Hi Pavel,
The SuperTIS has been built by Philipp Gutknecht, see this post. Also the PMP and MCP amps has been built by a few other people (with success).
Cheers, E.
I don't know what circuits you refer to, but more can be found here: http://home.tiscali.nl/audio/
\\\Jens
Hello Edmond
Well, I think I'd call that attempted-OIC, because I've yet to see an example of it that was stable in the real world.
Whereas I've always called that OIC that works. I do appreciate the transition introduced by the network (there's plenty about it in APAD6) but since ordinary Miller compensation already has a transition in that global NFB is transitioned to purely local NFB, I don't find TMC a helpful description.
Which brings me back to my question- what implications (if any) does the 2-pole issue have for practical design?
Cheers, E.[/QUOTE]
look at the loop gain with the probe cutting both the global and T resistor feedback loops around the output stage and you see the total loop gain around the output stage looks every bit as much "2-pole" "inside" TMC as with "conventional" 2-pole
you then design the 2nd order slope loop gain breakpoint a sufficient distance below the unity gain intercept to not degrade the phase margin too much - exactly like compensating the ordinary 2-pole when all the gain is visible in the open loop, by just cutting the global feedback loop
you then design the 2nd order slope loop gain breakpoint a sufficient distance below the unity gain intercept to not degrade the phase margin too much - exactly like compensating the ordinary 2-pole when all the gain is visible in the open loop, by just cutting the global feedback loop
Attachments
Last edited:
Hello Edmond
[..]
Hello Douglas,
Following the same resaoning, calling it OIC is even less helpful, as that acronym is alraedy taken for something else. Anyhow helpful or not, TMC is the de facto and established name for many year. Substituting it for something else creates nothing but confusion.
Well, jcx have already explained that (i.e. stability issues). Hope that suffice. Anyhow, the saying: "There ain't no such thing as a free lunch" applies also to this case.
Cheers, E.
Well, jcx have already explained that (i.e. stability issues). Hope that suffice. Anyhow, the saying: "There ain't no such thing as a free lunch" applies also to this case.
Nice to see you changed your opinion. It took over 3 years, but better late than never.
Hi Franzm,
Have you seen this model? >BIGBT with MOSFET drivers<
MOSFET drivers make a lot of sense in Bryston-alike topology in my opinion.
I can tell you why, but let's better move to the right thread then (the one where the link points)
Cheers,
Valery
I hope you enjoy and find it useful.
Yes, the caption of Fig 8.11 is a known error. Somehow the caption for Fig 8.3 got duplicated there. For the description of Fig 8.11, see the bottom of p210.
Have mercy, Mr. Self.
Of Edmond's "invention" of several years ago, all that it's left today as original work is the TMC acronym. Give him that much credit.
OIC vs TMC vs BMC
C'mon Douglas. OIS stands for Output Inclusive Compensation. Maybe you have forgotten what it means, so here is a reminder:
"Audio Power Amplifier Design Handbook Fifth Edition", page 217: Including the Output Stage: Output-Inclusive Miller Compensation, and the schematic is on page 216, fig. 8.1-C. (btw, I've also the 6th edition)
I'd rather TLA's. So what about Baxandall Miller Compensation (BMC)?
Cheers, E.
Last edited:
Douglas,
Having skimmed through the new VAS section of the 6th edition a reoccurring question comes up with respect to DC BIAS point for all of your designs. It is regardless whether the analysis is about a single ended or balanced topology.
It seems that the DC current through the VAS is not easily found unless simulated.
It also seems that the designs have the same DC and AC gain (From input to output of VAS) – is there a specific reason for this choise?
Have you examined (but not documented) the input -> VAS implementation found below?
Best regards
\\\Jens
Having skimmed through the new VAS section of the 6th edition a reoccurring question comes up with respect to DC BIAS point for all of your designs. It is regardless whether the analysis is about a single ended or balanced topology.
It seems that the DC current through the VAS is not easily found unless simulated.
It also seems that the designs have the same DC and AC gain (From input to output of VAS) – is there a specific reason for this choise?
Have you examined (but not documented) the input -> VAS implementation found below?
Best regards
\\\Jens
Attachments
Last edited:
No problem there then. Simulators are free.
It keeps thing simple.
No. What advantages could it offer??
Are not those two peaks the same thing litteraly.?.
I mean if on OL you have a pole at the origin and another one at 1-10khz, shouldnt, once the loop is closed, this second pole be rejected at high frequencies by the virtue of the miller loop..?.
This second pole could eventualy produce a peak at high frequencies if the output stage is slow enough such that the return ratio is smaller at this frequencies.
With TMC the peak doesnt appear as at the relevant frequency the miller loop is already closed around the VAS while with TPC the global feedback loop frequency response is wider, that is, there s more GNFB at said frequencies in this latter case, the feedback (global + local) around the output stage being the same in both cases.
I don't believe so
When you say "Open Loop" you need to be bit careful about which loop is open and once any loops are closed you need to be careful about what gain you mean.
We can start with no compensation and the outer loop open.
When you close the Miller loop the poles so-called "split".
When you close the TPC compensation loop there is not a pole at the origin and one at 1-10 kHz, there are two poles at 1-10 kHz - and typically a peak. This is already equivalent to the "split" pole miller case.
Once you close the outer loop then any peak in the closed loop response is determined by the behaviour at and above the Global ULGF. This is unrelated to the 1-10 kHz behaviour.
The TMC case is similar but there is not the extra shunt load on the VAS/TIS that TPC imposes. That shunt load is (mostly) bootstrapped away by the OPS. I suspect that matters more than some people realize.
Best wishes
David
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.