well I will speak for my little experience.
I use reference Harbeth monitors and I have a rel subwoofer, I also have Usher speakers which are not as good but more sensitive. I had a S.E.T 845 hifi amplifier with single driver and the 12 inch, 8 inch, 3 inches 100% alnico setup open baffle, most realistic but not hifi!!!, it had the best soundstage and the most realism in the bass but impractical for home use. the baffles extended from asymmetrical plywood sheets of 6'x9' cornered in the room.
When experimenting with feedback, you find a middle point when designing an amplifier where it is in between euphoria and cold ice.
Again, the more feedback weapons you have the better the outcome, everything is allowed, Chokes, split loads, power supply feedback, cathode windings, dumping stages in ss.
The main difference is that a 1 watt amplifier won't have the details of a 100 Watt amplifier, however it will sound beautiful if it is well designed. And besides, some designers can really push the 300b tube to 20 watts by driving it with a EL34 pentode.
The trick is to make your amplifier stable and able to drive capacitative/reactive loads. that's all.
I use reference Harbeth monitors and I have a rel subwoofer, I also have Usher speakers which are not as good but more sensitive. I had a S.E.T 845 hifi amplifier with single driver and the 12 inch, 8 inch, 3 inches 100% alnico setup open baffle, most realistic but not hifi!!!, it had the best soundstage and the most realism in the bass but impractical for home use. the baffles extended from asymmetrical plywood sheets of 6'x9' cornered in the room.
When experimenting with feedback, you find a middle point when designing an amplifier where it is in between euphoria and cold ice.
Again, the more feedback weapons you have the better the outcome, everything is allowed, Chokes, split loads, power supply feedback, cathode windings, dumping stages in ss.
The main difference is that a 1 watt amplifier won't have the details of a 100 Watt amplifier, however it will sound beautiful if it is well designed. And besides, some designers can really push the 300b tube to 20 watts by driving it with a EL34 pentode.
The trick is to make your amplifier stable and able to drive capacitative/reactive loads. that's all.
Gabdx, you seem to have a balanced approach to feedback and amplifier design in general. Now it would be nice to hear what NFB camp thinks.
My latest construction, which I'm very pleased with is a class AB BJT amplifier that has a rather unusual design. I made three attempts before I was satisfied.
At first I followed my instinct and applied a very modest NFB of 20db. That wasn't too good. Tre treble was a bit harsh.
Then I went in the other direction and applied as much NFB as was possible without jeopardising stability. Now it sounded very controlled - like I imagine a Krell sounds - an interesting sonic imprint but I got a bit tired of it , it was simply too dense.
Then I applied an NFB that was between the extremes. Bingo. The sound that emerged was now very gentle, inconspicuous, and with a nice 3D soundstage, but yet very dynamic at high volume.
This is my personal experience in this matter.
My latest construction, which I'm very pleased with is a class AB BJT amplifier that has a rather unusual design. I made three attempts before I was satisfied.
At first I followed my instinct and applied a very modest NFB of 20db. That wasn't too good. Tre treble was a bit harsh.
Then I went in the other direction and applied as much NFB as was possible without jeopardising stability. Now it sounded very controlled - like I imagine a Krell sounds - an interesting sonic imprint but I got a bit tired of it , it was simply too dense.
Then I applied an NFB that was between the extremes. Bingo. The sound that emerged was now very gentle, inconspicuous, and with a nice 3D soundstage, but yet very dynamic at high volume.
This is my personal experience in this matter.
What about a variable NFB amplifier.
That comes easy with an op amp based amplifier.
One can apply total feedback ( over the op amp and output stage ) or partial feedback ( over the op amp only ) with the same feed back resistors, this is just a matter of where you take the signal. One can make a nfb in between feeding a mix of both output signals.
That comes easy with an op amp based amplifier.
One can apply total feedback ( over the op amp and output stage ) or partial feedback ( over the op amp only ) with the same feed back resistors, this is just a matter of where you take the signal. One can make a nfb in between feeding a mix of both output signals.
Nelson's in the entertainment industry, I don't think he makes any secret of the fact
Nelson's in the entertainment industry, I don't think he makes any secret of the fact
He's also designed commercial amplifiers throughout the decades, which use negative feedback. Some of them are two stage amplifiers, but they still use feedback.
At the I Have More Money Than Brains hi fi store, there is an enormous Pass power amplifier. It is the size of a piece of airline luggage. I do not know what makes it tick, but it sure sounds fantastic. I do not know if it uses minimum feedback, or if it showcases Pass Labs methods.
there is inherent feedback and global feedback. a single transistor or tube amplifier with a 20k collector resistor and a 4k emitter resistor will yield a gain of 5. the next stage can amplify that by 5 again yielding total a gain of 25. global feedback ( op amp ) can have an open loop gain of 100000 plus falling to 400 at 20khz, global feedback equalizies this a total gain of 25 regardless of frequency. if it is stable it works brilliantly.
I don't think it has anything to do with THD.
Most speakers do at least 0.5% THD to 5% THD, with a dynamic range of 40-60 db at most, look at all single driver THD data,
Now use those drivers in a normal room which contribute to double the THD with at least 20% phase shifts from the XO, and you find the compounding of the amplifier 0.5 or 1% or 0.01 or 0.0001 makes no difference, at all.
The amplifier just sounds better interacting with speakers at this level of feedback.
And 'edit' congratulation to the amp builder for choosing the best sounding combination to him.
Last edited:
It's not such a strange hypothesis. Usually audio amplifiers have shunt feedback at their outputs, resulting in a near-zero output impedance, so the output behaves almost like an ideal voltage source. Reducing loop gain then increases the output impedance.
Dynamic loudspeakers distort much less when they are driven with current rather than voltage, but this usually results in a bumpy small-signal transfer due to the reduced damping of the loudspeaker's resonance (unless the loudspeaker box has an impedance equalizing network built in). Maybe some intermediate driving impedance can result in a somewhat bumpy response, but not so bumpy that it gets annoying, with somewhat reduced distortion of the loudspeaker itself.
If this is the correct explanation, using a combination of shunt and series feedback or simply connecting a resistor between amplifier and loudspeaker would be a more well-controlled way to achieve the same.
I don't know if it adds up with the numbers in post # 103, though. Post # 103 suggests that the subjectively optimal loop gain is substantially greater than 20 dB. Assuming that that's measured or calculated with nominal load, the output impedance will be substantially less than a tenth of the nominal impedance of the loudspeaker. The bump in the response will then be substantially smaller than 1 dB and the reduction in loudspeaker distortion much less than a tenth.
Dynamic loudspeakers distort much less when they are driven with current rather than voltage, but this usually results in a bumpy small-signal transfer due to the reduced damping of the loudspeaker's resonance (unless the loudspeaker box has an impedance equalizing network built in). Maybe some intermediate driving impedance can result in a somewhat bumpy response, but not so bumpy that it gets annoying, with somewhat reduced distortion of the loudspeaker itself.
If this is the correct explanation, using a combination of shunt and series feedback or simply connecting a resistor between amplifier and loudspeaker would be a more well-controlled way to achieve the same.
I don't know if it adds up with the numbers in post # 103, though. Post # 103 suggests that the subjectively optimal loop gain is substantially greater than 20 dB. Assuming that that's measured or calculated with nominal load, the output impedance will be substantially less than a tenth of the nominal impedance of the loudspeaker. The bump in the response will then be substantially smaller than 1 dB and the reduction in loudspeaker distortion much less than a tenth.
Last edited:
Subject 1:
The NFB camp has still not come up with any personal subjective reports about various amplifiers they have listened to. I will take that as an indication that those guys usually are fixated to some idea of the "blameless amplifier syndrome" if you see what I mean.
Subject 2:
About Gabdx discussion about speaker THD.
In a normal room, even a very good speaker measures extremely bad in terms of THD and also when it comes to linearity. There are bumps and boosts at +-10 db, up and down.
Still, when we change to another amplifier, we usually hears clear differences. That's why I don't think the actual level of THD matters much, it's the way the THD pattern is formed that matters.
Please read this document about psycho acoustics:
YouTube
The NFB camp has still not come up with any personal subjective reports about various amplifiers they have listened to. I will take that as an indication that those guys usually are fixated to some idea of the "blameless amplifier syndrome" if you see what I mean.
Subject 2:
About Gabdx discussion about speaker THD.
In a normal room, even a very good speaker measures extremely bad in terms of THD and also when it comes to linearity. There are bumps and boosts at +-10 db, up and down.
Still, when we change to another amplifier, we usually hears clear differences. That's why I don't think the actual level of THD matters much, it's the way the THD pattern is formed that matters.
Please read this document about psycho acoustics:
YouTube
1. When you believe that all well-designed amplifiers sound the same when properly level matched, there is no point testing this subjectively using yourself as the test person, because you can always confirm your believe by (subconsciously) giving random answers.
2. Do you level match the amplifiers within 0.1 dB and use some kind of random switch box to switch between amplifiers without you knowing which one is playing? If not, then the whole test is invalid anyway.
2. Do you level match the amplifiers within 0.1 dB and use some kind of random switch box to switch between amplifiers without you knowing which one is playing? If not, then the whole test is invalid anyway.
1. That's exactly what I want to stress. Rationalists are subjects of some sort of inverted placebo effect. Since they are assured that anything they hear is dubious, they don't even bother listening with their own ears.
What a boring life! And why are these guys participating in this forum? It's absurd, the perfect amplifier is already made.
2.
Yes, if we are here just to try to conserve our beliefs. Audiophiles reports that they hear something at home in their living room. In a typical blind test the subject usually is a bit tense and cannot hear all the nuances he hears in his hifi sofa. Should we therefore dismiss him? Isn't there reason to carefully investigate whether this placebo fool maybe is in to something real?
We should have a free mind. And I think the whole diyaudio is pirated by a bunch of rationalists.
Blind test aren't useful when it comes to such a subjective matters as perception of audio, music, timbres, 3D, whatever you want to call it.
MAYBE, is say maybe all audiophile are subjects of the placebo effect, but we can't be sure.
OK Marcel, if you listen by your own ears, then the whole thing is unproven. Yes, but come on, be a bit personal! Isn't it boring to be so assured of something that you don't even dare open your ears!
What a boring life! And why are these guys participating in this forum? It's absurd, the perfect amplifier is already made.
2.
Yes, if we are here just to try to conserve our beliefs. Audiophiles reports that they hear something at home in their living room. In a typical blind test the subject usually is a bit tense and cannot hear all the nuances he hears in his hifi sofa. Should we therefore dismiss him? Isn't there reason to carefully investigate whether this placebo fool maybe is in to something real?
We should have a free mind. And I think the whole diyaudio is pirated by a bunch of rationalists.
Blind test aren't useful when it comes to such a subjective matters as perception of audio, music, timbres, 3D, whatever you want to call it.
MAYBE, is say maybe all audiophile are subjects of the placebo effect, but we can't be sure.
OK Marcel, if you listen by your own ears, then the whole thing is unproven. Yes, but come on, be a bit personal! Isn't it boring to be so assured of something that you don't even dare open your ears!
You make a lot of assumptions just because many find a subjective discussion about what they do and do not hear a bit of a waste of time and energy, it does not mean they are not passionate about it, it probably means they are happy to listen and enjoy privately and you should be grateful when they chose to share their thoughts with you, you should definitely not demand they do.
It's well known that amplifiers with rising bass and treble response, and amplifiers with plenty of 2nd order distortion, sound "pleasant". The only thing a subjective test will show is which amplifier sounds the "nicest".
Which is ridiculous. The aim of music reproduction is to faithfully reproduce the material the artist recorded, not to mess with it.
From a visual perspective, it'd be like covering an artwork in yellow tinted varnish, because that looks pleasing to the viewer's eye. Whether it's pleasing or not is immaterial. The fact is that it's destroying the original work.
ScottJoplin. You mean that the "rationalists" also are passionate but perhaps conservative and they don't want to stick out their heads and reveal their personal experiences?
Yes...
But why not share these experiences? We all know that you can't prove subjective matters. Actually, there are recent research that reveals that our hearing is more hard to grasp than one might expect. Why constantly stick to the old assumption that the level of THD is the only thing that matters?
Why not have some courage and accept that we all are subjects of the placebo effect, but therefore should not resort to the lowest level of reasoning. I mean, subatomic particles can exist at two places at the same time, why be so certain?
suzyj.
I think a nice sounding amplifier is the one that reveals all possible nuances to the music.
I will elaborate on that. Normally when you auditioning something for the first time, you are always impressed. What you her is a new and fresh sound. But by time it usually reveals whether it was a real approvement or not. It can take one hour or half a year, that depends. But usually things sorts themselves out.
That's why blind tests are difficult. It takes time to decide what you like. Can you decide in 30 seconds if you love a girl or not? ( Drastic example)
BUT.....
It's a fact that everyone hears a difference when changing amplifier, disregarding if it's an approvement or not. This indicate, I think, that it's not the THD level that makes the difference.
That old theory is simply to blunt.
Yes...
But why not share these experiences? We all know that you can't prove subjective matters. Actually, there are recent research that reveals that our hearing is more hard to grasp than one might expect. Why constantly stick to the old assumption that the level of THD is the only thing that matters?
Why not have some courage and accept that we all are subjects of the placebo effect, but therefore should not resort to the lowest level of reasoning. I mean, subatomic particles can exist at two places at the same time, why be so certain?
suzyj.
I think a nice sounding amplifier is the one that reveals all possible nuances to the music.
I will elaborate on that. Normally when you auditioning something for the first time, you are always impressed. What you her is a new and fresh sound. But by time it usually reveals whether it was a real approvement or not. It can take one hour or half a year, that depends. But usually things sorts themselves out.
That's why blind tests are difficult. It takes time to decide what you like. Can you decide in 30 seconds if you love a girl or not? ( Drastic example)
BUT.....
It's a fact that everyone hears a difference when changing amplifier, disregarding if it's an approvement or not. This indicate, I think, that it's not the THD level that makes the difference.
That old theory is simply to blunt.
I think it is quite the opposite.
Manufacturers spend a lot of time testing the public reaction and asking experts to review their products.
All serious designers spend critical time asking for people 'blind' perception and blind testing their creations, they also refine as much as possible the sound.
The first thing an audiophile will do is to ask others for opinions on the system, and AB blind test many components. The goal is the best sound, not the best 'typology'. You will find audiophile having 300watts and 2 watts amplifiers and they keep both if they can extract good music from it. People don't care about feedback, they care if it sounds for their taste.
The aim is not power/typology but only sound, by combination, matching, taste and constant critics from ourselves and friends. The forum is a great place to read others experiences, it's greatly entertaining and diy especially because people experiment beyond going to the shop to simply switch gears.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.