The only difference between TPC and TMC is where the resistor is connected to. Connecting the resistor to the output, what is being referred to as OIC, is slightly better, else it is the same.
TPC vs. TMC has been discussed before BTW:
While being very similar, TMC offers higher performance due to inclusion of the output stage.
While being very similar, TMC offers higher performance due to inclusion of the output stage.
Exactly. It’s that extra little bit of error correction through the semi-local feedback I found attractive.
I show what is possible in https://www.diyaudio.com/community/threads/tpc-vs-tmc-vs-pure-cherry.235188/
Though the amp(s) in that thread are virtual, I've built and tested similarly amps in Jurassic times (1990?) with similar results ... especially the 20kHz full power THD residuals
Though the amp(s) in that thread are virtual, I've built and tested similarly amps in Jurassic times (1990?) with similar results ... especially the 20kHz full power THD residuals
TMC is totally worth considering and offers some advantages over TPC.
It would be easy to add both resistor options to a PCB design to have a fallback.
Removing the resistor would be the fallback to plain Miller worst case.
I once built an amplifier with TPC, and performance was very good, as far as I could measure.
The amplifier topology required the resistor goes to the rail for good PSRR, else I might have considered TMC instead.
However, I believe there were issues with stability in some cases (not sure) and I reduced the aggressiveness of the compensation, i.e. reduced the capacitor ratio, and increased the resistor value. The presumed instability was around the closed loop gain peak frequency BTW, hence it is plausible that there could be a risk. Has anybody else observed any stability issues at the closed loop gain peak frequency with TPC or TMC?
In simulation, there was good margin, but this is small signal only, and reality is always a different story.
Does anybody know whether TMC shows the same amount of closed loop gain peaking like TPC?
While being so similar, they are still slightly different.
This certainly has been discussed before, likely multiple times even.
It would be easy to add both resistor options to a PCB design to have a fallback.
Removing the resistor would be the fallback to plain Miller worst case.
I once built an amplifier with TPC, and performance was very good, as far as I could measure.
The amplifier topology required the resistor goes to the rail for good PSRR, else I might have considered TMC instead.
However, I believe there were issues with stability in some cases (not sure) and I reduced the aggressiveness of the compensation, i.e. reduced the capacitor ratio, and increased the resistor value. The presumed instability was around the closed loop gain peak frequency BTW, hence it is plausible that there could be a risk. Has anybody else observed any stability issues at the closed loop gain peak frequency with TPC or TMC?
In simulation, there was good margin, but this is small signal only, and reality is always a different story.
Does anybody know whether TMC shows the same amount of closed loop gain peaking like TPC?
While being so similar, they are still slightly different.
This certainly has been discussed before, likely multiple times even.
I can answer that question for you!
TMC:
TPC:
The THD also increases a hair with TPC, but it's only about 10PPM difference.
Then there's the OLG.
TMC
TPC:
I doubt that it makes any difference in terms of real performance, but I'm still a TMC fanboy! 🙂
TMC:
TPC:
The THD also increases a hair with TPC, but it's only about 10PPM difference.
Then there's the OLG.
TMC
TPC:
I doubt that it makes any difference in terms of real performance, but I'm still a TMC fanboy! 🙂
also worth checking out is astx's huge thread https://www.diyaudio.com/community/...ance-class-ab-power-amp-200w8r-400w4r.235194/
The important bits are my earlier posts there when his earlier SA201x versions responded correctly to my SPICE world suggestions in 'real life'. As its performance increased, it got to a stage when this no longer held. But the earlier problems encountered, and their solutions give you an idea of the work needed to get excellent performance from a power amp.
I don't think 'pure Cherry' is any more difficult to get stable as other compensastion schemes (including the usual plain Miller at the VAS) but I'm biased. Done properly, 'pure Cherry' gives a HUGE advantage in HF THD
The important bits are my earlier posts there when his earlier SA201x versions responded correctly to my SPICE world suggestions in 'real life'. As its performance increased, it got to a stage when this no longer held. But the earlier problems encountered, and their solutions give you an idea of the work needed to get excellent performance from a power amp.
I don't think 'pure Cherry' is any more difficult to get stable as other compensastion schemes (including the usual plain Miller at the VAS) but I'm biased. Done properly, 'pure Cherry' gives a HUGE advantage in HF THD
