|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
72.219.92.97
In Reply to: RE: Simple answer! posted by E-Stat on June 05, 2022 at 08:31:48
the voltage-in to current-out transfer curves all deviate from a straight line in some way or another. But, if you can get the voltage gain you want with only small current pertubations then distortion can be low. This works for a preamp but not for a power amp!
Follow Ups:
FETs & tubes have square-law transconductance, BJTs are exponential - none are linear. In general, if you arrange things to allow the signal current excursion to be along only a small part of the transconductance curve that might be linear enough (i.e. very strong class A biasing). Or, add some degeneration (local feedback) - that will straighten things out.
Whenever audio designers say 'no feedback' they mean no global feedback around multiple gain stages. Any designer who eschews local degeneration for audiophile sensibilities is not really a designer! Except in the case of tubes - if you degenerated the circuit you would not be able to hear the differences in non-linearity between tube types so clearly. That is a win-win, poor circuit design* gets passed-off as voicing/tuneability and everyone is happy:)
*I am joking about poor design. If it is what customers want then it is the right thing to do.
Edits: 06/06/22
According to Boyk and Sussmann, the triode actually follows a 3/2 law and not a quadratic law like FETs. This makes them somewhat more linear.
I'm not technical enough to understand the mathematical nuances but I understand the gist of it: if a little feedback can be bad, a lot of it is great. Yes, I know, your device must have enough "open loop gain" to sustain a lot of feedback.Follow the link below to find it ...
Dmitri Shostakovich
Edits: 06/07/22
That is a good article but a bit esoteric. It is not applicable to our discussion here about the inherent linearity/non-linearity of active devices but it is about global feedback around a whole amplifier and how to add more. To keep stability, usually open-loop gain is rolled-off from a very low frequency so there is a lot of 'excess gain' at low frequencies to correct distortion but at higher frequencies less correction can be made. Mr Putzey presents a method to increase the gain at higher frequencies and still keep the closed loop system stable.
I believe the idea goes back to Bob Cordell for audio amplifiers but, I suspect, the concept itself is a lot older in the general field of system design and servo control.
Unfortunately my lack of technical and math savvy means most of the article is mumbo-jumbo to me.
Putzeys's final conclusions, viz. that is essentially that there is no such thing as too much feedback, and that includes multi-stage feedback loops as well as local strike me as valid based my listening to my Purifi amp. ("Degenerative" is term I hadn't hear before reading this thread).
E-Stat got on about the supposed inherent linearity of amplifier components which you, Duke, addressed. However E-Stat's fundamental agenda is anti-feedback in general; this is misguided, IMHO, based on Putzeys and other commentators of the last couple of decades. E-Stat's archaic views on feedback are the basis for his categorical hatred of typical, integrated op amps which is, according, also misguided.
Dmitri Shostakovich
My categorical hatred is based on hearing literally dozens of high feedback amps.
I have heard a number of Putzey's designs where of course he is putting his theories into practice. The sound? Not musical.
E-Stat is not misguided, no-feedback designs often (but not always) sound closer to real music. High feedback designs have yet to produce a single example I could point to where I say "yes, it can be done this way"
Listening to audiophile preference for over 20 years I understand there are many who prefer the "musicality" of tube amps (and some S/S amps). Descriptions often include sound that's full, smoother, and (sometimes) having layered or "holographic" imaging. I have heard these qualities myself from certain equipment and understand what those folks are talking about. I agree that people are entitled to like these attributes.
Your explanation is that negative feedback robs the sound of these qualities because it causes higher order harmonic distortion.
My hypotheses is what what mainly contributes to these attributes isn't exclusively the absence of high order harmonics but, rather, the presence of 2nd and/or 3rd order harmonics.
I believe that lots of feedback reduces all harmonics and that includes high level harmonics. But it also greatly decreases 2nd & 3rd order harmonics and I believe that is why many do not like ultra-low distortion amps such as Purifi or Benchmark.
Dmitri Shostakovich
Lots of feedback creates more and more harmonics, even if very low in level, that acts like an artificial noise floor, which is signal modulated...because it's not really noise.
Reminds me of what we jokingly called "accu jack" back in the engineering department in my early career. It was a pneumatic analog servo controlled throttle actuator for a car gas pedal. If the PID control wasn't set just right accu jack's shaft would cycle in and out rapidly and uncontrollably.
Edits: 06/08/22
the reproduction of music is what matters most. :)
Enjoyment/engagement are what I prize most and I'm not too worried if it means my system is coloring/supplementing the source. I suspect many feel that what they prefer must be objectively better too and that might not be the case. Audio is broad church that allows many philosophies and it is not like we're talking about aircraft landing radar where it either works or people die! So, some things in audio that don't work so well still have a place and our world would be duller without them.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: