Lecture 15 Module 01

IIT Kanpur July 201
20 followers
Oct 22, 2023
This is lecture 15, so in the previous lecture we were discussing how to how to bias our
common source amplifier right, so we were discussing how to bias our common source amplifier
at least on the on the side of the gate, we were not particularly bothered of what were
what was happening at the side of the drain, so let us let us shift our focus to wherever
we stopped in the previous lecture and what are the problem that we are trying to address
the problem that we are trying to address was the fact that if we want to apply our
signal like this then the problem that we encounter is we encounter a floating battery
then we saw that a sufficiently large capacitor a sufficiently large capacitor acts as a floating
battery, but if we have a unlike a unlike a unlike a battery a sufficiently large uncharged
capacitor right will not be able to do the job because ultimately what is a battery?
A battery is a stored storehouse of a of sufficient amount of charge and if you have a battery
with sufficient charge then you have a voltage across it, but if you have a capacitor with
no charge you do not have any voltage across it. So, in order to it is not only sufficient
to use just a large capacitor you will also have to charge it with a voltage and then
we said that is then if in order to charge it with a voltage why do not we apply why
do not we have why do not we put a battery of V gq right why do not we put a battery
of V gq here. So, it seems like I will be able to charge these two plates of the capacitor
with a voltage V gq in the absent in the absence of the incremental input, but then we saw
that what is the problem if we do this the gate the gate of the transistor always remains
fixed at V gq. So, that is of no use because ultimately we
want we want the signal input signal to appear at the gate of the transistor right we want
the Y 1 1 incremental Y 1 1 looking into the looking into the network to be 0. In this
case Y 1 1 is infinity because looking into the network I have a battery a battery incrementally
is a is a short circuit right. So, this is of no use. So, the problem number 1 here is
the problem number 1 is you have a you have a fixed battery and what is and there is another
problem. Problem number 2 is something that we briefly alluded to in the previous lecture
was that in generally in an integrated circuit or generally in any circuit you will see that
we are given one master voltage source and one master battery right one master power
supply from which we need to we need to derive all the other power all other voltages that
are necessary right. So, I mean right away here we see that we
are using two sources one is V gq one is V gd which is something that that we do not
have a luxury of simply because I mean this is a circuit with one transistor if you have
like 100s of transistor would you use 100s of voltage sources obviously not right. So,
the two problems the problem number 1 is gate is held constant right. So,
this is problems and number 2 number 2 is need multiple batteries right. So,
let us say I mean let us address the second problem first let us say we want to get rid
of multiple batteries we want to generate the voltage V gq from V dd assuming V dd is the
highest power supply and that we saw right I mean the requirement of V gq was far lesser than the
requirement of V dd when we took the examples a few lectures back right. So, how can I get if V
gq is less than V dd how can I get a value how can I get reduced value of V gq from V dd what
is the simplest way let us say V dd is 30 volts that we had that we had gotten in few lectures
back and let us say V gq is supposed to be 3 volts. So, what will you do you would simply say
that simply say that if this is V dd we will put a resistor divider let us call this R 1 let us
call this R 2 if I choose R 1 and R 2 in such a way that this voltage what is this voltage this
voltage is R 2 by R 1 plus R 2 and V dd if I choose this to be equal to V gq right then it
seems like I can use this voltage right. So, let me mark this as V g and then I can look
it looks like I will be able to use this contraption to bias the transistor right ok.
So, let us try that let us try that and see what is whether this indeed solves the problem. So,
what will you do I will redraw it in such a way that I can get away by drawing only one V dd. So,
let me call this R L right and let us assume we have a battery here of value V dd here I
have this capacitor here I have a voltage source V i and here instead of that battery V gq what
we are trying to do we are trying to replace the battery with a resistor divider R 1 and R 2 and
this is V g what do you think the value of V g will be the value of V g in this case will be
whatever we saw in the previous slide right. So, V g will be V dd times R 2 by R 1 ok.
So, now, if I ask you a question can I always replace a battery of a lower value than V dd
with a resistor divider like this. Obviously, you can think of multiple cases where we cannot I
mean one example that we saw was in the introductory lectures probably in lecture 2 or
something where we saw that if we load a resistor divider with another resistances. So, that it
draws current out of the resistor divider and the value of the voltage at the resistor divider node
will drop right. So, which essentially means that you cannot use a resistor divider always
to replace a battery because a battery is supposed to hold a voltage right, but how come
we are able to use it in this case we are able to use it in this case because there is no DC
current that is getting drawn from this resistor divider right. What is the DC current that is
getting drawn on the left hand side on the left hand side what do I see on the left hand side I
see a capacitor. What is the impedance of the capacitor at DC it says an infinite impedance
what is the DC current that is getting drawn towards the right what do I see at the right
at the right I see the gate of a MOSFET which is which in a sense is again a capacitor.
So, what is the impedance that I am seeing I am seeing infinite impedance which means this I am
not drawing any current out of this resistor divider since I am not drawing any current on
the resistor divider it really does not matter right I can essentially I can essentially use I
can replace the battery using this resistor divider ladder. Note that we are not drawing
any DC current right this is an important distinction to make right. Since no DC current
is being drawn by the resistor divider right we can replace the battery Vgq with the R1 R2 ladder
right. If I had a scenario let us say instead of instead of a MOSFET I had another device which
draws current right into its whatever is equivalent to the gate then this would have been a problem
we would have had to see what values of R1 and R2 would satisfy our constraint of keeping that
voltage at the node Vg constant right. So, but fortunately MOSFET does not draw any
current. So, we are good fine. So, this is this essentially solves one part of the problem right
if this solves this problem we do not need multiple batteries. So, this part is solved
what about the first part what about the I mean what about the signal part of the problem right.
So, is a gate held constant what do you think is going to happen. So, let us look at it again
from the signal point of view. So, let me just sketch the gate side of it I mean I will not
sketch the full MOSFET just for brevity. So, this is Vi you have this capacitor R1 R2 and you have
the MOSFET here and so on and stops there I am interested in the voltage Vg or let me say I am
interested in the total voltage Vg right. Let me just mark the node Vg here and then ok.
So, this C tends to infinity that is something that we have assumed ok. So,
what do you think will this voltage Vg be? I can as well say that I will do I will do
an incremental analysis right I can do an incremental analysis just like we did in
case of in case of small signal analysis earlier right ultimately we are interested
in the incremental increment in the voltage Vg when I have applied a input Vi right.
So, what will be the incremental equivalent of this assuming C tends to infinity and Vi is not
at 0 frequency right. So, Vi is at omega naught that is Vi is at omega naught, but omega naught
is not equal to 0 ok because if Vi is at 0 then anyway I mean no signal no DC signal will flow
from Vi to Vg ok. So, what will be the incremental equivalent of this? The incremental equivalent of
this will be what should I do? I should replace all the independent voltage sources with a short
circuit. So, this becomes the short circuit this R 1 remains R 1, R 2 remains R 2. What
happens to C? What is C? Infinite capacitor is like a battery if the infinite capacitor is like
a battery and what do we do with the battery? We short it. So, that goes off and what do I
have here? I have Vi. So, what is the incremental voltage here right? I mean it is like what am I
asking right it is a short between Vi and that point. So, the incremental voltage at this point
will be will also be equal to will also be equal to Vi. So, what will be the total voltage at Vg?
What will be the total voltage at Vg? The total voltage will be incremental plus quiescent. So,
incremental is Vi and quiescent is V dd R 2 by R 1 plus R 2 ok fine right. So,
seems like this is what this is what we exactly want right seems like this is what we exactly
want at least at the input side and seems like putting this contraption of a capacitor and
and this resistor divided ladder of R 1 and R 2 seems to have done the trick right ok.
But not so fast because you might have already noticed that I have I have made
slight change as to how I have represented the source Vi. I have not marked the source impedance
or the source resistance R s which I had initially marked at the beginning of the lecture right. So,
let us put back the resistance and let us put back the resistance and see whether it has any it has
any effect or not right. So, let us put it back. So, what I meant is let us say you have Vi the
source is always accompanied with some resistance R s, you cannot wish it away you have a capacitance
C and V dd R 1 plus R 2 this goes here and you have a you have this MOSFET and everything.
We are interested on the on the gate we are interested in the in the gate voltage right
assuming C tends to infinity again. So, what will be the incremental what will be the incremental
voltage at the gate the incremental voltage how what will be the incremental circuit same old
same old I will retain whatever I mean I will retain R s. Obviously, Vi is required because
Vi is I am trying Vi is the incremental source C infinite capacitance acts like a battery. So,
I can short it right R 2 goes to ground here R 1 also goes to ground here right because V dd is
V dd is constant voltage then you have then you have the MOSFET right then you have the MOSFET
and let us not bother about the MOSFET right now because we are concentrating on the gate voltage.
So, now, what will be this incremental voltage V g? So, V g note that V g will not be equal to
Vi right because of what because essentially if I redraw this circuit what am I seeing I am seeing
this Vi then you have R s then you have a parallel combination of R 1 and R 2 right
parallel combination of R 1 and R 2 and this node voltage here is essentially V g. So,
what is V g? V g is Vi times R 1 parallel R 2 by R s plus R 1 parallel fine. So,
in order to so, now, can you comment on how do you choose R 1 and R 2 if I want V g to be
almost equal to Vi because V g to be almost equal to Vi is the goal is not it? If V g is
almost equal to Vi then I know that the entirety almost the entirety of the input signal appears
at the gate of the transistor. So, in other words I do not want any loading I do not want.
So, this what is this what is this stuff doing? This stuff is loading Vi correct since this is
loading Vi it is drawing current out since this is drawing current out the voltage at V g is
is dropping right it is not equal to Vi anymore. So, you cannot you do not have any control on Vi
you do not have any control on R s. However, you have a control on R 1 and R 2 because you are the
designer you are in charge of choosing the values of R 1 and R 2 which essentially means that which
essentially means that in order to ensure to ensure V g is almost equal to Vi or Ie to prevent
loading what do we need to do? We need to choose R 1 parallel R 2 such that R 1 parallel R 2 is
much much greater than right. Mathematically also this turns out to be true if you see R 1 parallel
R 2 is much much greater than R s then V g is almost equal to Vi and from the assignments in
week 1 that we that we did you would appreciate that in order to prevent loading what should we
need to do? The part of the network that is loading the actual network should be the resistance of
that should be much much higher than the Thevenin resistance of the network that I am looking in.
In this case the Thevenin resistance that we are looking in this side is R s right and and
the loading part is R 1 parallel R 2. So, in essentially if you do not want to load
you have to ensure R 1 parallel R 2 is much much greater than R s fine. So,
if this is in designers control then we are good to go right, but there is one pesky problem what
is that problem from where do I get an infinite capacitor right. So, we have been we have been
liberally using this assumption that this capacitor that we have here is of infinite value right,
but in reality nothing is infinite. So, we will have to replace it with the finite value of
capacitor. Now, what will happen if I replace it with a finite value of capacitor what do you
think is going to happen? So, clearly if I use a finite value of capacitor and V i is again at
omega naught not at not at 0 frequency then there will be some voltage drop across the capacitor
right. So, let me still restrict myself to incremental equivalence and let me just simply
say that this is R 1 parallel R 2 instead of drawing the entire entire stuff right.
So, how do you how would you go about figuring out what is the what is I mean what value of
capacitance that we need to that we need to use? One way to think about it is that the value of
the capacitance should be such that the voltage drop across the capacitance should be much much
lesser than much much lesser than the voltage drop anywhere else in the in the loop right.
So, note that this is a loop this is a loop if the capacitor would have been of infinite value what
would have been the incremental voltage drop across the capacitor? Note that I am again asking
about the incremental voltage drop not the absolute voltage drop right. The absolute voltage
drop the DC the capacitor was holding a DC voltage to begin with that it is still hold right,
but I am not talking about that I am talking about the incremental voltage drop in if the
capacitor would have been infinity right the voltage drop across the capacitor would have
been 0 right, but in principle we cannot get a we cannot get any infinite value of capacitance
which means we cannot get any 0 voltage drop across the capacitance. So, what is the alternative?
The alternative is to say that we need to size a capacitor in such a way that the voltage drop
across a capacitor is much much less than anything the capacitor anything that is anything in the
loop in which the capacitor is a part of right. Now, what is a generic way of looking into it
instead of saying that I mean I can have multiple loops this is a particular scenario I can have
multiple loops in the network you cannot probably identify the capacitor to be a part of the one
loop right. It is quite possible this is a particular case where I could where I can
identify a particular loop, but in general if you face this type of problems of big trying
to figure out how large a capacitor supposed to be or maybe how small a capacitor is supposed
to be right for in order to pass a signal without causing any voltage drop or rather
without causing any significant voltage drop across it what would you do? You would you
would try to see what is the Thevenin equivalent resistance that is the capacitor sees right.
So, in other words what we are saying is if I let me mark the capacitor with a different
color in other words what we are saying is if I am looking if I am looking at the network
sitting on the capacitor right what will I see I will see in this case I will see a resistive
loop right. So, I can as well I can in order to figure out what is the voltage across the
capacitor what do you need to do? I will Theveninize the Theveninize the circuit and assume capacitor
to be the to be the load right if capacitor is the load say capacitor is the load and
I am trying to Theveninize the circuit right. So, I am trying to make the equivalent these
two equivalent as far as the voltage drops across the capacitor that is concerned right
I am trying to figure out what will be this Vc right what will be this Vc ok.
So, then what should I do I am trying to I am trying to figure out what will be the Thevenin
equivalent of this. So, in other words I am trying to figure out what will be Vth what
will be Rth if I know this then I can simply say that then I can simply say that this is
a first order RC circuit right and all I need to do is to channelize a first order RC circuit
right. So, I mean one might you might turn around and tell me that why am I spending
so much time on this because it is looks like it is pretty evident looking in the network
on the left what I am trying to get at here is that it might not be as this evident always
your network might have lots of other elements wrapped around the capacitor. So, then what
do you do right. So, use this please use this as a as an example of how to go about figuring
these things out ok. So, let us do it with a simple network. So,
what will be what will be what will be Vth? The Vth will be how will you figure what Vth
will be you open the you open the you open the load right and see what will be the what
will be the voltage that you are supposed to get what is then what is the Vth? Vth is
nothing, but if you open the capacitor the voltage across the capacitor in the circuit
on the left will be will be nothing, but Vi right because no current will be able to able
to flow. What will be Rth? Again let us open the capacitor
what will what will you do to find Rth? You short the voltage source you de-energize any
source in case of voltage source you short it right and you see what is the what is the
effective resistance. In this case if you short your voltage source right you short
your voltage source then what is the effective resistance? The effective resistance is nothing,
but Rs plus R1 parallel R2 right. So, Rth is Rs plus R1 parallel R2 ok.
So, what do you think your Vc is? Vc is equal to Vth times 1 by j omega c divided by Rth
plus 1 by j omega c right which I can simplify as Vth by 1 plus j omega Rth right ok. So,
I can say that mod of Vc right or let me let me not say mod of Vc let me just keep Vc as is.
I should also say this Vc of j omega this is not Vc of t right Vc of j omega Vc of j omega
will be equal to a magnitude and a and a phasor right. So, what is the magnitude part of it?
The magnitude part of it will be Vth by under root 1 plus omega squared I forgot a c here
omega squared Rth squared c squared right. I should not say omega because I think I
said omega naught there. So, this should be omega naught should be omega naught
because I have applied a sinusoid of frequency omega naught right and it will also have a phase
what will be the phase? Phase will be minus of tan inverse omega Rth c ok.
So, let us not bother about the phase for the time being all we need to all we need to ensure
is that all you need to ensure is that if your Vc right the voltage across the capacitor is
what? It should be as low as possible right. We want the capacitor to be almost like a short
which means what? Which means we want for c to act like a short circuit we want Vc
to be equal to 0 or Vc should tend to 0 i.e. the denominator should tend to infinity
right or tend to be as high as possible right which means you need to ensure omega naught
Rth c should be much much greater than 1 which means c should be much much greater than
omega naught ok. So, we why do we go through all this song and dance?
We are doing this all this song and dance just to just because in principle we cannot
get infinite capacitor which means we need a large capacitor. If we need a large capacitor
we need to figure out what is large right is 1 picofarad large, is 1 nanofarad large,
is 1 microfarad large the answer to that is it depends. What does it depend on? It depends
on Rth right that is good because Rth is in your control. What is Rth? Rth is
R1 parallel R2 plus Rs, Rth is R1 parallel R2 plus Rs, Rs is not in your control, but
you can you can increase the value of Rth by increasing R1 parallel R2 that is one good thing
and also I mean omega naught might not be in your control, but if you have if you are designing a
circuit for high high frequency let us say if omega naught is large enough then probably you
can get away with a smaller value of capacitance. If omega naught is small then then I mean nothing
doing you can either try to make Rth very high in order to keep the value of the capacitance in check
or you just have to go with larger value larger value of c ok.
Now, note that this is the same thing I mean we could have come up with the same
conclusion without doing all this all this detailed analysis right. What is another way of
coming at coming at it? I am sure you know this already, but let me just phrase it for the for
completion purposes. So, what is this circuit? This this circuit when we sketch this circuit
what is the order of this circuit? Note that the order of the circuit is 1. What will you do in
order to figure out the order of the circuit? In order to figure out order of a circuit is
independent of a source right. So, so you you short the source in this case again it is a voltage
source. So, I am shorting it had it be a current source I would have opened it. So, you short the
source and then then try to see then try to see how many first I mean you can obviously see in
this case you can see that you have one capacitor and a series connection of R s and R 1 parallel R
2. So, this clearly this is a postural circuit. So, in general the trick is to see how many
independent initial conditions you can you can you can establish in a circuit that has memory
elements like capacitor is a memory element if you had inductor, inductor would have been a
memory element, but the trick is to essentially see how many firstly how many memory elements
you have and even more important than that how many independent initial conditions that you can
establish in the in the network. In this case obviously I can establish only one independent
initial condition that is the voltage across the capacitor right. So, so the order of the
circuit is 1. So, in in a first order circuit what is the what is the most relevant parameter for a
first order circuit? It is a time constant right. So, so what is the time constant of this network?
The time constant tau is R s plus R 1 parallel R 2 times c right right. So, it is no doubt that
time constant is nothing but R th Thevenin equivalent resistance times capacitance and
what is the condition that we are looking for? The condition that we are looking for is
the impedance offered by the capacitance. What is the impedance offered by the capacitance?
The impedance offered by the capacitance is 1 by mod of the impedance offered by the capacitance is
1 by omega naught c right. So, the voltage drop if you want the voltage drop across the capacitance
to be much much lesser than anything that is present in the loop in a Thevenin loop right.
So, what is the trick then? All you have to ensure is that the impedance of the capacitor
right is much much less than much much less than the Thevenin resistance right. So, essentially
you have to ensure ensure let me write it in a new page. To ensure negligible voltage drop
across a capacitor, we need to ensure the the impedance of the capacitor
at whatever frequency omega naught right. So, in other words in this case 1 over
omega naught c which is the impedance of the capacitor is much much lesser than
the Thevenin equivalent right, the Thevenin resistance of the loop right which essentially
means that the constraint that it directly comes out the constraint is the constraint is
c should be much much greater than 1 by omega naught.
So, this comes out out of plain argument right or in other words I can I can also say that this
also implies that the time constant of the loop that is c times Rth should be much much greater
than omega 1 by omega naught. So, you can approach this problem from multiple dimensions, you can
approach it from the from the perspective of finding out the time constant and and then relating
it to 1 over omega naught, you can approach it from the perspective of sketching out a Thevenin
equivalent and then finding out the actual drop across the capacitor and trying to minimize it.
You can also say that once I have the time constant right once I have the Thevenin equivalent
resistance, then I do not even have to I do not even have to do anything I will just set
c to be much much greater than 1 over omega naught Rth and I am good. I will see you in the next
lecture.
