I thought I already answered that question in several ways. First, in Linear Audio --> I can not hear much below the masking of the system thd -- in the best case I could muster with Quad electrostatics was harmonics at or below 0.1% Anything below that was either masked by system distortion levels and/or room noise level et al or I just wasnt able to hear it. Maybe it is possible to hear .01-.05%% for someone under special conditions. Now .001% would be impossible to hear under the simple symetrical sine wave tests. OK? Lets move forward -- I have shown several possible tests that correlate with subject listening. There is more to come but actually, I have written about much of it already decades ago... unfortunately it wasnt in mass circulation mags, I guess. I described the DA affect at DSPx to a very techical crowd and got no rebuttle (!?). Silence. Then the talk is back to sine wave tests for harmonics again. Oh well.
Sad state of affairs, if you ask me.
Last edited:
I mostly rely on other people to do the listening..... it is sooo time consuming. Never-the-less, I think ps dielectric used now is technically a slightly better material than the Teflon ....there are several types of Teflon formulations... not sure how many of them are used in capacitors. Hard to tell without chemical analysis. Mfr wont tell their sources.
Never-the-less, there is now only one source left for polystyrene of high enough quality for cap use (low voltage dielectrics are very thin and thus can be porous and thus trap air and allow for lower breakdown voltages... reliability. That source is in Asia. I would use REL polystyrene if high temperature was not an issue. They anneal thier ps caps and use ps for the outer wrap. Maybe Teflon in tube circuits. Otherwise, if cost and size or value/voltage are not available (ps is limited to smaller values) use pp everywhere. -RNM
Last edited:
No, that was the wrong audience and yes it is hard to find the articles now. So above are two choices that are relatively inexpensive, I'm honestly interested in a mechanism that measures -154db for a sine wave but is above that on music (to reach your quoted audibility threshold several orders of magnitude above). I find it hard to believe the manufacture of the end terminations would match that well (if the rectifying effects indeed exist).
Remember several people here and their "friends" claim they can line Teflon, PS, and PP right up every time. If distortion and DA corellate I would expect the DA effects to be pretty small too.
I have a test pre-amp here with MKP's on the input and it's badly microphonic. So any reasonably scientific experiment MUST eliminate this from tainting the data. I am reminded of calorimetric measurements used to "prove" free energy reactions. It is extremely difficult to do calorimetry right.
Wow! Duelund 24uF @ 100V, $2100 MSRP definately the wrong business.
Last edited:
Thanks Richard for your informed opinion. I have come to much the same conclusion. 10 years ago, I would have chosen Rel. Teflon, as I have them for EQ in my present Vendetta phono stage. For future phono stages, I still have to select.
What I don't understand is why people put so much emphasis on a single tone in testing caps. I thought that decades ago we found that AC analysis was virtually useless for DA, and some sort of asymmetrical test waveform was necessary to bring it out. It seems that many people want to use the 'tools' easily available to them, and if they find little or nothing, they dismiss any problem with caps. I have found different, and you have as well. Perhaps we have to 'enlighten' these people a little. However, take care, Richard. Many here just want to see 'null results' as it eases their use of cheaper materials, and they just might try to 'null' your measurements and experience, like they try to do mine. '-)
What I don't understand is why people put so much emphasis on a single tone in testing caps. I thought that decades ago we found that AC analysis was virtually useless for DA, and some sort of asymmetrical test waveform was necessary to bring it out. It seems that many people want to use the 'tools' easily available to them, and if they find little or nothing, they dismiss any problem with caps. I have found different, and you have as well. Perhaps we have to 'enlighten' these people a little. However, take care, Richard. Many here just want to see 'null results' as it eases their use of cheaper materials, and they just might try to 'null' your measurements and experience, like they try to do mine. '-)
$2400 is sick ! Your question deserves a long and careful answere. I will get back to you on this. But, I strongly suspect that there could be other factors beside thd/harmonics involved.... If you look at this statistically.... a zillion people under the most varied of conditions, parts and time and equipment and brands all say the same thing.... what are the odds? It should all average out to a null result but it doesnt seem to. Odds are its probably true or else we are out of our field as to the possible explanation. yet, if you took a wider sampling of people who are not listeners to music and HD systems ... it might be a null. I am not convinced that thd is the whole issue of what people claim to hear. If they cant hear ultra low thd then what are they hearing effects from? Its something.
Last edited:
Absolutely spot on with the concerns about microphonics IMO. I mentioned a zillion posts back about Rich May expressing the view that the type of dielectric was not nearly as important as the details of construction, in particular for wound film caps the care with which the winding was done, the tensioning and its maintenance. SY quickly concurred and ironically also cleared up some misconceptions (the full story awaits some kick-@ss Rhone) I had had about Pons and Fleischmann having been expert at calorimetry.
Note (lest I be placing words in his mouth, and he is not in this forum that I know of) that Rich did NOT say that the dielectric was irrelevant --- merely that variations in manufacture were important.
When we deal with the issues around vacuum-state electronics I doubt anyone to speak of is unaware of microphonics. And the measures taken to reduce the effects are clearly not entirely effective (some may have seen subassembiles in old equipment with rubber grommets supporting things). And even if the electronics are in a stabilized zero-acceleration inertial frame, there is still the potential for acoustical excitation.
I worked with some of the high-perveance triodes in an equalizer. These have an astonishingly small grid-cathode separation, and the Q of certain mechanical resonances is quite high. I saw ringing at high kilohertz frequencies which, when excited, could be followed into the grass for seconds!
Somehow I'm not surprised; for some people, some things are slow to sink in – if at all.
For some people, the fact that certain phenomena aren't shown on sine waves harmonics tests, is a proof that those phenomena must be imagined.
Snubbers in some of Hagerman's tube amp kits are very noisy acoustically. A friend who built one complained of noise, and I assumed he meant in the electrical output. He brought it by. Wow!
Even film caps can be noisy. When the film caps in the surge generator discharge (6KV 3KA pulses) there is a very audible clunk. And speaking of DA, those caps are normally shipped with a short across the terminals since they can recharge to almost the last voltage when left open, which could be fatal.
Since DA will only show up at very low frequencies it is not surprising that you need an asymmetrical waveform - it probably contained some DC or at least very low freq LF. On a sine wave it would simply give a small phase shift - could be difficult to spot. A waveform with steps in it would show up some exponential decays which some might misinterpret as 'distortion'. Some things are slow to sink in!john curl said:
You'll have a fatter wallet, and you'll enjoy what you make just as much as someone with fancy caps.
Asymmetrical test signals, even generated at mid frequencies, show DA in relatively large amounts. It is easier to use a lowish frequency to maximize the deviations due to DA. Then 1-10% was measurable with the poorer performers. And that is from only one cap. How about 5 caps in series?
Morinix, no cap is the best cap. IF you have to use a cap, try to first use the best materials easily available. This is usually Polypropylene. For hi end we only use polystyrene and Teflon for EQ, but we might use Mylar for supply bypass, as, at least, it is linear enough, in tight spots where polystyrene is just too large. Acrylic has been suggested by a friend in Texas, and might be very good for this application.
Waveform suggestion, anyone? The cap null test showed residuals that confirmed the R/C ladder model for DA to me nothing more, why was that not what was emphasized? I even spent hours with SPICE hand fitting a simulation to the results driving the residual down into the mud.
What is interesting is that testing with an asymmetrical waveform, originally used by you, Scott Wurcer, does show a characteristic RESIDUAL, especially when the cap is really rolling off, as well as passing audio. However this RESIDUAL, even though it has a similar characteristic with the same loading, ranges from 10% to 0.001% or a factor of 10,000 to 1 in the amount of RESIDUAL, depending on the cap 'quality'. It just seems to me that such a ratio must change the audio fidelity to some degree.
2 tones: high and low frequency. Observe the result on FFT plot. See nothing. And relax.
Do a pulse test exactly like I did it. Otherwise, the DA and esr and thd et al are all too low to worry about - thats why they are used. They are the closest to the ideal cap. So what is the issue, really? That some say they can hear differences? I am more interested in what different test do show as non-perfect and secondary side effects in real applications use. -RNM
- Status
- Not open for further replies.