Digital System Design

Professor. Neeraj Goel

Hello everybody, in our last lecture we had a question that how we can represent negative
numbers. So, today we will try to answer this question that how negative numbers can be
represented and also we will try to answer the second question that if negative numbers can be
represented then how addition, subtraction or other binary calculations can be done.

(Refer Slide Time: 00:48)

So, let us start thinking in a way that how do we represent it in a decimal numbers. A negative
number in a decimal representation, we represented by putting a negative sign. So, can we put
the same negative sign in binary numbers? Possibly no, because whenever we are going to use
binary numbers they would be in a digital world and in a digital world, digital world only
understand the language of 0s and 1s. So, that means my negative number or that negative sign
has to be represented in form of 0 or 1. So, let us think of some strategies how we can represent a
negative number using 0s and 1. So one possibility, one naive possibility I would say would
come to our mind that we can represent a, we can leave one particular bit as a sign bit.
(Refer Slide Time: 01:55)

So, that kind of numbers I will call sign magnitude negative numbers. So, here 1 bit which bit to
be choose, so let us say most significant bit. Let us say most significant bit is designated now to
represent a sign bit and 1 would represent negative and 0 would represent positive. So, this could
be one easy sign representation.

If we see this method, this is good. So, let us say we have a 8 bit binary number and I want to
represent minus 35. So, if I want to represent minus 35, so I can put 8th bit or most significant bit
or left most bit as 1 and remaining 7 bits I can use, remaining 7 bits I can use to represent 35, fair
enough. So, 35 means 16, 32 plus 3 plus 1, 32 plus 2 plus 1 is 35, so I can represent minus 35
using this and similarly if I have to represent any number like positive number 127, so my most
significant bit is 0 because the number is positive and the rest of the bits I can use to represent
that number.

So, with this way we can clearly see the range of number could be minus, for 8 bit number it
could be minus 127 to positive 127. So it looks good, intuitive and it looks to work also, but
there are few challenges here. One challenge is that there are two 0s, one that in, this is a
negative 0 because the sign is minus. 1 represent minus sign, so this particular representation is a
minus 0 and this particular representation is a plus 0. So, having two 0s is complicated. We will
see how. So, the other thing is, let us say I need to do some arithmetic here.
(Refer Slide Time: 04:08)

So, what I need to do is, I need to see whether a sign bit is same or not. Let us say I want to do
addition, if sign is same then I have to add if signs are different then I need to subtract. Similarly,
if it is a subtraction then first I need to change the sign and then figure out whether it need to be
added or subtracted. So and addition and subtraction otherwise can be done like we remove the
sign bit, so how the process would be because sign bit have a special meaning here and the rest
of the number is a magnitude number.

So, if I need to subtract these two numbers or add these two numbers, so this is a negative
number and this is a positive number. So, what I need to do is, I need to subtract this positive
number from the negative number. So essentially here also, I need to see that which number is a
bigger number, if magnitude of this number is big, magnitude of this number is here, magnitude
of this number is big, so that is why we are subtracting this from this, this number from this
number.

And subtraction we have learned in our previous lectures, so we do the subtraction that way and
then we get this result. So, whatever result we get after that we add the sign bit, so sign bit would
be added based on the calculation, so because whatever number we found is a bigger number, so
sign of that number will prevail. So, this is how subtraction and addition would happen if the
signs are different.
So, here now finally we have added the sign bit of 0 because this number was a bigger number.
So, there are couple of steps involved, it overall, in summary I can say this steps are first I need
to find out which, I have to first remove the sign bit and then find out which number is a bigger
number. If subtraction need to be done then we need to do subtraction using the subtraction
method.

And then after that whatever result is achieved then we have to find which is the sign bit of the
result and then catenate that or add that result bit in the MSB and then that is the final addition.
Similarly, if subtraction has to be done then again if signs are different then we need to subtract,
if signs are same then we need to subtract, if signs are different then we need to add and then
find the corresponding result of the sign bit.

(Refer Slide Time: 07:17)

So, overall process is little bit clumsy because here we are doing a, like we first of all require a
different hardware for addition and subtractions and we do not know which hardware is required,
you need to require both adder, subtractor and a controller which is determining when to do
addition and determining which sign bit is required that is additional steps. So, along with this
arithmetic confusion there is also additional confusion because of two 0 bits, so two
representation of 0. Can we have something better? So, let us think of some other new
representation.
(Refer Slide Time: 07:52)

So, let us say we are representing it in a 1s complement manner. 1s complement means we are
taking in complement of that number. Let us say if u is a number and I need to find out the
negative of that u, then what I do is, I will flip all the bits. That means if the bit was 0, I make it
1, if bit is 1, I make it 0, so that way flip number is a negative number that is my assumption and
this particular representation I am calling 1s complement representation.

So, because for positive number we assume that MSB most significant bit is 0, so because we are
flipping, so last bit or most significant bit is 1, that means it is negative and if most significant bit
is 0, then it is positive and N minus 1 bit would represent magnitude, so in a way this is very
similar to sign magnitude, that 1 bit is reserved for sign and rest N minus 1 bit is representing
magnitude but the order is different, because we are here inverting the bits.

So, let us take some example, let us say this is plus 13, 0000 1101 it is in 8 bit and then minus 13
it is simply inverting all the bits, 1111 0010 or similarly any other number, so essentially
receiving or getting a negative number is easy I just need to flip all the bits or reverse the polarity
of all the bits and because I can represent N minus 1 bits of magnitude, the range of the numbers
would be minus 2 is to power N minus 1 to 2 is to power N minus 1. So, I can represent minus 1
to 2 is to power N minus, so for example if it is a 8 bit number I can represent minus 127 to 127,
so this is the number of bits I can represent.
(Refer Slide Time: 10:01)

Here also, we see that we have two 0s, so before that let us also see the mathematical
background, so when I am flipping all the bits, it is also equivalent to subtraction from 2 is to
power N minus 1 because 2 is to power N minus 1, so let us say the case is 8 bit number, for 8 bit
number, the 2 is to power N will be 256, 256 minus 1 would be 255, so that means all 1s.

And if I want to get a negative number, I am essentially doing inverting of all the bits which also
means I am subtracting from 2 is to power N minus 1 and I am subtracting this u or basically
magnitude of this u from the 2 is to power N minus 1, so this can be we can say this is the
mathematical representation of this number, negative number.

Now here also there are two issues, one issue is I have two 0s, one I call positive 0 because sign
bit is positive and the other is the negative 0 which is invert of all this 0, so this is a negative 0.
Addition and subtraction, if I am doing it on only positive numbers then it mostly works, if I am
doing only with the negative numbers then also it works, but I am doing addition and subtraction
using regular arithmetic or regular binary additions, then if one number is positive, one number
is negative, it does not work.

So for example, here if you see this particular number is of 13 and this particular number is 1, if I
invert this, this is 1101 so that means it is also 13. If I am subtracting 13 and adding positive 13
and minus 13, if I add plus 13 and minus 13, let us see the output 1 plus 0 is 1 0 plus 1 is 1 and
similarly all the 1s I am getting. So, minus 13 plus 13 is equal to minus 0, still okay. Still okay
because at least it is giving me 0.

But on the other hand let us say, consider another number, so this number is minus, this number
is plus 13 and this number is minus 12, so I am adding minus 12 and plus 13, the output is 0, so
that means there is this particular output is wrong. So, addition works for a negative number or
like adding two negative numbers or adding two positive numbers, adding negative and positive
numbers will mostly work but it does not work in some cases.

So, because it does not work in some cases, this particular condition is not good at all. Because
we believe computers, computer is supposed to do everything correct functionally. So, that
means we need to find out some other representation which is better than this, which does not
create this confusion and this particular error came essentially because we have two
representation of 0s, one representing all 0s, another representing minus 0 that means all 1s. So,
because of two representation of 0s we are getting this, this correction error of 1 in some cases.
So, what can be the other method of representing negative numbers?

(Refer Slide Time: 14:10)

Another alternative we can think or is we can add a bias. So everything is positive in our
numbers, so let us say if I want to represent any number, I add that particular, some particular
bias and for example, let us consider that for 8 bit number, I add a bias of 127. So, if I am adding
a bias of 127, let us see how it works.
So I want to represent 1, I want to represent 1 in binary, so first thing I need to do is I need to add
127, so it become 128 now I will represent this 128 in regular binary, so that means 100 0000.
So, similarly for any other number if I want to add, I want to represent in binary, then first I need
to add this offset or this bias. For example, I want to represent minus 127, so first I need to add
bias, so minus 127 plus 127 will be 0.

So, that means if this number is there 0000, it actually represents minus 127 in bias mode, so in a
bias representation. Similarly, if I want to represent plus 128, the output would be 128 plus 127
that means 255 and 255 can be represented using 1111 1111. So, there are couple of observations
here, now if I want to represent any number using bias method, then for every like if my number
is 6 bit, 7 bit, 8 bit, bias has to be defined, it should be known to everybody who is doing
computation.

So, it is better to standardize a bias if I am using a bias method of representing negative numbers.
The other thing is, for representation sake it is still okay but if I need to add or if I need to
subtract numbers, then I need to take care of bias. Because let us say I am adding two numbers,
so F u plus v, u and v are two numbers, I want to add two numbers, so biased number this F u is
bias number, this is also bias number because bias is added here as well as here. So, if I want to
add two numbers, I need to finally subtract bias out of it.

So, this representation using biased method even for addition or subtraction, I have to, in case of
subtraction I have to add bias, in case of addition I need to subtract bias, then only I can get
actual addition or subtraction. The other thing is let us say the number is 6 bit, number is 5 bit,
for 5 bit number bias would be different, for 7 bit number or 8 bit number bias is going to be
different. So, this bias further, as I said earlier it should be known whenever we are doing any
kind of arithmetic using this biased negative numbers.

This also looks crude, although we have solved one particular issue with this, that 0 will have
only single representation, but because addition, even addition and subtraction would require
additional step of adding bias or removing bias. For multiplication, it is even more difficult, we
will not be able to use this bias method for doing multiplication. So this leave it, leave us at a
crossroad that we are not able to find intuitively what, how should we represent negative
numbers.
(Refer Slide Time: 18:38)

So, to find out the solution let us go back and see that where why it is so challenging or what are
we actually looking for? What should be the ideal representation of negative numbers? So, one
this we would like is, it should be simple, two representation of 0s will give unnecessary hassle
and the results are not desirable because it will create at least some confusion. The other thing
the addition and subtraction should be easy and it would be because we are talking of a digital
design or a digital hardware, it would be even better, if addition and subtraction both can be done
using a same hardware.

Or in a same method it because we are sometime adding numbers, some of them are positive,
some of them are negative, adding negative and positive numbers, so that is why if there is only
single method of addition and subtraction that will make our life easy. So, again to look at things
let us, let us look at another representation of numbers that may help us in understanding how
should we represent negative numbers.
(Refer Slide Time: 20:03)

So, this particular method of representing number is called number circle. So let us say we have
4 bit, I want to, I want to find out 4 bit numbers, so I would like to represent I can represent it
like a circle. So, how do I write this, create this circles, because I want to create it using a 4 bit,
so I will first of all write four 0s that will be the top of this circle and every increment will give
me in a clockwise direction next number.

So, increment from 0 to 1 is next point in the circle, 1 to 2, 2 to 3, similarly, every addition is
giving me a circle and when I reach 15 and again add one then it will become 0, so there is a
discontinuity here. So, this discontinuity is also called breakpoint and if this essentially, this also
means that I cannot represent any number bigger than 15 in a 4 bit representation. So, this is also
called overflow.

So, there are couple of more things here, first of all I am moving in a clockwise direction. If I
want to add, let us say if I want to add 4 to any particular number I want to add 4 to 1, so I will
go four steps in a positive direction, 1 2 3 4 and let us say I want to subtract, let us say I want to
subtract 4 from 12, what I will do is, I will go four steps in anticlockwise direction. So, in
anticlockwise direction I go, so 11 10 9 8, so after four steps I got 8, I got at 8 so this is what it
is.
But if I crosses this particular line which is a line of discontinuity, then result will not be correct
here. Here, at least in the positive circle, the results will because we are reaching a condition of
overflow.

So, let us say I want to add 4 to 14, so because I have crossed this particular circle 14 15, 0 and
1, so I will get an output of 1 but because it is crossing this particular line of discontinuity or line
of breakpoint so I would get an overflow condition or similarly going in anti-flow, anticlockwise
direction can give it, give me a underflow condition. So, this particular diagram has been taken
from this website which we will visit after couple of slides. So, can I take some clue from this
number circle.

(Refer Slide Time: 23:06)

If I represent in such a way that let us say, so that I can represent some negative numbers also. So
in case of negative numbers, what I am doing is I am trying to do in the same way that going
clockwise will give me correct results, going anticlockwise will also give me correct results.

So, let us say I want to subtract 4 from 2, I want to subtract 4 from 2, so I will go in a negative
direction or in a anticlockwise direction four steps, 1, 2, 3, 4, so subtracting 4 from 2 is giving
me minus 2. Similarly, if I am adding 5 from minus 3, I am adding 5 to minus 3 then I am going
in clockwise direction because I am doing addition, I am going five steps, 1, 2, 3, 4, 5, so adding
5 to minus 3 is giving me 2. So this means, this number circle, if I represent in this way it can
give me some good characteristics that it is a intuitive or a standard representation.
(Refer Slide Time: 24:23)

So, we understand clockwise means addition and anticlockwise is subtraction. So, this is at least
one observation we got from here, the other observation we can see is if I want to see any, I want
to get a representation of plus a number or minus number, so let us say I want to see plus 4, so
what I can do I can do write clockwise direction from the 0 1 2 3 4, it gives me this particular
number.

So, that means this is the representation of plus 4, so similarly if I go in a anticlockwise direction
from the 0 then it will give me minus 4, minus 1, minus 2, minus 3, minus 4, so if I put my
numbers in this way, so we are all talking, I am right now talking only in the decimal form, in the
decimal form along with the decimal form, we can look at the binary also, so that means 1111,
adding 1 will give me 0000 and this addition is all correct.

But adding 1 from 1101 will give me 1100 and if I add 1 to this number then it would be 1101,
so that means again it is a natural representation that adding 1 in a clockwise direction is giving
me correct results and there is only one representation of 0, discontinuity would happen here, so
discontinuity would happen here, so if I am crossing this particular line, so let us say I am adding
1 to 7, then I will give me 1000, because I am crossing this line discontinuity then this would be
an overflow condition.

So, because 1000 does not represent 8 but it represent minus 8. 8 is essentially out of bound for
this particular representation, for 4 bit number, the number range is a 7 or minus 8 to 7. So,
whenever I cross this particular line of discontinuity or a line of breakpoint, then there would be
overflow or underflow; overflow if I am going in a positive direction and underflow if I am
going in a anticlockwise direction.

The other observation here is, let us say I am taking u steps or some x steps in clockwise, let us
say I am taking four steps from 0 to four steps here, this also means the same thing can be
achieved by going anticlockwise in twelve steps or 2 is to power N minus u anticlockwise steps
are equivalent to u clockwise steps. So, let us remember again, clockwise means we are adding u
numbers, adding a number u and anticlockwise means we are subtracting number u.

So, that means if I can represent addition and I can represent it using subtraction also that means
here addition can be represented using subtraction that is a very magical property, this particular
circle has. So, because now addition can be represented using subtraction, so that means I can do
addition using subtraction or subtraction using addition, this makes my life simple, easier, there
would be only single representation and there would be only single way of adding numbers or
subtracting numbers. So, I will quickly break this presentation to see this number circle and then
we will get back to this presentation.