This may be true, but when comparing TPC and "TMC", it makes sense to use exactly the same component values.
One then finds that the stabilty margins within the inner loop of a "TMC" compensated amplifier are precisely the same as those of the major loop in the TPC compensated amplifier.
Hi Waly,
All that has been said, as far as I know, is that TMC provides a first-order open-loop gain as seen by (only) the global feedback loop. I don't think anyone has said it is is overall a first-order system.
There is no free lunch, but some are tastier than others, at least to some
Cheers,
Bob
If Middlebrook probe was set as in this simulation then TMC loop gain shows similar behaviorr as TPC.
Damir
Attachments
I disagree about using exactly the same component values to compare TMC and TPC.
Adjust the values so that you get the same ULGF and them see what the gain and phase margins are. That's a better test of equivancy.
Th fact that TPC achieves lower distortion by making more global feedback available at HF vs TMC which applies it where the biggest issue is (the output stage) should tell us they are fundamentally different.
If you are going to use TPC, it's probably a better fit with class A (no cross over distortion) while TMC is a better fit in my view for class AB.
I have TMC'd two power amps and had no stability issues. That said, I like at least >60 degrees of PM and 12 dB gain margin and this may be a contributing factor.
Adjust the values so that you get the same ULGF and them see what the gain and phase margins are. That's a better test of equivancy.
Th fact that TPC achieves lower distortion by making more global feedback available at HF vs TMC which applies it where the biggest issue is (the output stage) should tell us they are fundamentally different.
If you are going to use TPC, it's probably a better fit with class A (no cross over distortion) while TMC is a better fit in my view for class AB.
I have TMC'd two power amps and had no stability issues. That said, I like at least >60 degrees of PM and 12 dB gain margin and this may be a contributing factor.
Last edited:
The notion that you get more loop gain about the output stage with "TMC" than with TPC is an outright myth.
This can be proved, as I have pointed out, by using exactly the same component values calculated for TPC for "TMC". You should then find that the total loop gain about the TIS and the output stage with both TPC and "TMC" is the same at all frequencies, and that the unity loop gain frequencies and stability margins are also the same.
This can be proved, as I have pointed out, by using exactly the same component values calculated for TPC for "TMC". You should then find that the total loop gain about the TIS and the output stage with both TPC and "TMC" is the same at all frequencies, and that the unity loop gain frequencies and stability margins are also the same.
Last edited:
It may not be appropriate to select TMC or TPC in isolation of the amplifier topology. Other schemes may even be necessary depending on amplifier topology, as such it may be unnecessary to stick to one technique. Simply use what works best for your topology.
But its good to appreciate that in any amplifier thread these days the debate for TMC or TPC sneaks in further educating us
. Keep up the good work.
kind regards,
Harrison.
But its good to appreciate that in any amplifier thread these days the debate for TMC or TPC sneaks in further educating us
. Keep up the good work.kind regards,
Harrison.
After all I got Bob's Amplifier design book ( used) and managed to read a few chapters.
Imho a new edition needs to cover additionally 2 subjects
amp speaker and acoustical environment as one system ( i.e. active speakers)
and
forward error correction of class D or switching amps
these can eliminate two major issues of switching amps
namely digtialisation errors of small signals ( noise becomes distortion)
and timing errors ( also become distortion)
Bob explains how the diffusion driven carrier motion in BJTs can be mimicked with a large virtual capacitance, that leaves the impression power BJTs can only be voltage driven.
That is not correct in general.
So there should be a discussion of an optimum generator resistance of drivers.
And eventually .... the oldies but goodies....QUAD's legendary triple cascade the only topology that
reads junction temp on the fly
I finally managed to get such a topology working with faster RET BJTs but I do not yet understand why that had been so tedious and cumbersome to accomplish ( oscillation and such) it appears it is very sensitive to reactive loads
Imho a new edition needs to cover additionally 2 subjects
amp speaker and acoustical environment as one system ( i.e. active speakers)
and
forward error correction of class D or switching amps
these can eliminate two major issues of switching amps
namely digtialisation errors of small signals ( noise becomes distortion)
and timing errors ( also become distortion)
Bob explains how the diffusion driven carrier motion in BJTs can be mimicked with a large virtual capacitance, that leaves the impression power BJTs can only be voltage driven.
That is not correct in general.
So there should be a discussion of an optimum generator resistance of drivers.
And eventually .... the oldies but goodies....QUAD's legendary triple cascade the only topology that
reads junction temp on the fly
I finally managed to get such a topology working with faster RET BJTs but I do not yet understand why that had been so tedious and cumbersome to accomplish ( oscillation and such) it appears it is very sensitive to reactive loads
Last edited:
Agreed.
IMHO, it is possible to design (as opposed to fudge a single example of) an amp to be unconditional stable under all loads .. IF you use an output inductor.
I don't know how to do this without one. I know amps which are unconditionally stable under all loads without one ... but can't explain how they achieve this. Unfortunately, they are not 1pp zillion THD.
I don't think you will ever get JC to admit to this heresy
And even if he did it in the dim & distant past, he will disclaim all responsibility for such designs. _________________________
Could you post example circuit or link to one please ... especially of the cascodes.Bonsai said:
_________________________
For QUAD type triples, the 'cure' is to reduce the transconductance of the first BJT with an emitter resistor. (Mike Albinson of QUAD later used an inductor instead, for the 'triple' in the 405).hahfran said:
This is similar to Prof. Cherry's emitter resistor in the VAS when using 'pure Cherry' compensation. In both cases, you don't lose OL gain if you are driving from Hi Z.
Dunno about "reads junction temp on the fly."
Last edited:
Dunno about "reads junction temp on the fly."
that is because of the feedback the triple cascade appears as a supertransistor
with the property theta junction equals that of the first BJT
Doesn't matter that much who wants comments on what new book on audio amps I am convinced the days of linear amps are numbered and so are the old stuff of amp passive crossovers n-way speakers.
Seems obvious.
Audio lies on electronic *industrial* technology. Electronic technology is more and more digitally based, and the analog part of it evaporates years after years like water under the sun.
With diminishing resources, it is more economical to replace pounds of copper by millions of transistors, the power of calculation of CPUs bring a lot of possibilities we we could not imagine with analog.
As an example, today, i find more simple to set a digital filter delay than to work on physical offset distance between my loudspeakers in enclosures.
There will be nostalgics building linear amps for a while, like tubes ;-), but it will be as anecdotal.
The active devices are faster and faster, so the digital evils are pushed farther and farther, while digital provide reproducibility and absence of those hidden perverse effects of analog parts, like sound of components, subjective differences we do not know how to measure, etc.
This will make some audiophiles sad, like they were when they discovered on CDs that the real sound of the recordings they loved in their vinyls were not so beautiful than they imagined
Last edited:
No. Linear amplifiers will be around long after you've become soil fertilizer.
Especially if you realise that class-D amps are both linear AND analog
jan
Mike,
This is one area where we strongly disagree. One should not constrain the comparison to using the same compensation component values for the two techniques. Let the best technique win in any given amplifier design, using whatever combinations of component values that give the best performance for that technique. The constraint should be that the two implementations yield the same degree of stability.
Cheers,
Bob
Indeed. Transitional technology.
Next-step: direct power DACs. Radio inter-connexions. One computer, wifi, enclosures, what else ?
Last edited:
Then don't use the same values and let the best technique for a given amplifier win. You are way over-constraining the problem.
It is outright foolish to assume that two techniques that are fundamentally different (I think we all agree to that now) should want the same component values for best performance.
Cheers,
Bob
Just as the finest LP equipment and tubes had been developed - at almost the peak of the technology, the limitations are known and replaced. Then as that replacement is nearly perfected, its limitations are known and replaced. As we now move more and more away from the limitations of all analog, to be replaced by almost all digital only equipment, (except microphone and loudspeaker and ears) I wonder what its limitations will be and what will replace it.
-RNMarsh
-RNMarsh
Last edited:
Hi hahfran,
These are good suggestions, and I will add them to my list for the second edition. Keep the suggestions coming.
Semiconductor physics is certainly one area where I could stand to learn more and at the same time be able to articulate the issues understandably to non-physics readers without giving wrong or grossly over-simplified impressions. This is always an area where we must make some difficult decisions about level and type of abstraction. For example, the hybrid pi model is an example of such a compromise.
I have long been an advocate of active loudspeakers and have actually given some thought to including some of that subject in the next edition. See my website (CordellAudio.com - Home) for the description of the Athena active speaker system and also the discussion of the EQSS technique for improved bass in a given cabinet size in an active loudspeaker context. The Athena has four 125-watt vertical MOSFET power amplifiers built into each cabinet.
Cheers,
Bob