Analog Electronic Circuits

Prof. Shouribrata Chatterjee

Department of Electrical Engineering
Indian Institute of Technology, Delhi

Lecture – 23
Op-amps, OTAs

Welcome back to analog electronic circuits. This is the 23rd lecture in this series.
We are going to continue with the Op-amps and the OTAs we were discussing in
the last class. We are also going to look at how negative feedback is going to be
used with these Op-amps and OTAs. That is pretty much the topic for today’s
lecture.

Figure 1

The circuits shown in Figure 1 are the two circuits with opamp. Analysis of these
circuits is very straight forward, all we have to remember is that, if we have drawn
the circuit right, then the voltage at minus terminal is going to be equal or almost
equal to the voltage at the plus terminal, which means that the voltage at the minus
terminal in the circuit I in figure 1 is 0.

In the other circuit II, the voltage at the minus terminal is more or less is going to
be equal to the voltage at the plus terminal, which means in this circuit the voltage
is Vin. There is one more thing that we need to keep in mind, that the current going
into the plus and minus terminals is always 0. Note, this is the gate, we mean
inside the Op-amp there are MOSFET’s, and these are gates of the MOSFET’s.
So, the current going into the gate is 0, which means that the current going into
the minus terminal is 0, the plus terminal is 0.

In circuit I, what is the current through R1? Left side is Vin the other side is 0. So,
𝑉𝑖𝑛
the current through R1 is , all of this current is forced to go the through resistor
𝑅1

R2 because no current can go into the minus terminal, which means that the drop
across the R2 resistor is going to be the current times R2, which means that the
𝑅
drop across the R2 resistor is going to be 𝑉𝑖𝑛 . 𝑅2. We start from 0, the potential on
1

the plus side is 0; that means, the potential on the minus side is going to be
𝑅
−𝑉𝑖𝑛 . 𝑅2.
1

𝑅2
𝑉𝑜𝑢𝑡 = −𝑉𝑖𝑛 .
𝑅1

𝑉𝑜𝑢𝑡 𝑅2
=−
𝑉𝑖𝑛 𝑅1

This is the standard analysis; this is the short cut analysis that we need to know
because these circuits are going to come often. These are all very famous circuits,
which are favorites of people. In the second circuit, what happens? + terminal is
𝑉𝑖𝑛
Vin the other terminal is Vin and therefore, the current through R1 is . Clearly this
𝑅1

current has to come through R2 it has no other way. Because the current going
into the minus terminal is 0, that current has no other way, it has to come through
𝑉𝑖𝑛
R2 which means the drop across R2 is . 𝑅2 . We start from Vin, that is the voltage
𝑅1
𝑉𝑖𝑛
at the minus terminal. The voltage at the plus terminal of R2 is going to be . 𝑅2
𝑅1
𝑉𝑖𝑛
more than Vin, which means Vout is going to be equal to 𝑉𝑖𝑛 + . 𝑅2 , we can take
𝑅1

Vin common, and that is our 𝑉𝑜𝑢𝑡 /𝑉𝑖𝑛 .

𝑉𝑜𝑢𝑡 𝑅2
=1+
𝑉𝑖𝑛 𝑅1

This is the standard way of analyzing this. We have one important issue over here,
that we were discussing in the last class, if the Op-amp that we made is not really
an Op-amp, but an operational trans- conductance amplifier. We did not really
make an Op-amp, because its output impedance was large. So, either the output
node is a drain, it has got to be a drain of some MOSFET; either it is a cascode
drain, or an ordinary drain; whatever it is; the output impedance of it is somewhat
large. And large output impedance automatically means that the circuit is going to
behave like a current source, instead of a voltage-controlled voltage source, which
the Op-amp should be. What we have unfortunately made is a voltage controlled
current source. The question that we were trying to answer is that, is this going to
affect anything. By the way, the OTA is drawn like a trapezium not a triangle, this
is a voltage controlled current source.

Figure 2

If we make the same circuit with an OTA instead of an Op-amp, will it work? What
is going to happen? This is what we were thinking about in the last class. Now,
what is the current going into the minus terminal? This is made up of MOSFETs.
It has got to be 0 perfect. The current going into minus terminal is nothing, but 0
which means whatever is the current through R1, all of that current has to go
through R2. Now, all that we need to establish is, what is the voltage at the minus
terminal.

We are going to do a brief analysis. Before we discuss about the voltage at the
minus terminal, we have one more issue at hand.
From the last discussion to this discussion one thing that we did not took up is why
the voltages at the 2 input terminals more or less equal? That is question 1. Next,
is there a problem, if we draw plus minus instead of minus plus? Question 3 what
is the voltage at the minus terminal actually in the OTA case? How does it work
out?

These are our questions right now and, we are going to try to answer all of these
questions with analysis. The idea is this, if we think long back towards the first or
second class, we had studied the load line, when we talked about diodes when
we talked about the MOSFET initially we talked about the load line.

We are going to bring the load line back into the picture to do our analysis and,
the way it works. Let us call the voltage at the minus terminal Vx, since we do not
know what it is; let us call it Vx. Now, imagine this OTA is not there.

Let us for once, throw it out of the circuit, let us not worry about the presence or
absence of this OTA. There is something over there clearly, we do not want to
know, what it is. Can we write some KCL KVL? We can write the KCL at the node
Vx and, that KCL will look like given below.

𝑉𝑖𝑛 − 𝑉𝑜𝑢𝑡 𝑉𝑜𝑢𝑡 − 𝑉𝑥

+ =0
𝑅1 𝑅2

In any case even after we plug the Op-amp or OTA back over there this current is
going to be 0. Our KCL is still going to be valid, with or without the Op-amp or OTA
whatever it is. Now this R1 and R2 are just causing extra stress over here, let us
assume R1 and R2 are equal without any loss of generality. For now, we are going
to assume R1 and R2 are equal just for this analysis because, we do not want to
write too much, not much is going to be changed, just the equations will look a
little simpler. The ideas remain exactly the same, if you do not like it, you can keep
R1and R2 in our notebook, we are going to remove R1and R2 and call it one general
R. So, if we call it R, then the R can safely cancel out because on the other side
there is 0, that is why we made it R. And we are going to get,

𝑉𝑖𝑛 + 𝑉𝑜𝑢𝑡 = 2. 𝑉𝑥
Now, Vout is an unknown we do not really know what is going to be Vout, Vx is the
input terminal to our device; let us call this the device, Vx is the input terminal to
our device, Vout is the output terminal of the device. Let us make a graph with x-
axis as Vx and y-axis as Vout.

Note the choice of axes, the choice of axes is the input terminal to the device and
the output terminal of the device, where my device is the Op-amp or the OTA
whatever it is. We have an equation over here 𝑉𝑖𝑛 + 𝑉𝑜𝑢𝑡 = 2. 𝑉𝑥 and, we would like
to plot this equation on the graph, where the x-axis is Vx, and the y-axis is Vout.

It is going to look like a straight line with the slope of 2. And if Vx is 0, Vout is going
to be -Vin for example. It is going to cut the y-axis at - Vin. Whatever Vin is we do
not know. If Vin is + 0.5, then it is going to cut the y-axis at minus 0.5, if Vin is +1
then this line is going to cut the y-axis at minus 1. If Vin is 0 then this line is going
to cut the y-axis at 0, if Vin is minus half it is going to cut it at plus half if Vin is -1
then it is going to cut it at plus 1. We get a family of curves, for different values of
Vin, we can draw different curves and, all of these curves are straight lines which
are parallel to each other as shown in the graph drawn in figure 2. They are all
straight lines with slopes of 2, our drawing is very bad over here, it looks like it is
45 degrees, it is not 45 degrees it should be something with the slope of 2.

We have not bothered with the device at all, we have only done one KCL and this
KCL is valid. Whether this device is an Op-amp, or this device is an OTA, or there
is no device at all it does not matter, the KCL is still valid.

This device is not a diode, diode would not work because it is going to take current,
and we took the current into the device equal to 0. So, it should not be that the
current into the device starts mattering. It should be a device such that that the
current is 0. So, op-amp OTA, air these are valid devices for us right now.

Now, we go to the market, purchase the Op-amp, or the OTA, we sit at the drawing
board design the OTA, and then we plug it in over here. Before we plug in, we
need to ask the question, what is the characteristics of this OTA, what is the input
output characteristics of this OTA.
Remember the input terminal is Vx the other input terminal is at ground. So, we
do not worry about the other in input terminal. The output terminal is Vout. Now, the
way an Op-amp or an OTA is going to work is that if the minus terminal is positive,
then the output is going to become negative. If the plus terminal is positive in this
case the plus terminal is ground, if the minus terminal is negative, the output is
going to become positive. This is the inverting terminal basically, then we design
the OTA for a large voltage gain, of let’s say 10000. If our input is minus 1 milli
volt, the output is going to try to shoot up to plus 10 volts.

Now, it depends on what the power supplies of this OTA are, if our power supply
is only 2 volts then of course, we cannot shoot the output up to plus 10 it is going
to stop at 2 volts. If the power supply is 1.2, it is going to stop at 1.2. If the power
supply is 12 volts than 10 volts is okay. This is what is going to happen, this is the
kind of characteristics.

Let us draw it, the drawing is going to look like the characteristics drawn in red
color on the graph in figure 2. We are drawing is the characteristics of the device.
And the gain is the slope in the transition region is, -10000. In this case this slope
is going to be -10000. Because we have designed the OTA or the Op-amp with
the gain of 10000.

So, we have got these blue lines are because of Kirchhoff, the red curve is
because of the device, the circuit has to obey both the blue as well as the red. It
implies that it is going to operate at the point of intersection of the 2 curves.
Suppose our Vin value is +1 volt, and it is going to choose the particular line, which
cuts the y-axis at -1. Now, we have chosen that line; the point of operation is going
to be the intersection point of that blue line and the red curve, which happens to
be the green point marked on the graph on the line which cuts the y-axis at -1. If
we change the input to -1 volt, it is going to choose the point marked on the line
which cuts the y-axis at 2. So, depending on the value of the input, we are going
to get a value of the output.

And clearly what is going on over here, the one thing that we should be able to
see visually over here is that Vx is going to be very close to 0. We mean it does
not matter what is Vout, but Vx over here has got to be equal to 0, because we are
cutting is so close to the y-axis. Even if it is not 0, it is something very small
because the gain of the Op-amp or the OTA is so high.

If Vx is equal to 0, does it matter if the device in circuit figure 2 is an Op-amp or an

OTA? because, the current through R1 goes in, no current goes into the OTA, the
same current goes out, automatically we have declared the value of Vout. It is the
same analysis whether it is an Op-amp, or if it is an OTA. We do not really need
to make an Op-amp, an OTA will do just fine.

But take a look at it, be careful, think about it before declaring victory. It looks as
if Vx is equal to 0, then the result is the same whether it is an Op-amp or an OTA.
Now the question is, is Vx going to be equal to 0, and that question is answered
by this graph. If the gain of the Op-amp is indeed high, then there is no problem.

Now, there is a problem, that the gain of the OTA, there is a chance that it is going
to reduce, because of the circuit. Why is it going to reduce? Remember the gain
of the OTA was the trans-conductance times the output impedance. Trans
conductance, we have got something. Looking into the output terminal of the OTA,
we had some output impedance and, we got the trans-conductance times that
output impedance was the gain. But if we look at Vout and look at the output
impedance, we will have the output impedance of the OTA and, also R2 because
R2 is right to ground. So, looking into the output of the OTA we see some
impedance, but we also see R2 in shunt with it.

Now, 𝐺𝑎𝑖𝑛 𝑜𝑓 𝑡ℎ𝑒 𝑂𝑇𝐴 = 𝑡𝑟𝑎𝑛𝑠𝑐𝑜𝑛𝑑𝑢𝑐𝑡𝑎𝑛𝑐𝑒 𝑋 (𝑜𝑟𝑖𝑔𝑖𝑛𝑎𝑙 𝑜𝑢𝑡𝑝𝑢𝑡 𝑖𝑚𝑝𝑒𝑑𝑎𝑛𝑐𝑒 || 𝑅2 )

There is a chance that the gain of the OTA is going to reduce. This is a reason to
worry; how worrisome this is; depends on how small the value of R2 is. If the value
of R2 is very large, then it is not a problem, because, the large resistance is in
shunt, nothing much is going to happen. But if the value of R2 is small compared
to the output impedance of the original OTA, then it is a problem, as the gain of
the OTA is going to reduce. And if the gain of the OTA reduces what happens to
our curve? The curve becomes a little weaker as shown in green in the graph in
figure 2.
Of course, this is a highly exaggerated curve instead of a slope of 10000, we might
have a slope of 1000. Our gm is still large. We will still have some gain instead of
gain of 10000 may be our new gain just 1000. In which case this green curve does
not look like green, we mean it will still look like the red curve, 1000 slope is still
very large.

But in any case, what we are going to see now is that we are slightly off from our
answer, but we are still more or less over there. We are still going to follow the
pattern that Vx is almost equal to 0. Just that we are a little worse off, how much
worse off will have to do numerically quantitatively. Graphically, it looks like we are
still not that much worse off. Vx is still something which is very close to 0.

This answers a few things- first it answers, why are the two terminal voltage is
more or less equal, this graph should answer that. Number 2 we should be able
to understand, why an Op-amp and OTA are very similar and what the difference
is. The Op-amp has low output impedance to start with. It is a voltage-controlled
voltage source. So, the output impedance is very low.

In case of an Op-amp it is a voltage-controlled voltage source, with a very low Rout.

If Rout is low, then putting an additional R2 in shunt with it is not going to change
anything it is going to continue remaining low. Nothing much is going to happen,
the gain is going to remain pretty much the same, because the gain is relying on
gm not on Rout. gm is very large; Rout is very small in case of the Op-amp. While in
case of the OTA gm is a nominal value Rout is very large because, it is a current
source. So, this is pretty much the difference between the Op-amp and the OTA.
Two questions answered, the third question was what happens, if we interchange
the minus and plus terminals, what happens if we make a mistake, what is going
to happen? Are the blue curves going to remain same, or are they going to
change?

The blue curve had nothing to do with the device; the blue curves were made
when the Op-amp was thrown out of the picture. So, they will do fine they will
remain exactly where they are, no change in the blue curves. The only change is
going to be in the red curve, the curve for the Op-amp, because now that we have
interchanged the minus and plus terminals.
The polarity of the red curve is going to change. So, instead of the way it is, it is
going to start looking the other way, just the polarity is going to change because,
now if Vx is slightly positive, then Vout is going to be positive.

So, that is what happened, this is the only change that takes place. And what
happens, we have to establish the point of intersection between the blue curve
and the device curve. Let us say we pick this straight line where Vin is 0, where
does it intersect the device curve. Earlier it was intersecting at the origin, it still
intersecting at the origin, but there are two more points of intersection.

We have got two extra two new points of intersection. This straight line is crossing
the device curve at three points of intersection, not 1, and this is a disaster. The
way it works is, it is either going to be at the top point, or it is going to be at the
bottom point, it is never going to show up at the center point.

It is almost like we have placed a football on top of a hill and, we expect the football
to remain right on top of the hill and not roll down on any one of the two sides. And
once it is rolled down on one side it is going to stay there. So, when we make such
a circuit if we happen to interchange the plus and minus terminals, the circuit is
going to remain at either the positive supply voltage or the other supply voltage,
and it is never going to move from there. Notice the DC operating point gain, the
small signal gain at the two extreme points of operation is 0 that means, the Op-
amp has no gain. If we apply a small input signal nothing is going to make a
change in the output, it is going to remain exactly where it is. Curve is totally flat
over here. This circuit does not work basically, if we flip the plus and minus
terminals. We will either be stuck at bottom point, or at top point we will never
come in the middle and, that is the reason why the plus and minus terminals are
important.

We have to draw it correctly, if we do not draw it correctly, and make a mistake

the circuit does not work. The circuit is not unstable. A lot of peoples come back
and say, the circuit is unstable; it is not unstable, it is a very stable, right there is
no stability issue over here. The circuit is rock stable, once the football has rolled
down on the other side; it is not even going to bother trying to come back to the
other side of the hill. It is there it is going to remain there forever. The circuit is
very stable, there is no instability, it is just that there is no gain and, this is not the
operating point that we were looking forward to.

So, what is the mechanism to make sure that the operating point of the Op-amp
is correct? This mechanism is called negative feedback.

Figure 3

This is the mechanism by virtue of which the operating point of the Op-amp was
set nicely. Let us now imagine we just have Op-amps. We are going to use this
triangle and the trapezium interchangeably, and do not worry too much about it,
whether it is a triangle or a trapezium right now.

For example, in the circuit drawn in black in figure 3, suppose the input voltage is
0, now we can do the analysis of this with the help of the curves in figure 2. We
choose the center blue line because the input voltage is 0, and we find that the
point of intersection is at the center and lo and behold everything works out. That
is fine. We already understand that analysis.

Now, let us think of it in a slightly different way, think of this node Vx, if Vx is a little
more than plus, if Vx is a little higher than the plus terminal, then the output terminal
of the Op-amp is going to be a lot lower, because the input is at the minus terminal
and the Op-amp has large gain. If Vx is slightly higher than what is required, then
the output is going to be a lot lower and, when the output voltage goes down it is
going to pull the voltage at the node Vx down through the resistor.
Now, think of it this way, the left side of resistor R1 is fixed at ground, right side is
at Vx. If the other side goes drastically down, R1 and R2 forms the resistive divider;
it is automatically going to pull down the point in the middle, like a lever. The
fulcrum is on one side, the fulcrum is ground. The center point Vx.

If Vx is a little higher then, the output voltage on the other side goes drastically
down, if this goes drastically down then because of the resistive divider, it is going
to automatically bring the voltage at Vx is also going to come down. If it is a little
higher than required, then immediately the output of the Op-amp is going to
respond and pull it back lower, if Vx is a little lower than required, then the output
voltage is going to go drastically up, and it is going to pull it back to the right place.

This beam is going to be balanced right at the center. This principle is called
negative feedback. If the voltage at Vx is slightly lower than required, then the
output is going to go drastically up, if it drastically goes up then, it is going to correct
the mistake that happened at Vx to start with. This principle is called negative
feedback.

Now imagine a situation where the Op-amp was connected the wrong way as
shown in the circuit drawn in red in figure 3. If Vx was slightly lower than required,
as it is connected to the positive terminal the output is going to go drastically low,
it is going to go drastically low and pull Vx down even further. Vx was a little low
the output is going to go drastically down and when it is going to pull Vx down all
the way. When Vx goes a little further down, again the output responds by going
further drastically down. This beam is not going to be balanced at all. We started
from here Vx was a little down immediately the output fell drastically down. This is
eventually going to collapse it is going to look like that drawn in green in figure 3.
If we draw mechanical beam over there it is just going to collapse all the way.

If Vx was a little up to start with, then it is going to push up. If Vx was a little up to
start with, then the output is going to be drastically up, and it is going to push it
upwards. We would not get balanced in this particular beam, when the output is
reinforcing the mistake, we started with an error, and then we are reinforcing the
error this is called positive feedback as opposed to negative.
We will see negative feedback everywhere in real life, it is there in all places, even
when we are trying to pick up an object. Suppose there is an object we keep it on
the table, close my eyes and now I do not know where the object is, I am looking
for it, I cannot find it. The moment I open my eyes; I see where our hand is and
where the pen is and, I command my hand to go a little bit forward by the error.
The error that I committed I try to fix it, this is also negative feedback. So, negative
feedback is there everywhere.

The circuit in black in figure 3 over here is a negative feedback circuit and, here
both terminals are at ground, now where we apply the input is irrelevant. We can
apply the input at negative terminal in which case it becomes the first opamp circuit
we had discussed; we can apply the input at the plus terminal instead; in which
case it becomes the second circuit. Both the circuits come from the same root,
just where we apply our input.

Figure 4

Let us do one more. Some of us might know how to drive; most of us know how
to ride a bike. Suppose we have a hilly road, there are ups and downs on the road,
and we are told that, we have to drive our bike at a constant velocity no matter
what. It is a straight road or just that there are ups and downs all the way. We have
to drive our bike at a constant speed. Now, the most important thing that we need
is a speedometer, if we are given such a task.
Something that is going to tell us our current speed, without a speedometer, this
is a no brainer we have to give up and say, we cannot do it. The first thing that we
need most important to do this job is a speedometer. We are going to draw the
speedometer as shown in figure 4. We are going to call the rightmost arrow as the
speed. It shows the current speed or actually this is the current action, and the
speedometer is going to measure our current action, and declare a value for the
speed. This box over there is the speedometer.

We have been asked to travel at a constant velocity that is 40 kmph. The

speedometer is always showing the current speed. Sometimes, when we are
trying to go uphill our speed is probably coming down. Sometimes when we are
trying to go downhill our speed is increasing and so on and so forth. So, maybe
this 39 maybe this is 41 some number is coming out over here.

Now, what is going to be our job? Our job is going to see that 40 kmph was my
target velocity, this is our current velocity what is the difference between the two,
what is the error, what error are we making? Maybe we are going downhill, and
our speed is 41. The error in that case is -1 kmph from the target, maybe we are
going uphill, and the speed is 38. The error in that case is + 2 kmph.

We are trying to measure the error and, then after measuring the error what we
have to do? Depending on the error we have to take preventive action. either we
press our accelerator even more or we ride our bicycle harder, we push more force
on the pedals, whatever we want to do right. If we are driving a scooter, we put
more petrol into the engine right, we turn the accelerator.

There is our actuate. We actuate whatever is to be done and, how much we

actuate depending on the error. If the error was 0 we do not actuate at all; if the
error was minus 1, we have to break. So, depending on the error we have to
actuate, we have to either accelerate or decelerate or keep our velocity just the
same. Here we have a gain.

This is the general scheme of things. Now this looks straight out of our control
theory book. This is the classic negative feedback schematic. The Op-amp is
going to come right at the place of actuate and, the job of the Op-amp is to make
sure that its two inputs are more or less equal.
In any control system we are going to use the Op-amp to do this job. Let us go a
little further. We already know the inverting amplifier, the non-inverting amplifier;
we can do additions, subtractions with Op-amps.

Figure 5

In Circuit I in figure 5, if there is negative feedback, then the minus terminal will be
equal to the plus terminal that is the idea. Number 1, if there is negative feedback
minus and plus are going to be equal. In this case there is negative feedback.

The voltage at minus terminal is going to be 0; that means, the current from source
𝑉2 𝑉1
V2 through resistor R is , the current from source V1 through resistor R is . By
𝑅 𝑅
𝑉1 𝑉
KCL, the current in the feedback path is + 𝑅2 . And this current going through the
𝑅

feedback resistor R and therefore, the drop across the feedback resistor is R times
the current, which means the drop across the feedback resistor is 𝑉1 + 𝑉2 . And we
started with a voltage of 0 volts at the minus terminal of opamp, that automatically
means that 𝑉𝑜𝑢𝑡 = −𝑉1 − 𝑉2 .

This is generally called an adder circuit; the reason why it is called an adder is
because we did the plus operation with KCL. Overall, there is a minus, but we
could also rewrite it has 𝑉𝑜𝑢𝑡 = −(𝑉1 + 𝑉2 ), that is why this is called an adder. The
other classic circuit is the Circuit II in figure 5. Whenever things get complicated
like in this circuit, we start using superposition as opposed to our traditional
analysis and, superposition is going to work very well with Op-amps.
We silence the voltage sources one by one. First, we silence V2, which means V2
become 0 volts, then the plus terminal which is in between is also going to be 0.
And then it falls back to the same old circuit. Plus terminal is at 0, we apply V1,
which means that Vout is just going to be - V1. And then in the next step we silence
V1, which means we make V1 equal to 0 and apply voltage at V2, if we apply
voltage V2 over there, then the voltage at the plus terminal is V2/2. And now it is
the other circuit, where we have applied V2/2 at the plus terminal and we have got
𝑅2
R to ground R in feedback and, this has a gain of 1 + which is 2 in this case
𝑅2
𝑅
because it is 1 + 𝑅 . We have got V2/2 and we have got R and R over there so, it

has a gain of 2 and, if it has a gain of 2 then the output is nothing, but V2. Because
V2/2 through a gain of 2 becomes +V2 and because of superposition these two
quantities come together, add up. This is pretty much oral analysis; we do not
have to do much. We have analyzed visually; we get our answer right away. So,
this is called a subtractor, because it really does 𝑉2 − 𝑉1.