Taming Ringing...Square-Wave Response

[EchoWars]

Oh....should have seen that sooner.

Playing with values that I have laying around here, I've changed C7 to 22pf, and R18 to 2.6K. This dampens the ringing much quicker, but it still needs help, so I also played with values to jumper across the feedback resistor. 12pf here calms it down a lot, but there is still a lot of overshoot on the rising edge. I have a few 39pf caps, but that seems to be too much. 22pf is almost there, but a bit more seems in order to make it right.

I have some 33pf caps on the way to use across the feedback resistor, and I think that may be what is required to eliminate the ringing without dampening the response too much.

Does that seem like an excessive value? Sure cleans up the wave nicely though, and doesn't round off the square wave too much, even at 20KHz.

 

[subwo1]

You can also try placing some in series to obtain lower values. BTW, did you try tacking one from the VAS collector back to the differential transistor base?

 

[EchoWars]

Yes I did, but it seemed to round the square wave off without reducing the ringing nearly as well as simply putting a cap across the feedback resistor. Spent a couple of hours experimenting with caps here and there, and whatever I do seems to be a compromise (and I knew it would be). So the adjustment of the cap and resistor at Q5 and a cap across the feedback resistor seems to do the least damage while eliminating the ringing.

 

[subwo1]

I was very curious. Thanks for the feedback.

 

[jcx]

if you have to play with the compensation perhaps looking at R23,24 C9,10, which gives a 1 MHz pole that you may not want to double up by adjusting C1,R18 without considering the pole/zero locations and possible cancelation

C9,10 termination to the ps (where they can feed tru ps noise directly to the output) instead of ground is just one of many poor design choices visible in this schematic

 

[ChocoHolic]

Hi Echo!
...sounds quite OK, now.
I think 33pF across the 33k feedback is perfectly fine,
also 39pf or up to 47pf would be OK for my taste.
No normal loudspeaker can follow such fast signals.
Even with 47pF in parallel to 33k in the feedback you would still end
up in a -3dB frequency response up to around 100kHz.....

...now enjoy the amp...
Cheers
Markus

 

[EchoWars]

if you have to play with the compensation perhaps looking at R23,24 C9,10, which gives a 1 MHz pole that you may not want to double up by adjusting C1,R18 without considering the pole/zero locations and possible cancelation.

Although the cap (I know it's hard to read, but it's C7...) has been slightly more than doubled, and the resistor changed to 2.6K, it seems to not have had any effect other than to reduce the ringing no matter what test I run on it. I'm not up to the calculations, but if you're game I'd certainly take a peek at your findings

ChocoHolic...I should have parts in today, so we'll see if the 33pf cap is the right compromise. I believe it will be.

 

[subwo1]

You could also try adjusting C5 and C6.

 

[EchoWars]

I did that...and it rounded off the square wave without dampening the ringing.

Got the 33pf caps in today, and put them across the feedback resistors. Up to about 7KHz, the square wave is sharp and neat. After that, the edges begin to round off, but even a 20KHz wave still looks better than what I've seen on a lot on amps.

Unless anyone has a rabbit to pull out of a hat, or point out some horrible side efect that I haven't consiidered, I believe I'll leave this one alone. Fini...

 

[subwo1]

good work