I am simply saying that a SUBTRACTIVE DIFFERENTIAL TEST will show DA as much as any nonlinear distortion and will have difficulty getting a null, no matter what the amp's quality is, IF it has a typical electrolytic or Mylar input coupling cap.
If we couldn't null it out with a simple cap difference test, how are we going to null it out with an entire amplifier, unless we make the simpler to handle? But then it is not truly representative of music, anymore, is it?
If we couldn't null it out with a simple cap difference test, how are we going to null it out with an entire amplifier, unless we make the simpler to handle? But then it is not truly representative of music, anymore, is it?
It's simple- put a signal into the amplifier. It distorts or it doesn't. Your test (well, Scott's test, really) is excellent for detecting DA. But for audio, I think it's more relevant to look at the component used as intended and see how the overall device performs- fitness for use, as it were.
It's not that he doesn't know how to measure anything. It's that he failed to account for certain things in his interpretation of the measurement's results.
se
Mr. Curl, I measured "distortion" caused by coupling capcitor, not DA. And if is her no voltage across capacitor, then here can by no DA, first must capacitor be charged to some voltage.
You claimed, capacitors caused distortion (at levels about -60dB.) what is prooved (by correct aplication) untrue, valid only if it is alowed to develop some voltage across cap (e. g HP or LP filtering aplications..). BTW, i read Your paper, time ago, and many others papers ,too.(I am 59, and making amps about 30years..)
BV's test is more appropriate for masking cap's imperfections.
My logic is the following. Interstage caps are VERY listenable, one can state this even with mid-fi system. Various caps affects sound to various degree.
If BV's test show very low distortions for any cap, this approach helps nothing in selection of better sounding cap by measurements.
If another method is much more sensitive to cap's distortions, and allows to distinguish various caps more reliably, we need just this method for design purposes.
If specific measurement setup, even that not resembling audio system, would allow to measure differences induced by various power supplies or PCB design, we need just such a setup, if usual spectrum analyser does not allow to distinguish the effects mentioned.
Testing distortion doesn't mask cap imperfections for the intended use. The cap distorts or it doesn't.
How about actual listening tests (ears alone)? I've tried that, but apparently not many others have.
How about actual listening tests (ears alone)? I've tried that, but apparently not many others have.
VERY strong effect of a cap in the signal chain is so obvious and so familiar even for beginners, that I do not see what do you mean.
In all my recent designs, I use SE topology, with an output cap. So, I tested many types of lytics, of film shunts and their possible combinations, and could discuss multitude of measurents and listenable effects. For preliminary sorting of caps, it is not needed to measure harmonics, enough to measure angle of dielectric losses at 10-20kHz. That says quite a lot. Acceptable values of capacitances of electrolytics, and final combination with shunts, up to now is a matter of empirical findings. Measurements of frequency dependencies of impedances do not help a lot.
I would not use cap measurement approach, that does not correlate with listening. Any strange method, that would correlate with listening, will fit the purpose.
I hope, nowdays nobody takes it in serious, that any measurements in audio must be based on spectral analyses.
A problem must be analysed from various angles, if it allows to solve it. If only INTENDED USE SETUP helps to solve nothing, why should we give support to it? We can get into discussion again and again, and the main point, again, will be, that of interest is not the whole basket of distortions, but only those, most sensed by hearing. If they are masked by orders of magnitude larger harmonics distortion signal, it is not a proof, that they are absent.
Last edited:
Just change the cap (input, NFB, output) and listen the difference. The difference is so huge, that I see no need in more sophisticated test setups. Compare a lytic alone, and film polyprop alone. Any unexperienced listener states a big difference in sound. And this is so, irrelevant of what harmonics measurements show. Usually the last say nothing interesting, since important effects are masked by unimportant harmonics.
You mean some method. But if, as a final result, it states that all caps sounds similar, all the pages, containing its description, better to be used in a way, other than reading them.
If we can not see bacteria via simple glass, it does not mean that bacteria does not exist. Invent a microscope, and it will be found. But no, somebody will insist, that simple glass is THE INTENDED USE approach, and microscope is forbidden. Hence, bacteria does not exist, in spite that people get ill and find it related to bacteria.
Last edited:
I don't understand your analogy. Either specific caps used in a certain position distort or they don't. Either specific caps used in a certain position sound different or they don't. You don't use a microscope for listening, just your ears.
Sure.
I was testing if a difference could be heard by changing the input cap of a Tripath amp. The Tripath needs a DC blocking cap because the inputs have a 2.5V bias. All caps were 2.2uF if possible, or 2uF if not.
I tested a least a dozen from boutique to ordinary, electrolytic to film to film in oil. Some sounded very different, some not much. My favorite were the Obbligato oil for mids, Obbligato film overall and Jantzen crosscap because it was very similar to the Obbligatos but much smaller. Some of the film caps weren't so great.
This was a specific test in a specific application. I kept notes. I picked what I liked on music. Testing was done over headphones.
If you would ask BV to test just the differently sounding caps, and he would not report measurable difference, what would be your conclusion ?
If you do diferential test (substraction) with two caps , with music signal with all transients and asymetry in signal, and audible results for your ears is ..nothing,silent, zero, and you at same time claim, you clear hear (with the same ears), diference between the those two caps, conected e. g as input capacitor in the same amp, what do you mean about it? Here is not black magic, or yes?
Last edited:
2Pano
Can You post here simplified circuit, how was this cap conected?
BTW, tripath is PWM amp, and You are talking about DC blocking, so no coupling cap in signal chain. And in Tripath amps, input coupling capacitor is part of HPF with corner freq. about 10Hz and more, so here is AC voltage across this cap, and so distortions too, depending cap type.
Can You post here simplified circuit, how was this cap conected?
BTW, tripath is PWM amp, and You are talking about DC blocking, so no coupling cap in signal chain. And in Tripath amps, input coupling capacitor is part of HPF with corner freq. about 10Hz and more, so here is AC voltage across this cap, and so distortions too, depending cap type.
Last edited:
- Status
- Not open for further replies.