Best signal path capacitor solution?

[D.A.R.R.Y.L.]

Hi all,
I have been modifying my denon 2900 with some very audible differences.

I also have removed the Elna polars from the signal path after final opamp to rca with Solen 9,1 µF 400VDC MKP, bypassed with two WIMA's 0,1 µF and 0,01 µF.

I was wondering what the exact result is by replacing the 220 µF Elna's with the 9,1 µF Solen's? I know a bit of electronics but not much.

Why isn't the signal dependent on the exact 220 µF. Why does the 9,1 µF Solen sound better? What is the exact use of these caps in this signal path config?

Is WIMA the correct choice or would one advise a better bypass cap?

Thanks in advance,
Greetz, Ramon
The netherlands

[D.A.R.R.Y.L.]

PS
Heart of the matter being that if it is possible to even go lower than 9,1µF to say 4,7 or 2,2 so I can even use better caps which would be to expensive in the 10µF variant.

[SimontY]

Hi D.A.R.R.Y.L.,

Those would be DC blocking capacitors, which are there just to stop any DC leaking out and into your amplifier, where it may be amplified (and then hurt your speakers). If you have DC blocking in your pre-amp or power amp then you can omit these caps and short across the space with a length of wire. This will be a large upgrade.

Also, most people seem to prefer unbypassed film caps when used in the signal path.

edit: I'm not the one to do calculations but I think you can go as low as 4-5uF without affecting the audible frequency response. If you make the caps too small they form a highpass filter, like the cap in series with your tweeter.

[D.A.R.R.Y.L.]

Hi Simon,

How do I know if my amp has DC-blocking?
I have an Electrocompaniet ECI-4.

[D.A.R.R.Y.L.]

PS Do you mean I can omit all and just put in a WIMA of Vishay unbypassed film caps from output opamp to rca? What value would be recommend for this film cap for all the signa to pass-through?

[SimontY]

Err, you'd need to take a close look I guess, but as it's a commercial amp it will almost certainly have protection against DC input.

The best thing to do is remove them for the timebeing and connect your multimeter to each of the outputs and check for DC. Connect red to center, black to the ground, and you'll get a reading of something like 0.003v or something I think. If the number is low enough to be amplifed 20x or so and not be a significant voltage then you can safely leave the caps out.

Please somebody confirm some numbers as I'm not sure what safe figures are.

[SimontY]

The signal should NOT go through ANY capacitors ideally. If you put it through a tiny WIMA it will sound very tinny - you'll only have the top end! I did that once accidentally, hehe.

[klitgt]

If you MUST use a DC blocking cap (signal cap, coupling cap, to use some other words for the same thing) I can recommend copper foil paper-in-oil caps. They have no sound of their own (almost).
The value depends on the input impedance of the next amp (pre or integrated) provided the source output impedance is low (it often is).
With power off, put the DVM red lead to the RCA input center and the black to the sleeve and see what OHMS read, normally 10K or more of input impedance.
To calculate the cap value, use the attached simple spread sheet. If you e.g want to go flat down to 8 Hz as frequency response, with an input impedance of 10K (10000 ohms), the cap has to be 2uF. So values in the range of 10-220uF are overkill.

[SimontY]

Yes, I have none in my cd player, but have them in my preamp, and they are 1uF iirc. They are LCR brand polypropylene with solid copper leads. These have virtually no sound to speak of. Another (more down to earth) recommendation for you

[D.A.R.R.Y.L.]

;-)
Very useful recommendations!!

LCR I can order via farnell, so that is very nice indeed.
To klitgt: would you advise to still bypass these lower values?