any person proclaiming niceness should be taken as being true to himself only...it will have no effect on how i perceive the sound either way...and so is meaningless to me....
the lesser you know about how circuits work, the more you are likely to fall for hypes and myths....
i know of many from other forums who are "cap changers" and go about proclaiming their so called "find" but having no clues as to how the thing works...
the lesser you know about how circuits work, the more you are likely to fall for hypes and myths....
i know of many from other forums who are "cap changers" and go about proclaiming their so called "find" but having no clues as to how the thing works...
I really find this very condecending.
You assume that because someone doesn't agree with you, they don't know what they are talking about. Sound engineering is the essential foundation on which we build our pieces - but then there is judicious choice of which sound principles to apply - that is a value judgement based upon our own personal experience.
Effectively people here are equating well respected engineers, who approach things in a different way to themselves, as snake oil sales men and assuming that anyone who accepts their engineering experience is a fool.
The only possible reason that you can make a statement about caps not differing in sound is if you so mask the components through excessive use of gNFB that you genuinely cannot hear a difference. Caps sound different but only in a pure simple open loop application can you readily hear those differences. Yet I have yet to hear a coherent decisive explanation for why caps do sound different. Are we to dismiss the evidence of our ears because one group of analytical engineers have yet to reduce the differences to a simple formula ?
So many unfounded insults in such a short post.
Shoog
Last edited:
No, there's lots of people who disagree without assuming the other fellow is ignorant. You should hear Morgan Jones and me going at it about the desirability of "speed" in voltage regulators (and you may be assured that neither of us considers the other ignorant)!
Sadly, that assumption DOES get made when sweeping pronouncements are put forth without a scintilla of backup data or informed physical plausibility.
Exaggeration is another debating technique which does not assist truth-seeking. When 'top designers' talk nonsense they usually sincerely believe what they are saying.Shoog said:
Caps can sound different because of reasonably well-understood dielectric and construction issues. I have never used 'audiophile' caps, but some report that they can sometimes have worse construction than ordinary caps. Thus, in some cases, it is clear that the perceived 'improvement' is either imaginary or may actually be a degradation.
Caps have different sounds. I use the best caps possible without spending silly money on audiophile caps. My prefered solution is to use no caps in coupling situations - and this seems the best approach to avoid colouration. Some caps seem more important than others when considering their impact on overall sound.
I extensively use silicon and CCS in my designs and find that they definately improve matter - so there is no irrational bias from me here.
I still find it amazing that people could criticise people like Nelson Pass and Allen Wright. I have indeed disagreed with things that Allen has said usually based upon a cost benefit analysis, but I am certain that every design decision he made was based upon years of careful evalutation and refinement of his designs. Its arrogance to believe that because he chose a zero gNFB approach he is a conn man selling expensive snake oil. He evolved his thinking away from gNFB over years of experimentation - that I respect. These are men who have built some of the best received amplifiers and preamplifiers of the last 20years - that certainly doesn't happen through pure myth.
There are reams of audiophile snake oil sales men, but please be careful where you spread that insult to as it certainly doesn't improve the credibility of your arguments.
Shoog
I extensively use silicon and CCS in my designs and find that they definately improve matter - so there is no irrational bias from me here.
I still find it amazing that people could criticise people like Nelson Pass and Allen Wright. I have indeed disagreed with things that Allen has said usually based upon a cost benefit analysis, but I am certain that every design decision he made was based upon years of careful evalutation and refinement of his designs. Its arrogance to believe that because he chose a zero gNFB approach he is a conn man selling expensive snake oil. He evolved his thinking away from gNFB over years of experimentation - that I respect. These are men who have built some of the best received amplifiers and preamplifiers of the last 20years - that certainly doesn't happen through pure myth.
There are reams of audiophile snake oil sales men, but please be careful where you spread that insult to as it certainly doesn't improve the credibility of your arguments.
Shoog
Well, since the only person who mentioned Nelson and Allen is you, you're arguing with yourself, not with any of the other participants in this thread.
And I can certainly disagree with things they say in their marketing documents without criticizing them personally; I often do the former without ever having done the latter- Nelson is a friend who has been nothing but kind and generous, and I've spent some incredibly fun times drinking and discussing design, life, the universe, and everything with Allen (whom I miss greatly).
And I can certainly disagree with things they say in their marketing documents without criticizing them personally; I often do the former without ever having done the latter- Nelson is a friend who has been nothing but kind and generous, and I've spent some incredibly fun times drinking and discussing design, life, the universe, and everything with Allen (whom I miss greatly).
All of the influences I have mentioned are advocates of a mininal or zero global Negative Feedback design philosphy, they are proof of the concept. There has been a definate suggestion made (not necessarily by yourself) that anyone who advocates such an approach is a snake oil salesman, so join the dots here SY.
Shoog
Shoog
You keep mentioning these gentlemen, not us. I have not criticised them. I have expressed a view which you say they disagree with. I don't feel they are insulting me by disagreeing with me (if your report of their views is correct) so why should they feel insulted because I disagree with them? You keep mentioning snake oil, not me.Shoog said:
You should not get so emotional about what, in essence, is a technical argument about apparently disputed facts. In my experience most real experts are much more aware of the caveats and limitations of their views than their enthusiastic supporters.
Lets just look at where the trouble started on this ^^
There were other examples along the way.
Let it also be said that I use gNFB around opamp and still have two gainclones (with high negative feedback). However I moved into valves because it allowed me to move away from this approach, and the improvement in clarity was like night and day. The only global feeback amp in my main system is a gainclone subwoofer driver for my main speakers. Loss of detail isn't that important in that support role.
Shoog
Last edited:
You will note that no names were mentioned by me (as I had none in mind). I merely offered an explanation why some people might prefer something which is inferior to alternatives.
I am getting bored with this 'discussion', so I will leave it at that. I will ignore any further provocation.
I am getting bored with this 'discussion', so I will leave it at that. I will ignore any further provocation.
Last edited:
The issues of gNFB appears more difficult. Some of the better designs in my own experimentation are ones that run near the point of oscillations, or the maximum feedback possible. When these networks were adjusted for greater stability, or reduced feedback, the quality of the sound deteriorated.
The problem with maintaining maximum feedback is that this is dependent upon some value of speaker load in a power amplifier. It isn't practical. Despite those issues, this suggests that zero feedback and maximum feedback can be both better than feedback in the middle ground.
The problem with maintaining maximum feedback is that this is dependent upon some value of speaker load in a power amplifier. It isn't practical. Despite those issues, this suggests that zero feedback and maximum feedback can be both better than feedback in the middle ground.
Johan Potgieter over in This Post observed the same thing, as have I, and he even mentions the same point where gNFB starts to ruin the sound: 13dbv. Others mention the same thing.
I also do SS design, and have pure silicon amps that sound really good, and definite ideas as to how I do it (and Doug Self doesn't agree at all: all I have to say to that is he doesn't have to listen to my designs, and I don't have to listen to his) and have gotten results that beat the living daylights out of any "Big Box" store offerings I've ever heard. The design I came up with, I never claimed originality and it was years later before I found something quite similar in an Hitachi application note.
Connect a BJT to an active collector load and you get EEEEENORMOUS voltage gain, use DC coupling, a good layout, and you can really pour on the gNFB. Way in excess of 13dbv. It would seem there is some sort of "gray area" where too much is a bad thing, and way more is better. It's NBD to get those feedback factors when you have what's basically a large, discrete op-amp. Not such an easy thing once ferro-magnetics enter the design.
May be find out with a preamp design? No ferromagnetics involved, so no reason not to make a "glass op-amp" that can handle larger amounts of feedback.
Oscillators are created by engaging in the use of excess feedback becoming positive feedback, that by such a condition the feedback is not therefore negative. If we consider that if the feedback amount is set to a point on the edge of wanting to oscillate, where the positive and negative feedback amounts are equally fighting each other, it seems that the amplifier can't then decide to go in the positive or negative direction.
In other words the amplifier appears can be set at two points of having no corrective feedback and therefore two states exist without any negative feedback. This then leads to the conclusion that using feedback gains between the two end points are both increasing of negative feedback as perhaps leading to undesirable results.
In other words the amplifier appears can be set at two points of having no corrective feedback and therefore two states exist without any negative feedback. This then leads to the conclusion that using feedback gains between the two end points are both increasing of negative feedback as perhaps leading to undesirable results.
True. The more marginal the gain is above what is required to sustain oscillation the more pure the oscillation frequency can become. In other words setting the gain of feedback as least necessary to support the oscillation.
The argument that I am trying to present is that I am referring more to the most exact gain as equal to neither preventing or encouraging oscillation. My suggestion is that if a condition exists, whereupon the connection of a feedback signal cannot be concluded as creating a negative or positive feedback result there is no overall negative feedback.
Ultimately if you can't assign a polarity to the direction of the feedback it ceases to be negative. My suggestion is that this occurs when the gain is increased to the point of the onset of an oscillator, whereupon the amount of positive feedback neutralizes the negative. If we now consider this to be true and we are not trying to build an oscillator, it follows that negative feedback exists between these ends, and whereupon changes in sonic's occurs from either end toward the middle.
No, I still can't grasp what you are saying. Feedback has an amplitude and phase. Whether it can be regarded as positive or negative depends on the sign of the in-phase component. You can always assign a polarity except for the situation where the feedback is exactly in quadrature, which is usually only at a few specific frequencies.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.