Re: Re: Re: Re: TPC vs. TMC
See: http://www.diyaudio.com/forums/showthread.php?postid=1599540#post1599540
syn08 said:
See: http://www.diyaudio.com/forums/showthread.php?postid=1599540#post1599540
Dave said:
PGP=Pretty Good Poweramp www.synaesthesia.ca
HEC=Hawksford Error Correction http://www.essex.ac.uk/dces/researc...s/C39 NS and nested differential feedback.pdf
TPC=Two Pole Compensation see D. Self book
TMC=Transitional Miller Compensation e.g. http://www.diyaudio.com/forums/attachment.php?s=&postid=1161190&stamp=1174224139
OPS=Output Power Stage
PCP=Pretty Complex Poweramp, Edmond's current project (unpublished yet).
salas said:
Don't know yet. It will be evaluated when the construction will be completed, most likely following the same pattern and equipment as for the PGP amp: http://synaesthesia.ca/listening.html
Edmond,
it`s my pleasure.
(When you find that the misuse of simulation software has reached an excessively detrimental point, you should consider to irrecoverably delete it and start over with a clear mind. It`s an explicitly inappropriate tool for predicting perceived sound, misleading and fooling you completely, thereby preventing you from ever designing good sounding circuits).
it`s my pleasure.
(When you find that the misuse of simulation software has reached an excessively detrimental point, you should consider to irrecoverably delete it and start over with a clear mind. It`s an explicitly inappropriate tool for predicting perceived sound, misleading and fooling you completely, thereby preventing you from ever designing good sounding circuits).
a different cattle of fish
You are totally right. I should also throw away my calculator, slide rule or even my left hemisphere. Such attributes are useless for building musical instruments and predicting the perceived sound of them. However, you missed my point, I'm not interested in such instruments. Rather, I'm interested in designing amplifiers.
You are totally right. I should also throw away my calculator, slide rule or even my left hemisphere. Such attributes are useless for building musical instruments and predicting the perceived sound of them. However, you missed my point, I'm not interested in such instruments. Rather, I'm interested in designing amplifiers.
The only exact & appropriate tool to design amplifiers are the human ears.
Especially those sitting on the head of older audio gentlemen.
This is how it has been ever since before the transistor & computer was born.
And this is how it will always be.
Because no calculator can ever perceive sound of music.
Now, the very exact human ears can perceive sound better than any measuring instrument made by man.
Especially when the sound signals are routed to bypass the mind and thoughts of the person who carry them perfect ears.
This is how we avoid any bias, that eventually would be left in the clever brain of the test person listening.
Doesn't matter they can see galaxes beyond the beginning of time in those skys. Human eye can see such things much better, at least if I use my imagination.
Same thing with molecules, I can see them better than any electron microscope. Human sensory perception is something so unique in the nature of creation. There is no animal alive that can hear things and see things better than one man.
Not even my little domestic cat.
Because we people have this fantastic ability to dream and make use of fantasy.
Lineup
audio regards - threw his calculator out his window this morning.---------------------------------------------------------------------
His PC will probably go the same way including all simulation spice software.
Unless he is too addicted to his internet connection and www.diyaudio.com
where modern audio science reaches higher & higher levels, day by day.
Yesterday seems to gotten a temporary never seen before peak in insights and logical thinking.
Edmond,
You are arbitrarily separating the functions of amp and instruments here; but from the POV of the listener, their combined effects only are perceived.]
There is a huge amount of good engineering in a properly designed amp, of course, and you have demonstrated this well. But the laws of engineering alone do not apply; if the amp is to please, it should be musical too, whatever that means...... I would posit that EEs should understand something of the psychoacoustics of music as well.
To me this means that any distortions introduced by the amp should not be even faintly discordant; this would imply the lofty goal of removing all odd order distortions, where possible, or at the least taking them below -130dB for 5th, 7th, 9th, etc. The nature of global negative feedback makes this difficult because of the mixing at the error amplifier, but oddly enough the choice of components is also an issue.
This is not to refute you, or anyone else bent on designing a zero distortion amp, but given that some distortion is inevitable it could point the way to better engineering design.
It is also my contention that you do not need to be a degreed EE to design an audio amplifier. Plenty of designers come from other disciplines; JLH was a nuclear scientist, for example, and plenty have no relevant qualifications in this area. As has been pointed out here, this is not rocket science, at least at the basic design level. And a surprising number of basic designs - NP is the classic example - sound pretty damn good.
Hugh
You are arbitrarily separating the functions of amp and instruments here; but from the POV of the listener, their combined effects only are perceived.]
There is a huge amount of good engineering in a properly designed amp, of course, and you have demonstrated this well. But the laws of engineering alone do not apply; if the amp is to please, it should be musical too, whatever that means...... I would posit that EEs should understand something of the psychoacoustics of music as well.
To me this means that any distortions introduced by the amp should not be even faintly discordant; this would imply the lofty goal of removing all odd order distortions, where possible, or at the least taking them below -130dB for 5th, 7th, 9th, etc. The nature of global negative feedback makes this difficult because of the mixing at the error amplifier, but oddly enough the choice of components is also an issue.
This is not to refute you, or anyone else bent on designing a zero distortion amp, but given that some distortion is inevitable it could point the way to better engineering design.
It is also my contention that you do not need to be a degreed EE to design an audio amplifier. Plenty of designers come from other disciplines; JLH was a nuclear scientist, for example, and plenty have no relevant qualifications in this area. As has been pointed out here, this is not rocket science, at least at the basic design level. And a surprising number of basic designs - NP is the classic example - sound pretty damn good.
Hugh
Aska has it right , no rocket scientist's are needed here..
(ahhhh ,maybe a couple of them)
An Amplifier designer's main purpose is to please the listener.
Most amplifier design's I see at DIYaudio are just modifications
of classic designs by Douglas Self , Leach , and Sziklai.
(he has the patent on push/pull output designs.)
http://www.pat2pdf.org/patents/pat2762870.pdf
The speakers you listen to make a MUCH bigger difference
than the amp's sonic qualities. Not to say the amp doesn't
have a impact on the listening stage ,(A hitachi MOSFET lateral
amp sounds much different than a leach amp) but most
of this is the amps design and the radically different
outputs.
A little simple math, access to a database of designs/articles,
and with a little help from other crazy DIYer's there is no reason
one can not have safe ,reliable, and sometimes superior
equiptment at a fraction of the cost.
OS
(ahhhh ,maybe a couple of them)
An Amplifier designer's main purpose is to please the listener.
Most amplifier design's I see at DIYaudio are just modifications
of classic designs by Douglas Self , Leach , and Sziklai.
(he has the patent on push/pull output designs.)
http://www.pat2pdf.org/patents/pat2762870.pdf
The speakers you listen to make a MUCH bigger difference
than the amp's sonic qualities. Not to say the amp doesn't
have a impact on the listening stage ,(A hitachi MOSFET lateral
amp sounds much different than a leach amp) but most
of this is the amps design and the radically different
outputs.
A little simple math, access to a database of designs/articles,
and with a little help from other crazy DIYer's there is no reason
one can not have safe ,reliable, and sometimes superior
equiptment at a fraction of the cost.
OS
Hugh,
Maybe it's a disappointing confession for you, but I have no intention to listen to my designs. In the past I've listened to several low distortion amps AND speakers as well and, regrettably, it wasn't an exiting experience. They sounded just okay, but nothing special. And, as long as we are talking about hi-fi, opposed to 'high-end musical instruments', I think it should be that way. An amplifier should do just one thing: amplify without adding (or deleting) things that were not present in the original. Therefore, I have no high expectations when it comes to listening to one of my almost distortionless amps. That's the reality of true hi-fi. Nevertheless I enjoy designing that kind of stuff.
Regards,
Edmond.
Maybe it's a disappointing confession for you, but I have no intention to listen to my designs. In the past I've listened to several low distortion amps AND speakers as well and, regrettably, it wasn't an exiting experience. They sounded just okay, but nothing special. And, as long as we are talking about hi-fi, opposed to 'high-end musical instruments', I think it should be that way. An amplifier should do just one thing: amplify without adding (or deleting) things that were not present in the original. Therefore, I have no high expectations when it comes to listening to one of my almost distortionless amps. That's the reality of true hi-fi. Nevertheless I enjoy designing that kind of stuff.
Regards,
Edmond.
AKSA said:
Hugh,
Maybe it's just me, but I feel there's some inconsistencies in your post.
For example, getting those odd harmonics below -130dB, there's no way you can do that without feedback.
OTOH, those NP basic designs you admit sound pretty good, they have MUCH larger even and odd harmonincs than that.
So, I'm a little confused of what I should strive for now as a designer ...
Jan Didden
AKSA said:
Hugh,
First, I am having huge difficulties in accepting that a human can hear -130dB of 5th, 7th, 9th, etc... of distortions through speakers that have 1% (-60dB) distortions. Every way I'm looking at this issue, from a hearing physiology perspective to commons sense, I still get a :bs: alarm. I don't care if 1 out of 10 million can (after all, there are people with exceptional abilities everywhere, some can see the Mars satellites with the bare eye). It is perfectly fine to design an amp for these exceptional fellas, but clearly state this in your design goals. Otherwise, you are selling to the rest of 9,999,999 an illusion rather than something they could enjoy.
Secondly, your second statement about global negative feedback and mixing at the error amplifier is in general flat wrong. While it is certainly possible to build an amp in which closing the loop increases high order distortions here's where EE comes to the rescue; the analysis of a GNFB amp shows the designers how to avoid such pitfalls. In fact, all this :bs: about GNFB increasing high order distortions is coming from the JC's golden times (the 70's) when high speed devices were barely available, in particular in the US. It was then difficult to build a high speed amp, that would allow GNFB to properly do it's job. To his merit, JC understood this and successfully built high speed amps and without any Spice tools or aid. But there were a gazillion of poorly designed amps out there that brought GNFB a bad rap that continues, without any grounds, today. I dare to say it would be darn difficult today to build an amp with the same poor characteristics as in the early 70's.
Third, your statement "you do not need to be a degreed EE to design an audio amplifier" is probably true but dangerous to generalize. I'm suspecting the very aggressive (and sometimes insulting) statements coming from the subjective team are coming from a deep frustration of not really understanding what's happening in their amps. Their human need to be in control makes them invent reasons they can understand and follow (like the effect of the silver foil capacitor) as a compensation of their inability to understand the real limitations (e.g. transistor capacitance nonlinearities) and the ways to circumvent them. Unfortunately, there are a good number of snake oil merchadiser that know how to take advantages of these otherwise in perfectly good faith fellas, and here you get the $1000 power cable and $2000 capacitor markets. This species is mostly interested in spreading FUD, urban myths and black magic and this makes them in my eyes one of the worse kind of salesman. To add insult to injury, some are qualified engineers which are perfectly aware they are spreading :bs: which doesn't even give them the benefit of ignorance.
hi Hugh,
There is a lot of emphasis about spectra of harmonic distorsion and its correlation with subjective appraisals of amplifying circuits. I've seen some designers degrading their otherwise good designs by deliberalty favouring some kind of second order non-linearity in order to get H2 distorsion higher than H3. This is in accordance with a school of thinking around the eighties which said that it was better that way.
However, a second order non-linearity does not produce H2 distorsion only. It also induces a DC component which modulates the operating conditions of the whole amplifier. This can only enhance intermodulaton products. To me, it is an overlooked perverse effect of second order non-linearity.
There is a lot of emphasis about spectra of harmonic distorsion and its correlation with subjective appraisals of amplifying circuits. I've seen some designers degrading their otherwise good designs by deliberalty favouring some kind of second order non-linearity in order to get H2 distorsion higher than H3. This is in accordance with a school of thinking around the eighties which said that it was better that way.
However, a second order non-linearity does not produce H2 distorsion only. It also induces a DC component which modulates the operating conditions of the whole amplifier. This can only enhance intermodulaton products. To me, it is an overlooked perverse effect of second order non-linearity.
NFB vs. HEC vs. hoax
1. Please, no more subjectivistic babbles on this thread, as this one is about YAP.
2. I almost forgot to mention the perhaps most striking and important aspect of Ovidiu's YAP: This amplifier clearly demonstrates that simple NFB around the output stage can easily compete with error correction, or even produce better results (3 times less THD20 compared to a well known HEC amp). QED.
1. Please, no more subjectivistic babbles on this thread, as this one is about YAP.
2. I almost forgot to mention the perhaps most striking and important aspect of Ovidiu's YAP: This amplifier clearly demonstrates that simple NFB around the output stage can easily compete with error correction, or even produce better results (3 times less THD20 compared to a well known HEC amp). QED.
I was putting a different POV, expecting a huge caning, and instead I received nothing but courtesy......
Thank you gentlemen, this forum is certainly improving, I'm most grateful.
I do not choose to argue with anything said here; I do not know all the answers. However, a few points might be pertinent.
Years ago I designed an amp which deliberately introduced tiny amounts of H2, the AKSA, just as Sebastien alluded. People loved it, it sold like hotcakes. Distortion at full power into 8R at 20KHz was 0.045%, not too high, but much of it H2 and H3.
Then I felt I'd have a shot at a very low distortion amp. I worked hard, did my homework, lots of testing, careful work with the oscilloscope, checking stability margins, etc, and this amp was cleaner, but lost that ol' tube sound. Most people liked it, a few did not. I went further, and then further again, and now I have an amp with vanishingly low distortion, and the resolution and quietness is a revelation. I have not yet measured it, but I would expect around 0.003% at full power into 8R; every circuit block has been radically improved for linearity.
Some sort of threshold has been crossed. This amp has almost supernatural detail, and since the VAS is common base, it's very fast and very quiet. But I find it difficult to explain, particularly with my limited knowledge. I use nested feedback, phase lead, charge suckout in the output stage; it's fairly complex. I'm actually beginning to think there is something in this zero distortion approach; it has its merits, no question.
Thank you again Edmond, Jan, Syn08, Sebastien, for nice responses. I think it is still possible to make serendipitous discoveries in this technology; I honestly feel privileged to have taken aim at a zero distortion amplifier.
Hugh
PS. Edmond, I quite understand that you are not particularly interested to listen to your amps. That's cool. For my part, I will happily continue to design amplifiers even if I go deaf.......
Thank you gentlemen, this forum is certainly improving, I'm most grateful.
I do not choose to argue with anything said here; I do not know all the answers. However, a few points might be pertinent.
Years ago I designed an amp which deliberately introduced tiny amounts of H2, the AKSA, just as Sebastien alluded. People loved it, it sold like hotcakes. Distortion at full power into 8R at 20KHz was 0.045%, not too high, but much of it H2 and H3.
Then I felt I'd have a shot at a very low distortion amp. I worked hard, did my homework, lots of testing, careful work with the oscilloscope, checking stability margins, etc, and this amp was cleaner, but lost that ol' tube sound. Most people liked it, a few did not. I went further, and then further again, and now I have an amp with vanishingly low distortion, and the resolution and quietness is a revelation. I have not yet measured it, but I would expect around 0.003% at full power into 8R; every circuit block has been radically improved for linearity.
Some sort of threshold has been crossed. This amp has almost supernatural detail, and since the VAS is common base, it's very fast and very quiet. But I find it difficult to explain, particularly with my limited knowledge. I use nested feedback, phase lead, charge suckout in the output stage; it's fairly complex. I'm actually beginning to think there is something in this zero distortion approach; it has its merits, no question.
Thank you again Edmond, Jan, Syn08, Sebastien, for nice responses. I think it is still possible to make serendipitous discoveries in this technology; I honestly feel privileged to have taken aim at a zero distortion amplifier.
Hugh
PS. Edmond, I quite understand that you are not particularly interested to listen to your amps. That's cool. For my part, I will happily continue to design amplifiers even if I go deaf.......
Jan,
I would correct you - I am extremely uncool. I have three skills; topology, layout and listening, and I push all to the limit. But it is pretty much my vocation and obsession and I enjoy it. I do sometimes wish I had a decent income, though.
I admire much of the analysis here, but I would enjoin talented people to build and listen, and take their pure intellectual pursuit to some sort of consummation - it is enormously satisfying, and beats argueing, though some debate is necessary to exchange ideas.
Hugh
I would correct you - I am extremely uncool. I have three skills; topology, layout and listening, and I push all to the limit. But it is pretty much my vocation and obsession and I enjoy it. I do sometimes wish I had a decent income, though.
I admire much of the analysis here, but I would enjoin talented people to build and listen, and take their pure intellectual pursuit to some sort of consummation - it is enormously satisfying, and beats argueing, though some debate is necessary to exchange ideas.
Hugh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.