I'll hazard a guess you didn't get this info from the back issues of Scientific American. 🙂 This caught my eye though:
Do these systems achieve such spectacular performance with a single circuit path, swept narrow frequency band measurements and heavy gain control? The electronics can't possibly be wide open all the time, can they? Remarkable performance either way.
Do these systems achieve such spectacular performance with a single circuit path, swept narrow frequency band measurements and heavy gain control? The electronics can't possibly be wide open all the time, can they? Remarkable performance either way.
rdf said:
Even then, SOMETHING has to be capable of dividing the signal into a couple of million channels before scans are performed (simultaneously, BTW).
And of course, first stage amplification is shared.
Just to get things straight. I don't think there is cable distorsion, or at least not any such phenomenon on a level that could possibly matter. There seems to be nothing in the commonly accepted physical models and our common ways of using these that suggest such a phenomenon should exist. However, that per se is no proof that there cannot exist some form of distorsion, however unlikely it seems. So from a strictly scientific point of view, there is no reason from what we know today to believe that cables cause distorsion, but on the other hand, neither is it possible to disprove it. My remark to John earlier was meant to say that from a neutral scientific point of view, the empirical results we have from the two sources are not sufficient to draw any conclusions either way, which should not be interpreted as any support for the claim that cable distorsion exists. What Bratislav said makes a lot of sense. Still, if I am to play the devils advocate, it could still be possible that a phenomen that causes audible distorsion for audio would not cause sufficient harm to the sensor signals to matter for them. It could be that the two types of signals are sensitive to entirely different parameters.
Quoting from a post I made myself a couple of years ago:
Quoting from a post I made myself a couple of years ago:
About 22 years ago, I needed to find a 'super cable that would be coaxial, small in diameter, relatively flexible, and able to withstand 5000V or so. I had a friend who made cables at Lawrence Berkeley Labs (maybe you have heard of them) and he recommended a particular cable that they used for similar projects, as I was wiring a laser control system. Just for the fun of it, I made a set of audio cables from this pretty, white teflon coax cable, and it sounded lousy! Gee, it must have been my imagination. However, I did find it very useful for lab work and it still connects my main analyser to the outside world.
john curl said:
I recall cables like these sound a bit bright and a bit harsh in the mid range.
In essence, correct, but one change is required.Christer said:
He did not find distortion in the cables..he found distortion as a result of using specific cables on a specific test machine.
Christer said:
Correct, it is not sufficient proof. However, the REASON for the differences in measurement lies with the equipment and it's reaction to the cables. That avenue of thought was ignored by JC when it was raised.
It also suggests that the newer equipment was simply designed built and tested to the levels being tested, whereas JC's was not. JC tried to update an old tester via some improvements in electronics, but did not consider the ramifications of the old test setup construction. Baselining the equipment with respect to it's improved capabilities was also rejected by JC.Christer said:
Concur.Christer said:
hermanv said:
john curl said:
LBL...hmmm, name rings a bell...I know I've heard of them before..don't they make vitamin supplements??😉
john curl said:
Why must it have been your imagination? Just because you could not correlate any physical parameters of the cable to sound?? As I said years ago, shield ground...
Cheers, John
PS...Actually, the correct name is Lawrence Berkeley National Laboratory, one of 17 national labs supported by the DOE..😉
Hi jneutron,
With this in mind, a scientific method would be valid for a specific case only. One combination of output and input characteristics and grounding between two pieces of equipment. One can see the variabilities (combinations and permutations) are endless. That's probably why the cable thing will never end.
You may be able to assemble some guidelines. Then the audio community would need the characterize each piece of gear. Most can't even get resistance right.
-Chris
That's been my basic point as these relate to audio equipment. It's all part of a system and they interact. Some more than others. Test equipment reacting to cables is therefore just as valid.
With this in mind, a scientific method would be valid for a specific case only. One combination of output and input characteristics and grounding between two pieces of equipment. One can see the variabilities (combinations and permutations) are endless. That's probably why the cable thing will never end.
You may be able to assemble some guidelines. Then the audio community would need the characterize each piece of gear. Most can't even get resistance right.
-Chris
believers (or not)
It's a shame that the fundamental method of deciding if competently designed (as in reasonable, non-corroding connectors, low capacitance, reasonable conductivity) cables have an inate sound continues to elude (either purposefully or out of ignorance) those who propose that they must and do.
Discussions about DBT's, ABX, etc. really miss the point, which is:
Null hypothesis -> cable group "A" and cable group "B" have no differences in their means or variances
Alternative hypothesis -> cable group "A" and cable group "B" exhibit significant statistical differences in their means and variances in the property of interest.
There is a tremendous amount of literature available to support this format, eg.
"Practical Experimental Designs" William J. Diamond, 1981 Lifetime learning Publications
(aka experimental design). I believe much of the inertia against this valid approach expressed by the subjective camp comes from 1) Ignorance and or laziness to comprehend the methodology and 2) the fact that proper design of the experiment to support one or the other conclusions often takes a great deal of effort and time to "get it right" and include all of the confounding 2nd (and possibly 3rd) order interacting variables.
One could set up either a single sided alternative or a double sided alternative using, for instance, various opinions as to why cables do sound different, as the criteria and then use a Hadamard matrix at some chosen resolution to determine the validity of conclusions drawn from such a test sequence.
In close to 20 years reading arguments, tests, etc. about "cable sound", I've never seen a correct statistical analysis evaluation support the alternative hypothesis.
While it's true one cannot prove the null hypothesis, it's relatively easy to prove that the null hypothesis is incorrect. That's the purpose of using a proper decision making procedure.....
It's a transcendant method to reach agreement on experimental (and/or practical) results that the high end cable industry is vested NOT to pursue, and that much pseudoscience effort is directed at discrediting...
John L.
auplater
john curl said:
It's a shame that the fundamental method of deciding if competently designed (as in reasonable, non-corroding connectors, low capacitance, reasonable conductivity) cables have an inate sound continues to elude (either purposefully or out of ignorance) those who propose that they must and do.
Discussions about DBT's, ABX, etc. really miss the point, which is:
Null hypothesis -> cable group "A" and cable group "B" have no differences in their means or variances
Alternative hypothesis -> cable group "A" and cable group "B" exhibit significant statistical differences in their means and variances in the property of interest.
There is a tremendous amount of literature available to support this format, eg.
"Practical Experimental Designs" William J. Diamond, 1981 Lifetime learning Publications
(aka experimental design). I believe much of the inertia against this valid approach expressed by the subjective camp comes from 1) Ignorance and or laziness to comprehend the methodology and 2) the fact that proper design of the experiment to support one or the other conclusions often takes a great deal of effort and time to "get it right" and include all of the confounding 2nd (and possibly 3rd) order interacting variables.
One could set up either a single sided alternative or a double sided alternative using, for instance, various opinions as to why cables do sound different, as the criteria and then use a Hadamard matrix at some chosen resolution to determine the validity of conclusions drawn from such a test sequence.
In close to 20 years reading arguments, tests, etc. about "cable sound", I've never seen a correct statistical analysis evaluation support the alternative hypothesis.
While it's true one cannot prove the null hypothesis, it's relatively easy to prove that the null hypothesis is incorrect. That's the purpose of using a proper decision making procedure.....
It's a transcendant method to reach agreement on experimental (and/or practical) results that the high end cable industry is vested NOT to pursue, and that much pseudoscience effort is directed at discrediting...
John L.
auplater
Actually folks, I always knew the Lab on the hill as Lawrence Radiation Laboratory, which it was for many decades,and which we informally called the Rad. Lab. It overlooks the apartment where I now live. They changed the name to something more politically correct. ;-)
For the record, my ST1700B analyser had been repaired and calibrated within the last 10 years of the test by the manufacturer of the design. The initial design is about 30 years old, BUT I upgraded it myself with better IC's to lower the noise and distortion further.
The measurement residuals that I actually measured were at the level of approximately -150dB, as I measured the cables with a -30dB signal. This was important, because above the -150dB level, the distortion was not seen. It could well be a ground return problem that is not present in the AP analyser, compared to the ST analyser that I use, BUT it does not mean that the cable does not produce distortion itself, the only difference is the the distortion might be generated in the shield ground return path, rather than the direct signal path. Either way, there are differences in distortion measurement between cable components.
For the record, my ST1700B analyser had been repaired and calibrated within the last 10 years of the test by the manufacturer of the design. The initial design is about 30 years old, BUT I upgraded it myself with better IC's to lower the noise and distortion further.
The measurement residuals that I actually measured were at the level of approximately -150dB, as I measured the cables with a -30dB signal. This was important, because above the -150dB level, the distortion was not seen. It could well be a ground return problem that is not present in the AP analyser, compared to the ST analyser that I use, BUT it does not mean that the cable does not produce distortion itself, the only difference is the the distortion might be generated in the shield ground return path, rather than the direct signal path. Either way, there are differences in distortion measurement between cable components.
Multi-factorial designs
her's a pretty good intro to confounded variable experimental design, if anyone cares...
http://www.statsoft.com/textbook/stexdes.html#2a
many chemistry related examples... as us chemistry types are always throwing various things together to see what happens... ;-)
auplater
her's a pretty good intro to confounded variable experimental design, if anyone cares...
http://www.statsoft.com/textbook/stexdes.html#2a
many chemistry related examples... as us chemistry types are always throwing various things together to see what happens... ;-)
auplater
I was funnin ya, John..john curl said:
Here, the verbage is typically...I work at "the lab"...or, do you work at "the lab".
john curl said:
Your efforts to upgrade your equipment was noted years ago, and it is worthy of praise for the intent.
Changing the electronics within the box, of course, still leaves you with an old box and it's associated issues.
john curl said:
It is actually not the shield that is producing the distortion. It is, however, one of the consequences of the shield.
For audio, it is the input shield loop that needs to be considered.
Amplifier design which does not consider input shield current sensitivity, is not adequate to the need. It is that way simply because amplifier designers do not use Faraday's law of induction with low impedance circuitry. They do not know it is relevant..yet. That will take time.
Cheers, John
Bratislav said:
I don't know where to begin:
Lets go backwards: I and no one else on this thread ever said that audio was the most critical and demanding application, what is the point of this statement except ridicule without content, it sucks.
Since the statistical analysis I reffered to is used in every catagory mentioned above, that answer is distrubuted thoughout my responses.
Please show me how to measure a nanaovolt or even smaller directly. No amplifiers, no filters, no autocorrelators no choppers, I do not think you know how this stuff works, there is no such device. In fact, small signals are usually measured by stretching time. If you read a voltage often enough and you filter out noise long enough and you sum and divide the results again and again eventually tiny signals are revealed, there is nothing either real time or direct in this technique. It is perfectly valid and most instruments do it internally and automatically, by extending this basic idea more and more resolution can be found, but it takes longer and longer and more and more samples it is not a "direct" measurement at all.
Measure directly? Clearly you can't think astronomers use a tape measure to get distance? You can't think they use a scale to get weight(mass)? Let me provide an example: Lately one area of big astronomical news is the finding of planets, one common technique was to measure the relative brightness of two stars (relative because the detector just isn't good enough to measure just the one stars' brightness, only a comparative measurement to remove errors and noise could do this job), the brightness of the two stars is compared over a very long time with many many readings, when one star shows a periodic dimming, the astronomers calculated how large a body would be needed to obscure that much light and proposed that a planet of size X and year Y passing in front of the star would account for the tiny change in light. Even then this light differential is so down in the noise that many thousands of readings are needed to resolve these tiny signals and remove the noise. There is nothing direct in this measurement, it is statistical averaging of enormous numbers of samples. The second technique concerns gravitational wobble of the suspect sun. Again, the best telescopes do not magnify a distant star so it produces a significant disk on the telescope display. You don't put a picture of the stars disk on a CRT and then measure how far it moves, you're lucky to get a displacement of a pixel or so. No, the image is mechanically wobbled and the number of times one pixel is lit vs the number of times another is lit is averaged over long sample times, summed and divided again and again and then maybe you can find the tiny perturbation in the posiition, hardly a direct measurement.
Even your desktop Volmeter averages signals over time to get an answer. The smaller the voltage the longer it takes to measure (assuming the same number of decimal places).
Good Digital to Analog converters resolve around 50 nanovolts directly (the slow ones) faster D to A's resolve closer to a few hundred microvolts. Measuring small things is all about time and math. Perhaps we have a differnt understanding of the term direct.
So, I haven't been in school for long time and maybe I've showed how dated my knowledge is, but to the best of my understanding these techniques of sampling and working the samples with a great deal of math is how tiny signals are measured. The cables used are far from the weakest link in the system.
jneutron said:
Yes, my formulation was sloppy and I did of course mean what you say. Nothing more than that can be concluded from a single experiment.
Yes, it obviously has something to do with either the test equipment or the way the test was set up. Unless something was done very wrong in the other measurements, the cables do obviously not always cause distorsion. If the distorsion John measured is because of unreliable test equipment, methodological error in the measurements, or that the cables actually can cause distorsion under certain circumstances remains unknown. There is also a possibility that some methodological error was done in the other measurements, which is why it is important to have several independent experiments that correlate.
I addition to this, we may have our own beliefs about what is the truth about cable distorsion and the measurements done, but that is beliefs, and differs from what can be scientifically concluded from theory and measurements.
nanomeasurements
http://www.atecorp.com/Equipment/Keithley/410C.htm
http://www.testequity.com/products/967/
http://www.evaluationengineering.com/archive/articles/0898dmm.htm
http://www.emelectronics.co.uk/spec/P13.html
ain't google great??
auplater
http://www.atecorp.com/Equipment/Keithley/410C.htm
http://www.testequity.com/products/967/
http://www.evaluationengineering.com/archive/articles/0898dmm.htm
http://www.emelectronics.co.uk/spec/P13.html
ain't google great??
auplater
Hey John C., I see at the top of the thread, replacing the IC's in the ST1700B for a lower noise floor. Do you mind telling me what you used?? I've (still) got a ST1701A and see a little "popcorn" on the nulled distortion traces, ....even on a loop-backed signal. I've also got a ST1200A Test panel that had uA2605's for God sakes. Some were bad (full DC offset) and I replaced them and found the pinout could not be duplicated on subs. But any info on your quiet replacements would be greatly appreciated.
Thanks
Thanks
- Status
- Not open for further replies.