19)

,4 V2^
48 ^o 2 '

_ V2^r '
9 2
8o«/>
 "

358 Bulletin of the Bureau of Standards. {Vol. 6, No. j.

4. EFFECT OF INDUCTANCE.
By reference to equation be seen that the denominator
(7) it will
can be decreased (if -^ is large in comparison with rp) by a decrease
in cos o, even though in so doing the first term is increased, as it
necessarily is when the instrument is thrown slightly out of res-
onance. If at the same time cos o> can be kept near unity, the
deflection will be increased. The analytical expression for cos <o
is complicated, since it involves the back electromotive force and

consequently the amplitude of the vibration. However, referring

to the vector diagram, Fig. 1 it will be seen that if a large induct-
,

ance is placed in the galvanometer circuit, so as to make the

angle rj nearly 90 °, and the frequency of the impressed electro-
motive force is increased slightly, so that the angle a assumes a
negative value somewhat less than 90 °, it is possible for the back
electromotive force to be much larger than the impressed electro-
motive force and the latter still have the larger component in phase
with the current. The combination of the inductance and the
slightly high frequency of the impressed electromotive force result
in bringing the back and impressed electromotive force somewhere
near quadrature and the latter somewhere near the phase of the
current. If the frequency of the impressed electromotive force is

not too much above the resonating frequency, the result may be
an increase of the component of the current in phase with the
impressed electromotive force. This means an increase in the
power supplied to and consequently an increase in the amplitude
of the vibration of the moving system. A condenser in series with
the galvanometer and the lack of tuning in the opposite sense
should give corresponding results.

5. THE USE OF A TRANSFORMER.

If the terminals of the galvanometer are connected to one of the
windings of a transformer and the other winding is connected to
the source of electromotive force, the equation for the current in
the galvanometer is

7T °
XE-I Xr
-x
-E x h XE-E b , ,
= 7+xv"- (2o)
P
 . •

Wenner.i The Vibration Galvanometer 359

Here assumed that the currents in the transformer windings

it is

are in opposition and have the inverse ratio of the number of turns
in the coils. The requirements of a transformer to meet these
conditions will be considered later. A comparison of equation (20)
with equations (5) and (7) gives

^X^E COS (O

(p+ X 2
r x ) ^//3 2
p + a (p
2 2 2
-p 2
)
2
+ pyjr 2
COS a

or if p =p and taking cos co =1

»

4> = ^X^E
pKp+x^p+f 2
]

and

v =
^X±

VT is a maximum when X = - 2
— 7^— (22a)

6. THE SYSTEM OF UNITS.

In expressing the sensibility or deflection it will be convenient

to use as the unit of current the microampere, of electromotive
force the microvolt, of power the micro-microwatt, of deflection a
millimeter broadening of a line image one meter in front of a plane
mirror (=25X10 " 5 radians) and cgs , units for the other quantities.
The resistance must then be expressed either in ohms or elec-
tromagnetic units to correspond with the quantities with which it

is associated.
In the following work when the sensibility is expressed in the
arbitrary units or the resistance in ohms, it is indicated by a line
over the symbol representing the quantity.
Referred to these conditions and writing 2irf for p , we have
from equations (10), (11), (12), and (22), respectively,

A
=ii3^ =AX4Xl ° =9- OXlo "°p (23)

K = FX4Xio^ 9 .oX.o^L (23a)

360 Uletin oj trie tfureau oj standards. [Vol. 6, No. 3.

A'=A' 4Xio- = 4 Xicr ^ 4 4

(24)
•

(25)

^ * f r /« 1 V^2~ \ yO 1 -».2l' (26)

7. METHOD OF DETERMINING CONSTANTS.

If it is desired to make a change in one or more of the working
constants, it is well to determine the intrinsic constants thenfirst;
a comparison with the above equations will suggest the changes
to make and give the effect of any particular change. The intrinsic
constants can be determined from five independent measure-
all

ments. Measurements which are easily made and from which the
constants can readily be calculated are: (a) the resistance, (b) the
direct -current sensibility, (c) the current sensibility, (d) the electro-
motive force sensibility, and (e) the resonating frequency.
The resistance can be compared directly with a standard
(a)

by means of a Wheat stone bridge or by other convenient methods.

(b) The direct -current sensibility can be obtained by passing

a small known direct current through the coil and observing the
change in deflection when the current is reversed.
(c) The galvanometer is then connected through a high resist-

ance to the terminals of a suitable low resistance forming a part

of the external circuit of an alternating -current generator. Means
should be provided for regulating the speed of the generator. The
frequency of the generator or the free period of the instrument is
then adjusted so the tuning is good. The resistance shunted, the
current through the shunt, and the resistance in series with the
galvanometer gives the current, which, with the observed deflec-
tion, gives the current sensibility.
The electromotive force sensibility is obtained by shunting
(d)

a smaller resistance, and the removal of the resistance in series

with the galvanometer. The electromotive force is then known
9
Defining the direct-current sensibility in the usual way, the relation follows from
equation (1), since when the system has come to rest By or <f>y=\j/i. The deflection
used is that produced on reversing the current or the double deflection to correspond
to the broadening of the image produced by an alternating current
Wenner.] The Vibration Galvanometer. 361

in terms of the shunt and the current through it. This, with the
observed deflection, gives the electromotive force sensibility.
(e) The resonating frequency can be obtained from the speed of

the generator, considering, of course, the number of pairs of poles.

Reference to equations (23), (23a), and (24) and remembering
that /> 2 = y/#, then gives, in terms of the five quantities given
above, the intrinsic constants, as follows:

a^
.
9 i(A-VP ) 8.i(A_-VP )
j
3
AA'V '
f
2
A V
2 '

(27)
3 6(A-Vp) 9 oooo(A-Vp)
UAA'V '
Y ~

AV
f

8. BASIS OF THE EQUATIONS.

In the derivation of the equations, it has been assumed that
the moving system experiences an angular acceleration propor-
tional to the resultant torque, a retarding torque proportional to
the angular speed, a restoring torque proportional to the angular
displacement, and a displacing torque proportional to the current
and independent of the displacement and the rate of change of
the current. That is, (a) the moment of inertia, (b) the moment
of damping, (c) the moment of restoration, and (d) the moment
of displacement are all constant.
It is further assumed that harmonics in the electromotive force
or current wave produce no effect on the behavior of the instru-
ment toward the fundamental, but cause an additional vibration
depending upon their magnitude and frequency. It should be
mentioned that the effect of the time constant of the circuit,
which is generally small in comparison with the time constant of
the instrument, has been neglected. If the time constant of the
circuit is large it may materially effect the rate of change of the
amplitude of the vibration.
The moment of inertia and the moment of restoration can better
be considered together. If the vibration takes place about the
center of the mass the moment of inertia is constant, and for small
displacements the restoring moment is at least very nearly pro-
portional to the displacement. When, however, the system is
unsymmetrical the vibration, because of an unbalanced lateral
13260 — 10 5
 .

362 Bulletin of the Bureau of Standards [Vol. 6, No. 3.

force due to the tension of the suspensions, takes place about some
point or line which does not coincide with or pass through the
center of mass and whose position is a function of the frequency
of the driving force. Under these conditions the moment of
inertia and the moment both functions of the
of restoration are
frequency of the driving force or current and the system has two
10
free periods. In the article referred to above, Rosa and
Grover found that as the frequency of the current was changed,
one of the galvanometers used by them gave maximum deflections
at about 110.5 and 120 cycles per second, and a very small deflec-
tion at 115 cycles. This shows that the instrument had two free
periods differing from each other by about 10 per cent. This
large difference in the periods was caused by a lack of symmetry
introduced on the repair of a broken suspension.
If care has been taken to keep the system symmetrical, or if it
is carefully balanced, the two free periods may be very nearly the

same, but when a few hundredths of a per cent change in the fre-
quency produces a material effect on the deflection, it is possible that
the difference in the periods may still be sufficient to produce an
appreciable reduction in the sensibility and flattening of the curve,
showing the relation between the deflection and the frequency.
A little experience enables us to tell, in most cases, from the vibra-
tions produced on tapping the instrument, if the two periods are
nearly the same, and, if not, what changes should be made. If,

then, the suspensions are bifilar and provided with a means of

independent fine adjustment, the two periods can usually be
adjusted in a few minutes to very nearly the same value.
The moment of damping is made up of a number of different
factors, among which the air pressure and friction, eddy currents
and hysteresis are probabfy the more important. The retarding
torque is therefore only approximately proportional to the angular
speed, so the moment of damping may depart appreciably from a
constant.
The displacing torque is, to some extent, a function of the dis-
placement and is effected by hysteresis and eddy currents. The
latter makes the value smaller for alternating than for direct cur-
rents, so the moment of displacement is not quite a constant.
10
This Bulletin, 1, p. 291; 1905.
Wenner.-} The Vibration Galvanometer. 363

The equations given above, derived on the supposition of these

quantities being constant, therefore should not be expected to give
exact relations. They should, however, indicate fairly well how
an instrument of good design and careful construction will behave
under any ordinary set of conditions. A comparison of observed
and calculated values in the following work gives in many cases
differences no greater than the experimental errors.

9. METHOD OF TUNING.
The method of tuning first used consists of applying to the ter-
minals of the galvanometer a small alternating electromotive force
and adjusting its frequency or the free period of the moving sys-
tem until a maximum deflection is obtained. This is the method
usually used and for some purposes it is very satisfactory. In
general, however, the change in amplitude of the vibration with a
change in tuning is small, especially near resonance. If the tuning
is made in this way, and observations taken for the electromotive
force and current sensibility, a correct value is obtained for the
former, but there may be an error of 50 per cent or more in the
latter.
however, a high resistance (of the order of 100,000 ohms) is
If,

placed in series with the galvanometer, an electromotive force of

suitable value used, and the adjustment made for a maximum de-
flection, the tuning is very much better. This gives a sufficient
accuracy in the tuning for most purposes, including the determina-
tion of the constants.
However, even with the high resistance a sufficient precision in
the tuning could not be obtained for some of the work. It was
therefore necessary to use some other method, and the one chosen
makes use of the phase angle between the displacement of the
moving system and the driving force. On the shaft of the gen-
erator is mounted a large disk back of and slightly above which
is placed the line source of light. The disk cuts off about half
of that portion of the line which it is desired to use except when
exposed by notches, one for each pair of poles, cut in the edge.
These notches are so placed with respect to the windings that
they expose the lower half of the line only at the instant when
the electromotive force is a maximum. Relatively large resist-
364 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

ances are used, especially in the galvanometer circuit, so that the

resonance curve is sharp and the current may be assumed to be
in phase with the electromotive force. One observes then a band
of light above and a line of light below the edge of the disk. When
the tuning is good, the line appears below the center of the band.
When it is disturbed, the line moves toward one side or the other,
depending upon the sign of the disturbance, and the displacement
is considerable before the shortening of the band is appreciable.
A glance at the image tells at once the sign and gives an indication
of the magnitude of the lack of tuning so that the adjustment can
easily and quickly be made. By using a large resistance in series
with a galvanometer having a small moment of damping, this
method indicates a lack of resonance when the difference in the
frequencies is of the order of 0.0 1 per cent of the resonating fre-
quency.
10. GALVANOMETERS USED AND THEIR CONSTANTS.
The galvanometers used were one of the Rubens type made by
W. Oemke, one of the Campbell type made by Robert W. Paul, and
one built in the Bureau of Standards' instrument shop. In what
follows the galvanometers are referred to as Nos. 1, 2, and 3 in
the order here given. The latter was designed with the idea of get-
ting an accurately balanced system of known moment of inertia
and of investigating the effect of air pressure. The general con-
struction is somewhat like the Rubens type, except that a single
magnet is used. The pole pieces are of transformer iron, and the
winding consists of four coils. The moving system consists of a disk
of transformer iron 10.5 mm in diameter and 0.36 mm in thickness,
a circular mirror 10 mm in diameter and 0.58 mm in thickness,
and bifilar suspensions of fine steel wire, the latter provided with
a means for the independent adjustment of their tension so that
any lack of symmetry may be compensated. As arranged, the
suspensions form a coil of a single turn which, vibrating in a strong
magnetic field, causes a considerable damping. This, however, is of
consequence, since for the present purpose we are not con-
little

cerned particularly with the sensibility. The case is arranged so

that it may be closed tightly and the air exhausted. The fine
adjustment of the tuning is made by a soft iron shunt to the mag-
net. This shunt is on the outside of the case, so that a considerable
 .... ;

Wenner.] The Vibration Galvanometer, 365

change in the period can easily be made even while the moving
system is in a vacuum. Unless otherwise stated it will be under-
stood that the case is open.
The constants of these galvanometers are tabulated in Table I.

TABLE I.

Constants of the Galvanometers. 11

Galvanometer No.

Alternating current sensibility. . 1.5 7.3 4.2

Direct current sensibility .018 .010 5.5X10-*

EMF sensibility .0014 .0061 .0075

Resonating frequency 100 100 100

Resistance 234 30 74

3d harmonic current sensibility. .0030 .0017 9.2X10"* 6.7X10"*

3d harmonic EMF sensibility. . 13X10-6 5.6X10-6 9.6X10-* 8.5X10-*

5th harmonic current sensibility. .0010 5.6X10-* 2.2 X 10-*

5th harmonic EMF sensibility. . 4.3X10-6 19X10-6 2.8X10-6

Power sensibility .0021 .045 .032

Resonance effect 39 116 2700

Ratio of back to impressed EMF 78% 97.5% 87%

Resonance range (current) 1-1% •14% •10% .10% .12%
Resonance range (EMF) 2.5% 9% 4.6%
Time constant (current) .2 1.7 2.5

Time constant (EMF) .09 .045 .23

Moment of inertia .012 .014 .015 .024 .020

Moment of damping .12 .018 .016

Moment of restoration 4500 5700 7700

Moment of displacement 20X10* 14X10* 11X10*

Those values obtained from measurements of the current sensi-

bility, the direct current sensibility, the electromotive force sensi-
bility, the resonating frequency, and the resistance are placed in
the second column, while those values obtained by direct measure-
ment or by an independent method are placed in the first column.
Where there are values in both columns, the first serves as a check
on the theory. Additional data for galvanometer No. 2 are
given by Figs. 3 and 5, and for galvanometer No. 3 by Fig. 4.
11
The units used in this table are the same as given on page 3 59, that is, the unit of amplitude of the
vibration is a millimeter broadening of the image of a source of light 1 meter in front of the mirror
the unit of current is the micro-ampere and the unit of electromotive force is the micro-volt.
 —

3 66 Bulletin of the Bureau of Standards. [Vol. 6, No. J.

The generator used in obtaining most of the data given was

directly connected to a motor operating in synchronism with a 12

tuning fork whose frequency could be varied over a wide range by

weights and continuously over a range of about 1.5 per cent by
a spring. This gave the necessary control and regulation for the
work at and near 100 cycles. The electromotive force and current
of three times this frequency was obtained from another generator
whose speed was neither very constant nor definitely known.
11. CHARACTERISTICS OF THE INSTRUMENTS.

An inspection of the curves or a comparison of the constants

will show that resonance has a marked effect upon the amplitude

GALVANOMETER NO.2

kit

cc °
HI
\-

5 5 SI

O
\-
O
-J
U.
o A
ID
O 2
<|

^"MILLIMETERS PER 100 MICROVOLTS

—8- °r~

100.0 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.9 101.0 101.1 101.2 101,3
FREQUENCY-CYCLES PER SECOND

Fig. 2. Showing- the Effect of Tuning on Amplitude of Vibration with Constant Current and
Constant Electromotive Force.

of the vibration. seems to have been the idea that the effect of
It
13
resonance was to increase the sensibility by about ioo times.
Taking the ratio of the alternating to the direct current sensi-
bility gives: for No. i 83 for No. 2, 730; and for No. 3, 7600.
, ; If,

however, by the effect of resonance we mean the ratio of the ampli-

tude produced by a certain amount of power supplied at the reso-
nating frequency to the deflection produced by the same amount
12
Hough and Wenner, Physical Review, 24, p. 535; 1907.
13
M. Wien: Ann. der Physik, 309, p. 443; 1901.
Campbell Phil. Mag., 14, p. 497; 1907.
 —

Wenner.] The Vibration Galvanometer. 367

of power supplied by direct current, these figures are too large,

since no account has been taken of the back electromotive force.

A
matter closely associated with the resonance effect is the
change in the ratio of the amplitude of the vibration to the cur-
rent, with changes in the frequency of the latter. This ratio in the
neighborhood of the resonating frequency is shown by Fig. 2, for
galvanometer No. 2, and it will be seen that the ratio has a sharp
maximum at 100.5 cycles per second and falls to one-half for a
change of 0.14 per cent in the frequency.
It will also be noticed that a change of 0.5 per cent from the
resonating frequency results in no noticeable reduction in the
ratio of the deflection to the impressed electromotive force. Fig.

3 shows the ratio of the deflection to the electromotive force at

CO
GALVANOMETER NO.2
b
EFFECT OF 500 OHMS SERIES RESISTANCE
£ 7
o
cc
o
C7 \T O

cc u
LU
a.
CO
cc 4
LU
h-
LLl

z *
o
y-

S 1

100.0 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.9 101.0 101.1 101.2 101.3
FREQUENCY-CYCLES PER SECOND

Fig. 3. Showing Effect of Resistance in Decreasing Amplitude of Vibration with Constant

Electromotive Force in the Circuit.

the terminals of the instrument and when 500 ohms are placed in
series. The 500 ohms in series makes the resistance of the circuit
18 times larger, yet at the resonating frequency the ratio of the
deflection to the electromotive force reduced by only about 45 is

per cent. The reduction calculated from the constants given in the
table is about 32 per cent. It is evident from the behavior of this
instrument that when it is used in bridge work with a constant
impressed electromotive force the resistance of the galvanometer
circuit, if under 300 ohms, has but little effect upon the sensibility.
 1

368 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

The effect of air damping on galvanometer No. 3, is shown by

Fig. 4. be noticed that the air reduces the resonating fre-
It will
quency by about 0.75 per cent. This change in frequency is
accounted for if we assume that the system drags with it a quan-
tity of air whose moment of inertia is equal to that of a sphere of
air having a radius slightly larger than the radius of the disk. It
is also to be noticed that the removal of the air makes the reso-

nance curve much sharper. The effect which is of real importance

is the reduction of the sensibility by the air friction.

100.0 100.1 100.2 100.3 100.4 100.5 100.6 100.7 100.8 100.9 101.0 101.1 101.2 101.3
FREQUENCY-CYCLES PER SECOND

Fig. 4.- -Showing Changes due to A ir Damping, in Resonating Frequency, Sharpness of Tuning,
and Current Sensibility.

Since the back electromotive force is proportional to the ampli-

tude the power supplied per microampere is, according to the
curves, about 9 per cent less when air damped. For the same
deflection, then, the power supplied, when air damped, is about 1
per cent greater. While the adjustment of the instrument was
changed between the time when the constants were determined
and when the data for the curves were taken yet we may use, with-
out appreciable error, the power sensibility given. This shows
that the power in micro-microwatts necessary to maintain the
vibration is 31 times the square of the amplitude. Taking 11 per
cent of this gives as the power dissipated in air damping 3.5 times
the square of the amplitude.
 —

Wenner.] The Vibration Galvanometer. 369

Fig. 5 galvanometer No. 2 the effect of a large induc-

shows for
tance on the ratio of the amplitude of the vibration to the electro-
motive force. The inductance used was the 480- volt winding of a
600- watt transformer and the increase in the ratio was 125 per
cent. While the resonating frequency was 100.52, as shown in
Fig. 2, yet the maximum amplitude of vibration occurred with a
frequency of 101.06 cycles per second. The effect observed was
much more marked with an inductance about a quarter as large.
No effort was made to get an inductance of the most suitable value.

CO
y-
GALVANOMETER N0.2

01.4
EFFECT OF INDUCTANCE
o
or

ih.2

crl.O
u
a.
co
oc .8
ui
\-
UJ

O
I-

S.2

100.0 100.1 100.2 -fcOO.3 100.4 100.5 100.6 100.7 100.8 100.9 101.0 101.1 101.2 101.3
FREQUENCY-CYCLES PER SECOND

Fig. 5. Showing Shift of Apparent Resonating Frequency and Increase of Amplitude of

Vibration due to Inductive Circuit. Resonating Frequency 100.52, Apparent Resonating
Frequency 101.06. Amplitude with Practically Noninductive Circuit .6 mm per 100
Microvolts.
12. AN ILLUSTRATIVE PROBLEM.
The consideration of the precision which may be attained in the
balance of any one of the alternating current bridges will show
the bearing of some above work on problems of this kind.
of the
On account of the simplicity of the equations, the problem chosen
is the comparison of the product of a capacity and self-inductance

with a frequency by the "series" method. Of course here only

those factors which have to do with the precision of the balance
will be considered. We are therefore concerned only with the
case where the bridge is very nearly balanced and the galvanom-
eter tuned to the frequency of the test current to be used. Most
 — .

37Q Bulletin of the Bureau of Standards [Vol. 6, No. 3.

of the equations following are only approximate and can not be

used except under these conditions.

Fig. 6. Series- Inductance Bridge.

The arrangement of the circuits is shown in Fig. 6, and the con-

ditions for a balance are

r r
—i
= — andJ Lp=^rr.
s T L
l
(28)
r2 rx Lp

For simplicity we shall take

r3 = r' (i + a) and r
f
=r =r =r x 2 4t

also
L =L(i+6),
,

where a is the lack, in proportional parts, in the resistance balance,

and b is the lack, in proportional parts, of the inductance balance.
Then if E 1 is the electromotive force impressed on the bridge the
electromotive force between and M
and between N
and when M
the bridge is balanced is 2 E 1
y When the bridge is unbalanced by
.

a resistance r 3 in place of r and an inductance

f
in place of L the U
electromotive force between and M N
is, if the galvanometer cir-

cuit is open,

£»[r.+V-i(L^-^)]

r*+r

and since Up — ^rr = Lbp this reduces to

Wenner.] The Vibration Galvanometer. 371

2/+ /a +7^1 Lbp

Substracting y 2 E 1
gives as the electromotive force of the gal-
vanometer circuit

_ £ (/q+V17
1
77
! Lbp)
'

4/ +2r'a +2^ — Lbp 1

or since for an approximate balance 4/ is the only significant

term in the denominator,

E = ^r(r'a+TJ~=riLbp). ( 2 9)

If the points N and O are connected by a winding of any electrical

instrument in which the ratio of the impressed electromotive force
to the current is m or simply by a conductor whose resistance is m
and if the circuit is practically noninductive, the current through
the instrument is given by the equation

r +m (30)

and the power delivered to the instrument is given by the equation

Em 2
, .

w is a maximum when m = r f
or
E
Wm ~ '

4r'

If the points N and are connected by a vibration galvanometer

having a resistance approximately equal to its reactance, the cur-
rent is given by the equation

I = (33)
-7Ti
where E h =D(E-r'I) (34)

and -Eb =DE2 .

(35)

E 2 is the electromotive force impressed at the terminals of the

37 2 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

galvanometer. These three equations give

r_ E(i-D)
{3b)
r'(i-D)+p
and

E*=r'{i-D)+p (37)

The power delivered to the terminals of the galvanometer is IE 2,

therefore

' w =>'(i -D)

+pY
m
To get the power converted w', we may multiply w by D or

^y, E*p(D -D*)

(l -D) +tf
-
«39)

The condition for the maximum value of w f

is that

D=P±f„ (40)

or
p=r'{i-D). (41)

Of these equations (29) gives the electromotive force of the gal-

vanometer circuit, and this value substituted in (36) gives the
current through the galvanometer or in (38) gives the power
delivered to the terminals of the galvanometer. Equation (39)
gives the power converted by the galvanometer and equations
(40) and (41) the condition under which this power is a maximum.
Where a transformer is to be used, equation (41) serves to deter-
mine the ratio which makes the sensibility of the whole combina-
tion a maximum. Since in well-designed and constructed instru-
ments the power necessary to maintain a vibration of any specified
amplitude depends only slightly upon the resistance or the ratio of
the back to the impressed electromotive force (provided this ratio
is over 75 per cent) the condition for maximum power converted is
,

at least very nearly the condition for a maximum sensibility of the

combination. Equations and (41) will therefore be of service
(40)
in determining conditions which will give the best results with a
particular instrument.
 —

Wenner.] The Vibration Galvanometer. 373

I^et us now consider the case where the resistances of the bridge
arms are each ioo ohms, the self-inductance .25 henry, the fre-
quency 100 cycles per second, and the electromotive force impressed
on the bridge 20 volts.
In order that the bridge may be balanced the capacity required
is a little over 10 microfarads. If the bridge is out of balance by 1
part in 100,000, both in the resistance and in the self -inductance,
the electromotive force of the galvanometer circuit is, according to
equation (29), 93 microvolts and the maximum power which the
galvanometer may receive is, according to equation (32), 22 micro-
microwatts. galvanometer No. 2 is used, it will be seen from its
If
power sensibility that unit amplitude of vibration is the maximum
which can be obtained.
According to equation (38), the power actually received by the
galvanometer is 6.1 micro-microwatts, and this gives .52 as the
amplitude of the vibration. From equation (36) we get as the
current through the galvanometer .071 microampere, and this mul-
tiplied by the current sensibility gives .52 as the amplitude of the

Fig. 7. Series-Inductance Bridge showing transformer in Connection with the Galvanometer.

vibration. Again, using the electromotive force sensibility given

by equation (23a) , when
taken as the total resistance of the
p is

galvanometer circuit (130 ohms), and multiplying by the electro-

motive force, 93, we get the same deflection.
If, however, we use a transformer connected as shown in Fig. 7,

we see from equation (22a) that the step-up ratio of the trans-
former should be a little less than 3.5. Using 3.5 as the value of X
in equation (26) gives .0105 as the electromotive force sensibility
374 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

of the combination and this multiplied by the electromotive force

93 gives .98 as the amplitude of the vibration. This, we have

just seen, is about the maximum value attainable with this gal-
vanometer.
The use of a transformer in this way, in effect, reduces the- resist-
ance of the galvanometer inversely as the square of the step-up ratio
without changing its power sensibility or the ratio of the back to
the impressed electromotive force. Reference to equation (41)
shows that in effect the resistance of the combination of trans-
former and galvanometer should be 2.5 ohms to receive the
maximum power available. Since the resistance of the galvanom-
eter is 30 ohms (that of the transformer is assumed to be negli-

gible), the step-up ratio should be ^12. Taking this as 3.5 or 2.45
as the effective value of p, gives from equation (?>&) 22 micro-micro-
watts as the power received by the galvanometer, and from the
power sensibility we get again .98 as the amplitude of the vibration.
For a frequency of 300 the impedance of the arm containing the
self -inductance and capacity is large in comparison with that of the

other arms. In calculating the current through the galvanometer,

we may consider this arm to have an infinite impedance, and,
further, since the back electromotive force is so small in comparison
with the impressed electromotive force, it need not be considered.
The current, then, through the galvanometer is approximately the
same as would flow through a resistance of 130 ohms in parallel
with 100 ohms and with the combination in series with 100 ohms
when the electromotive force is equal to that impressed on the
bridge. This gives
F
"
350
If, then, the electromotive force impressed on the bridge has a 3d
harmonic amounting to only per cent, I 3 will be 570 micro-
1

amperes. Multiplying this by the current sensibility for the third

harmonic we see that we have a third harmonic vibration of about
the same amplitude as produced by a lack of balance of one part in
one hundred thousand for the fundamental. Under these con-
ditions, then, the use of a test voltage having a 3d harmonic of
1 per cent limits the precision to about one part in one hundred

thousand.
 .

Wenner.] The Vibration Galvanometer 375

Ifa transformer having a step up ratio of three and one-half is

used, the 3d harmonic current is reduced to about one- third of the
former value, while as we have seen the sensibility of the combina-
tion to the fundamental is about doubled. This extends the pre-
cision of the balance to two parts in one million. It is thus seen
that in certain cases the use of a transformer not only increases
the sensibility of the combination but also permits the use of a
test electromotive force having larger harmonics. If the relations
given above for the use of a transformer are to hold with a fair ap-
proximation, it is necessary that the secondary have a reactance
large in comparison with i-s-(i— D) times the resistance of the
galvanometer, and that the resistance of the windings be small
in comparison with the resistance of the circuits. In calculating
the reactance it should be remembered that the permeability of
the iron core is probably not over 200 for the very low magnetiz-
ing force used. For a frequency of 100, a 2-kilowatt 60-cycle
"potential" transformer having the secondary wound for 15 times
as many volts as the galvanometer has ohms will generally fulfill
the conditions sufficiently well to give results within 20 per cent
of those calculated from the equations. If bridges of different
resistances are to be used the primary winding should be in
sections to give the different ratios needed.
l
13. DESIGN.

While it is not our purpose to discuss fully, in this paper, the

problem of the design of vibration galvanometers, yet a brief con-
sideration, from this standpoint, of the theory and data given
above will bring out a few of the principles involved. The con-
sideration will be limited to instruments to be used in circuits of
fairly low resistance, generally under 500 ohms and low electro-
motive force. This is the condition met with in alternating cur-
rent bridge work which at the present constitutes the main field
in which the vibration galvanometer is used. The instrument
may be connected directly in series with the circuit, or to the secon-
dary winding of a transformer the primary of which is connected
in series with the circuit.
When the galvanometer is connected directly in series with the
circuit, the relation between the amplitude of the vibration, the
376 Bulletin of the Bureau of Standards. [Vol. 6, No. 3.

impressed electromotive force, and the various intrinsic constants

of the instrument is given by equation (9). It should be noticed
that the an plitude is a maximum for a particular value of the mo-
ment of displacement, that is when
^ = r/3.
2
(43)

This is same condition as is given by equation (40) and

really the
both signify that the mechanical power received by the moving sys-
tem is a maximum when the back electromotive force is equal to
one-half the electromotive force impressed on the circuit. With
galvanometer No. 2 used on a bridge having a resistance of 100 ohms
in each arm it will be seen that ty is about eight times as large as
2

r/3. As ^ is proportional to the field strength of the magnet, it

follows that a reduction of 60 per cent in the field strength results
in an increase in the sensibility of the combination of about 70
per cent. A comparison with equation (10) shows that this
change results in a decrease of about 40 per cent in the power
sensibility. Where yjr2 = r/3 equation (9) takes the form
E
*= ~P^ (44)

As r is made up of the resistance of the galvanometer and the arms

of the bridge in a series parallel combination but little is to be
gained by making the resistance of the galvanometer excessively
small. In inductance and capacity bridges p is generally fixed
from other considerations, so we have 13 as the only factor under
this condition, which materially affects the sensibility of the
combination. The power necessary to maintain a vibration of a
given amplitude is proportional to the moment of damping. The
design then should be such that the back electromotive is approximately
equal to one-half of the electromotive force impressed on the total galva-
nometer circuit, and the mechanical power necessary to maintain at the
chosen frequency a vibration of any particular amplitude should be as
small as possible. It is necessary also to take into consideration
the resolving power of the optical system to be used, as this often
limits theamplitude of vibration which may be detected. Where
the galvanometer is to be used regularly with a transformer the
ratio of the back electromotive force to that impressed at the
Wenner.i The Vibration Galvanometer. 377

terminals should be large, though there is little to be gained by

making 90 per cent. Here, as before, the moment of
it over damp-
ing, or the mechanical power necessary to maintain at the chosen
frequency, a vibration of any particular amplitude, when the
circuit formed by the winding is open, should be as small as
possible.
It will be seen that the moment of inertia of the moving system
does not affect directly the sensibility. It does, however, affect
very materially the resonance range and the sensibility to the
various harmonics. Where the balance depends upon the fre-
quency or the wave-form the moment of inertia should be rela-
tively large. On the other hand where the balance is independent
of the frequency and especially if no special device is available
for controlling the frequency, the moment of inertia should be
very small to make the resonance range large.
We have seen (page 386) that the power lost by air friction by
galvanometer No. 3 for a vibration of unit amplitude is 3.5 micro-
microwatts. If we assume
that 75 per cent of this is lost by fric-
tion and pressure against the mirror we see that the power sensi-
bility of agalvanometer having a frequency of 100, a mirror 1 cm
in diameter and operating in air is limited to .4, or about 10 times
the sensibility of galvanometer No. 2.
The sensibility may be made higher than this by changing the
shape and reducing the surface of the moving system, yet if a very
marked improvement is to be made in the sensibility it may be
necessary to do away with the air damping by placing the instru-
ment or at least its moving system in a vacuum.
14. SUMMARY.

i. The importance of the electromotive force developed by the

relative motion of the magnet and the winding of a galvanometer
is pointed out and the general theory of the vibration galvanometer
is developed.
2. The fundamental equation is stated and, under certain condi-
tions, a solution derived giving the various working constants in
terms of the intrinsic constants.
13260— 10 6
378 Bulletin of the Bureau of Standards. [Vol. 6, No. s.]

3. The advantage of using a transformer withan instrument

developing a relatively large back electromotive force is shown.
4. The double period sometimes observed is shown to be due to
an unsymmetrical system.
5. A method of tuning is described which is more sensitive than
the method generally used and which is applicable in some other
cases where the vibration is forced. This method consists in
adjusting the frequency of the driving force or the moving
system so that the phase angle between the driving force and the
displacement produced is 90 °.
6. Equations are derived giving the intrinsic constants in terms

of five easily measured quantities, viz, the resistance, the direct-

current sensibility, the current sensibility, the electromotive-force
sensibility,and the resonating frequency.
7. Intrinsic and working constants for three instruments are
given, some of which were determined by two independent meth-
ods and thus serve as checks on the theory.
8. The principles to be followed in the design of a high-sensibility

instrument for bridge work are developed, viz: (a) On account of

the back electromotive force, heretofore overlooked, the resistance
of the galvanometer should be very much
than the resistance
less
of the bridge (b) the back electromotive force should be one-half
;

of the impressed electromotive force; (c) the mechanical power

necessary to maintain at the desired frequency a vibration of a
unit amplitude should be as small as possible.
Washington, May 25, 1909.