B+ filtering - why not just use properly voltage rated caps?

[dms]

Hi - I am really curious about the large cap filtering on the A1a Express- I had assumed that the B+1 stacked caps and bleeder resistors were the chosen configuration ONLY/Primarily because the B+ 1 voltages required a much higher voltage cap than Mr. Fischer had in his parts box at the time, as I've read he typiclly used what he had around...

Now looking at the Komet pictures, I keep seeing that series capacitor/bleeder resistor thing, and I'm wondering why.

In the A1a, 2 series 160uF are supposed to give a bit less than 80uF, and you can easily get an 80uF Sprague atom @450 volts... that is what I'm at risk for using unless someone says - "Stop - this won't give you the same result!" Can anyone explain? or explain how critical this filtering is with regard to sound character?

Same goes for the 40 - 20 - 20 - 20... you can find a JJ 4 element @ 500v that covers this in one can.

thanks to anyone taking a shot at this - I'm curious and thrilled to learn about why! - Davids

[doctord02]

At the time, I'm not sure that Ken had ready access to 450v rated 80uF caps.

As for the multi section JJ 40/20/20/20 can, it should work, tho you may not be as effectively decoupling between stages with the common ground that is shared with each section. Might make no difference. The wreck Paul Ruby built for me uses the multi section can and I dont have ripple/hum issues with the powersupply.

[dms]

Yup - agreed, but from looking at the Komet pics, it looks like there are 4 mallory 80uF caps (2 on top of the board, 2 under) with a small army of 220k escorts. For Komets, the cost of a capacitor is probably not an issue, and I think the TC80s are already rated at 450volts (?), so they are stacking capacitors for some other specific reason. Any guesses?

[markd]

Here's a guess. Even though hooking two 80uF caps in series gives you 40uF, it doesn't physically limit the ability of those caps to hold electrical charge. An 80 uF cap has larger plates than a 40uF cap. That larger plate area can hold twice as many electrons as the plates of the 40uF cap. Two 80uF caps have the ability to store four times as many electrons as a single 40uF cap. I don't see how hooking them in series changes that. Even though the capacitance value is reduced to 40uF when hooking two 80uF caps in series perhaps there is a larger amount of electrons stored, and available, than when using a single 40uF cap. That would mean the caps would discharge slower and the power supply would be more solid.
Whatever it is, I'm sure it made an audible difference to KF and the guys at Komet amps. I doubt it was because those were the only parts in the box. Any more guesses? A definitive answer would be even better.

It's like the 25watt wirewound "choke" resistor. The notes with the Liverpool schematic in the Files of this site:
http://maximoaudio.com/download/schemat ... matic1.pdf
show a voltage drop across that resistor as 19v at maximum output (I would have expected more). It only takes 19mA to drop 19v across a 1K resistor. Any old 5watt resistor would be able to comfortably handle that with room to spare. A 25 watt wirewound resistor certainly isn't required there. My guess is he used that because of the added inductance, that type and size of resistor has, made an audible difference to him.

[MikeKCA]

Yup - agreed, but from looking at the Komet pics, it looks like there are 4 mallory 80uF caps (2 on top of the board, 2 under) with a small army of 220k escorts. For Komets, the cost of a capacitor is probably not an issue, and I think the TC80s are already rated at 450volts (?), so they are stacking capacitors for some other specific reason. Any guesses?

It's my understanding that Komets run at or close to 450V B+.

[markd]

I checked the voltages on one that was biased very cold(the owner felt it sounded best at that setting) and the plate voltage was well over 450v. If it ws biased at 70% it would have still been slightly above 450v, IMO, but I didn't rebias it. A Komet is set up where a non technical person can rebias it easily. And this one had been biased by ear. The bias current at idle was about 40%. And the plate voltage was around 480v.

[Dai H.]

doesn't KF have preferences of some filter caps over others? In Vintage Guitar mag(?), there was something he wrote about that and IIRC, he thought IC brand filters sounded horrible.