It is true only when your observation is applied to mass marketers such as Pioneer, Sony or Onkyo. It is not true for the high end boutique brands such Audio Research, Bryston etc. Even mid-fi brands like Marantz, NAD or Parasound invest in resources and have access to component sources that ordinary DIYer cannot afford or aware of.
Throwing money at small component vendor who tends to make unsustained claim of their parts will not necessarily make DIY audio equipment better. Many of us also do not have the trained ears of people like Nelson Pass etc.
I know what I like in term of sound stage, tonal balance, etc. But I am not good enough to tell the sonic difference of different capacitor types. My ears are very sensitive to both hum and white noise. I have equipment to measure THD distortion down to 0.01%, which is usually good enough for me.
The kind of test that Monte McGuire did in very valuable to DIYers. Thank you, Monte. May be analog_sa can do the same for his favorite capacitors.
Great, you are half way there, just add music to that sensitivity range
I cannot measure beyond -120db and do not believe that harmonic distortion in capacitors has even a slight correlation to sound quality.
Not sure why you mention specifically ARC and Bryston. Bryston use only commonly available parts and so do ARC with the exception of their Reference range where they use custom ordered coupling capacitors, made i believe by Solen. Yawn.
Great, you are half way there, just add music to that sensitivity range
I cannot measure beyond -120db and do not believe that harmonic distortion in capacitors has even a slight correlation to sound quality.
So what does? Why dont you measure that. Oh yea cant measure imagination.
Just curious, have you measured a Nichicon Muse ES? Those are the bi-polar and easily available green ones.
I did not say they measure the same. If you think you can hear the difference between -140db and -155db distortion than waste your money. Caps can be audibley different in some cases where they are taxed, like speaker crossovers and maybe feedback networks but the audiophool belief that same spec caps are audibley different in most circuits is BS. If its true you should be able to prove it and explain the physics.
Last edited:
I have measured the UES1C221MPM1TD, which is a 16V 220µF Muses ES cap. It's not the same value as the 470µF 35V so a direct comparison is not possible, but still, the distortion using the same test (with no DC bias voltage, since it's a bipolar) is below the FFT bin noise floor, -163dBV for both 2nd and 3rd harmonic. This 220µF cap has about 2x the signal voltage across it but basically the same current through it, so one might expect this to make the distortion numbers greater than if a 470µF cap were used if voltage is the key, but if current is the key, the results are comparable. Regardless, the ES Muse series seems to be very clean under test.
I also tested a UFG1V471MHM1TO, which is a 470µF 35V Muse FG series polar cap, using 22V DC bias and the same 2kHz signal, and measured -151dBV of 2nd harmonic and less than -163dBV of third harmonic. Second harmonic might be hard to understand, since a cap should be symmetric, but DC bias alters this, since the field strength across the dielectric will be stronger for one half of the cycle when using DC bias than the other half of the cycle. If the field strength ends up saturating the dielectric, an even order nonlinearity will develop. So, while 3rd harmonic is what we'd expect from a cap. a biased cap can produce 2nd harmonic by this means if the dielectric is close to saturation.
Just a note of caution: the levels I quoted were voltage levels, not decibels relative to a carrier. The actual signal voltage across the cap, due to the RC filter action of the test jig, places the signal around -40dBV with a +15dBV drive. So, it's hard to say what this harmonic level should be referred to - the -40dBV signal?
Also, we measure distortion with a pure sine wave because it's easy and it's the most sensitive method, not because it represents what music does. A complex waveform will also produce cross modulation products, which should be a lot more audible than pure harmonic distortion.
At the end of the day, my conclusion is that we don't know how bad we can make a circuit before it annoys us, and every decade, circuits seem to be getting better by about 20dB less crud. In the 60s, a distortion level of 1% was deemed to be the "i don't care" threshold, but by the 80s, that seemed to be reduced to 0.01% Now, we can get things down to -140dBc if that's your goal, and lower is possible. So, I'm not ready to declare that any number has any actual merit or not, and that lower is better if you want clean. I cannot find a way to refute that 1% is better than 0.0001% distortion, and the only evidence we have is that circuits with more distortion produce outputs more different than their inputs. Can we hear it? I don't know.
Andersonix, small mistake here, EKZN350ELL471MJ20S is not the Nichicon KZ/MUSE , is actually a General Purpose Capacitor. I think the nichicon KZ/Muse will measure at least as good as the silmic caps.
Thanks for the kind words!
I have not, since the original circuit required 35V rated caps, preferably around 470µF and using 25V polymers in this particular circuit, with 22V across them, was risky - I had one "run away" with excess leakage. Good thing for current limited lab supplies!
I just did a new search and it seems that things have improved w.r.t. HV polymer caps, so I will revisit this. I will try to order some this week and let you know soon!
Nichicon es measurements Nichicon Muse ES bipolar caps measured: <-120dB THD, <-140dB IMD
I expect Nichicon KZ to be very similar if not better .
I expect Nichicon KZ to be very similar if not better .
@Monte McGuire:
When I saw your post I remembered that some time ago we exchanged a bit here on diyaudio - also about capacitors I think ...
Thanks, Monte - I shall be interested in hearing how they perform ... (and since you actually buy & test some I do hope this information is also interesting to you yourself ... )
Cheers,
Jesper
You are welcome
When I saw your post I remembered that some time ago we exchanged a bit here on diyaudio - also about capacitors I think ... Thanks, Monte - I shall be interested in hearing how they perform ... (and since you actually buy & test some I do hope this information is also interesting to you yourself ...
)Cheers,
Jesper
To Monte Mc Guire-are you measuring these as signal or bypass caps? I always preferred the Nichicon es and muse series but I know why many may prefer the Similac II. The Similac II's I've bought tend to have quite a bit of capacitance over their rated value vs the Nichicons many hitting that 20% or above range. Many modern nichicons are right at or slightly under rated value. When used in bypass situations the extra capacatance may be perceived as better bass. If you are measuring signal caps many manufacturers use the Nichicon FW in the signal path and Use FG as bypass caps. It would be interesting to see what the FW's measure. Also are you burning caps in for several hundered hours? I know most sound different after a few hundred hours. Thanks for the measurements- have never seen them directly measured before..
If you believe a $200 Sony A/V receiver and an ARC $20,000 amplifier sound the same, you are entitled to your opinion.
Even Sony do not use "only commonly available parts" in their higher end A/V receivers ($500 and above). The main rectifier capacitors are all custom ordered and not available to DIYers. The Sony mass market store A/V receiver is a very different story.
BTW, Audio Research do not make audio equipment for mass market store. Their Reference 150 amplifier is no better or worse than the Classic 150 in term of parts used and manufacturing, but it is newer.
I am ok with the lack of knowledge, but the lack of logic really puts me off. Best of luck.
Has anyone ever successfully measured break-in in capacitors? That would finally prove to subjectivists that measurements are useful and make objectivists eat humble pie