Sorry for the late answer, i have had a busy personal/professional life these last weeks. It will be more calm during summer, I hope.
Quote:
However, adding C to the end of an RCRCRC or whatever it is, is not going to lower the total R along the way at all.
If your end game caps (closest to the audio circuit) can deliver very fast, the current that resupplies the caps must go thru that earlier system with it's impedances and parasitic properties (slower caps).
The overall idea is not to lower the total R, but to "decouple" as much as possible the Signal loop (Energy provider > Tube > Headphone) from the PSU.
The PSU is there only to refill the Decoupling Banks. As long as the PSU can provide more energy than what is used in the Signal loop (which is the case obviously), the Energy Balance of the cap banks between what goes in and what goes out is OK.
There's no clear scientific consensus on this, but a largish minority of DIY guys think that parallel caps is a bad idea. I'm somewhat in this camp as well.
The resoning goes, that in the cap system you create overlapping areas, and when they discharge, the smaller cap has a phase error in discharging.
Exactly the same as in a speker crossover. The overlapping areas are problematic. Some people prefer full range speakers despite them having problems in high treble and low bass. Also many people cite the lack of a crossover system as a big benefit in using headphones.
Did you try out and compare single cap and multicap solutions and come to find you prefer this system?
My opinion on this matter is undecided, leaning strongly in favour of single caps where possible.
Yes I have tried a lot of different setups :
Setup 1 :no caps : original
Setup 2 : 1 cap : 33µF Clarity Caps
Setup 3 :2 caps : same as above + 0.33µF RIFA PME271
Setup 4 :3 caps : same as above + 10nF polystyrene (not good at all ... maybe not enough burning)
Setup 5 :3 caps : 0.15µF FKP1 + 4.7µF MKP10 + 100µF TPC
Setup 6 :4 caps : my current setup (with the White Monster)
And yes, I have heard significant changes between each of those setups, not all were good. The current setup (the 3 and 4 caps) is the one which provided the more benefits by a large margin.
I have been trying to see if a "one cap" could be enough too. It would have make my life much easier. But I could not reconciliate the few parameters I wanted to play with in a single cap.
I wanted :
- fast transient
- low impedance
- high energy
- form factor that could fit the amp and allow to put the bank as close as possible to the tubes (not that easy)
- not a bank breaker
The problem I had with the "single cap" is that it could not satisfy all the criterais at the same time.
You can't have a low impedance with a small cap, especially on the audio freq. band due to the "Zc = 1 / jwC" part of the impedance law.
You can't have a fast transient on a very large cap which will most probably be a Lytic
A middle ground between those could have been some large industrial FKP "laser pulse" caps, but they are incredibly hard to find (B2B most exclusively) and expensive.
And "large" means that you'll have to put then some inches away from your tubes, adding some inductance in the process and killing your "high freq response".
And even then, you can't expect to have the almost infinite transient of a silver/mica and the low impedance at 100Hz ...
So, maybe the "phase error" you're talking about can have an sonic impact, I honestly don't have a clue. If this is an area for improvements (better values scale ?), I'm really open to suggestions/discussion.
But I suppose this is always a mater of cost/benefit. I'll have my transient and my low impedance at the cost of some phase error and maybe oscillations too.
But overall, when I listen to it, it's a clear improvement for me.
I don't say it's final or it's the best way, just saying that in this context it works, it's easy to implement (sort of) and it's cheap.