Global Feedback - A huge benefit for audio

Jump to Latest
#2,102
bimo said:
It is not true. It depend on all design. Example: https://www.diyaudio.com/forums/solid-state/338815-anistardi-peletuk-4.html#post5835050

This amplifier have high loop gain and have H2 dominant.
Click to expand...

It does depend on topology. But conventional topologies are dependent on enough global feedback to bury the higher harmonics. It's orthodoxy, but it works.

I would like to remind everyone of the old days, when the gain of the stage was largely dependent on the tube. When the tube starts to go bad, local degeneration decreases, overall gain decreases, and it is gradual but clearly audible. You keep turning the balance control until one day you get fed up and take all the tubes to the drug store for testing. Then you bought a couple tubes right there, went home, and fixed your rig. Modern circuits, no matter how humble, always match to a fraction of a decibel.
 
#2,103
I would like to remind everyone of the old days, when the gain of the stage was largely dependent on the tube. When the tube starts to go bad, local degeneration decreases, overall gain decreases

This reminds me of when I worked in a TV repair shop when I was at school more than 40 years ago, a customer came in to complain that his picture would not stop rolling and he could no longer adjust the hold control far enough as it was at its end, he thought that perhaps the chain had slipped

Stuart
 
#2,104
The situation may be even better than indicated. The Fourier transform is blissfully unaware of the circumstance that modern (de riguer) amplifiers do not produce any even order harmonics.

Use the Scientific Method to investigate the validity and accuracy in perception. Prove to yourself that you indeed like what you hear and that what you hear is not just an illusion but a fact.
 
Last edited:
#2,106
Without Fourier's contribution to harmonic analysis, Class D amplification and Inverters would have been much harder to improve to their current quality. For instance, Class D amplification relies heavily on the result, that the fourier series for a particular signal, contains very small coefficients for the first twenty harmonics. Afterterwards, the coefficients start increasing, but the series still converges to a finite value. Such a result, allows the use of a low pass filter to blow the unwanted and more determinant harmonics greatly improving performance. The same analysis applies to high efficient pulse width modulated inverters: these are possible using the same principle as used for Class D amplification.

If some wealthy clients prefer an amplifier built around other principles besides Fourier Analysis and Negative Feedback, there is nothing wrong in that. Some manufacturer, large and small, will provide for that market share to satisfy their needs.

Market shares are built around satisfied clients' needs and desires. As long as an amplifier satisfies its buyer, the supply-demand chain continues to exist.
 
Last edited:
#2,107
jan.didden said:
You should go one step further, starting with G=K+d where K is a constant gain and d is the amp distortion without feedback.
Then run your calculation and show that the distortion d is reduced to d/f.
Jan
Click to expand...
This is not an accurate model unless the amplifier is linear. In a linear system, which isn't what we are talking about, you can decompose the gain into a sum of components. Such as gain + noise. In a non-linear system the summation assumption inconveniently breaks down and the math literally doesn't add up.
 
#2,108
billshurv said:
And how would you suggest we do that? One person glomming components and claiming wife in kitchen improvements is not scientific method.
Click to expand...
I'm imagining you in the kitchen with your wife when she tastes your stew and tells you it lacks salt. You tell her she is not being scientific and you arrange a 20 person focus group to taste your stew and ascertain "scientifically" whether there is a reliable lack of salt or not.

I don't think so. I don't think you would consider that necessary.
 
#2,110
Fast Eddie D said:
Second, I'd like to point out that you're both right. Global feedback does change the harmonic profile and it does skew it towards higher order harmonics. But the design theory (which I realize isn't always achieved) is that if there's enough loop gain in the passband, and the higher order harmonics are below the threshold of audibility (even mediocre modern amplifiers achieve -80 dB which is pretty good), then the distortion is negligible. I've seen specs that claimed the higher order harmonics were -110 to -120 dB which is almost certainly negligible.
Click to expand...

Indeed, but this little side-discussion was specifically about amplifiers with very small amounts of negative feedback. In some special cases that can do more harm than good. Basically when you have a quadratic term in the open-loop gain, you get a square-root-like error signal when the loop is closed.
 
#2,114
traderbam said:
This is not an accurate model unless the amplifier is linear. In a linear system, which isn't what we are talking about, you can decompose the gain into a sum of components. Such as gain + noise. In a non-linear system the summation assumption inconveniently breaks down and the math literally doesn't add up.
Click to expand...

This is factually incorrect; for low levels of nonlinearities (as in every competently designed audio amplifer before closing the loop), the problem can be handled as quasi-linear. Whatever mathematical method you may want to use, from expanding in Taylor series to the perturbation theory, the result is the same: the gain can be treated as a sum of gains, plus a small additive nonlinear term, which can be analyzed separately and the result added to the overall response.

Start with 20% open loop distortions and this may indeed not work. Not what I call a "competent designed audio amplifier", in fact it may take some heroic efforts to make it that bad, at least for solid state.
 
#2,119
It must be difficult for the DIY hobbyist to understand the reality of amplifier design when there are so many conflicting statements from those who supposedly know what they are talking about.
Click to expand...

Normally I don't agree with much of your posts, but this one certainly is 100% correct and points up that in any area where the experts disagree there are likely to be religious wars rather than logical discussion.

This has been disagreeably gladiatorial over the years, and all I've learned from this has been SJ's observation; and further, that when there is vehement argument it might just be possible that the answer is somewhere in between.

Particularly enjoyed Scott Wurcer's wonderful cartoon. Reminds me of 'Glutton' from John Cleese.
 
Last edited:
#2,120
Bonsai said:
All this anti feedback nonsense was given wings in an article written by Martin Colums in c. 1996 in Stereophile (‘A Future Without Feedback’) which regrettably presented a lot of false information. It’s been regurgitated ad infinitum for the last 25 yrs.
Click to expand...

Plenty of designers that prefer zero FB designs wouldn't know who M Colums
is. I was talking to one yesterday: Day gig - services complex high power, PA
gear such as 20kW Lab Gruppen amps, D & B, Adamson etc etc. One of the
best in the country. Preferred replay system -> home built, ZFB valve gear.

Contrary to what most believe, many (not all) designers that use
predominantly ZFB topologies are not technically incompetent fools, they just
prefer the sonic outcome.

TCD
 
You must log in or register to reply here.
Top Bottom