[barrows]

So I wonder what people think about using coupling capacitors and their "best" implementation.  I am talking about the Twisted Pear Audio Legato 3 DAC I/V stage here, but we can keep the discussion broad reaching.

Currently I have 1 uF Clarity MR caps, bypassed with .01 Jupiter Copper Foils (into 100000 ohm impedance), I added the Jupiter bypass as an experiment to see if the more expensive caps make a difference, and they do.
So now I have to decide what way to go forward, I have a strong belief that bypassing coupling caps is a bad idea, as the two capacitors will have a time difference, and so signal passing through one cap will be phase differentiated from signal passing through the other cap, right?

Therefore: perhaps it is better to have a single high quality cap at 1 uF and no bypass at all?  Is 1 uF really big enough (into 100000 ohms) or do I really need a bigger cap to be sure deep bass remains in phase?

Or, what about this: a 10 uF or so electrolytic to be sure all bass always passes, even into lower impedance loads, and then a 0.1 UF high quality film cap bypass, which would still handle almost the entire audible range of frequencies?

[JohnR]

the two capacitors will have a time difference, and so signal passing through one cap will be phase differentiated from signal passing through the other cap

I can't help wondering what result would show if you were to simulate the various combinations. (A phase difference is not something that requires listening to detect, and should therefore show up in the sims - ?)

[*Scotty*]

How much DC voltage do you have block. If isn't too much, say above 5v,  I would recommend a solid polymer electrolytic from Nichicon. This is the Sanyo OSCON technology. I think that they are the best coupling cap you can use if the DC offset isn't too high. The caps have a fair amount of leakage current, unlike the Blackgate caps.
 The nice thing about solid polymer caps is that there is no smearing, period. They also seem to be pretty much not there as far as having a signature. They are very transparent to source. One more cool thing is that they don't seem to go through the awkward phase that the Blackgates go through, they sound good from the getgo and just get better from there. This type of cap is available from both Digikey and Mouser, with 3 brands making surface mount parts but only 2 making through hole parts.
Scotty
Don't bother to bypass these caps as it creates a discontinuity and basically cruds up the sound.

[*Scotty*]

Technically speaking the 2 different value caps will also have a resonant frequency that is different from one another with the smaller one having the higher frequency, the combination of the differing resonant frequencies may or may not have negative sonic consequences depending on the system and how far apart the two frequencies are. I have heard that bypassing a film cap with film cap that is more than 10 times smaller is preferable to using one that is only 1/10th smaller. I haven't used film caps for solid state coupling applications because the partnering electronics always had an input impedance that made a film cap completely impractical.
Scotty

[barrows]

Scotty: I am well aquainted with using polymer caps for decoupling the power rails on high speed circuits, and I prefer the Murata ECAS parts for this purpose, but you are now the second person who has suggested using polymers for output coupling a solid state circuit and blocking DC, interesting as clearly an cap like this is much smaller and much more affordable than good film caps!
My circuit has 8-9 VDC on both phases, I guess one would orient a polymer cap appropriately to the DC levels: with +8 VDC on the signal line, the plus side of the cap would be the signal input and the minus side of the cap would be the signal output?

Can you share any subjective views on the tonal qualities of using a polymer cap for this puspose, in your view is it just neutral?  lack of smearing sounds great to me, I have always been a bit skeptical using huge (physical size) film caps for coupling circuits running at just a couple of volts.  Is the D/A of the polymer caps lower than that of an MKP?

[Folsom]

Careful, if the sound is good with electrolytic/polymer at first but then gets worse quickly as the music plays, then the equipment is trying to polarize them. It's a real thing, and sneaks up on you.

[*Scotty*]

8 to 9volts DC is a lot to block with this type of cap, I would go ahead and put a 16 volt 180 uF solid polymer in the hole and measure the DC leakage in milliamps to see if you have too much.
 When a polarized electrolytic cap is used as a DC blocking cap, it has to be put in with the +pole looking at the positive DC voltage polarity with the negative connected to ground as Barrows stated. When a polarized electrolytic capacitor is used as a DC blocking/coupling cap it only really works well after the charge carriers are pushed around and the cap is fully formed, it does a poor job,sonically speaking, of allowing the AC signal to pass until this occurs, Generally an electrolytic capacitor will show a certain magnitude of DC leakage current until a certain number of hours have elapsed with a DC voltage applied to the capacitor. The amount of leakage current continues to drop, until hopefully it meets the manufacturers stated spec for this parameter.
 As the electrolytic capacitor is naturally polarized due to its chemical makeup, I don't think I would ever describe them as being unpolarized, only unformed. It is this "forming up" process that causes the observed sonic variability between when the signal is first applied and the subjective assessment of it sonic qualities after a few hundred hours of being in the circuit with signal passing through it.
 Also when aluminum organic polymer capacitors are used in power supplies they don't have the typical 10% overvoltage rating that a conventional chemistry electrolytic cap has, on the same subject, neither do the Jensen 4pole electrolytic caps, you hit the rated voltage on them, they go bye-bye.
Scotty

[Folsom]

Bi-polar electrolytics have the issue of changing sound on certain equipment, too.