Tip:
Highlight text to annotate it
X
Hello everybody! In our series of lectures on basic electronics learning by doing we
will move on to the next topic. Before we do that it will be better to recapitulate
what we learnt in some of the previous lectures.
You might recall in the previous lectures we discussed about the frequency response
of amplifiers. In order to do that we actually looked at some simple networks like the lead
network, the lag network and then looked at the frequency response of these two circuits
because we understood that the frequency response of an amplifier can be understood in a simple
manner by looking at the contributions to the frequency response from different capacitive
reactances that we have in a given circuit on the basis of equivalent lead-lag network.
That is the reason we discussed about the frequency response of the lead network and
the lag network independently in the last lecture.
Before we go on to the actual thing I will again show you the general graph of the band
width of an amplifier. You can see on the x-axis the frequency scale is there in logarithmic
scale, 10, 100, 1000 etc, and in the y-axis you have the gain, magnitude of the gain which
is basically the output voltage by the input voltage.
Approximately over the mid frequency range, over a wide range of frequencies you find
the gain is almost a constant straight line, flat straight line corresponding to gain of
voltage at mid frequencies and as you come below you find at very low frequencies the
gain starts falling of almost to zero and similarly at high frequencies also you find
the gain starts falling of beyond certain frequencies. When we discussed our h parameters,
equivalent circuits for the basic transistor amplifier, common emitter amplifier we ignored
the contributions due to the various capacitors like the coupling capacitor, the bypass emitter
capacitor, etc, because the reactance offered by these capacitors will be very, very small.
We can almost take them as a short, for practical purposes, at those frequencies. But when you
want to focus your attention to the frequency behavior of such amplifiers at very low frequencies
then you cannot ignore any more the contributions from these capacitors and that is why here
the arrow shows this fall in the frequency at low frequency regions is due to the contributions
of the reactances of the various capacitances like the C1 the coupling capacitors, the output
capacitor and the emitter capacitor.
Similarly if you look at the high frequency, the contributions to high frequency fall comes
from several other new contributions like the parasitic capacitances about which I will
perhaps explain in a moment and the various other dependents for example of the gain of
the transistor beta or the …... This is the general scheme of the band width of amplifier
which takes into the account all range of frequencies; the low frequency, the high frequency
and the mid frequency.
Let us start looking at the general discussion on the frequency response of an amplifier.
You see on the screen the transistor, common emitter amplifier with a voltage divider network
R1 and R2 and you have a coupling capacitor C1. You have the signal source Vs and Rs is
actually the source resistance of this Vs and you have the RC the collector resistor.
You have RL the load resistance and the C2 is the coupling capacitor at the output connecting
to the load and CE is the bypass capacitor across the emitter bias Re.
This is a very familiar circuit. I am sure we have discussed this earlier when we discussed
about the biasing circuits as well as later on and this is very familiar to you. But when
we drew the equivalent circuit for this circuit earlier we ignored the contribution to C1,
C2, CE, etc. That we cannot do any more because we want to look at the low frequency, high
frequency response also of these transistors. The best way to do that is try to find out
the contribution due to each one of these capacitances alone ignoring the presence of
others and then try to look at the equivalent circuit and then try to see whether it belongs
to a very specific lead like network and then we have already understood the performance
of the lead network and from that we can discuss the frequency response of the transistor.
By this method we will be able to know the contribution to the cut off frequency. What
will be the cut off frequency corresponding to each one of these contributions and finally
we will take out the one which is most dominant, the one which is most responsible among these
things and that will be the one which will decide the overall bandwidth of the amplifier.
So this is a very standard simple scheme that we will adopt.
So we have to see the contribution to C1, C2 and CE. These are the capacitors that will
contribute to the low frequency response of the amplifier.
Let us consider one by one. Let us take C1 which is the coupling capacitor at the input.
I have here drawn the equivalent circuit ignoring the rest of the information and only showing
the amplifier with reference to its input resistance Ri and you have the C1 and you
have the source resistance and the signal source. This only is shown and you can see
that it is a very simple circuit that we have already discussed earlier. The total resistance
now is Rs plus Ri and the cut off frequency, we know already, f1 is 1 by 2 pi Rc; here
R is Rs plus Ri. So f1 is 1 by 2 pi Rs plus Ri into C1. So this is the contribution. This
is frequency corresponding to the point at which the mid band gain will fall to 0.707
times the normal maximum gain.
If I look at mid band frequencies this C1 will be very low. The contribution from this
will be very low and it can be taken as the short, the reactance component and we ignore
this. But at low frequencies we have to take it into account. Similarly at the mid frequencies
the actual voltage applied at the input terminals of the amplifier Vi will be equal to Ri Vs
divided by Rs plus Ri because the Rs and Ri will act as a simple potential divider and
the voltage across the Ri is the one which is going to be applied across the amplifier.
So Ri into Vs divided by Rs plus Ri gives you the input voltage at the mid band frequencies.
If you look at the frequency response corresponding to that network it will have very low gain
at very low frequencies and as you increase the frequency the gain slowly increases and
finally around the mid band range it becomes a constant.
This is the contribution due to the capacitor C1 which is called f1 which is given here
again as 1 by 2 pi Rs plus Ri times C1 and the Rs plus Ri are the two resistances. Ri
is the input resistance of the amplifier, Rs is the signal source resistance that we
have already discussed. What is that Ri? We all know what that contribution to Ri is.
The Ri is contributed in the case of a rc coupled amplifier that I showed you already,
a common emitter amplifier with the voltage divider bias. The R1 parallel R2 will be one
contribution and the hie which is equal to beta times re which we have already discussed
this is the other contribution.
The resistances at the Ri will be the parallel resistance of these two resistors. So Ri is
nothing but R1 parallel R2 parallel beta times re. So that plus Rs will be the total resistance
in the circuit with the C1 and the output voltage at these low frequencies will be Ri
multiplied by Vs divided by the potential divider which is Rs plus Ri minus j XCs. (XCs
seen in slide; audio says XC1- ?)
What is that –jXC1? That is the contribution due to the reactance component corresponding
to the capacitors. That also has to be this reactances. It cannot be ignored any more.
The contribution is there at low frequencies and we should take it into account. That is
what we have shown here.
If we want to move on to the contribution due to the other coupling capacitor at the
output which is C2 the network can be now redrawn corresponding to that.
You have RL, the load resistance and C2 the coupling capacitor which is of interest to
us now and Ro the output Thevenin’s resistance of the amplifier. This is the circuit that
we should now look at to find out what will be the contribution to the low frequency response
corresponding to this coupling capacitor. The total resistance now is R0 plus RL at
the output circuit and the cut off frequency due to C2 two is f1 prime.
Why I have said f1 prime is to distinguish from f1 which is the contribution due to the
capacitor C1 and f1 prime is a contribution due to the capacitor C2. f1 prime is 1 by
2 pi the total resistance in the circuit which is R0 plus RL multiplied by C2 the coupling
capacitor and the graph will almost look identical to what you saw in terms of the response.
But the value of f1 and f1 prime can be very different depending upon the magnitude of
the capacitance as well as the output impedance R0, the input impedance Ri, etc. When you
discuss this C2 we would like to ignore the contribution of C1 and CE and take into account
the contribution due to C2 alone. Like that we discuss the contribution to each one of
capacitances that we have in the circuit and then look at the overall response of the amplifier.
In the ac equivalent circuit for this circuit when the input voltage is zero, Vi is equal
to zero. is something like that We have Ro, you have the Rc that will also come as a shunt
parallel to R0 and you have the RL and C2. The value of Ro now is decided by Rc and small
ro. Small ro is the output resistance of the transistor.
Ro, capital Ro, which is the effective Thevenin’s resistance, is equal to RC parallel ro which
we can see from the circuit.
Let us move on to look at the contribution to the low frequency response due to presence
of the emitter bias or the bypass capacitor in the emitter bias circuit. There is a bypass
capacitor CE.
CE, the general form of the R-C configuration for this is very similar. We can see that
CE is there and the amplifier contribution is put in terms of the RE which is the effective
emitter resistance as seen as the Thevenin’s resistance at the emitter point of the transistor,
which I call capital RE and because it is again having the same type of a resistor and
capacitor the cut off due to CE can also be determined in a similar manner and the response
will look very similar.
At low frequencies the gain will fall and the F1 double prime is the cut off frequency
corresponding to the contribution from CE which I call here F1 double prime is equal
to 1 by 2 pi Re CE where Re is the effective emitter resistance that is to be considered.
CE is the emitter bypass capacitance. If you look at the equivalent circuit of this section
including the transistor the re will be at the emitter point, the contributions due to
all the resistors on the base side will be given by Rs prime by beta plus re. This is
the basic emitter resistance. This Rs prime basically is the contribution from the R1,
R2 and the Rs. So it is Rs parallel R1 parallel R2 is what I call Rs prime and Rs prime divided
by beta because you are now translating that emitter base resistance contribution to the
base resistance on to the emitter side. So you have to divide by the beta of the hfe.
Rs prime divided by beta is the contribution of those resistances at the emitter side plus
the re the intrinsic emitter resistance plus the capital RE which is the resistance we
have introduced for biasing purposes and the CE the bypass capacitor. So this will be the
equivalent circuit of this part and we can evaluate the total Re as this Re and this
entire thing comes in parallel. Therefore RE parallel Rs prime divided by beta plus
re.
This will be the contribution to re. Once you evaluate this thing re you can calculate
the contribution to the frequency f1 double prime as 1 by 2 pi Re CE which we have already
seen. So it is possible to get the cut off frequency due to each of these contributions,
the contribution due to C1, the contribution due to C2 and the contribution due to CE that
is what we have …. and I have also told you how it can be evaluated for a given common
emitter amplifier. Now it would be good to take a very simple example and then try to
work out the various frequency responses.
I have a problem here. Determine the low cut off frequency for the network shown in figure
below using the following parameters. There are values given C1 is given as 10 microfarad,
the emitter bypass capacitor CE is 20 microfarad, the coupling capacitor at the output is 1
microfarad, the Rs source resistance is 1 kilo ohm and R1 is 40 kilo ohm and R2 is 10
kilo ohm, RE is 2 kilo ohm, the emitter bias and RC is 4 kilo ohm and RL is 2.2 kilo ohm.
It is also mentioned that the beta value for the transistor is 100 and the internal resistance
ro of the transistor is assumed to be infinity for simplicity. Most of the time it is true;
very close to that and VCC is 20 volts. All the information is given. Now we have to find
out what is the low frequency cut off of this circuit? What is the procedure? The procedure
as I explained to you is very simple. Take the contribution of each of the capacitors
that we discussed. The 10 microfarad C1, the 1 microfarad at the output C2 and the contribution
due to the 20 microfarad CE and find out which one is the worst in terms of the frequency
response and choose that as the dominant contribution to the low frequency. Before we do that we
have to get certain parameters. First let us check
RE. How does it compare with beta times capital RE? Beta is 100, capital RE is 2K that comes
about 200 K and that if it is very, very large compared to 10 times the R2 value then you
can ignore the contribution from beta RE.
So that is coming to be 100 K. Let us calculate what is VB? The voltage at the base is R2
divided by R1 plus R2 times VCC. We all know it is potential divider. It’s a voltage
divider bias and if you substitute the value of the resistors 10 K and 40 K multiplied
by 20 volts the value comes to be around 4 volts. So at the base of the transistor the
voltage is 4 volts. What is IE, the emitter current that is given by VE the voltage at
the emitter divided by RE, that is the dc current; so 4 volts. VE we can calculate by
knowing the VBE drop across the base emitter junction. So 4 volts – 0.7 volts divided
by 2 kilo ohm is the contribution to IE and that is 3.3 by 2 kilo ohms. That is equal
to 1.65 milli amperes. You have calculated IE and then once you know IE you can calculate
the small re the intrinsic emitter resistance of the transistor because we know re is equal
to 26 milli volts divided by 1.65 milli amperes. This is standard formula we have used number
of times and the value of small re comes out to be around to be 15 ohms when you do this
for this circuit and once you know the small re you can calculate what is beta re which
is 100 times that value and that is around 1.5 K.
You know beta re and you can also get mid frequency gain. The mid band gain is minus
RC parallel RL divided by small re and RC is 4 kilo ohm given in the problem. RL is
2.2 kilo ohm. Therefore the parallel value of 4 kilo ohm and 2.2 kilo ohm you have to
evaluate and divide that by the small value 15 ohms which is the re value and the answer
is nearly -90 degrees. The minus indicates that there is a phase inversion between the
input and the output. That is the output voltage is 180 degrees out of phase with the input.
We have also discussed that earlier. What is the contribution to the input resistance
Zi or Ri is equal to R1 parallel R2 parallel beta re. We have already discussed that and
if you substitute the values R1 is 40 kilo ohm, R2 is 10 kilo ohm, beta re just now we
have calculated is 1.576 kilo ohms and therefore the effective resistance should be smaller
than the smallest and you do them in parallel and therefore you get 1.32 kilo ohm. That
is the Ri. Once I know the Ri, I know what is the actual input which is applied across
the input terminal of the amplifier which is nothing but Ri divided by Ri plus Rs multiplied
by Vs. We can calculate the Vi by Vs as Ri by Ri plus Rs the same thing modified 1.32
ohms divided by 1.32 kilo ohm plus 1 kilo ohm. That is 0.569 is the actual ratio of
Vi and Vs.
We can get AVs the output voltage Vout by Vs. This is actually capital AVs, the gain
corresponding to the signal source Vs and that will be -92 into 0.69 so -51 approximately.
That is the overall gain of the common emitter amplifier. Having found the mid frequency
gain we can now find out what is the contribution to the frequency response at low frequency
and high frequency corresponding to each of the component. I will take first R1. Ri is
equal to R1 parallel R2 beta re that is found to be 1.32. That we have already done. What
is f1? f1 one is 1 by 2 pi Rs plus Ri into C1.
You substitute for Rs and Ri and then C1 is 10 microfarad. You evaluate this and the frequency
f1 comes out to be 6.86 hertz that is very low frequency. So at 6.86, 70% is the mid
frequency gain or voltage gain. That is what we understand from this expression f1 is equal
to 6.86. Let us calculate the contribution due to C2.
C2 is also f1 prime. That is what we call C2’s contribution.
1 by 2 pi Rc plus RL times C2. Rc is 4 kilo ohm, RL is 2.2 kilo ohm and C2 is 1 microfarad
and if you evaluate this expression f1 prime is found to be 25 hertz. C1 is around 6.86
hertz and f1 prime due to C2 is about 25.6 hertz. This is slightly higher compared to
that. Let us come to CE. Before you calculate CE you should find out what is Rs prime which
is the combination of all those things Rs parallel R1 parallel R2 and that value is
found to be 0.889 very low value, about 889 ohms. What is Re, the effective emitter resistance
and that will be capital RE the one which you have connected parallel Rs prime by beta
plus re. We have seen that earlier and re is 2 K given in the problem and Rs prime just
now we calculated to be 0.889. Therefore 0.889 by 100 plus this small re is 15.76 ohms and
so the effective resistance is 24.35 after simplification.
This is my re. Once I know the effective emitter resistance we can calculate the f1 double
prime. f1 double prime is actually 1 by 2 pi Re CE where Re is the effective emitter
resistance and 1 by 6.28 which is 2 pi multiplied by Re which is 24.35, just now you evaluated
multiplied by 20 microfarad and if you evaluate this expression you get 327 hertz.
So what do you have? You now have three results. One is 6.86 the other one is about 25.8 and
the third one is 327. Which one will you choose as your low frequency cut off? The best method
is to choose the highest of it. The highest of the three is 327 hertz. That means the
CE capacitor is the one almost dominant among the rest of the capacitances in deciding which
is the low frequency cut off, cutoff frequency. So the low frequency, the cut off frequency
in this problem is 327 hertz which is basically due to the CE, the emitter capacitor. So I
hope this example will help you to understand how the various capacitances contribute to
the cut off frequency at the low frequencies and how to choose the one which decides the
effective low frequency response of the amplifier.
Now we will have to look at the high frequency response. Before we go into the high frequency
response it is useful to learn about another theorem which is called the Miller’s theorem.
The high frequency contribution comes from the inter electrode capacitances. That is
the capacitance between the emitter and the base, the collector and the base and the base
and the collector. Because of that we have to have an understanding how we can simplify
some of the configurations and Miller’s theorem is very useful in this context. I
have taken a general amplifier ‘A’ that is the box at the center and you have a capacitor
which is linked to the input and output. So it becomes a feedback capacitor because output
will be fed back to the input through the capacitor at certain frequencies when the
reactance becomes small and therefore this will have some effect on the overall performance
of the amplifier or the frequency response of the amplifier. But how do we understand
this contribution? How do we simplify the circuit because it seems to be having contributions
both at the input and at the output? For this we try to understand the Miller’s theorem
and apply the Miller’s theorem.
What is the Miller’s equivalent circuit for this? Miller said that one single capacitor
which is connected between the input and the output can be split into two capacitors one
exclusively connected at the input side the other connected at the output side. That is
what is shown in the picture that you see here.
You have got one capacitor or capacitance? coming between the input terminals and another
capacitor comes at the output terminals and these are the two capacitances that you get
when you apply Miller’s theorem and split one feedback capacitance that is showing from
the input and the output. What he says is the input capacitance the effect of the C
at the input will be C into 1- A where A is the gain of the amplifier. Usually the gain
of the amplifier will be very, very large. So C into 1- A is almost equal to A and that
means what? A times C. So the capacitance of the input side will be a very large contribution.
That is the capacitance between the input output which is the bridging capacitance multiplied
by the gain of the amplifier when the gain is very large. The input capacitance will
be very large in magnitude because it is multiplied by the gain whereas if you look at the output
capacitance you find C into A minus one by A. That is what the Miller’s theorem says
and when A is very large compared to one, A minus one by A is almost equal to one and
so the output capacitance will almost be the same as the Miller’s capacitance which is
bridging the two inputs input and the output. The C input due to the Millers theorem is
C into 1-A and the C output due to the Miller’s theorem is C into A minus one by A and this
when A is very large compared to one is almost equal to one and C Miller is equal to C the
capacitance between the input output.
This theorem is very useful in understanding the high frequency response of the amplifier.
If you look at the circuit that we have here you can see I have put some capacitances between
the base and the collector. I call it Cbc the capacitance due to the base collector
and similarly between the collector and the emitter Cce and the base and the emitter Cbe.
These are new capacitances that we have introduced in the circuit. What are they? They are actually
due to the device itself. The transistor itself can be looked at as a capacitor. I already
mentioned to you a simple p-n junction has got capacitance. So you have a p type and
n type which are conducting regions separated by a very small depletion region which is
like a dielectric or an insulator and therefore this can be effectively a capacitor. You do
have similarly at every junction a capacitance, the junction capacitance. In this case between
the base collector junction you have a capacitance Cbc and you have a capacitance CE, Cce and
Cbe between the base and the emitter out of which, which one is the Miller capacitance?
To which I should apply Miller’s theorem? If you look at you can see the Cbc this collector
point is the output point, base point is the input point.
The capacitor which is connected between the base and the collector which is given by Cbc
is the one which is going to be the most crucial in deciding the frequency response at high
frequency. You must also imagine that the circuit is built with number of wires and
these wires can be very close to each other and due to this wiring also there could be
a capacitance that is introduced into the circuit. That is called wiring capacitance.
So all the capacitances due to the wiring, due to the positioning of the various devices
etc., I have put all together and call them as the Ci the input capacitance due to the
wiring. Similarly at the output also we can have a contribution to capacitance due to
various wires. That I call Co. What is Ci and Co? Ci and Co are the wiring and other
stray capacitances at the input and the output side. Cbc is the capacitance which is actually
between the output and the input. Therefore we will apply the Miller’s theorem to the
Cbc and then try to evaluate the high frequency response. Now if I drop the equivalent circuit
of the network or the amplifier and taking into account the presence of the Ci, the Co
etc., you find what you get on the screen.
You have Cs, Rs, Vs and Rs. You have the R1 + R2 parallel resistor coming here. You have
the Ri, the input resistance the Ci the stray capacitances all of them put together. Then
you have the Ro at the output Rc, RL and Co which is the stray capacitances of the output.
Now you can see the capacitances C1, C2 and CE are not shown in this picture and that
is because at very high frequencies the contribution due to C1, C2 and CE will be very, very small
and therefore they are almost ignored or taken as short at these frequencies. What is the
Thevenin’s equivalent circuit in order to evaluate the frequency response of these contributions?
The Thevenin’s response you can see on the input and the output side, I have shown separately
on the screen, it is nothing but Rs parallel R1 parallel R2 parallel Ri the input resistance
of the transistor and you have the Ci which is the contribution due to the stray capacitances,
this form the ones in full circuit.
Similarly at the output you have the Rc parallel to RL and also parallel to small ro which
is the transistor output resistance and Co is the stray capacitances in the circuit.
Once we know that you can immediately calculate what will be the f1 or in this case it is
f2.
This should be Ci here. Ci is the wiring capacitance, base emitter capacitance, the
Miller capacitance and that is nothing but CWi plus Cbe plus the Miller capacitance is
actually Cbc and when you connect it according to Miller’s theorem it should be one minus
A times so that is why it is written as one minus Av times Cbc, the base collector capacitance.
For the output network the f2 prime is 1 by 2 pi R Thevenin’s at the output into Co
and R Thevenin at the output is RC parallel RL parallel small ro and the Co is nothing
but the wiring capacitance, the capacitance between the collector and the emitter and
also the Miller contribution to the capacitances.
As we increase the frequency the contribution due to these things will decrease and therefore
the gain also will fall. I have taken another example here for looking at the high frequency
response. Determine the high frequency cut off of the network shown in the figure which
is basically the same amplifier using the following parameters. C1 is 10 microfarad,
C2 20 microfarad, Cc 1 microfarad all the thing are same as this but in addition we
have given the other capacitances like for example Cbe is 36 pF; Cbc is 4 pF; Cce the
capacitance with collector emitter junction of the transistor is 1 pico farad and Cwi
is the contribution due to wiring at the input side and Cwo is the contribution to the capacitance
at the output side that is around 8 pF.
What is pF? pF means pico farad. Pico farad means 10 power -12 farads. Micro means 10
power -6; pico is 10 power -12 farads. So it’s a very, very low value and the low
value becomes significant only at very high frequencies otherwise it is very low. What
is the solution? The solution we first calculate Ri. We have done that earlier.
Ri is 1.32 kilo ohm which is actually due to the parallel combination of R1 parallel
R2 and parallel Ri and we will also take into account the contribution due to the source
resistance Rs. What is the Thevenin’s resistance at the input? Rs parallel R1 parallel R2 parallel
Ri and if you substitute all the values 40 kilo ohm, 10 kilo ohm, etc., you get the Thevenin’s
equivalent resistance to be 0.531 kilo ohm and the 0.531 kilo ohm is nothing but 531
ohms. So with Ci is equal to Cwi plus Cbe which is the base emitter capacitance plus
the Miller capacitance at the input side which is 1-A times the Cbe, the base emitter capacitance.
If I substitute the values Cwi is 6 pF given in the problem. Cbe is 36 pF given in the
problem and 1-90 where 90 is the gain of the amplifier multiplied by 4 pF. The capacitance
between the base and the collector that is between the input and the output is now enhanced
or multiplied by the beta value and that is why you get such a huge value 406 pico farad.
This is the Miller capacitance at the input side.
What is the fi? fi is the high frequency response due to the presence of Ci.
fi is equal to 1 by 2 pi R Thevenin i into Ci and if you calculate the values R Thevenin
is 0.531 kilo ohm we calculated; Ci is 407 pico farad. If you calculate the frequency
it is 738.2 kilo hertz at the high frequency. It is in kilo hertz; very large frequency.
Now if you want to calculate the contribution to the cut off frequency due to the output
side then we have to calculate R Thevenin at the output side. That is Rc parallel RL
and that is 1.419 kilo ohm and with Co the contribution to Co will be coming from the
wiring capacitances in parallel with Cbe in parallel with Miller capacitance. CWo in the
problem is given as 8 pico farad. Cce is given as 1 pico farad and the Miller capacitance
is 1-A. That is 1 minus 1 by 90 degrees, 90 which is the gain into 4 pF. This is the Miller
contribution and if you evaluate the total is found to be 13.04 pico farad.
What will be the cut off frequency? The cut off frequency now due to the contribution
from these capacitances is 1 by 2 pi R Thevenin’s at the output stage multiplied by Co and if
you substitute the values you get that to be 8.6 mega hertz. Mega means 10 power 6 hertz.
So the high frequency cut off is really very high corresponding to nearly 8 to10 mega hertz.
I have put all these values in the form of a table. The cut off frequencies for low frequencies
the contribution comes from C1, C2 and CE and the high frequency contribution is 738
kilo hertz corresponding to fi and fo is 8.6mega hertz.
If you look at this table you can tell the worst case condition corresponding to the
input and the output. What is the overall f1 and the overall f2? If I call the high
frequency response as f2 what is the overall input and the output cutoff frequencies? I
must take the largest value in this; 6.86 hertz, 25.68 hertz, 327 hertz. The cut off
frequency which is most important, dominant is given by 327 hertz which is the largest
frequency among all and that is the worst case. If you look at the high frequency it
should be the other way. In a high frequency the contribution due to the input capacitances
are 738.24 kilo hertz and the contribution to the output capacitances is 8.6 mega hertz
and the worst thing here is 738.2 kilo hertz. Therefore you can take that as the upper cut
off frequency. So in this problem we know what is the lower cut off frequency? What
is the upper cut off frequency? The lower cut off frequency is 327 hertz and the upper
cut off frequency is around 738 kilo hertz. So these two are the ones which will decide
what is the band width of the amplifier? Actually the band width of the amplifier is the difference
between these two frequencies. That is 738.24 kilo hertz minus 327 hertz gives you the band
width. If you are working with the dc coupled amplifier then there is no low frequency coupling
capacitor involved and the amplifier response will start almost from 0 frequency that is
dc and you don’t have to worry about the presence of f1 and if you don’t have f 1
then the band width is given by the upper cut off frequency itself. That is if I don’t
have the input low frequency contributions if it is a dc coupled amplifier then the bandwidth
of this amplifier, we have worked out the problem, is 738 kilo hertz. So that is the
bandwidth.
Incidentally when we discuss bandwidth there is a related thing that we have to worry about.
Is there a short cut method of measuring the frequency response of an amplifier? What is
the normal procedure? The normal procedure is you construct the amplifier, you apply
the signal source and keep the amplitude low because it is a small signal amplifier and
monitor the output. You keep on increasing the frequency from very low value to very
high value and every time you record the output voltage from which you can get the gain and
at very low frequencies the gain is very small and as you increase the frequency the gain
increases and it comes to a maximum value and it maintains the maximum value constant
over wide range of frequencies and if you further increase the input frequency it starts
falling off and that is how the response curve that I showed comes about.
This will take long time. If I start from very low frequency to very high frequency
checking at every frequency what is the gain and what is the input output voltage etc.,
it will take a long time. Is there a quick method by which I can measure the frequency
response or get an idea about the response of the amplifier? The answer is yes and that
is by using what is known as a square wave testing of the amplifier. Before I go into
that I will briefly explain to you the concept of the rise time and the fall time and the
corresponding relationship between the rise time and the bandwidth.
Here I have shown a very simple circuit with a voltage source, a switch and a resistor
and a capacitor; something like our equivalent circuit.
The only difference this is a dc voltage. If I now switch it on by moving this switch
then suddenly there will be a large current flowing trying to charge the capacitor. Initially
the charge on the capacitor will be zero. Once I switch on then the voltage will start
building up very quickly to very high value and once it is completely charged it will
be charged to the full voltage V and it will reach the voltage V and then it will remain
constant if I don’t discharge for long time. What is the rise time? The rise time is shown
here. It is defined as the time taken between the point when the voltage comes to about
10% of the maximum V plus and 90% of V plus. So the time between 0.1 times to 0.9 times
the total voltage applied is what is called TR the rise time. That is by definition. You
can also do that by using a square wave instead of having a dc source.
You can have a square wave source that is a function generator with a square wave output.
You connect the RC and you switch on you would find this increase from zero to plus V will
not be instantaneous at the output depending upon the bandwidth and so there will be a
finite time delay in reaching the full V value, voltage value and that can be defined as the
rise time, the time taken to rise from 10% to 90% of the total amplitude V of the square
wave; I can take it as the rise time. If I have a simple RC circuit and give a square
wave you would find the output will not be exactly square.
It will show some distortions from the input square wave. These distortions give us lot
of ideas regarding the overall frequency response of the amplifier. That is what we should try
to understand. For example I have here a square wave which goes from Vm as the maximum voltage
to -Vm as the minimum voltage and has a period T.
What I do is I give this as the input to an amplifier whose bandwidth I want to evaluate
and monitor the output wave form on an oscilloscope. From the type of waveform you get if it is
a very wide band width amplifier what would you get? You will get an exact reproduction
of the same square wave at the output. In that case the bandwidth is very, very high;
reasonably high. But you will always get some distortions in the square wave and those distortions
will indicate to you quickly what will be the bandwidth; whether it is having good low
frequency response or good high frequency response etc., one can get by quickly analyzing
using a square wave input. Why is it so? What is special about the square wave? You may
recall there is a Fourier theorem or the Fourier series of expansion for a square wave. I have
given it on the screen.
It is having contributions due to several frequencies sin 2 pi fs t which is the fundamental
frequency; fs is contribution due to the fundamental frequency and 1/3 sin 2 pi 3 fs which is the
third harmonic, 3 times fs, contribution; similarly the fifth harmonics 5 fs, nth harmonic,
etc.
If I now take all these sine waves separately one at fs, one at 3fs, one at 5fs and monitor
the amplitude to be 1/3rd, 1/5th etc and then superpose all of them together what will I
get? If I keep on doing it up to infinity I should not stop with n but I should keep
on doing it up to infinity. Then I would get a perfect square wave which means to say a
perfect square wave can be considered as due to several sine wave contributions which are
basically odd harmonics like fs, 3fs, 5fs, 7fs, etc up to infinity and there is a gradual
decrease in the amplitude also of these components and if you superpose all of them what you
get will be the total. Instead of doing spot frequencies you are actually doing it by using
a square wave and because the square wave is contributed by number of frequencies you
will be able to understand the contribution. That is what is shown in this figure here.
You can see there are number of sine waves coming in. This is the fundamental one, large
frequency. Then you have the third harmonics that is coming here and the contribution due
to all of them if you add you will get a perfect square wave if it is for infinite sequence.
In the next picture I have shown you the various types of square wave that we will get at the
output.
For a clean square wave that you give at the input you will get different types of square
wave at the output and by looking at the wave form you will be in a position to say what
type of an amplifier it is. What is its frequency response? For example a poor low frequency
response of an amplifier will be shown by this shallow region that you see here. There
will be a curved portion at the top which should normally be flat. For a good square
wave it will be curved and this shows that it has got poor low frequency response if
you have an output like this. If you have output still bad very poor, low frequency
it would fall by a very large extent. This fall will be very large. When that happens
the low frequency response of that amplifier is very, very poor. What about the high frequency?
If the high frequency response of an amplifier is very poor then you get a curve of this
type. This will be corresponding to the poor high frequency and if it is very poor it will
still be curved on both sides and this shows very poor high frequency response. If I just
want to get quickly information regarding the performance or the bandwidth or the frequency
response of an amplifier the quickest way is to just give a square wave at the input
and look at the wave form that you get at the output. If it is any one of these you
would able to decide whether it is having a poor or very poor low frequency response
or high frequency response.
There is also a relationship with reference to the bandwidth and the rise time and that
relationship is 0.35 by the rise time; tr is rise time.
0.35 by tr is actually the bandwidth fH1 which is actually the cut off frequency at the high
frequency and for the dc amplifier there is no low frequency cut off and this itself becomes
the bandwidth. In general to get a quick idea of the bandwidth people will just do 0.35
by rise time by giving a simple square wave and looking at the rise time which is actually
the time taken between 10% to 90% of the maximum you can find out the value of tr and you will
be able to get the estimate of the bandwidth also. That tilt that you see here in the graph
for a square wave can be percentage tilt which can be obtained as p% is equal to - missing
in audio V minus V prime by V where V is this voltage at the top and V prime is the lowest
value that you got for the horizontal portion multiplied by100. This also can be used and
from that also the P the tilt divided by pi into fs the frequency of a square wave will
give me the bandwidth idea. These are the two different ways in which you can quickly
get, by using a square wave, the frequency response of an amplifier. What we have so
for seen in this lecture is to look at the contributions from the coupling capacitors,
the bypass capacitors which are all contributing to the low frequency response of the amplifier
and also the inter electrode capacitance between the base and the collector and by applying
Miller’s theorem you can split them into two contributions at the input and the output
and then evaluate their performance that will correspond to the high frequency cut off and
from that we will be in a position to evaluate the bandwidth of the amplifier. We also saw
how in a very simple manner we can use a square wave to get a quick idea about the bandwidth
of a given amplifier from the rise time as well as from the type of wave form you get
because the square wave due to the Fourier theorem is basically contributed by number
of fundamental, 3rd harmonic, 5th harmonic, odd harmonic. All of them together if you
combine you will get a square wave therefore is equivalent to performing the frequency
response at different spot frequencies like f3, f5, f, etc and what you get is the result
of all those things and so you can get a quick idea about the frequency response of amplifier.
Thank you!