GRollins said:
Hi Grey,
This all makes sense. Very plausible cause and effect. The only question is how big is the effect. Everything you have described, right out of electronics 101, should be eminently measurable. We probably all need to try harder to measure this stuff.
Has anyone here done any such measurement experiments? As a start, maybe expose an amplifier to high SPL over a varying frequency range, and see if any effect is een as a signal voltage at the output.
Cheers,
Bob
Bob Cordell said:
I stumbled on this effect by coincidence a few years back,
when testing an amp like Didden did. I got a bit curious, and tried to figure what made all this noise, as I could simply hear the music through some component. It seemed to come from the capacitors.
Later on I sent a 100Hz sine through and replaced the dummy load with speakers.....tapping the capacitors with a pencil gave a ticking sound, even with the amp sweating over the 100Hz sine at full blast.
So, my conclusion was that a chassis must protect the circuit to some extent against vibrations....
Magura 🙂
fizzard said:
One of our biggest problems in sending 10 Gb/s or 40 Gb/s per wavelength across the ocean for up to 10,000 km is the nonlinearity of the glass optical fiber. Over that distance, when you have, say, 64 wavelengths on a single fiber, the nonlinear effects of various kinds do add up and are very measurable.
On the other hand, we also seek dielctrics that are optically nonlinear in order to make modulators and the like. Lithium Niobate is one of them.
Cheers,
Bob
Many thanks to Grey in his post #3673 and all who replied. I encourage anyone to speak with Grant Carpenter of Gordon Instruments on the subject. He works very hard on designing to minimize induced currents, and has the results to prove it: nearly unmeasurable channel separation on his stereo micpre (<-140dB @ 200kHz!), as well as other great measurements.
OK here is my $.02: I've spoken to many estimable folks who've sworn up and down that microphonics and induced currents haven't shown up on anybody's scope or AP. Also I've talked to a number who have seen effects. I haven't searched all of AES papers on the subject because life goes on. But I'll be astonished if the general effect isn't measurable. I see 3 big suspect types: 1) transformer mechanical hum, 2) external vibrations reaching circuit components, and 3) the sum of all the modulation of the B and E fields (Grey's point) from typical operation of the circuit.
My old company put its money (enough to kill the outfit 🙁 ) into chassis construction to reduce internal and external vibrations. We also used enormous quantities of $$ monolithic ceramic C0G ceramic caps not merely for their temperature stability, but for their sound quality, esp. when used as coupling caps. These caps which we built into ladders (as many 80 per) of nearly zero inductance. C0Gs (=NPOs), unlike X7Rs and Z5Us, are generally considered not to be self-microphonic. Why? Well, they are thick and heavy relative to the others, though the dielectric's other mechanical properties may also affect that. Also we potted some components like transformers, said cap ladders, emitter resistor ladders for amp output stages, whole crossovers, and entire amp stages in Hysol (PC12?) to add inertia and thermal stability to those parts. I would really love to hear the AES presentations. Again thanks to all for tackling a too-neglected topic.
GRollins [/i]...3) What's the definition of a generator? A magnetic field and a conductor in relative motion. Congratulations said:
OK here is my $.02: I've spoken to many estimable folks who've sworn up and down that microphonics and induced currents haven't shown up on anybody's scope or AP. Also I've talked to a number who have seen effects. I haven't searched all of AES papers on the subject because life goes on. But I'll be astonished if the general effect isn't measurable. I see 3 big suspect types: 1) transformer mechanical hum, 2) external vibrations reaching circuit components, and 3) the sum of all the modulation of the B and E fields (Grey's point) from typical operation of the circuit.
My old company put its money (enough to kill the outfit 🙁 ) into chassis construction to reduce internal and external vibrations. We also used enormous quantities of $$ monolithic ceramic C0G ceramic caps not merely for their temperature stability, but for their sound quality, esp. when used as coupling caps. These caps which we built into ladders (as many 80 per) of nearly zero inductance. C0Gs (=NPOs), unlike X7Rs and Z5Us, are generally considered not to be self-microphonic. Why? Well, they are thick and heavy relative to the others, though the dielectric's other mechanical properties may also affect that. Also we potted some components like transformers, said cap ladders, emitter resistor ladders for amp output stages, whole crossovers, and entire amp stages in Hysol (PC12?) to add inertia and thermal stability to those parts. I would really love to hear the AES presentations. Again thanks to all for tackling a too-neglected topic.
SY said:
You can file down a lead to a point on a high k ceramic and play old 78's with it.
BTW, recognize my new avatar 🙂
Bob Cordell said:
Bob,
Too many people try to isolate effects, then claim that they're unmeasurable, hence wrong-headed. They then criticize the person who advanced the proposition in the first place, calling them names, mocking them, saying that they made up their minds what they'd hear before they even listened...all the usual foolish BS.
Let me make a medical analogy. Since the advent of what you might arguably call modern medicine, it has been the express intent of researchers to nail down one and only one cause for any given disease or condition. While it's clear that a cut on your finger was caused by a slip of the knife as you were preparing dinner, something like cancer isn't so simple. It's beginning to seem as if you start with a genetic predisposition inherited from your parents. That, in turn, is triggered by environmental factors, such as exposure to chemicals (tobacco being a well known example) and a host of other things, any one of which would not work alone. The human genome studies have turned up a number of genetic anomalies that serve as precursors for various diseases without any one of them being sufficient, in and of itself, to bring about the disease.
Audio is where medicine was fifty or a hundred years ago. Many still blindly assume that there is one and only one cause for any given audible effect and if a proposed mechanism isn't measurable, then it isn't valid. This mindset needs to fall. THD as a single indicator has done its job. It alone cannot account for all things audible. Nor can IM, slew rate, clipping, residual noise, or any of the other things that people tend to obsess over to the exclusion of all else. It's not necessarily one thing--it's multiple things at once.
My EM field proposal above (a simple, rather obvious mechanism that I've proposed numerous times to varying degrees of acceptance) may or may not be measurable by itself. If it is...great! If it's not...great! I don't mean it to be taken as a stand-alone bugaboo. It's not "Grey's EM field theory" alone, it's Grey's EM field theory, plus microphonic parts, plus that French heat-distortion theory (can't remember the guy's name at the moment--begins with an L), plus cable direction, plus cap ESR, plus THD, plus inductance in the internal wiring of the circuit, plus...ad infinitum
The era of simplistic one-for-one cause and effect in audio is over. It has been for twenty or thirty years. It's time for a more sophisticated approach. The way to design a better sounding circuit is to realize that attention to myriad details--any one of which may not be measurable or even necessarily audible (to everyone--like sensitivity to "corked" wine, sensitivity will vary from one person to the next--this is a proven medical fact) alone--will sum to something that is audible.
Don't assume that "Grey's EM field theory" is measurable--but I guarantee that it exists. Given the mechanism I've proposed, it is utterly impossible for it not to exist. You would have to violate the laws of physics for it not to be there. But measurable? Don't bet on it. You're welcome to try to measure it. I encourage you to do so. But don't blindly assume that if it's not measurable that it's not relevant. In combination with other factors, it just might be important. Doctors are discovering that their view of disease has been unacceptably narrow. In the same sense, it's time that audio folk consider that multiple low level factors could sum to audible effects in the real world.
Grey
There is strong evidence that a compact object in the Cygnus X-3 binary system produces an intense beam of ultra-high-energy cosmic rays. Here, we examine the effects of such a beam hitting the JC-4(gr) case and of the subsequent production of secondary neutrinos. We consider how high a beam luminosity is allowed and how high a neutrino to gamma ray ratio can be obtained from such a system. We will also attempt to measure the effect of the enclosure compression by the beam leads to pion cascading and a neutrino spectrum peaking at 1−10 GeV energies, which may affect the sound through internal interactions.
Given the mechanism I've proposed, it is utterly impossible for it not to exist. You would have to violate the laws of physics for it not to be there. But measurable? Don't bet on it. You're welcome to try to measure it. I encourage you to do so. But don't blindly assume that if it's not measurable that it's not relevant. In combination with other factors, it just might be important. Doctors are discovering that their view of disease has been unacceptably narrow. In the same sense, it's time that audio folk consider that multiple low level factors could sum to audible effects in the real world.
Given the mechanism I've proposed, it is utterly impossible for it not to exist. You would have to violate the laws of physics for it not to be there. But measurable? Don't bet on it. You're welcome to try to measure it. I encourage you to do so. But don't blindly assume that if it's not measurable that it's not relevant. In combination with other factors, it just might be important. Doctors are discovering that their view of disease has been unacceptably narrow. In the same sense, it's time that audio folk consider that multiple low level factors could sum to audible effects in the real world.
GRollins said:
I'm with Bob on this one. If you can hear it, it can be measured. An incredible amount of subtle detail is being extracted from the optical spectrums obtained through monitoring the plasma in wafer processing equipment. Optical/audible it's all just part of a spectrum. It's just going to take a fresh look to address it.
Now whether or not it's hearable is another question.
Of course, it will be measured and understood, eventually. So what? Most of the time, typical engineers drag their feet, don't integrate a total number of effects such as caps, and whitewash it all over until their circuit is deemed unsuccessful, and then they move forward with great reluctance to admit what we have known for decades.
One example is coupling caps. There are only pretty good to awful coupling caps, so it is best to design them out. That is why I design them out of my circuitry, as best I can, and then use what might be considered obscenely priced devices, where they are absolutely necessary. Vibration control is one major factor in my cap selection.
For example, many decades ago, I designed in an input coupling cap for the JC-3. None of my more modern designs use this cap, nor would I design it in, any more than a sports car designer might use a cast iron head and block assembly. It might have been OK decades ago, but not today. If someone had a problem with the JC-3 power amp, for example, the first thing I would look at would be the input coupling cap. Next, I would look at the power supply bypass caps, and third, I would try to minimize and/or eliminate the output coil. That is what we call progress. I would not have worried so much 35 years ago, about these factors. Today, I know better.
One example is coupling caps. There are only pretty good to awful coupling caps, so it is best to design them out. That is why I design them out of my circuitry, as best I can, and then use what might be considered obscenely priced devices, where they are absolutely necessary. Vibration control is one major factor in my cap selection.
For example, many decades ago, I designed in an input coupling cap for the JC-3. None of my more modern designs use this cap, nor would I design it in, any more than a sports car designer might use a cast iron head and block assembly. It might have been OK decades ago, but not today. If someone had a problem with the JC-3 power amp, for example, the first thing I would look at would be the input coupling cap. Next, I would look at the power supply bypass caps, and third, I would try to minimize and/or eliminate the output coil. That is what we call progress. I would not have worried so much 35 years ago, about these factors. Today, I know better.
MikeBettinger said:
In theory, yes. In practice, you have to know what to look for and develop an appropriate approach for measuring it, so the practical answer is--at least for the time being--no. Take a look at the post above yours. Do you really believe that someone with an attitude like that is going to stir himself to do something useful? When pigs fly, maybe. Probably not even then. Too lazy. It's easier to criticize others than to exert yourself. Golly, might break a sweat holding that soldering iron. Much safer to just sit on your butt and yap.
With the limited amount of time that I have, I'd rather build circuits and listen to music than try to develop measurement systems. This really rankles certain folks who seem to think that I'm supposed to drop everything and come up with numbers (which they would reject anyway) just for them. My position is that if they're so bloody obsessed with the numbers they can damned well develop their own measurements. Which they can't be bothered to do...natch. So they remain in a self-reinforcing belief system wherein they won't consider the possibility of other distortions, so they don't develop measurements, so they don't find out anything new, so they remain in their comfort zone of THD measurements. Nice DO LOOP, there. Meanwhile the rest of the world moves ahead, albeit painfully slowly.
Even if you develop a system to measure one of these multi-cause things, you're still going to have a difficult time teasing apart the various contributing factors...exactly the same problem they're facing in medicine now. I live in the here and now, not the year 2050. What they will know then does me absolutely no good today, so I don't fret about it.
Grey
There's another side to this too, if you can measure it it HAS to matter. I once demonstrated that one could put a small piece of steel near a crossover coil and create thirds at the -140dB level.
It goes without saying that there is no amplifier with unmeasurable distortion so if people hold themselves above any accountability for what thay claim to be audible there is no end or point to the argument. You are always going to be able to say well this amp has -121dB thirds while this one has -123dB but adds -117dB sevenths. As I said before there are those that prefer signal paths with gross levels of distortion that would bury any measurement anyway.
BTW have you seen Baxandall's paper on how small GNF around low gain stages not only increases distortion but creates higher harmonics? Sussman and Boyk recently did it analytically but I don't think it got published. Baxandall built simple amplifiers and measured it.
Yup that's old R. S. I figure he's an appropriate watchful eye over these proceedings.
It goes without saying that there is no amplifier with unmeasurable distortion so if people hold themselves above any accountability for what thay claim to be audible there is no end or point to the argument. You are always going to be able to say well this amp has -121dB thirds while this one has -123dB but adds -117dB sevenths. As I said before there are those that prefer signal paths with gross levels of distortion that would bury any measurement anyway.
BTW have you seen Baxandall's paper on how small GNF around low gain stages not only increases distortion but creates higher harmonics? Sussman and Boyk recently did it analytically but I don't think it got published. Baxandall built simple amplifiers and measured it.
Yup that's old R. S. I figure he's an appropriate watchful eye over these proceedings.
I'm more curious why wine can taste so different. I'm sure there is a very subtle level of tweaking known only to the vineyard, but I could be wrong. Know anything about it?
Scott, your talk with RS must not have gone too deep. I can only lead a horse to water, not---
Scott and Bob, you both should read Cherry's paper in 'IEEE Circuits and Systems Magazine' first quarter 2008, pp. 55-71. He has discovered higher order distortion generation from feedback.
Scott, your talk with RS must not have gone too deep. I can only lead a horse to water, not---
Scott and Bob, you both should read Cherry's paper in 'IEEE Circuits and Systems Magazine' first quarter 2008, pp. 55-71. He has discovered higher order distortion generation from feedback.
syn08 said:
Not sure where they are coming from, but I'm watching those gamma rays come through every now and then on my work bench Lecroy LT344.
They saturate my scintillation counter PMT preamps, currently excited with a Cs137 pill and some radium foil (handled with pliers).........
Not sure about their impact on audio amplifier sound, but they do cause at lest some degree of statistical error in my detectors count rate 😀
Nelson Pass said:
This is, of course, always affected by the needs of the interpreter.
Scott wurcer said:
Very good point! I personally believe if something is audible it's definitely above the -100db floor, it's just that the standard tests don't create the conditions to excite the effect and it's more of a combination of energy distribution and amplitude that contribute.
The real problem comes in separating this "distortion" from the true signal, considering it is caused by and related to the full spectral envelope of the music we're trying to enjoy.
Single/multi frequency tones into resistive loads at full power are meaningless. Although it's fun to listen out on the edge, the audible problems in most systems are apparent at civilized levels where nothing is being taxed and the noise floor is a subtle background contributor.
The whole idea is that a measurement system needs to be able to detect differences in broadband normal operation; maybe the answer will come from the up and coming generation of engineers Scott mentioned.
john curl said:I'm more curious why wine can taste so different. I'm sure there is a very subtle level of tweaking known only to the vineyard, but I could be wrong.
Scott and Bob, you both should read Cherry's paper in 'IEEE Circuits and Systems Magazine' first quarter 2008, pp. 55-71. He has discovered higher order distortion generation from feedback.
Again the wine and food analogies are weak, there is no perfect wine or "live" performance as one could say. SY has grabbed what I thought were perfectly good wines out of my hands because he is more sensitive to corkiness. People love wines all the time that are "technically" defective (I do).
I hope you meant to put quotes around "discovered".
I really don't need to have another conversation about how our op-amps are only good for offset servos or supply regulators or that they really need 10mA to Vee on the output. Bruce Hofer personally thanked me and uses them in the Audio Precision. BTW I have never listened to one myself. For me it's not a hobby unless it's different from work. I like to see how little you need to get the job done.
boloney
PS I use your AD797 op amps in the front end of my Sound Technology, as well. They work great for instrumentation.
I am still trying to use them in a cheap phono stage, but they are overpriced, and I may go to the LT1115. Walt Jung says they should work about as well, for my present needs, and they are $2 cheaper than the AD devices.
PS I use your AD797 op amps in the front end of my Sound Technology, as well. They work great for instrumentation.
I am still trying to use them in a cheap phono stage, but they are overpriced, and I may go to the LT1115. Walt Jung says they should work about as well, for my present needs, and they are $2 cheaper than the AD devices.
scott wurcer said:
Scott,
There doesn't have to be a perfect wine. My point in mentioning corked wines was to point out that there are scientifically verifiable cases of human senses varying in sensitivity to certain stimuli. If high end audio was as large a market as wine represents, we'd probably be able to find out some really interesting things about human hearing. Unfortunately, it's a niche market and it doesn't generate anywhere near the cash flow that alcohol does. We'll just have to muddle through.
And in precisely the same sense as you've said, there are people who are quite content with stereos that are, to some extent, "defective." That doesn't bother me in the least. I don't criticize people who can't hear things. I criticize people who can't hear things and conclude that they've heard all there is to hear.
Grey
P.S.: I taste and smell the tri-whatsit that gives wine the corked flavor and aroma, but I don't find it as offensive as others do. I file it under "barnyard" or "rustic" and go on. If it's really bad, I let the wine breathe for a while and that generally gets it down to a level I can live with. There have been a few instances, however, where the wine has gone beyond what I can tolerate.
My wife seems to be one of those who does not detect corkiness at any level--even wines I can't be in the same room with.
- Status
- Not open for further replies.