[SOLVED] How to make an amplifier stable?

For understanding the concept of phase margin i need to design bjt amplifier with poor phase margin (oscillatory output) and then do compensation to make the amplifier stable. can anyone suggest me the method for doing this.

Re: how to make amplifier stable ?

ALL opamps and most audio amps that have negative feedback use a "frequency compensation capacitor" to provide a good phase margin.
Make an amplifier with a capacitor value that is too low or that has too much negative feedback for the small value of the capacitor and it will have a poor phase margin. Then increase the value of the capacitor or reduce the amount of negative feedback to make it more stable.

Re: how to make amplifier stable ?

Some op amps as the LM301 are not internally compensated. They need external components to make them stable. You can build an amplifier with one of these and then stabilize it.
See the datasheet for the suggested compensation methods.
https://www.ti.com/lit/ds/symlink/lm101a-n.pdf

Re: how to make amplifier stable ?

Amazing. The old LM101A, LM301A are still produced and used (replacements)?

Re: how to make amplifier stable ?

About your private message to me: "WHICH CAPACITOR ARE YOU MENTIONING"
Most tutorials about opamps mention the frequency compensation capacitor. The old LM101A and LM301A opamps do not have this capacitor built-in and their datasheet talks about adding it.

I copied parts of the datasheet of an old LM741 opamp and I circled its frequency compensation capacitor. I also copied its frequency response to show you how this capacitor reduces high frequency gain for good stability with any amount of negative feedback.

Re: how to make amplifier stable ?

i need to make simple amplifier using bjt with poor phase margin and then need to improve it using compensation .so as to avoid oscillations and to make the amplifier stable.

Re: how to make amplifier stable ?

Start by designing a simple amplifier with about 3 BJT gain stages, and apply negative feedback to set the voltage gain.

It will probably have stability problems. Now you can add compensation to fix the stability.

Re: how to make amplifier stable ?

Note that most ordinary opamps have a voltage gain of about 200,000 times.

Re: how to make amplifier stable ?

godfreyl said:

Start by designing a simple amplifier with about 3 BJT gain stages, and apply negative feedback to set the voltage gain.

It will probably have stability problems. Now you can add compensation to fix the stability.

Click to expand...

can i do it with single stage..why i should go for 3 stage amplifier circuit

Re: how to make amplifier stable ?

One transistor amplifiers are normally stable to start with, so they don't need compensation.

Here's a quick explanation:
Negative feedback is often used to reduce the distortion of an amplifier. However positive feedback can cause the amplifier to oscilate.

The problem is that the gain of the amplifier rolls of at high frequencies like a low-pass filter, and this causes phase shift. If the phase shift reaches 180 degrees at some frequency, then the feedback is positive instead of negative, and the amplifier will oscilate if the gain is still high enough at that frequency (i.e. loop gain > 1).

An amplifier with a single gain stage will have a slow roll-off (first or maybe 2'nd order) so the phase shift is unlikely to reach 180 degrees before the loop gain drops below unity.

When more gain stages are added, the gain is much higher and the high frequency roll-off is much faster (meaning more phase shift), so instability is more likely.

- - - Updated - - -

Anyway, here's an example. The circuit shows a 3 transistor amplifier with two common emitter stages and an emitter follower at the output. Without C3, the amplifier is not stable - it oscilates with a small capacitance (e.g. 100pF or 1nF) connected across the output. Adding C3 fixes the problem.



To improve it a bit more after adding c3, you can:
1) Add a 100 Ohm resistor in series with the output.
2) Add a 100pF capacitor between input and ground, then add a 1K resistor (or bigger) in series with the input.

btw, If you examine the circuit, you may notice something unusual - there are two separate feedback loops:

1) R6 provides feedback from the output to the emitter of Q1, setting the voltage gain to R6/(R6+R1) = 11

2) R3 provides DC feedback from the emitter of Q2 to the base of Q1, to stabilise the Q-point.

Because of the interaction of these two feedback loops, the ratio between C1 and C2 is quite important. If the ratio is too low, you get a small peak in the low frequenct response.

Re: how to make amplifier stable ?

Stability of single transistor amplifiers can be pretty problematic if you are talking about RF applications above some 100 MHz. But I presume you won't like to calculate it's stability because it involves more complex transistor models and parasitic circuit elements. So looking at lower frequencies and multi-transistor circuits will ease your life a lot.

Re: how to make amplifier stable ?

One of the first transistor circuits I made oscillated. It used a single transistor as an emitter-follower. It had a little stray capacitance from its output to ground and it was missing a supply bypass capacitor. It also might have had a little input to output capacitance.

A single common-emitter amplifier has capacitance from its collector to ground so its phase shift is 90 degrees max then negative feedback will not make it oscillate.
Two common-emitter transistors in an amplifier produce phase shift that is 180 degrees max but at this high frequency the gain is reduced so it probably will not oscillate when it has negative feedback.
Three common-emitter transistors in an amplifier will oscillate when they produce some gain and have negative feedback when each transistor has a phase shift of only 60 degrees.

Re: how to make amplifier stable ?



- - - Updated - - -

but no oscillations are found without c3 cap @godfreyl

Re: How to make amplifier stable?

There's something wrong with your circuit. With the values I showed, the voltage gain should be 11 (about 21dB), and the low frequency rolloff should be -3dB at about 15Hz, IIRC. Your results seem to show the response rolling off at about 15KHz. Also the gain looks too low.

Besides that, are you sure it's really stable? Try applying a 100mV 100KHz square wave signal to the input, and look at the output on the emitter of Q3 with no load connected. Then try again with different values of capacitor (e.g. 100pF and 1nF) connected between the output and ground. What do the waveforms look like? No oscillation? No ringing?

Re: How to make amplifier stable?

i'm using 2n3904 transistor,,what is the value of C4? and i can use only 5v supply

Re: How to make amplifier stable?

The value of C4 depends on the load impedance and the frequency range of interest. If the circuit is used as an audio pre-amplifier, then 10uF would be OK since the load is not likely to be less than about 5 or 10K. If I get to my home computer later, I'll do some more simulations. Different transistors will give different results. Also, the SPICE models used in simulators aren't always very good, so there can be a difference between simulation and reality.

P.S. Did you see the edit to my previous post? Sorry for cross posting.

Re: How to make amplifier stable?

thanks for your reply,i'm doing real circuit..

godfreyl.. i have not got the ringing output..(i'm having limited source like 5v supply and 2n3904 transistor)

kannan1 said:

5v supply and 2n3904 transistor)

Click to expand...

That will make it more stable, but I'm surprized there's no ringing.

Did you try the test I suggested in post 14, with different capacitors between output and ground?

For that matter, did you ever try simulating the circuit I actually showed in post 10, with the same supply voltage, same component values, same transistors etc? If so, did you try the test I suggested in post 14 with that?

godfreyl said:

That will make it more stable, but I'm surprized there's no ringing.

Did you try the test I suggested in post 14, with different capacitors between output and ground?

For that matter, did you ever try simulating the circuit I actually showed in post 10, with the same supply voltage, same component values, same transistors etc? If so, did you try the test I suggested in post 14 with that?

Click to expand...

as i told i'm having only 2n3904 ..tried with that but output was not found ringing..i'll try one more time using simulator