Both Richard Marsh and I, separately have been researching problems in capacitors, and we started this approximately 40 years ago. Originally, I essentially ignored DA, and concentrated on non-linear distortion. Later, I found, by listening tests (not done by me) that DA WAS important, so I started following Richard Marsh's contributions on DA.
Later, about 30+ years ago, Walt Jung and I decided that Scott Wurcer's cap differential subtraction test, although not perfect, was the easiest way to show the general audience that there were DIFFERENCES between caps that were beyond RC matching, or even series resistance or inductance. Our test, developed in parallel with Scott Wurcer, did careful RC matching, and we added series R to better match two different caps, where one cap had more series R than the other.
We controlled the rise-time of the test waveform reduce to negligible the differences in inductance, or we could add a certain length of wire (depending on the cap) to match, and cancel the excess inductance in one of the caps.
Was our test beyond criticism? Of course not, BUT it did show that caps had differences that could not easily be accounted for, which was mostly DA.
Richard stayed with the more 'pure' or formal way of measuring DA, and got similar results, all in all. We can (and have) quibble between ourselves as to the superiority of one measurement approach or another, but the essential results are the same.
It is only those who have little knowledge or experience, where this is a new problem, or the people who want to beat back any work that we did in the past, for their own reasons, where any debate is generated.
Later, about 30+ years ago, Walt Jung and I decided that Scott Wurcer's cap differential subtraction test, although not perfect, was the easiest way to show the general audience that there were DIFFERENCES between caps that were beyond RC matching, or even series resistance or inductance. Our test, developed in parallel with Scott Wurcer, did careful RC matching, and we added series R to better match two different caps, where one cap had more series R than the other.
We controlled the rise-time of the test waveform reduce to negligible the differences in inductance, or we could add a certain length of wire (depending on the cap) to match, and cancel the excess inductance in one of the caps.
Was our test beyond criticism? Of course not, BUT it did show that caps had differences that could not easily be accounted for, which was mostly DA.
Richard stayed with the more 'pure' or formal way of measuring DA, and got similar results, all in all. We can (and have) quibble between ourselves as to the superiority of one measurement approach or another, but the essential results are the same.
It is only those who have little knowledge or experience, where this is a new problem, or the people who want to beat back any work that we did in the past, for their own reasons, where any debate is generated.
But my question still is, how did you attribute the nonlinearity you observed to DA as opposed to some other phenomenon? If the answer is, "I just assumed and didn't look at any other factors," that would be useful to know as well.
OK I can accept that -- didnt know the answer and asked. So, it is less than .1dB. Thats cool. Its isnt detected as a amplitude variation.
Do you have any tests which place a voltage across the cap and is not a sine wave. More like music... with transients or pulses?
Do you know that the dielectric is Not responsible for any distortion when significant volatage is dropped across it?
Maybe deep down inside the material it is not DA. Like I said, it just have a strong correlation to what people say they hear when going from polar to film. My test conditions tried to better mimic the music waveform. If not something about a materials DA, lets find it and close the books on it. But, I think it is the material and the DA of it strongly follows the distortion. Thats about all I can tell you.
Happy hunting.
THx-RNMarsh
Last edited:
I already gave my answer..... you look at everything and measure everything with Z analyzers, micro-Ohm systems, make samples/prototypes, talk to cap makers (REL-CAP owner has PHD in material science and is co-patent holder with me on a very low esl cap) etc etc etc.... its by process of elimination at first until you find something that matches the listening experiences being told... which are pretty consistent and uniform over the planet of people and time. DA is that match. If its something else, then someone else will have to spend the time to find it.
Sorry i cant get down to the molecular materials level in this to seperate out the variables further. Maybe Bas Lim could do that (REL-CAP).
THx-RNMarsh
OK, so the answer is, "I just assumed and didn't look at any other factors," and the conclusion is unsubstantiated.
You don't need to get down to molecules (that's my job), you can determine the variables electrically to see if it's an artifact of the test method, in the dielectric, in the terminations, or somewhere else. If it's in the dielectric, you can go farther and see if it's VCC, mechanical, whatever. But to ignore any controls and variables but jump on DA as the culprit with no further evidence is perhaps creating more legend than clarification.
You don't need to get down to molecules (that's my job), you can determine the variables electrically to see if it's an artifact of the test method, in the dielectric, in the terminations, or somewhere else. If it's in the dielectric, you can go farther and see if it's VCC, mechanical, whatever. But to ignore any controls and variables but jump on DA as the culprit with no further evidence is perhaps creating more legend than clarification.
for the most part, No. Certainly with any Westrex cutter system, which is what I am familiar with. Westrex 1700 was the "classic" Scully/Westrex system used by Capitol and many other American cutters.
You can measure and adjust many things in that chain, but "knowing" that you have +/- 0.1 db delivered to the surface of the record is difficult. Certainly now, with access to accurate metering, its easier and they can be adjusted to very close, electrically. Going the next step and measuring the cut velocities means optical interference methods (well described in the lit) that are tedious and require special equipment. Not me.
I have not looked at this specifically as to what may have been considered reasonable, I am going to guess that it was +/- 0.5db, electrically.
too see the schematics of the 1700 system,
lathetrolls [licensed for non-commercial use only] / Westrex 1700 Schematics
Alan
ps I think the Neumann system was able to be more accurate, but still has the same set of physical limitations. ie its a mechanical system translating an electrical signal.
Last edited:
I would love to find an affordable source of large value polystyrene capacitors!
I think we agree most modern capacitors are at least adequate. It may be a bigger issue today is the dc/dt part of the equation.
But I thought the issue was validity of the 30+ year old tests. I see no reason to dispute high DA capacitors do affect the complex waveforms.
As never two objects produced by human industry are identical (neither two planets ;-), models will always be approximations.
The disappointing thing for me is the absence of correlation between measurements and reviewers. A recent integrated amp in Stereophile had a 6 (!) decibel deficiency in the phono response at 20kHz, and gobs of (nonlinear) distortion. Without being told about this, two reviewers thought the product sounded great.
It is emerging that a great many audio professionals have significant hearing loss, almost certainly due to overexposure. Certain forces wish to conceal this lest it lead to an impact on employment, but it should be nothing short of scandalous.
It is emerging that a great many audio professionals have significant hearing loss, almost certainly due to overexposure. Certain forces wish to conceal this lest it lead to an impact on employment, but it should be nothing short of scandalous.
did you see the sprout measurements with the 7dB bass boost?
I think it is time for 3 cheers for the PC industry and switched mode power supplies taking over the world. Without them some of the wonderful capacitors we can now get cheaply would not be available and we'd still be using beer can caps with gobs of extra inductance everywhere
I think it is time for 3 cheers for the PC industry and switched mode power supplies taking over the world. Without them some of the wonderful capacitors we can now get cheaply would not be available and we'd still be using beer can caps with gobs of extra inductance everywhere
I thought I made it clear as possible that I DID look at all the other factors. What isnt clear about that? I looked at Vcc, mechanical and tested for a host of things far beyond what has been discussed here and got many brands and did distructive testing and took them apart to see how they were constructed, materials , means and methods. I am just telling you -- i am NOT going to go into all that investigation and inquiry just to please you--- it was months and years in the testing and working with multiple mfr as well. I relied on mfr info from people who make them and learned the whole process and the potentials in the process for issues. I weent to the factories and watched and talked and the patent I got with them... involved purchasing a custom made machine (under nda) to make them. I'm not going to go so deep as to tell you trade secrets to please your curiousity.
If you are not satisfied with what I have told you so far... so be it... draw what ever conclusion you want from your limited info. Even if this sounds trite, I dont have the time for it.... you have plenty of info over the years... you will have to sort the engineering from folk lore on your own time. I'm not being mean about this... it just isnt good use of my time.
I am sorry that others took the info and applied it to making boutique brands at high price which are nothing but repackaged parts. I am sorry that cable companies added DA to thier literature and all the rest. Is there anything I can do about that which was created by others/marketing?
There are a host of parasitic elements depending on materials and construction.... the one not looked at was DA. I exposed that as more important to audio than series L or esr for regular low level use. You have listening tests and that will or will not convince if anything at all matters for cap audio apps. If there is distortion, my experience says it is from the materials/dielectric and I think it is the DA mechanism that is responsible.... the reason I think so, is not valid for some people..... it correlated best with listening comments at that time. I did not do any dblt of various parts myself.... others did, I understand.... I thought I was shedding some light on capacitor mis-behaviour. The people I was around... engineers mostly say they heard differences in caps. I correlated the misbehavior with that. I also saw the 2H on FFT to make me more confident. Why does it matter if you dont hear anything? I dont worry about it. Some have reported caps sounded like day and night different between films and mfr.... you read that constructors at DIY all the time... this cap sounds this way and that. It never ended. So, for me it doesnt matter, i just buy cheap films and I dont hear such differences. But then, I havent used a coupling cap in decades since developing direct coupled and servo. So, that only leave EQ and filters to argue as to what mechanism causes the distortion..... I have not read anything which tells me it is not DA of the film. But if there was good reason to think it is something else, I'll buy that.
It is an educated guess from everything I did and I can rank distortion with voltage across the cap to DA ... distortion follows the DA. Do the DA vs distortion tests with voltage across the cap. Whats so hard about that? I had access to an HP339A then and owned the same THD/IM gear that JC did at that time. More than that I dont know.
THx-RNMarsh
Last edited:
And bypassing them with six varieties of film caps in a golden mean sequence of values, or some such.did you see the sprout measurements with the 7dB bass boost?
bcarso: Yes, commented on it in fact in here---and the reviewer didn't notice it
I think it is time for 3 cheers for the PC industry and switched mode power supplies taking over the world. Without them some of the wonderful capacitors we can now get cheaply would not be available and we'd still be using beer can caps with gobs of extra inductance everywhere
Yes, I'm grateful, even if the computer supplies tend to run the suckers at about their maximum ripple currents!
Nope, from what you were saying looked like you measured something and that was that. Sorry, but that's how I read it.
I guess your capacitor is
?
bad news for them... didnt they know about it? If they did... what were they thinking?
-RNM
What you looked at, how each factor was measured to eliminated it, how you controlled the variables, repeatability, you know, the basic stuff that researchers put into technical papers. Just saying, "I looked at everything else" with no backup data or explanation isn't particularly encouraging.
Most of the ps type caps are no longer being made... most companies stop making them.... displaced by pp for most apps. The best source of raw material specifically for the use in caps are made in China now. They intend to keep making it forever. No I wont give their name/sources and what makes them best over others technically. Trade secrets.
The only cap maker I know in USA which uses that best film source is REL-CAP. But they dont stock parts and custom values/tolerences would be expensive.. just a warning. Plus 12 week ARO.
THx-RNMarsh
Last edited:
- Status
- Not open for further replies.