§§§
In post 149, I wrote -
"I recently got another idea which I have rarely seen done : a push-pull class A buffer using compound pair in each branch driving simple power NPN and PNP emitter followers. As there is a lot of current to charge and discharge the power transistors, the output should be fast and crossover distorsion low"
In fact my idea in a bit different. It is not a push-pull but a variation of the type III configuration described in Self's book, 3rd edition, page 113, fig 5.4C. In this configuration, the drivers are emitter followers with an emitter resistor connected to the opposite power supply (NPN with an Re connected to -, and collector connected to +, PNP with an Re connected to + and collector connected to -). Each driver is always conducting. What I was thinking of was to replace these emitter followers with Sziklai pairs.
MIKEB
"Most of the symasym is not based on science then rather changing and listening."
Most amplifiers mainly developped with listening show more distorsion than those develpped with science. There is only one explanation : their authors prefer a bit more distorsion, because it makes sound more alive or more euphonic. This approach may be perfectly but it's a philosophy of musical instruments makers. The result are amplifiers which are said musical but they are not stricly neutral as they should be. So, perfection cannot invocated about them.
~~~~~~~ Forr
§§§
In post 149, I wrote -
"I recently got another idea which I have rarely seen done : a push-pull class A buffer using compound pair in each branch driving simple power NPN and PNP emitter followers. As there is a lot of current to charge and discharge the power transistors, the output should be fast and crossover distorsion low"
In fact my idea in a bit different. It is not a push-pull but a variation of the type III configuration described in Self's book, 3rd edition, page 113, fig 5.4C. In this configuration, the drivers are emitter followers with an emitter resistor connected to the opposite power supply (NPN with an Re connected to -, and collector connected to +, PNP with an Re connected to + and collector connected to -). Each driver is always conducting. What I was thinking of was to replace these emitter followers with Sziklai pairs.
MIKEB
"Most of the symasym is not based on science then rather changing and listening."
Most amplifiers mainly developped with listening show more distorsion than those develpped with science. There is only one explanation : their authors prefer a bit more distorsion, because it makes sound more alive or more euphonic. This approach may be perfectly but it's a philosophy of musical instruments makers. The result are amplifiers which are said musical but they are not stricly neutral as they should be. So, perfection cannot invocated about them.
~~~~~~~ Forr
§§§
This doesn't surprise me at all. I've read more than one account relating to how consumer listening test show a preference for something decidedly less than ideal. In speakers its not just5 distortion but frequency response.
I suspect a great deal of this relates to what one has become accoustomed to. The most common example I've seen is when someone unfamiliar with true subwoofers (in contract to seperate "bass modules") is presented with the difference between exaggerated (or boosted) bass and extended bass. It can take a while to figure out "get it" with regard to what is actually going on. I would surprised if this doesn't work for distortion as well. If you are accustomed to significant low order distortion, it's absence may be disconcerting. You may or may not alter your preferences as a result.
Ho forr !
I think i never mentioned the symasym to be a "perfect" amplifier,
(i hope so) it's just a nice good sounding one, not sounding coloured
or obviously changing something like you realize it with the DoZ.
It's good possible that it has plenty of distortion, according to sims
not, but i should try measuring realworld.
Mike
I think i never mentioned the symasym to be a "perfect" amplifier,
(i hope so) it's just a nice good sounding one, not sounding coloured
or obviously changing something like you realize it with the DoZ.
It's good possible that it has plenty of distortion, according to sims
not, but i should try measuring realworld.
Mike
How to determine (or calculate from schematic) the output impedance of VAS? Doug Self said the output impedance of typical VAS is about 22kohm.
Is it possible to get like 20ohm-50ohm output impedance of VAS without using additional transistors / Emittor Follower?
Hi, FORR
Is it possible to get like 20ohm-50ohm output impedance of VAS without using additional transistors / Emittor Follower?
Hi, FORR
Is this driver is the one that Electrocompaniet have been using all this time? The latest Electrocompaniet is using CCS instead of R for biasing from the opposite rail supply. See AW120 schematic around here somewhere.
No!
You can decrease it, when increasing the current in the VAS.
It depends a lot on your VAS transistors. 22k seems quite low to me. My opinion is that it is more like 50-100k. The higher the better. This can obtained by low-beta drivers or say better selecting your transistors for a high early voltage. Sometimes you find the output resistance in the datasheet. Using a tripple EF is better than try to decrease the VAS output resistance. Or see the solutions FORR has given.
Actually the preference for less than ideal amps goes a little deeper.
Firstly, the "ideal" reference amp in all likelyhood is not ideal. It just has lower distortion, in level, but it will have *some* distortion *somewhere*.
While some may indeed prefer true euphonics, I find it entirely credible that some amps w/o obvious coloration may sound better on the basis of the distortion spectrum and nature. The crux is to find out why people have these preferences.
Research indicates that the ear itself is highly nonlinear and therefore creates its own distortion. Musical instruments of course generate harmonics plentiful. And there is evidence that certain distortion distributions simply mask each other because they mimic either the ear's or the typical musical instrument's harmonic better than other amps. So, the "more accurate" amp may have low level distortions that stick out like a sore thumb, while the "musical" amp may have slightly higher level distortions which, however, don't stick out. Given the level of speaker distortion, one would think that anything below 0.1% THD should be indistinguishable, it seems not so though. Hence Distortion spectrum should matter more than distortion level.
Ideally one would like both low level distortions (because any distortion will contaminate the audio band and contribute fuzziness) yet of a nature that doesn't make them stick out. Since we seem to have very competent engineers here on this forum that handle the pros and cons of the technology really well, what we need most is actually not more engineering approaches, but a better understanding of what is objectionable to the ear, and why. The engineers can then find a technical implementation that fits the bill.
Note: as I indicated in some post above, even very high order distortion of a very low level may be objectionable if produced in the bass region, simply because it will still be in the audio band and in a much more sensitive area of the equal loudness curve.
As to the figures we can see showing essentially no harmonics above the noise floor, I'd really like to see some cross correlation analysis of the said noise floor to check if it doesn't hide perceptible high order harmonics...
Just rambling (since I can't contribute formulas)
😉
Firstly, the "ideal" reference amp in all likelyhood is not ideal. It just has lower distortion, in level, but it will have *some* distortion *somewhere*.
While some may indeed prefer true euphonics, I find it entirely credible that some amps w/o obvious coloration may sound better on the basis of the distortion spectrum and nature. The crux is to find out why people have these preferences.
Research indicates that the ear itself is highly nonlinear and therefore creates its own distortion. Musical instruments of course generate harmonics plentiful. And there is evidence that certain distortion distributions simply mask each other because they mimic either the ear's or the typical musical instrument's harmonic better than other amps. So, the "more accurate" amp may have low level distortions that stick out like a sore thumb, while the "musical" amp may have slightly higher level distortions which, however, don't stick out. Given the level of speaker distortion, one would think that anything below 0.1% THD should be indistinguishable, it seems not so though. Hence Distortion spectrum should matter more than distortion level.
Ideally one would like both low level distortions (because any distortion will contaminate the audio band and contribute fuzziness) yet of a nature that doesn't make them stick out. Since we seem to have very competent engineers here on this forum that handle the pros and cons of the technology really well, what we need most is actually not more engineering approaches, but a better understanding of what is objectionable to the ear, and why. The engineers can then find a technical implementation that fits the bill.
Note: as I indicated in some post above, even very high order distortion of a very low level may be objectionable if produced in the bass region, simply because it will still be in the audio band and in a much more sensitive area of the equal loudness curve.
As to the figures we can see showing essentially no harmonics above the noise floor, I'd really like to see some cross correlation analysis of the said noise floor to check if it doesn't hide perceptible high order harmonics...
Just rambling (since I can't contribute formulas)
😉
LUMANAUW
"How to determine (or calculate from schematic) the output impedance of VAS? Doug Self said the output impedance of typical VAS is about 22kohm."
I think this is still valid : the output impedance of a simple collector is about 100 kOhm/mA, wich means 100 kOhm for 1 mA, 20 kOhm at 5 mA and 10 kOhm for 10 mA.
However, with the Miller feedback effect, this impedance is reduced as frequency increases. I am not going to dig in the Self's book again tonnight , but I think the mentionned 20 kOhm is the average load on the VAS collector with 8 Ohm at the output of the amp, not the output impedance (I am not sure I am right).
LUMANAUW
"Is this driver is the one that Electrocompaniet have been using all this time? The latest Electrocompaniet is using CCS instead of R for biasing from the opposite rail supply. See AW120 schematic around here somewhere."
In the schematics you sent, my idea is effectively to replace T12 and T15 by Sziklai pairs. Constant current sources can replace the resistors connected to the supply rails.
A variant wich was used in the first solid-state amps made by Audio-Research (yes, the tube star) : the PNP driver feeds the NPN output and the NPN driver feeds the PNP output. The load of each driver is a contant current source, with an insertion of a small value resistor before the base of the output power transistor. The voltage across the small value resistors determine the quiescent current in the output transistors. Same as above : the drivers could benefit of Sziklai pairs.
~~~~~~~~~ Forr
§§§
"How to determine (or calculate from schematic) the output impedance of VAS? Doug Self said the output impedance of typical VAS is about 22kohm."
I think this is still valid : the output impedance of a simple collector is about 100 kOhm/mA, wich means 100 kOhm for 1 mA, 20 kOhm at 5 mA and 10 kOhm for 10 mA.
However, with the Miller feedback effect, this impedance is reduced as frequency increases. I am not going to dig in the Self's book again tonnight , but I think the mentionned 20 kOhm is the average load on the VAS collector with 8 Ohm at the output of the amp, not the output impedance (I am not sure I am right).
LUMANAUW
"Is this driver is the one that Electrocompaniet have been using all this time? The latest Electrocompaniet is using CCS instead of R for biasing from the opposite rail supply. See AW120 schematic around here somewhere."
In the schematics you sent, my idea is effectively to replace T12 and T15 by Sziklai pairs. Constant current sources can replace the resistors connected to the supply rails.
A variant wich was used in the first solid-state amps made by Audio-Research (yes, the tube star) : the PNP driver feeds the NPN output and the NPN driver feeds the PNP output. The load of each driver is a contant current source, with an insertion of a small value resistor before the base of the output power transistor. The voltage across the small value resistors determine the quiescent current in the output transistors. Same as above : the drivers could benefit of Sziklai pairs.
~~~~~~~~~ Forr
§§§
About engineers and solutions for high end audio:
see here:
http://www.biline.ca/critic4.htm
Pls read to the end.
Where to start?
see here:
http://www.biline.ca/critic4.htm
Pls read to the end.
Where to start?
Try to click the 'next page' link at the bottom 😀
Two channel stereo is as dead as the LP!
Well i guess that's true in the sense that both are alive and kickin' 😉
I guess you should read this content, knowing this is not a scientific or organizational website, but the homepage of a commercial speaker manufacturer.
Two channel stereo is as dead as the LP!
Well i guess that's true in the sense that both are alive and kickin' 😉
I guess you should read this content, knowing this is not a scientific or organizational website, but the homepage of a commercial speaker manufacturer.
Jorge said:
I don't know Jorge. But I particularly noted this section:
"It is a reasonable assumption that a superior audiocircuit concept would slowly but surely recruit a much larger following than an inferior one, so that eventually there would be some kind of consensus among practitioners and a discernible convergence toward the superior topology as new designs emerge. Total randomness in the choice of topology would indicate that no single approach is clearly superior to any other, in much the same way as total randomness in the results of a double-blind ABX test indicates that there is no clearly audible difference between A and B. It’s basically the same statistical criterion. Thus, if a designer had discovered a unique topology that sounded better but did not measure better, we would expect that topology to spread to other companies through “reverse engineering” of the product with the superior sound. There is no evidence whatsoever that this is actually happening. ".
This is something that I have been turning over in my mind for some time. I guess it means we are just having fun rather than making any progress.
Jan Didden
janneman said:
Excellent article, a must read for serious people genuinely striving for progress.
Yes Jan, in a way you may conclude that. Another way of looking at it and from the perspective you suggest, is that many people are tweaking and experimenting to get a personaly satisfactory way of listening, not necesarily accurate reproduction. Given the wide diversity in tastes it is to be expected satisfactory solutions are equally diverse.
Rodolfo
I say again : all this searching for objective perfect amp is only searching for amp with the best " masking effect ", 'cos as signal source is used in most of cases signal which came from AD/DA process. If you will use for testing signal from mic preamp or from good mixing console, probably results will be quite different and differences will be not so great. Try it and you will see, whatabout I'm talking.
I know it's controversial for an audio designer like myself to participate with any kind of opinion about this article, because it expresses the 'we know how you should do your work' advice for us designers.
The article appears to be interesting, in the sense that it appears to be a scientific whitepaper (also with an unclear hint that it might be a PhD study).
However the paper is also incredibly worthless, in the sense that is just a continuation of the endless search for a scientific solution to the 'perfect amplifier'. The socalled 'Black Box' theory, that has been proven wrong many many times over the years.
In a way they have even proved in the very paper that there is no result that fit's their theory, about the best amplifier type should have achieved the largest number of followers. This is clearly frustrating to the authors, that their theory and the real world and all the testers and reviewers don't match up.
And so they conclude:
Our theory is correct, and therefore the real world and the testers and the users are wrong.
It is largely the same as if two rookie students try to find the perfect painting, based on statistic sales figures, and pepper it up with how the artists and art reviewers work. Some kind of technique and education skills is involved in every painting of course, but they will find that the worlds art critics don't agree with their statistic results every time. And so to give worth to their white paper study, they conclude that all the art critics are wrong, and unserious. And the people who buy and enjoy the art are ignorant. And the artists should always use red, yellow and black, because these colors match with the most interiors, statistically.
But this paper actually has one good piece of information in it, namely that Bob Carver knew very well what he was doing!
My comment to the whole thing is, the Black Box theory doesn't hold up, any more than there is one painting that is the best in the world.
It depends on people's taste, and preferences, that is why there are so many different amplifiers out there, and not one that can be pronounced the best of all. To pronounce the reviewers unserious, and the users all wrong i think is very wrong, and tells more about these two authors than the peoples credibility they try to smear.
The amplifier's sound can be designed by the expert audio designer, to match the user's preferences, and particular properties can be embedded in the amplifier's circuits. This is exactly what Bob Carver (in this paper) mastered to perfection, and what the authors of the paper completely ignores. They seem to be blinded by the quest to find a particular conclusion. But what the h**k many so called 'white papers' are written with a conclusion first, and then filled up with more or less solid arguments to prove the conclusion. So that's nothing special. Just don't trust anything you read, not even when it appears to be a scientific white paper.
The article appears to be interesting, in the sense that it appears to be a scientific whitepaper (also with an unclear hint that it might be a PhD study).
However the paper is also incredibly worthless, in the sense that is just a continuation of the endless search for a scientific solution to the 'perfect amplifier'. The socalled 'Black Box' theory, that has been proven wrong many many times over the years.
In a way they have even proved in the very paper that there is no result that fit's their theory, about the best amplifier type should have achieved the largest number of followers. This is clearly frustrating to the authors, that their theory and the real world and all the testers and reviewers don't match up.
And so they conclude:
Our theory is correct, and therefore the real world and the testers and the users are wrong.
It is largely the same as if two rookie students try to find the perfect painting, based on statistic sales figures, and pepper it up with how the artists and art reviewers work. Some kind of technique and education skills is involved in every painting of course, but they will find that the worlds art critics don't agree with their statistic results every time. And so to give worth to their white paper study, they conclude that all the art critics are wrong, and unserious. And the people who buy and enjoy the art are ignorant. And the artists should always use red, yellow and black, because these colors match with the most interiors, statistically.
But this paper actually has one good piece of information in it, namely that Bob Carver knew very well what he was doing!
My comment to the whole thing is, the Black Box theory doesn't hold up, any more than there is one painting that is the best in the world.
It depends on people's taste, and preferences, that is why there are so many different amplifiers out there, and not one that can be pronounced the best of all. To pronounce the reviewers unserious, and the users all wrong i think is very wrong, and tells more about these two authors than the peoples credibility they try to smear.
The amplifier's sound can be designed by the expert audio designer, to match the user's preferences, and particular properties can be embedded in the amplifier's circuits. This is exactly what Bob Carver (in this paper) mastered to perfection, and what the authors of the paper completely ignores. They seem to be blinded by the quest to find a particular conclusion. But what the h**k many so called 'white papers' are written with a conclusion first, and then filled up with more or less solid arguments to prove the conclusion. So that's nothing special. Just don't trust anything you read, not even when it appears to be a scientific white paper.
Lars,
Have you read another article??
The thing I recognise is your painting analogy. I agree, just like there are many good paintings, just like it is impossible to paint "the" best painting, so there are many different tastes in audio that all are served by a school of design. There is no "best" amplifier, I agree. Each designer has his own taste that he tries to match to an audience.
But that also means that we stop to say that this or that amp is the "best". We should stop saying that no fb is best, or high fb is best, because there is no "best" defined. Unless you define "best" as as little as possible signal coloration. That is measurable and if we would adopt that as quality criterium, it would be the end of audio specialty manufacturers. That criterium is what the authors used, and their conclusion is flawless USING THAT CRITERIUM. But, again, using a criterium "I like it" (or I don't like it) you get a different conclusion.
And an unfortunate byproduct of "anything goes" is indeed ruthless scams to take the money from the unwary.
Jan Didden
Have you read another article??
The thing I recognise is your painting analogy. I agree, just like there are many good paintings, just like it is impossible to paint "the" best painting, so there are many different tastes in audio that all are served by a school of design. There is no "best" amplifier, I agree. Each designer has his own taste that he tries to match to an audience.
But that also means that we stop to say that this or that amp is the "best". We should stop saying that no fb is best, or high fb is best, because there is no "best" defined. Unless you define "best" as as little as possible signal coloration. That is measurable and if we would adopt that as quality criterium, it would be the end of audio specialty manufacturers. That criterium is what the authors used, and their conclusion is flawless USING THAT CRITERIUM. But, again, using a criterium "I like it" (or I don't like it) you get a different conclusion.
And an unfortunate byproduct of "anything goes" is indeed ruthless scams to take the money from the unwary.
Jan Didden
Jan: I agree 100% with you. We should say : I think nfb is better than fb. for example.
But there is no absolute truth.
But there is no absolute truth.
- Status
- Not open for further replies.