My old Perreaux, which I still have, is a pretty good cooker - the heatsinks ran at about 60degC. In raw form, it did a very nice job while it stayed in class A, but the limitations of the power supply crippled it quickly once a bit of class AB was called for - I spent a lot of time redoing, and redoing, the power supply so that it had a decent reserve of energy - a technician would have a heart attack if he saw the insides ...
I just did that !
My girlfriend buys washing powder ( Daz , Dash in France ) . She is as poor as a church mouse and fiercely independent . When asking why not buy cheaper products in liquid form she said " because powder grinds out the dirt " . I laughed , she laughed back and said it showed how little I knew . This morning reading this thread I thought of this and said how do I know she is wrong ?
She is a brand snob . If I bought her the worst possible Champagne she will jump with joy . If I buy the best Cava she is not so happy . If I bought her Riesling she might poor it down the toilet . She has many bits of paper so is no fool . I challenged her to come on a blind sparkling wine tasting . She looked very worried . What if she likes the Riesling ! As for these !!!!!
This vineyard is near me . I won't be buying Colleen any I guess .
http://www.youtube.com/watch?v=mfMfHFoOhH4
Last edited:
Only by ear. Let's say, you have a recording of solo piano: I have heard many times the typical standard of reproduction of that instrument. We have a real piano in the room, I have numerous times experienced close proximity to other, nominally quite superior pianos, I have a strong idea in my head what the subjective impression should be, especially when played vigorously. If a system can't convey that impact and tonality then I would judge the 'performance' to be inadequate ...
With regard to connections, the problem seems to be the sensitivity of the ear, its ability to adapt extremely rapidly to changing levels, to pick out meaning amongst conflicting input. Instrumentation quite often effectively uses time averaging, and only needs to deal with accuracies of the order of 60dB within a range - once you start asking for 80 or even 100dB accuracies things start to become much, much harder - the slightest perturbation can dramatically affect the reading.
Unusual for amps to be unstable without global NFB? I think the opposite is more likely to be true, unless you have an open loop full power bandwidth below say 1 kHz. In my approach, that is WAY greater.
Also, you AIM for 20 dB of global NFB, but exactly how much you end up using has to be confirmed by live samples, no math can assure a safe amount except by fluke, in which case you never needed the math anyway. But in general, I do get it right to 20 dB +/- 2 dB.
As for "it says so in the books", well, it was said in books that the Earth was flat for much longer that it has been recognized that it ain't necessarily so, to quote Bronsky Beat.
You know, every time I learnt the answer to a question which bothered me, that opened up just more questions and made me realize how much more I do not know yet, and probably never will. That never stopped me from searching anew, though. Heck, that's what I do it for, to learn something new.
While you mention what is indeed standard book fare, it does not mean that it's all unconditionally true. I see it as the quick'n'dirty way of doing things, even if some amps built around that doctrine did sound wonderful. Exactly the same could be said of my approach: it makes better than average sound more probable than otherwise, but does not guarantee it.
Communications receivers and active antennas routinely cope with dynamic ranges of 90 or 100dB, with no time averaging.fas42 said:
The comparison seems to be between your sighted subjective impressions and measurable objective criteria. Which should we place more faith in? If your ears really are as good as you believe, then you ought to make your money by hiring them out to people developing sonar etc. They could then scrap all their expensive test equipment and expensive computer programmers etc. and just rely on your opinion to develop and debug the next generation.
For most SS amps the open loop bandwidth may be much less than 1kHz. In any case, are you confusing bandwidth with stability? Sometimes the output stage may be unstable with certain loads and global NFB might tamr this, but this is quite unlikely because the frequency of instability is often way too high for the loop to do much about. An amp like this could have very strange clipping behaviour. No, most SS amps are perfectly stable open loop but adding GNFB can reduce stability so compensation has to be added.dvv said:
Now you may design unusual amps which are unstable open loop but that is not the norm.
Actually, the Earth has been known to be spherical for rather a long time. Like servo/feedback theory, this is open to scientific investigation. In every field there are always a few people who think the basic facts are wrong, or don't apply to them, but they or their customers usually find that Nature cannot be fooled.
Last edited:
Everyone who has no trouble deciding, when they hear music playing and don't see what it comes from, whether it's the 'real thing' or merely reproduction is up to the mark. The little clues are easily picked up, register with your memory bank of past experiences, and decision made with barely the slightest twitch of the conscious mind.
Experiments could easily degenerate into a bizarre parody of instinctive listening, like forcing a person to fully engage the rational, thinking mind to walk -- okay, switch this muscle on for 0.2 secs, wait for that nerve trigger to fire, OK, twist this bone 5 degrees to the right, and apply 2.4 Newtons force to the balancing sinew ... oh dear, oh dear, I'm falling over ... why did that happen ...?!!
Experiments could easily degenerate into a bizarre parody of instinctive listening, like forcing a person to fully engage the rational, thinking mind to walk -- okay, switch this muscle on for 0.2 secs, wait for that nerve trigger to fire, OK, twist this bone 5 degrees to the right, and apply 2.4 Newtons force to the balancing sinew ... oh dear, oh dear, I'm falling over ... why did that happen ...?!!
Last edited:
Walking is a good analogy to listening. Can a person actually get from here to there while remaining upright and unaided? If they can only 'walk' from here to there while the tester has his eyes and ears shut then there will always be the suspicion that they actually crawled or used motorised transport, while vehemently maintaining that they walked it. Being asked to walk a short distance does not create stress in adults, except for those who actually can't walk!
Similarly, those who claim to hear things should be able to demonstrate this in the absence of other clues. If not we are justified in expressing doubts about their hearing, however sincerely they believe it themselves.
As it is, amps with poor circuits and poor performance but high price tags are often found to 'sound better' than other amps with good circuits and low price tags, but only when the extra information is available to the 'listener'. It doesn't take a genius to work out what is probably happening, especially as this phenomenon is so well-established that it even has a name.
Similarly, those who claim to hear things should be able to demonstrate this in the absence of other clues. If not we are justified in expressing doubts about their hearing, however sincerely they believe it themselves.
As it is, amps with poor circuits and poor performance but high price tags are often found to 'sound better' than other amps with good circuits and low price tags, but only when the extra information is available to the 'listener'. It doesn't take a genius to work out what is probably happening, especially as this phenomenon is so well-established that it even has a name.
Apples and oranges. The harmonics of a radio carrier are far outside the operating bandwidth of most devices. A radio station at 96.9 MHz occupies a max +- 0.1 MHz. The first rectification harmonic would be 193.8 MHz, invisible to receivers. Many broadcast transmitters as well have very dirty modulation schemes but the 'distortion', falling in a complete different radio band, is easily filtered by the output section/broadcast antenna system.
Precisely the opposite. If I can't hear something in an ears-only test, well, that's that. If someone else can, that's fine, I don't see the problem. It means that the phenomenon in question is generally audible, but not to me, and that I need to use a second pair of ears if it's a problem I want to tackle. If I can't hear something in an ears-only experiment and someone else merely asserts that they can, that tells me absolutely nothing about the general audibility.
My daughter can hear a 20kHz tone being switched on and off. I can't. This does not keep me awake at night worrying, it's just that I can't/don't take my own evaluations of 20kHz performance very seriously.
Sigh. Go and read about the effect of third-order intermodulation on RF communications systems. Harmonics are a much smaller issue.rdf said:
No. Broadcast transmitters usually have extremely clean modulation; often a statutory requirement. You appear to be confusing modulation distortion and RF harmonic generation.
I comopletely accept your above comment, SO LONG AS IT APPLIES TO YOU ONLY. If you feel the need, by all means, do experiment to your heart's content. It can't hurt and it may evetually be useful.
But if someone else CAN hear say 18 kHz, and does detect a difference between A and B, while measurements demonstrate all but identical behavior of both A and B, then I do not see the relevance of your tests except to your own persona. The end result will still be determined by one's hearing.
And indeed, there was a time I could reliably detect 18 kHz even if I was not a child any more (I was about 30 then). Today, I am 60 and I can reliably hear only 15 kHz. I am sure this will not last long, it will necessarily go down as I age. No-one escapes the biological grindstone.
If someone claims to be able to hear a difference between A and B, I will believe him/her only if she can repeatedly do so 8/10 times in several sittings.
Regarding what you called "other clues", such as prices and name tags, I completely agree with you that this can and does heavily influence results if revealed. That is the unfortunate truth, we are all subject to economic pressure exerted overtly or covertly by manufacturer advertising.
It has by now far gone from pure economics, it's all about "precieved value" these days. Like tall speakers with 3 6" "bass drivers" of highly questionable quality overall, but hey, they make the box bigger and they do look a hell of a lot cooler than just one such driver in another box. Which could just as well cost more than those 3 together simply because it is a high quality product. Nobody, not even professional testers, care about such things.
These days, it getting more and more what it LOOKS like it could be rather than what it IS actually like. Our entire culture relies more and more on props rather than substance. Very MATRIX like.
That's what I said. BUT... that "someone else" has to actually HEAR it, that is EARS-ONLY.
If you don't bother testing ears-only, there's absolutely no basis for any sonic claims and you're highly likely to come to totally incorrect conclusions. That's fine for the fashion audio industry (which runs on stories and belief), and gives one much fodder for writing long meandering posts full of pseudoscientific terms on internet message boards, but no serious researcher will give it any credence whatsoever.
I tend to make measurements with designs that have proven successful over the decades. I seldom 'listen' to my new designs, however, I have others do the listening for me, in comparison to what the rest of the world offers in performance and price range.
If I were to help a newbe designer, perhaps with some courses in engineering under his belt, I would have to explain the reasons for many more measurements, than I actually need for my recent designs.
This is because: MISTAKES in design can be avoided, if you know what to watch out for, through understanding and experience, so these potential 'mistakes' can be avoided in the design phase, and therefore there is little need to test for them.
Let me give an example: I have sometimes recommended the Hirata Test, developed in about 1980, by a Dr. Hirata in Japan, yet I do not test for this with my new designs. Why? Because, working with the Hirata Test, 30 years ago, I found a problem with one of my designs that created Hirata Distortion, and I now avoid it, through design choices.
Most of you don't know the Hirata Test, and therefore what to avoid, so the majority here, are taking a chance, IF they follow some of the many paths in amp design to have Hirata Distortion. That is your problem, because you don't take me seriously.
An area that I am pretty good at testing for is: higher order harmonic distortion. That is because I went out of my way to get the added test equipment to make the distortion more sensitive, and to remove the noise, that always comes in at low levels and wide bandwidth. I am particularly sensitive to 7th and 9th harmonic, but 10th,11th are probably just as bad, but in MY designs they don't really get generated enough to register on my test equipment. I have found this test, and the optimization of a power amp's output stage current to be a significant help in removing listening fatigue from power amps. Yet, I still get 'static' when I mention it here.
The same goes for TIM. My designs are TIM free, because they are too fast to create it.
Yet some 'smart guy' will tell you that since 20KHz is all you are supposed to hear, then you don't need a high slew rate, and some IC power amp is more than good enough in slew rate. These are the products that need the TIM test! '-)
Then, once a completed, it is reviewed by the marketplace, and I get my feedback from them. So far, in the last 15 years, since the CTC Blowtorch was developed, I have been successful with just about every design that I have produced in the audio marketplace. In fact, it is the CTC project that allowed me to try and 'fix' a number of minor mistakes that made my designs, 'B' at best, and not 'A'. It is NOT luck, or even 'image', it is just following design rules that I have found to work.
If I were to help a newbe designer, perhaps with some courses in engineering under his belt, I would have to explain the reasons for many more measurements, than I actually need for my recent designs.
This is because: MISTAKES in design can be avoided, if you know what to watch out for, through understanding and experience, so these potential 'mistakes' can be avoided in the design phase, and therefore there is little need to test for them.
Let me give an example: I have sometimes recommended the Hirata Test, developed in about 1980, by a Dr. Hirata in Japan, yet I do not test for this with my new designs. Why? Because, working with the Hirata Test, 30 years ago, I found a problem with one of my designs that created Hirata Distortion, and I now avoid it, through design choices.
Most of you don't know the Hirata Test, and therefore what to avoid, so the majority here, are taking a chance, IF they follow some of the many paths in amp design to have Hirata Distortion. That is your problem, because you don't take me seriously.
An area that I am pretty good at testing for is: higher order harmonic distortion. That is because I went out of my way to get the added test equipment to make the distortion more sensitive, and to remove the noise, that always comes in at low levels and wide bandwidth. I am particularly sensitive to 7th and 9th harmonic, but 10th,11th are probably just as bad, but in MY designs they don't really get generated enough to register on my test equipment. I have found this test, and the optimization of a power amp's output stage current to be a significant help in removing listening fatigue from power amps. Yet, I still get 'static' when I mention it here.
The same goes for TIM. My designs are TIM free, because they are too fast to create it.
Yet some 'smart guy' will tell you that since 20KHz is all you are supposed to hear, then you don't need a high slew rate, and some IC power amp is more than good enough in slew rate. These are the products that need the TIM test! '-)
Then, once a completed, it is reviewed by the marketplace, and I get my feedback from them. So far, in the last 15 years, since the CTC Blowtorch was developed, I have been successful with just about every design that I have produced in the audio marketplace. In fact, it is the CTC project that allowed me to try and 'fix' a number of minor mistakes that made my designs, 'B' at best, and not 'A'. It is NOT luck, or even 'image', it is just following design rules that I have found to work.
You're confusing what happens at the output flange with what happens in the modulation stages. The first is a regulatory requirement, the second isn't. Reread my post.
I have. You said:rdf said:
Poor modulation can create problems in adjacent channels in the same band, not different bands. The output filter is unlikely to stop this, unless it is unusually narrow (and therefore probably lossy). Problems in different bands come from other causes, such as RF harmonics and IM in the final PA - the output filter can deal with these.
However, this is a side issue. The main point is that RF systems are typically much more sensitive to IM than audio systems, yet don't have the claimed problems with less than perfect connectors. Bad connectors can create problems, of course, but that is not disputed.
- Status
- Not open for further replies.