Questions about the 2-pole compensation VAS

[gaetan8888]

Hello

In the Doug. Self Power Amplifier book there is a 2-pole compensation VAS, as you can see in the image I've upload.

In that image you have Cp1 (100 pF) and Cp2 (1 nF) capacitors, what would be the practical resulting cdom Miller capacitance value of that VAS ?

The values of Cp1 and Cp2 capacitors are quite high, how much lower can we go for those capacitors ?

Anybody have try this type of VAS, how good it is, what is the pros and cons of this VAS ?

Thank you

Bye

Gaetan

[jaycee]

I've not seen the two that far apart before. I simulated it though and it seems to work very well indeed. Here is what I simulated - I have even laid out a PCB for this (a compact one - for a limited size case) but have not yet got around to building it. The CFP drivers may be excessive

[gaetan8888]

Hello jaycee

The value of C4 and C17 are relatively high, is it because it's a beta enhance vas, have you try much lower caps value ?

Quite all beta enhance vas I've seen used high value cdom capacitors, this type of vas seem prone to oscillate.

Have you calculate the resulting miller capacitance value of your 2-pole comp vas ?

Thank

Bye

Gaetan

[jaycee]

lower values caused oscillation. You'll see also the R-C network in the feedback path, this was also needed. I plotted the loop gain. There is a peak at 500KHz but this doesn't seem to cause a problem. The 0dB point is at about 1.9MHz with a phase of -104 degrees.

In simulation I've thrown square waves and sines at it, and it seems stable. It's also stable into a 1u capacitor, although it needs an L-R parallel network on the output for that.

I think 5th element tried a similar circuit here too, with good results.

[jaycee]

Yes he did. Take a look at this thread:

Slones 11.4 'Blameless'

[gaetan8888]

Originally Posted by jaycee

lower values caused oscillation. You'll see also the R-C network in the feedback path, this was also needed. I plotted the loop gain. There is a peak at 500KHz but this doesn't seem to cause a problem. The 0dB point is at about 1.9MHz with a phase of -104 degrees.

In simulation I've thrown square waves and sines at it, and it seems stable. It's also stable into a 1u capacitor, although it needs an L-R parallel network on the output for that.

I think 5th element tried a similar circuit here too, with good results.

Hello

I agree, but high value miller cap slow down the vas.

thank

Bye

Gaetan

[wahab]

Cp2 value seems uselessly high...

Cp1 value is also high considering that this is a single
transistor VAS that is compensated.
Such a high value would make sense with an enhanced
VAS which would add an common collector bjt to drive
it, as in the blameless.

Typical values would be 100 and 330 pF for an enhanced vas,
56 and 180 pF for a single bjt vas;
Of course, these values are to be adjusted according to the first
stage current capabilities.

To answer your question, although a high value miller cap slow down
an amp, with TPC , its influence is reduced by the resistor that act as
an attenuator, thus limiting the frequency dependent local negative
feed back, and thus extending the vas -3db frequency response..

[Bonsai]

Jaycee, I think you have plotted the open loop response?

Have you looked at the loop response which will give you a truer idea of your gain and phase margins.

[janneman]

Originally Posted by gaetan8888

Hello

In the Doug. Self Power Amplifier book there is a 2-pole compensation VAS, as you can see in the image I've upload.

In that image you have Cp1 (100 pF) and Cp2 (1 nF) capacitors, what would be the practical resulting cdom Miller capacitance value of that VAS ?

The values of Cp1 and Cp2 capacitors are quite high, how much lower can we go for those capacitors ?

Anybody have try this type of VAS, how good it is, what is the pros and cons of this VAS ?

Thank you

Bye

Gaetan

The way I understand it, the equivalent Cdom would be the series equivalent of 100pF and 1nF (about 90 pF).

BTW, this is extensively discussed in the Bob Cordell thread on negative feedback.

jan didden

[gaetan8888]

Hello

Looking at a Andrew T formula for those two cap, it seem that we need anyway a high value cap.

Slones 11.4 'Blameless'

Anyone have check the slew rate of a amp with a 2-pole compensation VAS compared to the same amp with a standard VAS ?

I've take a look at the Bob Cordell thread on negative feedback, it's a quite long thread to read.

Wen I do amps, using a phase lead cap and a good pcb layout, I can go down to 19 pF for the cdom cap, using a 10 pF phase lead cap, I check it with a 10 khz square wave. It give fast and clean sounding amps (it still need an input section with a good linearity). But I don't use a beta enhance vas.

Thank

Bye

Gaetan