Considering a 3-stage phono amplifier with a DC servo, measuring at the output and injecting the correction signal in the first stage;
are there any concerns in an DC-servo arrangement as shown below, where the servo-loop includes 2(3) gain stages?
I don’t want to build this amp like this, so please ignore everything else.
are there any concerns in an DC-servo arrangement as shown below, where the servo-loop includes 2(3) gain stages?
I don’t want to build this amp like this, so please ignore everything else.
No problem with DC servo via all stages (with U1 inverting and U2-U4 noninverting, so total DC negative feedback loop is OK), but R3 must be at least 10 times lower because 2 kOhm has too big Johnson-Nyquist noise (compare to MM cartridge resistance). Also I recommend use for U4 eTrim opamp OPA192 with 5 uV offset instead 80 uV @ADA4625 . See my MM phono here [ https://www.patreon.com/posts/68518012 ] , circuit diag with DC servo on OPA192 [ https://c10.patreonusercontent.com/4/patreon-media/p/post/68518012/b69e205aaf424c299f98312a8fa273cd/eyJxIjoxMDAsIndlYnAiOjB9/1.JPG?token-time=1687305600&token-hash=YPzIjSf_if7I3S53HXGwhix2JTa0zw0yc1aDWK0vWn8= ] . Also recommend increase R1 to 150 kOhm [ https://www.patreon.com/posts/70415214 ]
Last edited:
The only issue is that you must keep the right polarity. Most servo's are inverting, and insert the correction signal at the inverting input side.
If you servo over two stages and each stage is inverting, you're net still inverting, but with three inverting stages in the loop the net is inverting so now your servo signal is no longer inverting and will not work as expected.
Remember: two consecutive inverting stages are net non-inverting!
Jan
If you servo over two stages and each stage is inverting, you're net still inverting, but with three inverting stages in the loop the net is inverting so now your servo signal is no longer inverting and will not work as expected.
Remember: two consecutive inverting stages are net non-inverting!
Jan
That's somewhat confusing Jan, as there are three non-inverting stages (+12x, +12x, +2x), so the servo should not invert, as its signal is connected to the first stage inverting input.
I did not comment on that particular circuit as the OP clearly said that he doesn't want to build that and we should disregard it. 😎
The 2nd poster also appears not to have read or understood that.
I gave a general description of how it works, as that was asked for.
Jan
The 2nd poster also appears not to have read or understood that.
I gave a general description of how it works, as that was asked for.
Jan
Just make sure the amplifier is DC-coupled internally if you're going to wrap a DC-servo around it(!)
When you put a DC servo around a phono amplifier, you have to keep in mind that the extra feedback loop you make will shift the correction poles (not the zeros, just the poles). You can either precorrect the poles such that they end up at the right spot once the loop is closed, or design the DC feedback loop to have very little loop gain at 50 Hz and above (say a loop gain with a magnitude < 0.01), or put an inverse RIAA network in the DC feedback.
That would be in front of R1 (π-network: 100k-330n-100k for 10Hz).
But it's beyond the op's request.
I'm more curious about the phase shifts throughout these stages.
MvdG already hints this issue in #9!
But it's beyond the op's request.
I'm more curious about the phase shifts throughout these stages.
MvdG already hints this issue in #9!
Thanks for all the comments!
I’ve already made simulations for the final version and one can clearly see how the DC-servo affects the correction poles -I didn’t expect that in this extent.
I adjusted the values of the RIAA correction accordingly. The whole amp has now -3dB point close to 5Hz.
I will post the whole project once the amplifier is running; it’s a follow-up to this one: https://www.diyaudio.com/community/threads/poor-mans-lr-lcr-riaa-experiment.383480/
I’ve already made simulations for the final version and one can clearly see how the DC-servo affects the correction poles -I didn’t expect that in this extent.
I adjusted the values of the RIAA correction accordingly. The whole amp has now -3dB point close to 5Hz.
I will post the whole project once the amplifier is running; it’s a follow-up to this one: https://www.diyaudio.com/community/threads/poor-mans-lr-lcr-riaa-experiment.383480/