Bruno Putzeys paper on Negative Feedback

Status
Not open for further replies.
Jump to Latest
#202
scott wurcer said:
You must have missed what started all this, the uA741. The very premises in the abstracts of the time were flawed. If you take any amplifier with very high DC gain and single pole compensated as a general purpose building block making a 20dB gain circuit with it requires feedback. The feedback is what makes it work not the cause of the problem.

Dobkin published some ideas to fix the slew problem in 1969 with op-amps of the day (the ones that brought out the necessary pins) and Roberge had this in his textbook/class-notes already in 1970 when he was my teacher. The audio community just was not listening.
Click to expand...

So, power amps are not fast enough?
 
#205
traderbam said:
There's always one deviant in a crowd...and you are "on-topic"! What were you thinking?

Bruno designs, or was designing(?), switching amps for Hypex. His problem was the switching frequency of a few hundred kHz did not allow enough NFB loop gain at 20kHz using a single pole roll-off. As you know, normally, the open loop bandwidth of a linear amp is well above 1MHz, so some 10x higher. And presumably, the switching amps needs a lot of distortion fixing...although I don't know what the OL distortion is compared to a competent linear amp. Maybe someone will tell us. So to get higher loop gain at 20kHz and below, he stacked a load of poles and zeros in the forward gain path to give a Butterworth filter sort of steep roll-off. This is mathematically similar to the roll-off of nested feedback loops as explained by Cherry in the 80s.

Bruno's paper draws a comparison between stacked (or cascaded) poles and zeros and nested differential feedback loops. He uses linear models to derive the equations of both and shows they are very similar. He then draws the sweeping conclusion that there will be no difference in practice. Err...no. Feedback systems do not work like maths equations when the system contains non-linearities and many other practical characteristics not captured by the simple maths.

It is an interesting comparison, though, and something to ponder. We know that in many practical, linear circuits, adding heaps of NFB can make them sound just awful. I made a Maplin clone once...2SJ48/2SK133 were they? Those were the days...paper clips across the pins. The simplicity of the maplin amp will definitely have helped it to not get totally screwed by increasing the feedback. Do you know how much loop gain you achieved?

I am curious to know what effect a Bruno'esque cascade of semi-integrators (I heard he may have used 5 in his amp) has on the music. A wild phase ride not to mention all the extra circuitry to implement it, although he could have done it digitally in theory; but I don't think his amps use DSP as far as I know.

Who knows? 🙂
Click to expand...

Would you be interested in doing a simulation of my "modification" to this amp?
This is the exact schematic of mine, except I substituted a 47K resistor for the 1K in the feedback circuit http://www.muzines.co.uk/images_mag/articles/emm/EMM_81_06_mosfet_ampli_5_large.jpg
 
#206
scottjoplin said:
Why is it that NFB is an issue in power amps when it hardly ever seems to be one in op amps?
Click to expand...
Who says it isn't an issue in op-amps?
As Jan says you can't really use an op-amp without NFB so it is impractical to listen to an op-amp open loop to compare. I don't use op-amps but I believe some high-end companies still avoid them in the signal path and they have had a poor reputation to overcome in audio. I believe there are very good ones available now.

Making a whole amp in a tiny space on a single wafer of silicon brings opportunities and constraints that discrete circuits don't have. Capacitors, resistors. Isolation between components, protection circuits. And it comes down to the skill and priorities of the designer: if they don't optimize for sound quality (as opposed to, say, THD) then the op-amp probably won't sound great. Same for a discrete design.
 
#208
Ha ha. Yep that's the circuit, beautiful isn't it? TO-3s, I built it in the 80's.
I understand (I think) that small signal circuit design is quite different to power circuits, but I'm not sure how or where the line is drawn. Thanks for baring with me.
I'm glad you'd like to see how I got away with it 😉
I shan't hold my breath...............................................:cheers:
 
Last edited:
#210
scott wurcer said:
You must have missed what started all this, the uA741. The very premises in the abstracts of the time were flawed. If you take any amplifier with very high DC gain and single pole compensated as a general purpose building block making a 20dB gain circuit with it requires feedback. The feedback is what makes it work not the cause of the problem.

Dobkin published some ideas to fix the slew problem in 1969 with op-amps of the day (the ones that brought out the necessary pins) and Roberge had this in his textbook/class-notes already in 1970 when he was my teacher. The audio community just was not listening.
Click to expand...

Lets not forget Bart Locanthi was designing amps with degenerated LTP's and sensible BW limiting filters and compensation in the mid to late 1960's. On the whole TIM/SID thing, I think he was ahead of Otala et al - he just never published the way they did.

As for his 'Locanthi T' OP stage - the work of a genius. People keep on trying to cook up better but his is still the best.

🙂
 
#211
DF96 said:
What engineering criteria would you use for sound quality, having ruled out "THD"?
Click to expand...

Any IC professional that says this op-amp was engineered for sound quality at the expense of traditional measures of performance is making it up or at least crossing their fingers behind their back. Noise, bandwidth, dynamic range, and THD anything else is a lie. Marketeers sneak stuff into product descriptions from time to time they are usually embarrassing. Look at tomchr's chip amp thread for instance.
 
#212
scott wurcer said:
Any IC professional that says this op-amp was engineered for sound quality at the expense of traditional measures of performance is making it up or at least crossing their fingers behind their back. Noise, bandwidth, dynamic range, and THD anything else is a lie. Marketeers sneak stuff into product descriptions from time to time they are usually embarrassing. Look at tomchr's chip amp thread for instance.
Click to expand...
I don't think it is a lie. Look at how poor the correlation is between THD and sound in power amps. I'm not saying little green men are involved; I'm proposing that the measurements used are not good enough. But they are good enough for the marketeers (who may well be little green men 😉 ).
 
#215
traderbam said:
I don't think it is a lie. Look at how poor the correlation is between THD and sound in power amps. I'm not saying little green men are involved; I'm proposing that the measurements used are not good enough. But they are good enough for the marketeers (who may well be little green men 😉 ).
Click to expand...

Without deviation (from the norm), progress is not possible. In order for one to deviate successfully, one has to have at least a passing acquaintance with whatever norm one expects to deviate from. Franzk Zappa.😀
 
#216
traderbam said:
Look at how poor the correlation is between THD and sound in power amps.
Click to expand...
Sound what?

Preference? Maybe yes, some people like a little distortion.

Reproduction quality? No, THD must be below some reasonable level as a necessary (but not sufficient) requirement for sound reproduction to take place. If you doubt this just listen to 30% THD and tell us it is indistinguishable from the original music.

I'm proposing that the measurements used are not good enough.
Click to expand...
To repeat my question: what measurements would you use? And on what basis would you choose them? Note that the traditional boring clinical measurements like THD, bandwidth, noise etc. are based on careful listening tests; they are not a dastardly plot hatched by evil engineers in smoke-filled rooms in order to annoy the true believers.
 
#218
jan.didden said:
If this is correct (and I am not saying it is or is not), then a data sheet stating 'designed for audio' can only be BS.

Jan
Click to expand...

More just an empty statement, "My low noise , low distortion, op-amp is good for audio". I might be interested in seeing other statements, " Our op-amp has only -80dB THD but it's constant with frequency", or "We left out the Ib comp so there would be less noise with your MM cartridge". THAT corp does make IC's with audio as the target (possibly the only) market, notice they don't make a single op-amp.

With some work I managed to find one part INA217 that was compromised for audio, the Ib comp is absent and in some apps it requires inductive snubbers. That one would not have left the building.
 
Last edited:
Status
Not open for further replies.
Top Bottom