Hi everyone,
I was using Miller dominant pole compensation and this kind of compensation, seems to produce high values of THD20 on the order of 0.0x %, so I decided to use something like TPC (two pole compensation) or TMC (Transient Miller Compensation), but I don't know much about this compensation methods, I don't know what to do to calculate the capacitors and the resistor. Using TPC I have overshoot on closed loop step response, using TMC I don't have, but it was told me that the amplifier could have other kinds of instabillities. I've made this designs using TMC, but I don't know if they are stable at high frequencies.
Can anyone with experience in this areas help me here.
Attached some attempts, possibly not good, using TMC comp.
PS: Some comments are from other amplifiers, output power for 4 ohm is on the order of 150, at much lower THD1.
Thank you all for your support,
Best regards,
Daniel Almeida
I was using Miller dominant pole compensation and this kind of compensation, seems to produce high values of THD20 on the order of 0.0x %, so I decided to use something like TPC (two pole compensation) or TMC (Transient Miller Compensation), but I don't know much about this compensation methods, I don't know what to do to calculate the capacitors and the resistor. Using TPC I have overshoot on closed loop step response, using TMC I don't have, but it was told me that the amplifier could have other kinds of instabillities. I've made this designs using TMC, but I don't know if they are stable at high frequencies.
Can anyone with experience in this areas help me here.
Attached some attempts, possibly not good, using TMC comp.
PS: Some comments are from other amplifiers, output power for 4 ohm is on the order of 150, at much lower THD1.
Thank you all for your support,
Best regards,
Daniel Almeida
Last edited:
slog through hundreds (thousands?) of post in the threads forum seach throws up
A community dedicated to helping everyone learn the art of audio. Projects by fanatics, for fanatics! - Search Results for tmc
whew - only 664 results
or wait for Cordell, Self next book editions - both have promised to expand their treatment of compensation, TMC
A community dedicated to helping everyone learn the art of audio. Projects by fanatics, for fanatics! - Search Results for tmc
whew - only 664 results
or wait for Cordell, Self next book editions - both have promised to expand their treatment of compensation, TMC
Ok, thank you jcx.
But about the design I wanted to know if I did something wrong and what I can do to improve the design performance.
But about the design I wanted to know if I did something wrong and what I can do to improve the design performance.
just pointing out its a little complicated and there isn't even a clear agreement/exposition of the tradeoffs
Right, about Bob Cordell and Douglas Self new books, I'm interested in them, do you know if there's some release date already defined?
Thank you for your great support.
😉
I wanted to know why you put a small capacitor in parallel with the fb resistor?
How do you do to calculate the values in TPC? I can't understand the procedure in Bob Cordell's book.
🙁
Thank you for your great support.
😉
I wanted to know why you put a small capacitor in parallel with the fb resistor?
How do you do to calculate the values in TPC? I can't understand the procedure in Bob Cordell's book.
🙁
Last edited:
@ danny92
I'm not an expert on TMC etc, but i can answer this.
The cap on the FB resistor attenuates frequencies higher than the RC breakpoint @ 6dB an octave. Often it's used to limit the top end to say 20kHz or so. This prevents RFI & other noise being amplified. But the lower the bandwith of the filter = lower Slew Rate. For that reason, it's been referred to as, the Slow Rate 😀
I'm not an expert on TMC etc, but i can answer this.
The cap on the FB resistor attenuates frequencies higher than the RC breakpoint @ 6dB an octave. Often it's used to limit the top end to say 20kHz or so. This prevents RFI & other noise being amplified. But the lower the bandwith of the filter = lower Slew Rate. For that reason, it's been referred to as, the Slow Rate 😀
20 kHz?? - you wouldn't want the feedback "lead" cap anywhere near that big/closed loop gain frequency roll-off so low in power amps - we are compensating for noticeable closed loop gain - these are not unity gain stable op amps here
the cap is small with higher closed loop gain - one part of it is to compensate the feedback network for the input C of the inverting amp input - much like scope probe compensation C trimming
the cap is often adjusted empirically watching the step response overshoot - trimming for best rise time with little or no overshoot - but that is only a reasonably valid approximate goal for dominant pole Miller compensation stability wise
it may be made a little larger to improve loop gain intercept phase margin - but at the same time it adds "phase lead" it also extends the feedback loop gain intercept frequency upward where the output device phase may change too rapidly or uncertainly with load to be safe
the cap is small with higher closed loop gain - one part of it is to compensate the feedback network for the input C of the inverting amp input - much like scope probe compensation C trimming
the cap is often adjusted empirically watching the step response overshoot - trimming for best rise time with little or no overshoot - but that is only a reasonably valid approximate goal for dominant pole Miller compensation stability wise
it may be made a little larger to improve loop gain intercept phase margin - but at the same time it adds "phase lead" it also extends the feedback loop gain intercept frequency upward where the output device phase may change too rapidly or uncertainly with load to be safe
Last edited:
High frequency oscillations, cause any noticeable notifications, when you listen to the sound produced by the amplifier?
Why it is so important to avoid oscillations, because it can cause overheating, excessive dissipation, or because of negative effects on linearity and increased distortion?
That is not well explained in audio amp books.
For the TMC topology in this post I've followed a rule of thumb proposed in one thread about this comp topology, and consists in calculating the CMiller for dominant pole and set the serie caps for TMC with the same value, for example I was using 33pF for Cm now I use 68pF for the TMC capacitors because Ceq = (68pF*68pF)/(68pF+68pF) = 34 pF, for the resistor I used 1 kohm and seems an acceptable value, the values I've seen are between 1 kohm and about 10 kohm.
Why it is so important to avoid oscillations, because it can cause overheating, excessive dissipation, or because of negative effects on linearity and increased distortion?
That is not well explained in audio amp books.
For the TMC topology in this post I've followed a rule of thumb proposed in one thread about this comp topology, and consists in calculating the CMiller for dominant pole and set the serie caps for TMC with the same value, for example I was using 33pF for Cm now I use 68pF for the TMC capacitors because Ceq = (68pF*68pF)/(68pF+68pF) = 34 pF, for the resistor I used 1 kohm and seems an acceptable value, the values I've seen are between 1 kohm and about 10 kohm.
Last edited:
all of the above?
there is a fun paradox with output Zobel - the parts should be designed to withstand the power dissipation if the amp bursts into oscillation - but the Zobel is included to prevent oscillation - so if it is successful you could use lower rated parts...
there is a fun paradox with output Zobel - the parts should be designed to withstand the power dissipation if the amp bursts into oscillation - but the Zobel is included to prevent oscillation - so if it is successful you could use lower rated parts...
The Zobel should withstand high power dissipation, and you should use an output inductor to prevent high frequency oscillations from "entering" in the speakers. The inductor can have a problem because it acts as a low pass filter and could "kill" the high frequency response of the amp. Zero-pole/lead compensation isn't used in audio amps? Why?
PS: I think that the problem is the Slew rate as happens with dominant pole comp.
PS: I think that the problem is the Slew rate as happens with dominant pole comp.
Last edited:
@ jcx
I only used 20kHz as an example so he hopefully got the general idea. I know 20kHz is not HiFi 😀 but some "supposedly" Pro gear rolls off not far above this !
I only used 20kHz as an example so he hopefully got the general idea. I know 20kHz is not HiFi 😀 but some "supposedly" Pro gear rolls off not far above this !
Hi everyone,
Does anyone, with some experience in amplifiers using TMC, could please, see if the amplifiers that I've posted in the first post of this thread are a good approach?
They have some problems, instabilities, etc... ?
In simulations the results are good, including square wave test at 20 kHz, gain/phase margins and there're no peakings in the responses of the amplifier.
Thank you very much for your attention, and for the support,
Best regards,
Daniel Almeida
Does anyone, with some experience in amplifiers using TMC, could please, see if the amplifiers that I've posted in the first post of this thread are a good approach?
They have some problems, instabilities, etc... ?
In simulations the results are good, including square wave test at 20 kHz, gain/phase margins and there're no peakings in the responses of the amplifier.
Thank you very much for your attention, and for the support,
Best regards,
Daniel Almeida
Last edited:
you haven't understood what I've said so far if you think
thats why I don't like seeing TMC promoted to those who can't design a "conventional 2-pole" compensation to start
is any sort of indicator with TMC vs 2-pole
thats why I don't like seeing TMC promoted to those who can't design a "conventional 2-pole" compensation to start
Last edited:
But in Bob Cordell's book, in the explanation on TMC, he says that TMC have the advantage of not producing the frequency response peaking and overshoot that necessarily accompanies TPC. The only way to reduce peakings and overshoot in TPC is using a high frequency bypass capacitor in parallel with the fb resistor. And I don't understand why you are only criticizing and why you don't explain me what to do to design a TPC, in the books TPC is not well explained, TMC is derived from one type of compensation called inclusive Miller compensation, that was proposed by Peter Baxandall, and the capacitors are calculated having in account that the two compensation capacitors, works as a series Miller capacitors at higher frequencies where the resistor has no influence, so I used this principle to calculate the values of the components, in the designs I've made.
- Status
- Not open for further replies.