John Curl's Blowtorch preamplifier part II

[scott wurcer]

It is true that in order to get a flat total response at the output there may be peaks at intermediate nodes, and some designers forget that. To get delays you need storage, and that means resonances (e.g. HF in OPT).

A problem in op-amp based FDNR filters.

 

[ThorstenL]

Hi Ed,

I have found when I drive anticipating the curves in the road works better than driving by feel.

Well, predicting music is not as easy as predicting the road (lack of signage to start with...).

Feedback should occur in units of time measured by phase. Of course there always is feed forward available.

Well, phase is another way of saying time. So if feedback is not zero or leading for all frequencies, we may conclude that feedback is "post hoc".

Feed forward can only be Ad Hoc if it is in fact Pre Hoc, as in the distortion reduction system used with LP (aka Tracing simulation).

Ciao T

 

[jcx]

digital source in principle allows as much "look ahead" as you want, up to whatever is acceptable delay after pressing start to load the buffer - could calculate "pre-distortion" to compensate for the known distortions of your amp

of course engineering knows how to treat "delay" and phase shift in feedback systems - the "objection" is the Sophomoric use by popular Audiophile press/marketing to bash "negative feedback" as inherently flawed, unsuitable for audio amplification


I just got a look at Quan's paper that John so highly recommends - it would have been OK if published around 1980 - even then the objection would have been that no "contemporary" recommended "audio" op amp was tested - not TL071 nor SE5534 - for a paper presented in 2010 one has to wonder if the SCA http://www.sca.org/ has a audio electronics subgroup

I strongly suspect that his paper's FM distortion measurements would read - "not visible above instrument noise floor" for most of the past ~ 5 years op amps that are recommended for audio

 

[simon7000]

Hi Ed,

Well, phase is another way of saying time. So if feedback is not zero or leading for all frequencies, we may conclude that feedback is "post hoc".

Feed forward can only be Ad Hoc if it is in fact Pre Hoc, as in the distortion reduction system used with LP (aka Tracing simulation).

Ciao T

DC does not exist as I am not immortal.

Therefore all signals only have phase shifts, just with very long periods.

But from a practical perspective when the delay is a small fraction of a cycle the correction spectra is significantly reduced and as the shift approaches zero, the spectra also does.

So we can play with semantics, but the issue should be clear.

 

[scott wurcer]

I just got a look at Quan's paper that John so highly recommends - it would have been OK if published around 1980

There you go Jan, ask Barrie to "clear" the air. The sparks would fly.

 

[scott wurcer]

digital source in principle allows as much "look ahead" as you want, up to whatever is acceptable delay after pressing start to load the buffer - could calculate "pre-distortion" to compensate for the known distortions of your amp

of course engineering knows how to treat "delay" and phase shift in feedback systems - the "objection" is the Sophomoric use by popular Audiophile press/marketing to bash "negative feedback" as inherently flawed, unsuitable for audio amplification


I just got a look at Quan's paper that John so highly recommends - it would have been OK if published around 1980 - even then the objection would have been that no "contemporary" recommended "audio" op amp was tested - not TL071 nor SE5534 - for a paper presented in 2010 one has to wonder if the SCA Society for Creative Anachronism, Inc. has a audio electronics subgroup

I strongly suspect that his paper's FM distortion measurements would read - "not visible above instrument noise floor" for most of the past ~ 5 years op amps that are recommended for audio

The usual suspects have launched the poo-poo hoping it would not hit the fan but they missed.

 

[rsdio]

the "objection" is the Sophomoric use by popular Audiophile press/marketing to bash "negative feedback" as inherently flawed, unsuitable for audio amplification

To borrow and extend Simon's analogy:
An audio amplifier without feedback is driving on a curvy road with its 'eyes' closed.

 

[john curl]

Yes, I, too, was disappointed that Ron Quan neglected to put some measured results with some more COMMON, up-to-date IC's. He has tried a few, at my suggestion, perhaps a year ago. He does not design audio products, and he did his paper as a self-funded project to get down to what is important in audio design, which has been an ongoing hobby. He has a professor at Stanford University to vet his work. What I would be really interested in is the 4458-62 series of IC op amps that seem so universal in mid fi design, as they are cheap.
I would like to remind you Scott, that Dick Sequerra would like to talk to you sometime about PIM. He certainly is of another opinion than you appear to have here.

 

[john curl]

Why don't you find out what changing damping factor does, subjectively, by just adding a series resistor of 4 ohms or so, in series with your loudspeakers? Give it a shot. I have!

 

[Bob Cordell]

I just got a look at Quan's paper that John so highly recommends - it would have been OK if published around 1980 - even then the objection would have been that no "contemporary" recommended "audio" op amp was tested - not TL071 nor SE5534 - for a paper presented in 2010 one has to wonder if the SCA Society for Creative Anachronism, Inc. has a audio electronics subgroup

I strongly suspect that his paper's FM distortion measurements would read - "not visible above instrument noise floor" for most of the past ~ 5 years op amps that are recommended for audio

Hi jcx,

You are absolutely correct. Quan is a nice guy and quite intelligent. I spoke at length with him at the AES convention. But his paper was a diappointment, and adds nothing to the PIM debate, nor does it change the results that have been previously articulated.

The mechanism of feedback-generated PIM is well-understood. Feedback-generated PIM is ALWAYS accompanied by easily-measured intermodulation distortion. It is merely an amplitude-to-phase conversion that occurs via NFB. PIM was just a new name for differential phase and gain, long understood by video engineers. Even amplifiers without feedback have intrinsic PIM, and that PIM is usually actually reduced by NFB. Feedback-generated PIM depends on closed loop bandwidth, NOT open-loop bandwidth.

Cheers,
Bob



Cheers,
Bob

 

[scott wurcer]

I would like to remind you Scott, that Dick Sequerra would like to talk to you sometime about PIM. He certainly is of another opinion than you appear to have here.

That's nice but PIM in MHO is of no interest at all. I can't imagine wasting any time on this issue.

 

[john curl]

You know what is interesting is that Dick Sequerra made a phono stage with the AD797, except that he ran the device(s) OPEN LOOP! Why would he do such a crazy thing? If I were the designer of the AD797, I would be, at least, curious of why he did it that way. He is still happy with his results. I, being a first rate solid state designer, decided to use negative feedback, with my 'initially' AD797 type based phono preamp, the JC-3. For the life of me, I can't figure out exactly what he did to get it to work, but I do understand his motivation. Of course, Dick Sequerra's preamp directly challenges my Vendetta or Constellation designs, not my JC-3. Different class of product.

 

[Salas]

Maybe he run it head amp to passive Riaa and then took it down further with an LPAD and buffered it to output.

 

[john curl]

I'm sure it was something like that, perhaps with the added pin. Ingenious though.

 

[Chris Hornbeck]

The AD797 has its very high impedance junction at the output of the VAS and the input of the unity gain buffer (output stage) exposed. Loading this junction with an RIAA network would do the trick, but without any control over the gain stage's transconductance, gain is set by this (RIAA) load impedance alone. Might be kinda tricky getting a low enough gain and a high enough output voltage (from the given VAS current supplied) for the same choice of load impedance. But then again, that's why he's Dick Sequerra and I'm not.

Thanks,
Chris

 

[Salas]

It wouldn't surprise me if he would exchange many chips until he gets same gain on both channels and then discard some or relegate to other, loop defined duties. Cost no object for sure.

 

[hitsware]

It is my contention that 'stereo' can
be created visually ........
Born This Way - Lady Gaga (Cover by Karmin) - YouTube

 

[jan.didden]

There you go Jan, ask Barrie to "clear" the air. The sparks would fly.

I'm working on it

jan

 

[jan.didden]

The Roddam paper is interesting. However, I suspect that the analysis he presents is only true for a minimum-phase system, as only then would the 'real part' transfer function contain all the necessary information. Incidentally, he is using one of the wonderful things about analytic functions in the complex plane: if you are told something about the function you can probably deduce all the rest.

His talk about delay in feedback is rather misleading, as he assumes an average group delay across the whole audio spectrum when in fact almost all of the phase shift takes place in a narrow HF region. He admits that his treatment is incomplete, but that seems a bit weak having just misled readers. It is true that in order to get a flat total response at the output there may be peaks at intermediate nodes, and some designers forget that. To get delays you need storage, and that means resonances (e.g. HF in OPT).

His mentioning of 20uS of delay in the forward path seems quite high by contemporary standards but may have been a reasonable value at the time; he mentions a phase shift of 180 degrees at 25kHz, errr, 25kcs in the forward path.

Anyway, most designers today are aware that internal nodes may have relatively high error signal values due to phaseshift around the closed loop, and a competent design will have no problem with it whatsoever. In a forthcoming Linear Audio Classic publication ('Baxandall and Self on Audio Power', collected papers) Peter Baxandall pointed this out 2 decades ago.

jan

 

[fireworks]

Why don't you find out what changing damping factor does, subjectively, by just adding a series resistor of 4 ohms or so, in series with your loudspeakers? Give it a shot. I have!

So, what does it do ? The suspense is killing us !