I many times noted that in a dispute in fora, often every word is dissected and used against the opponent, no matter how far out of context (or, as in this case, even quoted) it is.
Unfortunately, it doesn't add to a fruitful discussion...
Here are measurements of an amp which runs the output stage open loop (only the front end is in the SuSy feedback arrangement), and contrary to the Ayre amp, it has MOSFET output devices (quite many AFAIR).
Pass Labs X1000 monoblock power amplifier Measurements | Stereophile.com
Pass Labs X1000 monoblock power amplifier Measurements | Stereophile.com
Positive, negative and local feedback, including error correction, follow the same rules. Differentiation to claim one is better than the other involves a rabbit hole. I know of one very expensive amplifier that does not use GNFB, but does use error correction for the output stages with a derivative of Peter Walker's bridge method. Is this method a handicap and fall into the bad feedback camp? JLH used positive feedback in at least one of his designs and it does not necessarily increase distortion. Positive feedback can be used to linearize a circuit that exhibits current fluctuation, the root cause of distortion in that particular configuration.
Therefore, the action of any kind of feedback includes current and voltage behavior. Blanket claims are best avoided. There is also the issue of false attribution when parameters cannot be separated.
Therefore, the action of any kind of feedback includes current and voltage behavior. Blanket claims are best avoided. There is also the issue of false attribution when parameters cannot be separated.
Good points. I think it is extremely rare to see an amp without ANY sort of feedback. It can be hidden as local emitter degeneration, or the positive feedback involved in cascoding and bootstrapping, or whatever, but I have yet to see one with no feedback at all.
Jan
Jan
I quote from the article (right at the end): "There is nothing to complain about in the Pass X1000's test-bench performance.—Thomas J. Norton".
...while at the same time directly above an IMD plot is shown that is everything but clean (to say it politely). Thanks stereophile! 🙂
Of course, and this has already been beaten to death. I would say applying feedback in the context of this discussion means applying it deliberately (and being able to open the loop too).
Something possible for many building blocks, but not for emitter followers. (We had a discussion about the loop gain=OLG minus CLG in an emitter follower and how it can be determined, remember..?)
Something possible for many building blocks, but not for emitter followers. (We had a discussion about the loop gain=OLG minus CLG in an emitter follower and how it can be determined, remember..?)
No, it wasn't current feedback, it was voltage feedback. In that commercial amplfier for example, the PFB stage was a Schmitt trigger-like circuit.
Sorry, I cannot agree with the 'cold, clinical thing'.
I like the way Stereophile does a subjective review and then a set of measurements and puts them up side by side.
I've seen plenty of amps with feedback get very good reviews - they have measured reviews going back 15 or 20 yrs IIRC.
Well we are not discussing Stereophile opinions here, but I'd like to point out that
should be read in context (again, I know).
19+20kHz @ 455 W "before clipping is evident on a scope trace" is not exactly what I would call normal at all, especially since they mentioned also "the X-1000 cause the line voltage to sag".
I would not judge the cleanliness of an amp on such a measurement. That was in 2004, and in the MX-R test in 2015 they used a more relaxed output level. The X-1000 was already in the region of 0.1% distortion at 455 Watts, while the MX-R was well below that at the 200 W IMD Test power.
"This resulted in a high output impedance of course, but by understanding which speakers were going to work and which ones were not we were able to avoid tonal coloration"
Seems you have to adjust an amp designed this way for the proper tone and sound by changing speakers. Really bad idea as most of us live within a budget.
Seems you have to adjust an amp designed this way for the proper tone and sound by changing speakers. Really bad idea as most of us live within a budget.
One thing that nobody can argue against, is the fact signals fed from a negative feedback potential divider are never in phase with the input, unless the former, is a constant DC voltage. That little nuisance of a phase difference, where it is ideal not having any, can cause serious amplifier misbehaviour like destroying an output stage in a wild burst of destructive self reinforcing amplification. But such bursts, have the pleasure to normally occur at frequencies well beyond the limit of human hearing perception. Luckily, a VAS transistor can be tricked into behaving like a slower one using a shunt capacitor arcross its base and collector. No matter what high frequency tries to drive a tricked VAS into destructive dancing, it is shunted away from producing any noticeable signal, thanks to brilliant engineers who transformed crudely behaving transistors into amazing finely behaving amplifiers.
In short, Negative Feedback, is an engineering control concept, worthy of admiration.
In short, Negative Feedback, is an engineering control concept, worthy of admiration.
It is rather lack of understanding on your part. Reducing the number of gain stages wasn't the purpose of Miller's circuit, if you read the patent. The issue here is PFB improving overall distortion performance.
I think no one is disputing that, isn't it?
What I dispute is the thread title "Global Feedback - A huge benefit for audio".
I still can't see why it should be a huge benefit, if no one is able to hear a difference between deliberately applied H2, H3 at -60dB and -80dB in a controlled test.
Why should I apply feedback to a competently designed circuit, if it does only lead to better measured but not perceived results?
depends if you are after fidelity or a pleasant sound. Sounds like you want the latter, which is fine, but don't project your preferences onto the rest of us!
You must have overseen that I wrote several times, that I'm not aware of someone able to discern between -60dB and -80dB of H2 and H3 in a controlled test listening to music.
Why should that lead to a preference of "pleasant sound" if I cannot distinguish between 0.1 and 0.01% H2 and H3 anyway?
(apart the fact that competently designed open loop circuits are well below -60dB, rather -80dB at reasonable listening levels, and that my LP's and speakers have an issue to keep up with anything less than 0.1% distortion over the frequency range)
If you are able to hear such low distortion levels with music, that is fine to me, although some may ask you to prove that with a controlled test.