You added C6/47pF from output to vas compensation which is some kind of cherry comp. I never got an amp stable which was using a cherry cap from output to VAS as you can read in my 2stageef thread ...
It is not about THD but harmonic profile of the distortion. You can ask Hugh Dean, for the explanation.
Of course harmonic profile is changing with level, but not because the compensation used but because increase distortion in an amp before feedback was applied. Or shall I say, depends of the circuit linearity without feedback.
Last edited:
Hi Bimo, OITPC makes sense with the traditional global feedback approach.
In the case of ODNF, only the local feedbacks are in place in each stage, so OITPC is not relevant anymore.
Although ODNF is, in fact, global in its nature (as it takes into consideration the whole amplifier, from its input to its output), the way this feedback is applied is rather special (the error signal is subtracted from the total signal at the VAS output).
If you look at the Simpelstark schematic, there are local c-e shunting caps at the drivers and a small capacitor across the front-end, limiting its bandwidth at HF. Now, the only thing we need to take care of is the error channel - the phase correction capacitor I mentioned earlier, does the job.
You must prove it of the same amplifier with different compensation can have same harmonic profile.
Damir, I think, the overall compensation approach is fine - what Thimios sees is some local issue. I have experienced this kind of situation a few times. One issue was in the VAS (local instability), the other one was in the output stage (caused by the local parasitics).
I don't need to prove that, because different compensations have different Loop Gain plot and thus different harmonic profile, but change of harmonic profile with the input signal level does not depends on the type of compensation used.
Bimo could we stop this discussion in this thread?
I agree.
Global stability is not enough to guaranty amplifier not oscillating. Example: using cascode if not carefully designed, it can oscillating, or parasitic inductance or capacitance of pcb layout can make instability.
Valery, yes I think it is local issue, problem in cascode, bonsai described that nicely.
No, oscillation was 21.7MHz look here. https://www.diyaudio.com/forums/solid-state/338814-lateral-cfa-120w-bsa-13.html#post5832304
Have a look here: 2stageEF high performance class AB power amp / 200W8R / 400W4R
and the below following posts. The compensation was a bit too agressive. Increasing the PM/GM cured the problem.
and the below following posts. The compensation was a bit too agressive. Increasing the PM/GM cured the problem.
Sorry Tony can you name the caps?
I see that always is unstable when a voltmeter connected at speaker output.
When i run an fft test and connect a voltmeter, immediately artifacts apeared
Last edited:
Toni I followed that thread before, but I use different compensation, you can look for OITPC. OITPC - Output inclusive TPC (not TMC)
Dear Thimios,
for a quick test I would remove C6 and increase C9 and C10 to e.g 47pF (schematic from post 1)
Hopefully no holy smoke...
BR, Toni
I would say a mix of cherry and tpc and ?
The 110pF (2 x 220pF in series) shunt from positive to negative vas input - has this been checked too? What does this do?
I can remember Dave Zan has jumped in to check some maths - but is there a result?
BR, Toni