I certainly believe this is a true statement, or at least, much closer to the heart of the matter than feedback (or absence of it).
Poor decisions, in design, or in manufacturing, or component selection are usually to blame, IMO.
would be curious about difference between class d made from:
1 dyi dc-dc switching ic with audio signal injected to fb
2 dedicated class d chip
imho, quite close soundwise😛
As I don't rate Pass as a designer, I have not spent much time analysing his designs. I would describe his designs as generally naive and eccentric, over-simplistic. He has a following, but I suspect this is mainly among people who would rather follow a guru than learn some circuit theory.ginetto61 said:
There is nothing wrong with high feedback. Wrongly applied feedback, or a poor open loop design, can damage sound. As I said, the right amount of feedback is the right amount of feedback.ginetto61 said:
Right. It is not a mistery that those designs measure worse than other commercial product
Better, there is no measurement that gives evidence of their superiority
so a little and humble op-amp can be much better than an extremely exotic and expensive tube stage because it measures better ... actually a lot better
Why not everyone seem to agree with this reasonable and rational approach ?
I read once ... it is difficult to fight voodoo with science 🙄
Thanks and regards,
gino
This is "The Question" Gino
I have my own answers, one of which has been strongly rejected in the forum. 😀
It can be shown mathematically, at least in devices that follow the 1/2 power law, that excessive negative feedback, favors the appearance of higher order distortion.
In the 80's, manufacturers began "The race of measurements" competition would be won by the amplifier that exhibited more zeros after the decimal point, was then the resurgence of valve amplifiers.
Recently completed the construction of a preamp and two class A monoblocks for a client, all valve, of course.
The simulations and the few measurements I could do, they were not at all impressive.
But the sound is sublime, such transparency as if they were not there, fortunately, both the client and I do not want a "sound"
I do not possess the knowledge to perform minimalist designs, on the contrary, I tend to complicate everything.
Maybe we are not measuring what we should measure...
When I knew all the answers, they changed the questions ! 😀
It was strongly rejected because it's not particularly true (I think you meant square law, not square root law). See the excellent discussions by Bruno Putzys in Linear Audio and Morgan Jones in Valve Amplifiers. While you're at it, it's worth reading the full, in-depth discussion by Baxandall, who (AFAIK) originated this idea but was careful to point out the limitations.
Hi ! very interesting indeed
But now the next question is .. how much negative feedback is excessive ?
And then ... is it not possible to build single ended zero feedback amps also flexible enough to be used not only with horns or similia ?
I think so ... I am not able but I think so
They should not have hundreds of watts clearly ... just 20-30 nice watts
and then with the right speakers selection the outcome could be very satisfying
I understand but I would not limited the discussion to tubes
Even bjts have some potential in these kind of designs, and are much cheaper, common and robust
There is one think that I do not like ... mosfet as output devices
I understand that they have some advantages but I have often hear them sounding vaguely soft
I like the right punch in the bass ... a thing that bjts are usually good at.
Maybe an hybrid with the voltage gain stages with tubes and the output stage with bjts is the best solution 🙄
Actually I have listened to some hybrid amps and they have always impressed me a lot indeed. I like them really a lot
They have always sounded from good to extremely good indeed
when this happen it means to me that the approach is a right one
You get the space the tone and the power ... everything
Thanks and regards,
gino 😀
Last edited:
Sorry it´s a typo, I mean square law, thanks SY !
But, what is a power of 4 between friends ? 😀
Here an example
http://digilander.libero.it/paeng/cathode_follower_follows__audio.htm
The strongly rejected is another...😀
OK, at the request of the rostrum, this is the strongly rejected
The guilties of that feedback doesn't work perfectly are Einstein and Boltzmann.
With the observer at the amplifier output, can not be corrected in the past (input), an event in the present (output)
This argument favors minimalist designs.
The guilties of that feedback doesn't work perfectly are Einstein and Boltzmann.
With the observer at the amplifier output, can not be corrected in the past (input), an event in the present (output)
This argument favors minimalist designs.
Fundamental misunderstanding of feedback. Do you manage to keep your car in its lane when you drive?
Go read the Putzys article, it's available (I think) for free at Linear Audio's website.
'Can be better', yes. Does not guarantee that it will be better, because opamps are not ideal components and people can misuse them. Same for valves, but valves may sometimes be a bit more forgiving of poor design.ginetto61 said:
Enough to create instability. or the verge of instability. Enough to expose common-mode distortion at the summing stage. Enough to degrade clipping performance in an amplifier which is sufficiently underpowered that it might frequently clip. Enough in a vain attempt to paper over a poor design.ginetto61 said:
No. Any feedback, however small, will create re-entrant distortion unless the forward path is distortionless. In almost all cases this extra distortion will be smaller than the intrinsic distortion of the forward path. In almost all cases the total distortion can be minimised by adding more feedback. The lesson we should learn from this is not that feedback is bad but that insufficient feedback is bad, and that non-linear amplifiers are really bad. Make it as linear as you reasonably can, then add the right amount of feedback (which in some cases may be 'none', and in other cases may be 'lots').popilin said:
As it is a false argument it actually undermines minimalist designs. If the issue really was causality then modern electronics would simply insert a delay and achieve perfection. Nobody would object to hearing their music a few 100ms late if this could allow the elimination of all distortion. The real issue is that we do not have infinite gain, therefore we can never reduce the error to infinitesimally small. The best we can do is make the error sufficiently small, and based on the inverse of the forward function of the amplifier. There may therefore be some scope in investigating amplifiers with a low-order inverse function, which will necessarily require the forward function to contain terms up to infinity. Counter-intuitive, and definitely non-minimalist!popilin said:
Last edited:
That's funny because it's counter intuitive to many people. Opamps are often represented as simple building blocks of gain that are easy to just plug in. Tubes are more mysterious, use high and low voltages, are temperamental and are from a mystic past.
Of course, no one in this thread would ever feel that way. I'm talking about the "other" circuit designers 🙂
That is a semantic point. What about opamps, where the feedback is mostly quadrature? Is that negative or positive?SY said:
People who understand opamps don't think of them in this way. People who don't understand opamps seem to fall into two groups:Pano said:
- 'idealists' think all opamps are ideal so can be swapped but why would you want to?
- 'chip rollers' think all opamps can be swapped without redesigning any other part of the circuit, and are convinced that many opamps are evil when the reality is that some may simply be misapplied. Many 'chip rollers' would not recognise a stability criterion if it dropped on their foot, and believe that as they can't hear 5MHz it doesn't matter if the circuit is singing at that frequency.
We agree as to the THD, but my point is about higher order components, which it is said, are the most annoying.
OK. Show me please a way to implement such a delay, without involving a trip to the past.
Semantic, yes, but I think it's of great pedagogical importance when exorcising the demons of audiophile legend.
True - but it will probably sound different, which is the whole point of opamp rolling, right? 😉
- Status
- Not open for further replies.