Audible benefits of larger coupling caps than required?

[DrewP]

Is there any sonic benefit to be had by using a larger coupling cap than the smallest one required to give your desired LF rolloff between couple stages?

One would assume (there's that word) that if you double the size of the cap, you effectively halve its ESR across the working frequency range.

Obviously there's a cost/size penalty to be paid, is there a sonic upside or penalty?

In the amp I'm building the calculated interstage coupling cap is 0.47uF and while I have a 0.47uF Wima MKP4, I also have a couple of 2.2u Auricaps sitting around doing nothing. If using the Auri's is going to sound markedly better then I figure why the heck stick with the Wimas?

[Ty_Bower]

Originally Posted by DrewP

...is there a sonic upside or penalty?

It could depend on the circuit. In some cases, particularly those involving multiple stages and significant amounts of negative feedback, the values of certain coupling caps may be critical. Changing them could move the dominant pole and result in instability (oscillation). That would definitely qualify as a sonic penalty.

[DrewP]

Circuit has zero global loop feedback and no designed in feedback within stages. Input gain stage stage cap coupled to a concertina phase splitter and then cap coupled to triode connected EL84's.

[Ty_Bower]

Low frequency roll-off developed by the coupling capacitors might help avoid output transformer saturation under certain conditions. That sounds nasty. There are certainly advocates of bandwidth limiting out there. You can't expect all miracles from iron. On the other hand, if your transformers are capable you should deliver as much of the spectrum as they can handle.

You say you've got the caps in hand. Try it both ways, and see. Let us know how it works out.

[Ty_Bower]

Originally Posted by DrewP

Input gain stage stage cap coupled to a concertina phase splitter and then cap coupled to triode connected EL84's.

Dat's a lotta caps. Why not AC couple the AF amp to the PS, ala Dynaco ST35?

[DrewP]

Ty, really, the only two caps I could potentially delete are the ones between the gain stage and the phase splitter where I could direct couple. The ones coupling the phase splitter to the output tubes are a given. I'm using fixed bias on the output tubes, so there's no cathode bypass caps there and the input and concertina are LED biased so the cathode bypass caps are gone there too.

That being the case I've only got 3 caps sum total in the audio section of the circuit.

It's a test bench amp anyway so I'll build it using the LED bias and the three caps and see how we go, I was just curious to know if there was any audible benefit in using a larger cap size or if anyone had done tests to see.

[disco]

Originally Posted by DrewP

... I was just curious to know if there was any audible benefit in using a larger cap size or if anyone had done tests to see.

Imperfections in a capacitor: inductance, series resistance, dielectric absorbsion. All these result in losses, changing with frequency. Although the phenomena are measurable capacitor losses are rarely published.

That's why cap A can do a good job in amp topology 1, while not necessary true for cap B or topology 2. From experienced hifi hobbyists it's understood coupling caps better have the smallest value one can come away with.

[DrewP]

Originally Posted by disco

Imperfections in a capacitor: inductance, series resistance, dielectric absorbsion. All these result in losses, changing with frequency. Although the phenomena are measurable capacitor losses are rarely published.
From experienced hifi hobbyists it's understood coupling caps better have the smallest value one can come away with.

While the coil length will be double in a cap with twice the capacitance, the number of turns will be somewhat less than double due to the increasing diamater of the winding, so twice the capacitance won't be twice the inductance. And with twice the surface area across which to conduct, I'd expect the ESR of the larger cap to be significantly lower (nearly half that) of the smaller value cap.

As far as hobbyists using smaller caps, I'd always figured it was down to lower cost, rather then better sonics.

[Miles Prower]

Originally Posted by DrewP

Is there any sonic benefit to be had by using a larger coupling cap than the smallest one required to give your desired LF rolloff between couple stages?

When using gNFB, there sure is. You can have a billiard table flat frequency response from 20Hz to 20KHz. However, if you're relying on gNFB to do that, then there is a definite sonic penalty. If the open loop response rolls off too soon, then the NFB decreases, forcing the closed loop gain higher to flatten the over all response. Less gNFB means less effective gNFB, and you might have a higher Zo at the lower frequencies, and therefore poorer damping of the woofer(s). This could lead to sloppy bass, as the woofer(s) will tend to produce their self-resonant note rather than the actual notes the musicians played. That tends to sound like monotonic thumping, and is definitely a sonic penalty.

When using gNFB, you also don't want to see the cutoff frequencies of your RC coupled stages clustering around the same frequency, as this will make the phase shift greater than it otherwise would be, and that's an open invitation to instability, or at the very least, a rising amplitude at that frequency that will compromise the sonics.

OTOH, you may not want too low a frequency response. Since vinyl platters are not completely flat, or guitar players don't neatly pluck the strings, you can get low frequency noise that you'd rather not be sending to your speeks.

[DrewP]

Yes, but does a bigger cap in the same position sound better than a smaller one?