Mundorf M-Cap Supreme as bypass cap?

[fas42]

Tsk, tsk, I've just seen your pursuing of advice on Audiogon also, you're really making sure you're getting a good cross-section of opinions!

But that's OK, it's worthwhile trying to achieve maximum understanding of a situation, though it would have been of value to have known from the beginning that the main caps were of the type you mention. They obviously are of very high quality and the number in parallel means that the 20 or 22uF unit is completely redundant, unnecessary, but the other 2 smaller units will still be of value, because of their very small size. In which case, the 50 times rule means that they're completely "safe" ...

There is some excellent advice in that Audiogon thread ...

Frank

[MGH]

Your advice is more clear to me and makes more sense. I thought I mentioned having film power supply caps early on, but my mistake sorry. However, I'm still wondering if the two small value bypass caps you recommend has higher ESR than my power supply caps. Even if they did, would it matter in this application?

[fas42]

Thanks for the thumbs up!

It's not the ESR that's at issue but the associated ESL, the parasitic inductance that is associated with the construction of the device that's the problem. The combination of that inductance with the rated capacitance of the device means there is a resonant frequency intrinsic to the cap, some spec's tell you what it is. And only at that special frequency does the extremely low ESR actually mean anything, on either side of this frequency the impedance seen by the circuit shoots up dramatically.

Now that I know the nature of what you're using, let's look at somewhat similar film caps made by ecicaps, a range called Unlytic:
Unlytic-400V.jpg

Look at the curve corresponding to 650uF, pretty close to what you've got. Yes, it gets down to 0.005 ohms, but then it shoots up as the frequency increases, hits 0.3 ohms at 1000kHz, 1MHz; not so impressive now! This is because the ESL is dominating now. Those 10 caps in parallel mean that you have theoretically improved all the good stuff by 10 times, so at 1MHz the impedance is 0.03 ohms, BUT, long lengths of wiring are needed to connect all those caps together, because of the physical size of them, and hence most of the theoretical advantages, at higher frequencies, disappear!

Now consider a cap in the same range, 10 times smaller:
Unlytic-1000V.jpg

The 50uF curve here has a key characteristic, the minimum impedance, where the ESR has a chance to do its job, has moved to a much higher frequency, roughly 5 times greater, and still looks pretty good, about 0.007 ohms. Something that those 10 560uF caps would be struggling to do, see what's going on here? It's the combination of the rated capacitance, the ESR, and the ESL that tells you where a cap will do a good job of bypassing, in terms of frequency, which is why paralleling several of different values works for a frequency range ...

Frank

[MGH]

Got it! Do you teach electronics? You should have an online course and charge a fee. I would be willing to take it.

Now, only if 0.47 microF ceramic caps at 1kV could be found...I would have the perfect bypass caps.

[fas42]

No, but thank you for the compliment!

There are people on this forum with vastly more knowledge and experience than myself, I probably have a good knack for picking up information that is directly relevant to what I'm strongly interested in, understanding the subtle factors that contribute to good sound.

Now, only if 0.47 microF ceramic caps at 1kV could be found.

Done! SV09AC474KAR - AVX - CAPACITOR CERAMIC 0.47UF 1000V | element14 Australia, but a bit pricey, $22. A film cap should still do the job perfectly well, but if you want to experiment, go for it! Let us know how it goes, what sort of improvement, or otherwise, you get ...

Cheers,
Frank

[MGH]

Thanks again Frank! Unfortunately, $22 is not considered pricey by "audiophile" standards. The high cost of "audiophile" components drives me nuts because I know most of it is just marketing BS.

BTW, you read the Audiogon thread where some recommended 1/100 ratio when selecting cap values for bypassing. Is this any better or worse than 1/50?

[fas42]

No, once you get past the 50 to 1 ratio then there won't be any problems with resonance, which is why you need at least this gap. So, 100 to 1, 200 to 1, etc, are all fine. But, remembering back to the curve of the impedance with frequency, that classic V shape, if the gap is too large then there is insufficient overlap of the characteristics, the 2 together will look like a W, you're not getting the best result.

Again, don't get paranoid about it: anywhere in the area of 50/1 to 100/1 will be fine, you don't need to get the precise value to make the ratio just so, possibly will give you more options when buying!

Frank

[MGH]

Thank you Frank for all your advice. You have made this a fun learning experience. I learn more from thoughtful, giving individuals as yourself than any book.

[fas42]

Happy to help out ...

Cheers,
Frank