Hi Hans,
Yes I built 100W CFA and 200W CFA amps and sold some PCBs here in this forum. 200W MOSFET CFA amp
I've just start to use QA400 for THD measurement, but still not very familiar with it, look some result here(but you have visited that thread too): https://www.diyaudio.com/forums/the-lounge/349926-black-hole-519.html#post6369533
Don't take it as whole true about the caps.
Damir
Hi Anistari,
When you say interesting, what do you mean, good or bad harmonic profile?
BR Damir
In a good way. You should do simulation yourself and try to explain why it happen.
OK, but how you came to conclusion that I did not simulate? How I could designed it with no simulation. My question was, did you measured it?
Of course I knew you simulate it. But you do not talk about it (THD vs output leve vs harmonic profile). It is not only you, maybe someone can explain it, too.
So many simulation of OITPC amplifiers I did, and I never build it myself.
Yes. Very good in imaging and attack time in general (CFA OITPC and an amplifier like AKSA 55 but with OITPC).
What to your feeling should be the point.
As I have shown and also by Damir, improvement in settling time is not at all guaranteed with OITPC, but can be achieved at no cost with carefull selection of components, so why not.
With TPC alone this is not possible but Damir’s CFA sets a milestone in this respect.
Distortion with OITPC is not better than with TPC so what else is there that I’m overlooking ?
Hans
Last edited:
Exactly, or at least a marginal improvement. Simulated very good THD performance of these OITPC compensated amps is coming from the (practically, huge) ULGF (6MHz or more).
In practice, as also Dadod's early measurements are showing, THD is much more affected by the layout and magnetic effects from the wiring than any TPC or TMC compensation can help with.
Something that I could never understand, early Dadod OITPC simulations were showing a double pole at high frequencies, then mitigated with a zero before the gain crosses the 0dB axis (sorry I don't have a link to those posts). That would indeed be an improvement in the HF loop gain (although a double pole is not easy to manage, since they may have the nasty habit to migrate in different directions when the loop closes, and/or when the output approaches the rails).
I am also not convinced on the worst case stability of this configuration; way too many loops to analyze, the overall feedback simulation doesn't always tell the whole story. A capacitive feedback from the output node was in depth analyzed by Cherry and, while proven effective in lowering the distortions, it always came with a stability penalty (which is also intuitive). So strong that Cherry suggested some crutches for this idea, like picking the output feedback from after the resistor in the output Zoebel.
It is of course different in the OITPC schema, but I still fail to understand how a capacitive shunt feedback at the amplifier output node can improve the THD without sacrificing the stability margins.
Improvement in Harmonic distortion... also settling time is not that important in audio since we do not deal with square waves. Settling time is more important in component video (obsolete) and in data acquisition.
Distortion with OITPC is not better than with TPC so what else is there that I’m overlooking?
If you have the time, could do a side by side comparison of Miller, Transitional Miller, TPC and OITPC?
Thanks!
Unfortunately that would be very difficult, since it needs a base amplifier that exposes ideal properties (read: impedances) at all feedback pickup points; for example, Transitional Miller may look like a better compensation method under different circumstances than TPC, although it was already mathematically proven that they are otherwise almost equivalent.
Using a CFA topology as Dadod did doesn't make the comparison easier, since the compensation options are less than for a VFA. In fact, other than shunt, Miller and the Ahuja compensation I can't think of any other CFA compensation topology (I could be wrong, though).
No, it does not, at least not always. As above, tying a type of compensation to a specific amplifier topology and implementation doesn't tell much about the properties of the compensation method.
In general, it is the compensation order that correlates with higher loop gain in the frequency domain of interest, that is, N poles and N-1 zeroes to bring the phase back, and for each compensation order there is a maximum well defined amount of loop gain that could be extracted from the open loop gain and OLGF. Unfortunately, practical considerations (realizability, sensitivity) are severely limiting the options over N>3 (N=5 is the most aggressive compensation I've seen), a good example of N=3 is Cherry's NDFL.
In general, it is the compensation order that correlates with higher loop gain in the frequency domain of interest, that is, N poles and N-1 zeroes to bring the phase back, and for each compensation order there is a maximum well defined amount of loop gain that could be extracted from the open loop gain and OLGF. Unfortunately, practical considerations (realizability, sensitivity) are severely limiting the options over N>3 (N=5 is the most aggressive compensation I've seen), a good example of N=3 is Cherry's NDFL.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.