What type of feedback cap is the best?

[K-amps]

I have heard many opinions on this from going to NP (non-polar not Nelson Pass) caps. Has anyone actually heard an audible difference? Which of the following (or more) do you think is the way to go.

1) Going single NP.
2) Blackgate VX
3) Elna Cerafine
4) Metalized Polyproplene
5) Polystyrene
6) 2 standard lytics back to back?
7) Bypassing the main FB cap with smaller cap?

How about voltage, whats a good base voltage to use? 1x the rails? 2x?

[MikeB]

Hi !

If you are speaking about dc-blockingcap in feedback, i use option 7.
I use normal electrolytic bypassed with wima-mks 100nf.
The bypassing definitely improves sound, takes out some colorizing.
For asymetric a NP is not necessary, as the polarity of the offset
is known. As voltages are not very high here, i used 10volt-types.
This can get critical if the amp fails...

Mike

[K-amps]

Thanks Mike, what kind of coloration did you perceive prior to the bypassing?

[MikeB]

Hmm, difficult to describe in words, it just sounded like freq-response
was not flat, the sound was to sweet, just somehow wrong.
But the electrolytic i used was very old, maybe it just had bad quality ?

Mike

[jcx]

any film, even mylar is lower distortion than any electrolytic but dc blocking caps are usually large so Al electros are commonly used

Cyril Batemann measured capacitor distortion in his "Capacitor Sound" series in EW/WW - his recommendation for large values is non-polar electrolytics - your #6 should be considered obsolete on low distortion grounds, Batemann found some advantage to going to 2 series non-polar Al electrolytics when blocking several V

I have measured low frequency distortion while exploring a sony ce-775 sacd player - surprisingly the main L-R channels with Nichicon Muse series caps actually had more low frequency distortion than the surround channels which used lower spec Nichicons

Doulas Self also has measured Al electrolytic coupling/dc blocking cap distortion and recommends oversizing them - by making the RC time constant large enough the distortion in the audio band drops

[Leolabs]

Depend on your applications and circuitry.Each type of cap has its own characteristics in term of sound and performance.For signal coupling,avoid aluminium but most of the time is "unavoidable" unless you willing to spend more money and PCB space!As for power supply bypassing,I just use anything I have,doesn't care whether they are ceramic or film,even people told me that they(the caps) do have something to do with sound.Regarding RIAA feedback or crossover feedback network,use something that are precision.

What brand???Well,me often stick to estabilished brands,doesnt care it is Japanese,Taiwanese,European,American or even African But dont be surprised if one day Chinese caps beat your audiophile grade caps!!!!!!!!!!

[Upupa Epops]

To k-amps : Use instead caps DC servo with low frequency ( below 1 Hz ) and good opamp. Isn't thruth, that servo make sound worse. Voltage of caps : lowest voltage, which you can buy ( 6-10 V by elyts, 40-63 V by others ).

[K-amps]

Pavel and others, the question was around upgrading the caps in an Adcom 555ii. The amp already has DC servo.

The cap in question is a Mylar, I am using a metal polyproplene type, I can also go Paper in Oil which many believe is sonically better.

The cap in question is C103 . I am thinking they have an effect on sound because the ground the feedback loop.

Any ideas?

Thanks!!

[Upupa Epops]

C 103 insulate input ground from chassis ground, which is good for suppresion of ground loops. Because is connected with low value resistor in parallel, parametres of this cap isn't critical. Nominal voltage of cap at this position is only a few volts.

[K-amps]

Makes sense Pavel, thanks. I got useful info from this thread thanks guys!