High Speed Devices and Circuits

Prof. K. N. Bhat
Department of Electrical Engineering
Indian Institute of Technology, Madras

Lecture – 23
MESFET
Velocity Saturation Effect
On Drain Current Saturation

We have been discussing the drain saturation current using the Shockley’s model.

(Refer Slide Time: 01:15)

There assumption is that the drain current saturates when pinch off occurs at the drain
edge. That is actually the voltage across this, the total potential drop there is V p0. When
Vp0 is the voltage across of that, when you have VDS(sat) here.

VDS(sat) plus Vbi minus VGS is equal to Vp0,that is the condition.

(Refer Slide Time: 01:50)

Now, what we have discussed later on is that the just couple of things. So, assume it
actually Low Field Mobility that is OK, if the fields are high and I D saturates due to the
channel pinch off at the drain end and just putting reverse whatever I said now and that is
the equation that we used for saturation voltage.

Vbi minus VGs plus VDS is the potential drop across the depletion layer that is managed Vp0
to current saturates.
(Refer Slide Time: 02:24)

Now, we discussed yesterday that such a situation is the questionable because; as you
move from the source end to the drain end, the current density keeps on increasing
because of the current density increases, velocity increases. In fact, electric field
increases, because after all what we have used is the equation: j is equal to qn into v.

Now area comes falls, so v increases, the electric field increases and the region is at this
end, you will have actually velocity saturation. What you are telling is, the real saturation
in current terms of we discussed in the last time real saturation is due to velocity
saturation at this end. Whether pinch off takes at that point or not? That issue we will see.

It could be show that, when the velocity saturation takes place, it is almost closing off at
the drain end that is the pinch off. So let us see that, what we did is the electric field
effect can play a very dominant role in saturating current.
(Refer Slide Time: 03:45)

Strictly speaking, the velocity field characteristic is actually something like this; the
electric field verses velocity. In silicon, if you take you know that it is not going up like
that, but it goes up like that. After 20, 30 kv per centimeter of volts then it saturates. That
can be met very easily.

If the channel length is 1 micron, if the voltage drop is 2 volts; V is equal to 2 volts and if
L is equal to 1 micron then V by L is equal to 20 kv per centimeter that can be met very
easily when you go to channel length which is of the order of 1 micron and if you go
down below that, it is sure that you have very high fields. But what you are trying to find
out is how to incorporate this curve saturation result? It becomes very complicated if you
take the entire distribution. If you take Gallium Arsenide, it will go like that; it will
saturate may be even at a small value and if it comes like this, it may be matching close
to that.

It will really complicate in those situations to get an idea to bring out the effect of
velocity saturation that is first order think to do analytically you can bring all this effects
by numerical methods later on. What you do is, instead of taking all these things; take
one of the equation characteristics as shown in the graph there. We will go back to this
graph here.
(Refer Slide Time: 05:41)

Instead of plotting it all those variations, dupe field approximation that is what done here.
Award all those speaking and everything; take it linear up to certain electric field.
Velocity is proportional to the electric field or E less than Es and till E equal to Es, E is a
saturation field and E is equal to saturation velocity for electric field greater than
saturation electric field.

So, at this point, the velocity is mu n into Es, saturation velocity is obtained right from
there. So, to make this assumption, now you find out what is the Saturation current? What
is the Saturation voltage? Would it correspond to the channel pinch off? And when
channel pinch off occurs, definitely there must be velocity saturation.

Other question you are asking is, when the velocity saturation takes place, will thereany
channels pinch off at all? So, those are things which we are going to examine. But now,
in the case of Shockley analysis, you had that advantage of assuming what is the
saturation voltage is.

That is VDS(sat) plus Vbi minus VGS is Vp0

Vp0 is thickness and doping qnd a squared by epsilon zero is Vp0 that is a known quantity
For a given VGS and Vbi immediately from that equation, you know that VDS(sat). Once you
know VDS(sat), you substitute in equation from the current where the V GS is there, that is all
you got that things.

There are two unknowns IDS and VDS(sat) both are there. If you have one of them, you can
find out the other one or there must be two equations which we can solve to that I D and
VDS(sat), both.

Here, we do not know what point to VDS(sat) is taking place but we know that it would
saturate from the velocity saturates at some point, that concept is VDS(sat).

Two unknowns IDS and VDS(sat). What we do now is, divide the channel into two regions.

(Refer Slide Time: 08:15)

Let go back to that. Here, that is exam in the two cases. This case, I have drawn that
depletion layer; Shockley’s model, IDS is a pinch off. Drain current saturates when the
pinch off takes place. What we want to examine is, whether we get the same condition for
the situation where the velocity saturation takes place. In fact, there is no doubt that
current saturates because of velocity saturation. Otherwise, if it is closing completely as
sum of your raising some time bad, current is found out to be zero.
We said qualitative leads a dynamic equillibrium and that dynamic equillibrium have
come due to velocity saturation here. It could happen so that this width is closed to A.
Now other condition is IDS due to v equal to vs at the drain end. Please understand once
again. In both the cases, the current saturates due to velocity saturation. I pointed out on
yesterday even without a junction when you go to high fields, current can saturate in the
resistor. We examine that in last lecture. Here, what we tell is, I have plotted two things:
one, a channel which is long and other one, the channel which is shorter.

So, the field in this region should be higher compared to the field in this region. It could
so happen that velocity saturation can take place here even without pinching off.

What you are telling is, VDS drops across the channel divided by the length is a field that
could be closed to the saturation field. The difference between these two is no difference;
both of them current saturate due to velocity saturation. But in this case, the current has
saturated before pinch off because before pinch off itself, there is enough voltage drop
and there is enough field; whereas, in this case it has allowed the pinch off. That is the
difference.

Now, we will take that equation which is there.

(Refer Slide Time: 11:10)

What we are trying to do now here is this. We will take the equation which we used from
this end to this end same as what we have used with v is equal to mu into E because we
are now taking the case like this; that is what we are doing. That means, till the electric
field is equal to saturation field, I can write v is equal to mu into E and from the point at
which E is equal to saturation field beyond that point, velocity saturation is there. Beyond
that point, the channel cannot string. This is the condition which is equivalent to the
pinch off that is what you are telling is from here to here as you move, field is not here as
you gone for larger and larger. At a particular point at which velocity saturates, E is equal
to Es and from that point onwards we cannot write E is equal to mu into T and from this
point you write, current density is equal to qn into velocity saturation. In this point, you
write current density is equal to q1 into v1 into v and that is the equation at you derived
(Refer Slide Time: 12:30)

So, in the channel region not at the wavelet end, other portions; we can still write this
equation which we have derived in the previous two lectures that is G0 into VDS(sat), we are
talking of the saturation current. In fact, what we are doing is that will go back to that.
What you are doing is, you are taking a situation where this velocity saturation is present
here and in this portion, there is no velocity saturation.

Up to this portion, I write the equation. Under that condition, current has saturated
because velocity saturation has taken place at the drain end. Now, it is the very familiar
equation for you.

It is written by writing j is equal to qn into v into area.

I am rewriting that equation. All that we are trying to do is writing the two equations
because after all we need to find the IDS and VDS(sat,) we need two equations.
(Refer Slide Time: 13:48)

So, we write one equation in that region, let me put that instead show it to you on the
board. That is a source, that is the drain and that is the gate. This is V GS; so what we are
trying to say is, you have a depletion layer like that, it may or may not pinched off;
whatever saying is, for this condition that we have drawn here velocity has saturated here
and beyond this point, v is equal to mun into v that is what you are trying to point out.

So in this equation, I write the same equation which we are used for Shockley’s model
where v is equal to mun into E. I get one equation for ID and when the velocity is saturated
here, the current has saturated. In that equation; I say whatever VDS is there is VDS(sat) and
whatever ID is there is IDS.

But I do not know anyway both of them.

To know both of them, I write one more equation at this point. At this point, I do not use
this equation but we use is, this portion. So, at this point current density is equal to qn
into velocity saturation. We will see how it works out.

And current is equal to qn into velocity saturation into this area and once you find out this
area that is, this height is equal to A minus H and H is related to V DS(sat). To get one
equation from this portion writing v is equal to mu n into E for the current equation and
other equation between ID and VD writing to the current especially for current here using
that portion. That is all what we are doing; we are trying to get two equations.

(Refer Slide Time: 16:15)

The first equation is like this; now instead of keeping all writing this long equation again
and again, what we do is normalize IDS, saturates the current. Please note I do not know
what is IDS is;, I just set IDS is the value you get when V DS is equal to VDS(sat) that is all what
I have done. So, IDS is equal to G0 I pull out this, multiply by Vp0, so I take inside also
divide by Vp0, multiply and divide, so you get VDS(sat) by Vp0.

2 by 3 square root of Vp0 to the power of 3 by 2 because this is pulled inside.See,

multiplied here by Vp0, that we divide everywhere by Vp0, so you get 2 by 3 Vp0 to the
power of 3 by 2 that is taken inside.

So, what is happening now is, the terms inside the bracket each one of them is divided by
Vp0. So, you normalize this with respect to Vp0 because Vp0 is the quantity which is
decided by the particular device doping and thickness of the layer. So, you have got that
terminal, I call it as uds. Instead of writing VDS(sat) by Vp0, I call it as uds normalized value of
drain source voltage and ID, I can normalize. What is this quantity? All these are
normalized with respect to these ratios.
These are dimensions of G0 into Vp0 it terminates the current, voltage to the conductance.
So, IDS divided by G; this quantity that is normalized value of the current, normalized
value of drain source voltage or normalized value of voltage at the source end; you call it
as uGS, the whole thing and this, you call it as uGS plus uds.

(Refer Slide Time: 18:32)

I rewritten that equation, go back to the equation and see whether we written correctly.
IDS by G0 Vp0 that is normalized value of IDS is equal to uds which is nothing but VDS(sat)
divided by Vp0, the first term (Refer Slide Time: 18:58). Minus 2 by 3 into u ds plus uGs to
the power 3 by 2, just go back once and see. Minus 2 by 3 into u GS is this quantity and
this is uds, VDS by Vp0. That is why you get that. Plus Vbi minus VGS divided by Vp0 to the
power of 3 by 2 which is that. All that you have done is the whole thing comes in one line
now. The whole thing it comes to be so terrifying with all these half a page occupying
here.

Now, single line but the identity is maintained normalized with respect to that, you can
write simpler equations in the simple form by normalizing.

This is the one equation that we have written for a current in this portion
(Refer Slide Time: 19:48)

I have both IDS and uds, both are unknowns I write other equation here. How do you write
the current? Because here the velocity is not given by this, that is given by velocity
saturation. So, what we do now is, write the current at the drain end; it is due to the
velocity saturation at the drain end.

What is the current actually? J or I D is equal to qn carrier concentration into velocity

saturation into area. This I am writing it as, what is n? Electron concentration which
actually equal to doping concentration that is N D velocity saturation of course which is
mun into Es, we can write it afterwards. Area, this height into W and that height is actually
equal to, this is h at the drain end and this is a. So, this is actually h minus a into W. So,
(21:33) that is equal to qND into W into Vs into W into a minus h that is the area. That is
what you have written there.

This equation put down here where h is actually the depletion layer width at the drain
end. Shockley’s analysis said h is equal to very close to a and it cannot be a. You can see
the current becomes equal to 0. Now, we can further simplify. I can write all that we do
here is qND is retain velocity, I can write it as mu n into Es because we are looking into
saturation velocity which is mun into Es, if you are looking to this particular graph.
Vs is mun into Es and we are substituting in that particular velocities saturation mu n into
Vs.

(Refer Slide Time: 23:00)

Let us go back to this slide now and see we have got velocity is that V s is not there
actually may be I think I should cut it off right away here.

(Refer Slide Time:23:25)

I will take a unit for mean but I think we will take it off right away here. Because we
substitute already for that, so it is alright now. We have got this, qNDmunEs W into a into
1 minus h by a. There is a problem of cut and paste; you know cut and paste and add that
mun into Vs anyway now you removed that. qND into mun into W into a. I retain all those
terms. Now, the rest is just manipulation. You have written by f 2 expressions for current,
one is this and other one is this. Here you wrote velocity saturation go back to that (Refer
Slide Time: 24:28). So, what you have done is you have done this substituted V s is equal
to mun into Es and then what I do is, go ahead further I just remove that E s there pull all
this together qND mun into W into a; just I remove that and put it here, multiplied by L
divide by L and this is obvious to you, why I am doing, what is this particular term?

This is the channel conductance G0, this is actually sigma, this is area of the channel
divide by length; sigma area by L that is the conductance and this is L Es into 1 minus...
(Refer Slide Time: 25:05)

Now, we write for an expression for h by a. What is a? Just to recall once again, I will put
it down here, Vp0 is just standard formula. So, a is proportional to root of V p0 and h here is
proportional to whatever voltage is present across that, square root of that. So, ratio of h
and a, h divide by a is square root of whatever voltage is present across that divided by
Vp0. What is the voltage present across that? Vbi minus VGS plus whatever voltage is
present across that and we are saying that is the voltage activity current is saturating. So,
that is why at the drain end, you got h is equal to V bi minus VGS plus VDS(sat) divided by
Vp0. That is simple.
(Refer Slide Time: 26:35)

Now, you are writing a normalized equation. What I did is, this is G0 multiplied by Vp0
divide by Vp0. G0 Vp0 into L Es by Vp0 and terms in within the bracket, 1 minus, I will
write that so that it is not clear to hear.

(Refer Slide Time: 27:08)

See, IDS is G0 into L Es into 1 minus square root of V bi minus VGS plus VDS(sat), not over
step into that, divided by Vp0. What we do is the quantity is G 0 into Vp0 into L Es divided
by Vp0, multiplied by that quantity and what is inside is equal to 1 minus V bi minus VGS
divide by Vp0 is VGS and VDS(sat) divide by Vp0 is uds That is what you have written there.

(Refer Slide Time: 28:21)

Now, I think we should have reason to be happy now, why? Not because of the lecture is
over because you have got two equations A and B. A is this one which relates i DS with uds
and B is other equation which relates ids with uds. Two unknowns, we can evaluate them.

This is the statement which is understand; we are computed the current here by using the
linear relationship. You have computed the current here using velocity saturation. A is
written for this portion, current. B is equation written for this portion. Now the current
computed using that and this are the same thing. So, equate both of them, we can find out
VDS(sat). So, that is right hand side of equation B and this is right hand side of equation A,
that portion. So, currents are same thing in a same device.

Now, what is alpha here? It is again writing one term for this particular quantity, alpha is
L Es by Vp0. This quantity is refer to alpha just for writing purpose; there is no, its not like
the current gain nothing like that. It is actually channel length to the saturation voltage,
how much it is compared to the pinch off voltage? So that place, this term alpha which is
the ratio of this channel length into the critical electric field or saturation field divided by
Vp0 place a very important role in deciding whether whatever we have writing down
earlier, Shockley’s model is correct or not? Come straight from here.

Alpha therefore is equal to this quantity, entire big term. Now, if u ds plus uGS is equal to 1,
what happens to alpha? Infinite.

(Refer Slide Time: 31:26)

uds plus uGS equal to 1, it gives alpha, actually alpha is equal to infinity, I am putting alpha
tends to infinity or if you want to be soft on that, alpha is very much greater than 1 that is
the meaning of that or you can put it other way, if alpha is very much greater than 1, u ds
plus uGS is equal to 1.

This indicates L Es divided by Vp0, he is very much greater than 1 that is the meaning of
that. What is this condition? This condition that you have written here turns out to be true
and that is satisfied or in closer and closer to be true, this is larger and larger.

What is this quantity? Write down, put it in that and remove the normalized thing you
recognize it. uds is VDS(sat) divided by Vp0 that is uds. The second term is, what is that? V bi
minus VGS; there is a potential drop at the channel and this potential drop across the
depletion layer at the source end divided by, that is what you have written. What is this
condition? This condition is, take this Vp0 to the other side, Vbi minus VGS plus VDS(sat) is
equal to Vp0. This is what you recognize. So, this condition turns out to be V bi VDS(sat), I am
write it an angle VDS(sat) plus Vbi minus VGS equals Vp0.

That is the condition. What we are telling is, when alpha is very much large compare to 1
that is L Es by Vp0 is very much greater than 1; velocity saturations takes place, that
condition may use. But when the velocity saturation takes place, this condition is
satisfied. That means long channel devices; the channel length is long and whatever
Shockley’s has been telling or writing the condition for saturation outs good, pinch off
takes place or at least which closed to pinch off. Its real pinch off takes that that will be
infinity but now this larger than 1 May be 3 or 4 like that, so what we are telling again is,
if the channel length is long, by the time you read that field there.

Let us say Es is a field that is average field, L into E field. So, that is long that means it is
very large by that time the high field is there. That voltage becomes large and it reaches
pinch off. Either way it is true, supposing pinch off voltage is small even if a channel
length is small; if this is very small, you can satisfy that condition.

So, what we are trying to just go back to that graph and see it once just maybe I do not do
that. So, what we are trying to say is that drop here is sufficient enough so it is closed
this, if L is long we know that, if the channel length is small, that will what happen but if
the channel length is small, if the small voltage your h will (36:26) achieve it may not
close at all that is the other condition. Obviously, when the L E s is smaller compare to
Vp0, you will not have this condition satisfied. We will see what happens there.

If that, we will have a situation where the general equation can be derived and people
have played games in the sense game is not real game; played at the equations and
estipulate interpolated etc., caught from analytical expressions with look same as that mun
Es W by L into VGS minus uds whole squared with the factor, we will see that.
(Refer Slide Time: 37:08)

This is Shockley’s model we are discussed just now that condition one; alpha very much
greater than 1. The other condition see, if this becomes equal to 1 that is Shockley’s
condition. If this term when the voltage reaches the saturation and it is less than the V p0,
the channel is not closing down. So, if this is less than Vp0, VDS(sat) alone will be even less
than Vp0.

The VDS(sat) plus that quantity is less than Vp0. But we are telling is VDS(sat) plus Vbi is less
than Vp0 that is channel is not closing down; that is situation this plus that is less than V p0,
means this portion is definitely less than Vp0..

Three times are there out of the two knock out that is much much smaller, that is this
quantity is less than 1. Now what we are writing is V DS(sat) approximately we can put it as
L into Es. This is of course assuming electric field is everywhere. But actually it should
move from this end to that an electric field is Es closer to this point.

VDS(sat) actually is less than L into Es because E is less than Es here. VDS(sat) is voltage
across that and it is L into an average electric field there. The average electric field is less
than Es because Es is highest here. L into Es by Vp0 is less than 1, is it clear now? See we
arrived at, I will go through quickly once if you gone through fast on that.
These three terms equal to be Vp0 is Shockley’s condition. That is valid when alpha is
greater than 1, L ES by Vp0 is greater than 1. We are trying to see, what is the situation
when L into Es is less than 1? In fact, we are going over to other way because we
understand this term clearly. Out of these quantities, I remove this quantity. So, all the
three times together is less than V p0. Therefore this term alone should definitely much less
than Vp0.

Now, I am removing this taking into other side V DS(sat) may be by Vp0 is very much less
than 1. What is that condition we are trying to see?

(Refer Slide Time: 40:15)

If VDS(sat) is very much less that Vp0, I will put there on the board. So, we are taking a
situation where VDS(sat) by Vp0 is very much less than 1. What is VDS(sat)? It is a voltage drop
across this portion and I can say it is channel length into the average electric field and
what is the average electric field here? That is less than saturation field because peek
electric field is maximum here, it is less than that so average electric field is less than E s.
See, this is less than that which means actually L into electric field average by V p0 is less
than 1, average electric field is less than that. So, this is less than 1, average electric field
is less than Es, so L into Es is definitely if I replace this by this, that will be given from
that. An average what you can say is VDS(sat) by Vp0 is less than 1.
So, L into Es even if it is a maximum thing, L into E s is less than 1. At the best case, you
can have this quantity less than 1. In other words, alpha less than 1 or L E s less than 1
correspond to that situation. We are indirectly going backwards. See, I could written
straight away this and according to that but I have done is, I have taken that equation
where the pinch off does not take place and I write this condition pinch off does not take
place velosity saturation, that condition corresponds to that.

We can take the VDS(sat) is equal to that maximum field itself L E s that is less than 1. Go
back to that velocity saturation throughout, now what is the situation? We say, this is
much less than 1 and now let go back to that.

(Refer Slide Time: 42:42)

An ids is actually equal to, what is i DS?. This I hope agree because this equation that you
have derived by writing the current here.
(Refer Slide Time: 43:12)

Now, what we are telling is, uds is very much less than 1 because we are replacing VDS(sat)
by almost equal to L into Es. So, this is less than 1. Now, when you say u GS plus uds, this
term is that quantity and this quantity is very much less than 1. I am knocking it out
because there is very much small compare to 1.

So, what we are telling now is, I can write an equation for i ds with involved through the
thing but not very complicated, all that you realize which is small quantity. So, I am
neglecting this if I write like this, what is the meaning of this? The meaning of that is, we
have got velocity saturation throughout the channel.
(Refer Slide Time: 44:15)

Let us put that down here now, things get little look bit blink because you are looking
into an extreme situation. The electrons when they enter, itself has reached the velocity
saturation. They already explained velocity saturation and they are not just getting into
the portion; they are getting from an inner portion that is the situation once reaches this
point if there is velocity saturation, what will happen to the depletion layer afterwards?

This is velocity saturation and you go beyond that point velocity cannot increased, the
saturation velocity. J is equal to qn into Vs, current is equal to qn into Vs into area. Vs
does not change, they cannot change and this will remain same thing. If it remains same
thing, what is the meaning of that? What is the drop here? Very small that is the meaning.
If there is no drop there may not be any current flow, it is very small. That means you get
very high fields with very small drop that is possible manner. You can get very high
fields here because field is in this portion it must be greater than the saturation of field;
we can get very high field, there is small voltage drops if length is very small sort channel
devices. So, we are driving ourselves into device which we have got channel length is so
small such that even with small voltage drops across that, you will have velocity
saturation then that will be the thing. Now you can thing (45:52) uGS that quantity is Vbi
minus VGS compare to that, that drop is small.
So, this is a situation you have got velocity saturation right through the channel that is
possible if launching the electrons right in to the channel with velocity saturation or they
just came immediately the velocity saturation.

(Refer Slide Time: 46:23)

Few things now; I think I am just now we can zip through because the concepts are over.
L Es that is alpha L Es by Vp0..

If it is very large that is long channel devices or pinch off voltage small such devices
velocity saturation takes place and pinch off takes place. They coincide. If velocity
saturation and pinch off coincide, that is why, Shockley’s theory survived so long
because we are talking of channel length is long or which are pinch off voltages are small
but the motive go to shorter channel devices; you will have velocity saturation but pinch
off voltage will not take place. Here, pinch off have not taken place at all, where there is
saturation right through, there is absolutely no pinch off. So, pinch off need not take place
of current saturation. There is this situation and we can write this entire thing, I am just
rewrite that, we can derive it in different approaches. I write the whole thing again that I
can skip all these because you get the same equation whatever approach you write.
(Refer Slide Time: 47:45)

VDS(sat) is actually L Es which is less than Vp0. In fact, there is no need of going through
and I just rewrote it in different approach but we have got the whole thing condition
satisfies before that.

(Refer Slide Time: 48:00)

Now, let us see the general case. The most general case is obtained you have to solve this
equations. You have got velocity is equal to mun into E, this portion at the drain end, this
is the portion. I have put those two things together. You have to equate these two and find
out uds from these equations. Numerical solution you have to do. If I have to find out u ds,
once I know uds, I can actually find out IDS by substituting either this or this. In other
words, we have two equations and equate two to get one equation. Alpha is this, you
notice of this. Next, what we can do now is, we are interesting in finding out, let me
quickly go through that; we are interested in finding out the value of u ds by using equating
these two. Wherever some clever people who did some solution but you do is to equate
these two to get(49:20) alpha. These we have used already, now alpha depends upon L
into Es by Vp0,, that is alpha.

(Refer Slide Time: 49:40)

What we are trying to see is, as a function of alpha what will be the u ds for a given VGS?
Alpha will be decided when you decide the device, E s is fixed. At the property of mid
device material, that is fixed.

You have taken that by lean arising thing 20 kb per centimeter or 10 kb per centimeter
depending upon whether you talk of silicon or Gallium Arsenide. So, that is fixed. This is
fixed from fifth device. What about this quantity? That is also fixed. For a given device,
if I am changing alpha for a given doping and thickness, I think change L and see.
So, one other way that peoples are done is, do that alpha verses u ds. Using this equation,
you substitute to do is for a given gate to source voltage uGS,, after all we cannot get very
all the things.

You find out alpha verses uds, how to do that? Change u ds and find out what is alpha is? It
is straight or you find alpha. To decide alpha and find uds if more complicated; you have
both numerator and denominator.

For given uGS, put for different values of uds verse alpha, you get that.

(Refer Slide Time: 51:35)

You can do numerically or you can do this substitution and find out and then that is
virtually hooking up in the sense fitting the parameter. So, what people did in fact to be
actual person who did Michele soore from redlich of techniques? We have published for
some time back with result because to simplify to get an analytic position. Numerically;
if I do not have to go for numerical, analytically we can do by substituting u ds different
values to find what is alpha?

For a given drain to source voltage, you found that you can fit into this expression. You
can fit in this expression because after all, each curve is started for one u GS. You can get
number of curves like that I have plotted one curve for one u GS and put that an analytical
expression corresponding that first fit and once get an expression for u ds in terms of uGS,,
you go back and substitute the equation 2.

What is the equation 2? Let go back this equation 2. i ds is equal to, see what we have got
now is, for a given alpha we know what is u ds is from that expression and in terms of u GS
so you get ids verses uGS. By substituting for uds in that equation,

(Refer Slide Time: 53:15)

I think propagate down here we will get i ds equals alpha into 1 minus root of uds plus uGS. I
have got an expression for uds from here which is alpha into 1 minus u GS divide by 1 plus
alpha minus uds So, I removed this quantity in terms of uGS, that means I get ids verses uGS.
That means, once I get no, what the gate voltage uGS? What is uGS? Vbi minus VGS divide
by Vp0 and ids is actually drain current, ids divide by G0 Vp0. So, what I am trying to point
out is, you know uds in terms of uGS by establish interpolation etc., you know ids in terms
of uGS. You can get the transfer characteristic that is what you have got.
(Refer Slide Time: 54:31)

Substitute there again an approximation is done, you get this. Here after, it is very simple
and you do not have to worry about the path is that of establish interpolation to get a fit
into that complicated formula simplified, you get this expression. Now, we can do that. I
multiplied by this, normalized value; IDS divide with G0 into Vp0. I remove the normalized
quantities and write it in terms of that. Go back to the original equation and then I get G 0
by Vp0 into. All that I have done is pull this Vp0 out.

So, I get Vp0 gets cancel of that 1 by Vp0 squared; I get this, just simple. I have not made
any modification there and same equation rewriting this, you get that. I multiplied this by
4 divide by 4, I get again. What is this quantity? Minus V threshold voltage, this whole thing
is VGS minus Vthreshold whole squared and this term is 4 alpha by 1 plus 4 alpha and G 0 by
4.
(Refer Slide Time: 56:04)

What is that? What is this first term? I multiplied again by that, this quantity is familiar
term for you and go through is again in next time because what you are showing is
ultimately you get the same expression that you get using the Shockley’s equation except
that you got this particular term to start right from here next time going into this.

Then we will see compare with some experimental results. In fact, it is a pair that these
things are described right in discussion when we bring the measurements of results.

So, I will stop on that in this discussion.