As the thread
http://www.diyaudio.com/forums/showthread.php?threadid=115513
was closed, though it has brought quite interesting discoveries, I would like to continue in a constructive discussion here.
http://www.diyaudio.com/forums/showthread.php?threadid=115513
was closed, though it has brought quite interesting discoveries, I would like to continue in a constructive discussion here.
I am suggesting several topics to discuss:
1) PSU, line stage and power stage solutions
2) passive components - resistors, capacitors
3) PCB, materials, number of layers, starr gnd and groundplane
4) case, shielding, point of connection of case and signal gnd, if any, connection to PE - if any
5) balanced or SE interconnection
6) EMI/RF content in audio signal and its audible influence
7) EMI throughput thru power supplies
1) PSU, line stage and power stage solutions
2) passive components - resistors, capacitors
3) PCB, materials, number of layers, starr gnd and groundplane
4) case, shielding, point of connection of case and signal gnd, if any, connection to PE - if any
5) balanced or SE interconnection
6) EMI/RF content in audio signal and its audible influence
7) EMI throughput thru power supplies
The list of what influences the "sound" of a "high" end amplifier will be infinitely long and highly contentious: NFB, resistors (and direction), etc,etc,etc,etc, all dependent upon the depth of the disorder of the particular "listener".
Why not just greatly simplify things and just ask what doesn't influence amplifier sound?
Hex vs Philips screws? White vs Black lettering? Blue vs Red LEDs? Wire insulation color? Sun spots?
Or perhaps even those alter sound for some? As evidenced by audiophiles hearing improvements while listening *very* carefully to loose screws.
cheers,
AJ
Why not just greatly simplify things and just ask what doesn't influence amplifier sound?
Hex vs Philips screws? White vs Black lettering? Blue vs Red LEDs? Wire insulation color? Sun spots?
Or perhaps even those alter sound for some? As evidenced by audiophiles hearing improvements while listening *very* carefully to loose screws.
cheers,
AJ
Hi,
I am pleased to see this 'resurrection', as between the noisy parts, there was a lot of interesting information.
In case anyone wishes to further the earlier comments on relays, and at the risk of accusations of self-promotion, I posted some of my own experiences here.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=115006&perpage=25&pagenumber=2
In an earlier post on that thread I also mentioned my personal preferences for Shallco-based switched attenuators, which I still consider to be the most transparent (and least damaging to the signal) method of attenuation for pre-amps etc.
I also agree wholeheartedly with Charles' earlier comments on shunt components being every bit as significant as any signal-path parts used in circuits, even though this may not seem intuitive. Careful listening tests tell me that this is so in every area, and not merely in potential divider situations.
Regards,
I am pleased to see this 'resurrection', as between the noisy parts, there was a lot of interesting information.
In case anyone wishes to further the earlier comments on relays, and at the risk of accusations of self-promotion, I posted some of my own experiences here.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=115006&perpage=25&pagenumber=2
In an earlier post on that thread I also mentioned my personal preferences for Shallco-based switched attenuators, which I still consider to be the most transparent (and least damaging to the signal) method of attenuation for pre-amps etc.
I also agree wholeheartedly with Charles' earlier comments on shunt components being every bit as significant as any signal-path parts used in circuits, even though this may not seem intuitive. Careful listening tests tell me that this is so in every area, and not merely in potential divider situations.
Regards,
Thanks. In case we overlook some kind of excessive attacks, I am sure we will be able to keep viable, constructive discussion.
There's a good information about components in Bruno's slide (http://www.hypex.nl/docs/Bruno Masterclass/slides.htm), #128-129. I don't know how far this will affect in analog amps, but the effect is clear in classD amps (about 500khz).
The point is, up until high frequencies, does a capacitor still a "capacitor", that ONLY do capacitance, or they have transformed into something else (like a capacitor behaves more like inductor in HF, due to (big) body construction), or have become an antenna?. Does a resistor still a "resistor" or have become inductor? Maybe this behavior in HF is audible in AF, interpreted as different sound from different component (while we cannot measure a difference, if the measurement is only in AF). We can just predict what happens if your CCT has additional red components like in slide#128.
In HF, a PCB track with groundplane on the other side transforms into "microstrip" with Mhz transmission-line properties. But how far this will be audible for AF?
The point is, up until high frequencies, does a capacitor still a "capacitor", that ONLY do capacitance, or they have transformed into something else (like a capacitor behaves more like inductor in HF, due to (big) body construction), or have become an antenna?. Does a resistor still a "resistor" or have become inductor? Maybe this behavior in HF is audible in AF, interpreted as different sound from different component (while we cannot measure a difference, if the measurement is only in AF). We can just predict what happens if your CCT has additional red components like in slide#128.
In HF, a PCB track with groundplane on the other side transforms into "microstrip" with Mhz transmission-line properties. But how far this will be audible for AF?
In post #4 I intended to say "shunt ..... being as significant as *series* signal-path parts" etc.
My apologies for any unintentional confusion, and it was too late to correct that post when I realised this.
Regards,
My apologies for any unintentional confusion, and it was too late to correct that post when I realised this.
Regards,
Well, I sometimes have a screw or two loose, but here's my take. It seems to be accepted that amplifiers which are technically close to perfect, e.g., Self's "blameless" designs, are not accepted as creme of the crop world class sonic perfection. Now, I'm not sure I could hear the difference in a fair fight, but let's say I could. Further, let's say the kilobuck design sounded better. The blameless design would probably do well when a differential measurement was made between its input and output- that's the essence of what makes it blameless. The kilobuck amp would have to fare worse, and should also show a difference between the two amps in a similar null test. If, in fact, it sounded different. If it passed the differential tests, any claim of sounding different would be highly suspect. FWIW, this how I read the summary of the Bob Carver TFM paper discussed recently on another board.
The only conclusion I can reach from this somewhat hypothetical situation is that the best sounding amplifiers achieve their goal by *not* going overboard on flat response, low distortion, and low output impedance. The designers have instead found some recipe that pleases the ear more than the proverbial straight wire with gain. As unsatisfying as it may be to us technical types, all perfect amplifiers sound worse than *some* imperfect amplifiers. Our targets may not be the right ones, but our training prevents us from changing them.
IMO, much of the so-called explanations for why certain components or designs sound better end up stretching fundamental principles beyond the breaking point, or are just deluded hand-waving. Why? Because either there isn't really a difference, or the difference isn't politically acceptable to the semi-technical crowd, so semi-technical BS is all that remains.
I also notice that various aspects of circuits, like regulators, have been analyzed in depth, and all the differences and pros and cons have been outlined. Jung has done some great work with this. Yet, in the end, nobody seems able to make a *proven* connection to why or how -90dB effects alter our perception of the sound. IMO, it's because they usually don't. The knowledge of what's in the circuit so contaminates the mind, that comparison becomes impossible. Those who believe they are somehow immune to this, IMO, don't understand the depth of the phenomena. Since it's inconceivable to them that the differences might not be real, the only way out is to discredit the test method- DBT, switch boxes, or whatever.
Audio is truly a hopeless endeavor.
The only conclusion I can reach from this somewhat hypothetical situation is that the best sounding amplifiers achieve their goal by *not* going overboard on flat response, low distortion, and low output impedance. The designers have instead found some recipe that pleases the ear more than the proverbial straight wire with gain. As unsatisfying as it may be to us technical types, all perfect amplifiers sound worse than *some* imperfect amplifiers. Our targets may not be the right ones, but our training prevents us from changing them.
IMO, much of the so-called explanations for why certain components or designs sound better end up stretching fundamental principles beyond the breaking point, or are just deluded hand-waving. Why? Because either there isn't really a difference, or the difference isn't politically acceptable to the semi-technical crowd, so semi-technical BS is all that remains.
I also notice that various aspects of circuits, like regulators, have been analyzed in depth, and all the differences and pros and cons have been outlined. Jung has done some great work with this. Yet, in the end, nobody seems able to make a *proven* connection to why or how -90dB effects alter our perception of the sound. IMO, it's because they usually don't. The knowledge of what's in the circuit so contaminates the mind, that comparison becomes impossible. Those who believe they are somehow immune to this, IMO, don't understand the depth of the phenomena. Since it's inconceivable to them that the differences might not be real, the only way out is to discredit the test method- DBT, switch boxes, or whatever.
Audio is truly a hopeless endeavor.
I think the EMI/RF issues are the key point today, given that airbourne and conductor-bourne RF hash has quite possibly increased at least one or two orders of magnitude in the last 20 years in the typical domestic situation.
And from experience (with musican's gear and industrial electronics, mostly) I know it is very hard to make something really HF-proof, especially when you only can modify the gear. But even when newly designed from gound up it is very hard and costly to get there, and (if one believes the rumours) the typical EMI "add-on" countermeasures like ferrites on signal/speaker and power lines may also have a negative effect on sound quality.
I have not found any typical consumer HiFi-gear so far (no access to the top-level stuff, though) that didn't fail with the simply cell phone test, showing the typical TDMA buzz. And way long before the buzz is clearly heard by itself it will affect sound quality.
- Klaus
And from experience (with musican's gear and industrial electronics, mostly) I know it is very hard to make something really HF-proof, especially when you only can modify the gear. But even when newly designed from gound up it is very hard and costly to get there, and (if one believes the rumours) the typical EMI "add-on" countermeasures like ferrites on signal/speaker and power lines may also have a negative effect on sound quality.
I have not found any typical consumer HiFi-gear so far (no access to the top-level stuff, though) that didn't fail with the simply cell phone test, showing the typical TDMA buzz. And way long before the buzz is clearly heard by itself it will affect sound quality.
- Klaus
Hi, KSTR,
That is a good example 😀 It proofs that it CAN enter the soundsystem. Because the "packets" are below 1khz, we can hear them, but what if the packets just slightly above 20khz, cannot be heard by themselves, but make massive IMD to AF signals.
That is a good example 😀 It proofs that it CAN enter the soundsystem. Because the "packets" are below 1khz, we can hear them, but what if the packets just slightly above 20khz, cannot be heard by themselves, but make massive IMD to AF signals.
Haha. Then I'm not alone 😀Conrad Hoffman said:
Conrad Hoffman said:
Accepted by who? The self appointed chosen? "World class sonic perfection" according to who's/what standard? Your friendly neighborhood hearing superhero, like those here?
Conrad Hoffman said:
Here's how I read it. Bob's a very sharp guy. He knows that his amplifier cannot improve the input signal, only degrade it. He knows audiophiles love/crave distortion. He knows that audiophile "SOTA" sound = distorted/colored output. That is what they crave. He knows given enough time, he can make his amplifier degrade the input signal and output a distorted/colored signal equivalent to that of the "SOTA" amplifier, so that the sound is "SOTA" as determined by the self analysed experts. He did it using null measurements. He degraded the sound to the point where the self chosen hearing gods liked it as much as the $20k piece of fake jewelry (Hey, at least it didn't blow up). By knowing a great deal about distortion/coloration mechanisms.
Like I said, he's a pretty sharp guy.
Conrad Hoffman said:
Once again, the "best" sounding according to who? Where did their expertise accreditation come from? What tests have they taken and passed? Their advanced degree in *Careful Listeningology came from where?
Conrad Hoffman said:
They have found that audiophiles have deep pockets and a deeper love of varying euphonic distortion. Eureka!
Conrad Hoffman said:
BS sells. Very well. Pay offs to the disseminators of media don't hurt either.
DBTs? You mean JLWSIs right? As is Just Listen Without Scientific Intervention. DBTs are worthless when you have a product to sell.
It should be mandatory to list what blameless loudspeakers and details of the blameless recordings used during any "test". You know, merely the two things most responsible for the sound passing through the amplifier before reaching your ear. Ok, the room too.
cheers,
AJ
AJ, I suspect we're quite in agreement on a lot of points! I do question one thing though (I question a lot more than one thing about life in general, but I digress...). Say a source is -8dB at the high end, and the amplifier happens to boot it by 8dB in just the right area. I say the amplifier has, in fact, improved the sound. So degradation is not the only possible outcome- the amp can improve the signal. Given all the unknowns in the recording process, strict adherence to the conventional definition of perfection may not be optimal, especially if there are any systematic errors that tend to be common to a lot of recordings. Not that I relish that situation. Maybe the RIAA should be spending more time and money doing what they started out doing- standards, and wasting less time and money in court.
They're not wasting time in court, they're making money in court and (mainly) with out of court settlements. That is, de facto, the function and structure of that organization, lawyers and accountants; I wouldn't want a group of lawyers and accountants making strategic engineering standards decisions, so thank goodness they're NOT involved.
I found another big source of RF gunk came from CCFL lamps. Especially these small screw in types. A lot are made in China and to be honest, they really radiate some crap.
While I was working on a proto amp a year or two ago, I went mad trying to find the source of the HF hash. After about an hour I remembered I'd just replaced my overhead bench light with a 15w CCFL. I turned the lamp off and the problem disappered.
Problem is, when you try to make a circuit RF hash proof, it probably affects th e sound (no proof of this, just my thought on it). Speaker cables make great a great input antenna for the amp. If your circuit uses a small cap across th e feedback resistor (now there's a reason for zero GNF!) you can get garbage straight into th e front end.
While I was working on a proto amp a year or two ago, I went mad trying to find the source of the HF hash. After about an hour I remembered I'd just replaced my overhead bench light with a 15w CCFL. I turned the lamp off and the problem disappered.
Problem is, when you try to make a circuit RF hash proof, it probably affects th e sound (no proof of this, just my thought on it). Speaker cables make great a great input antenna for the amp. If your circuit uses a small cap across th e feedback resistor (now there's a reason for zero GNF!) you can get garbage straight into th e front end.
There are alot of details that is not evident when we make classAB amps, suddenly becomes very important when we make classD selfoscilating amps. Try to make a DIY selfoscilating classD, it will bring new perspective on designing classAB amps. Try to shoot all the problems emerged in the process, that's where the lessons are.
When making selfoscilating classD, I study that a component lead, even only 1cm, have internal inductance, can affect the scope picture. This would seems crazy for analog design, but in 500khz oscilation, it does make a difference. I become more open-minded when reading about component (sonic) difference after knowing these things.
Maybe the physical construction is different, the lead material is different, all can make different properties to each component.
When making selfoscilating classD, I study that a component lead, even only 1cm, have internal inductance, can affect the scope picture. This would seems crazy for analog design, but in 500khz oscilation, it does make a difference. I become more open-minded when reading about component (sonic) difference after knowing these things.
Maybe the physical construction is different, the lead material is different, all can make different properties to each component.
- Status
- Not open for further replies.