M Huss cap across bleed resistor

[JoeK]

Hi I'm in the middle of building the 6V6 plexi and just trying to work out what purpose the "470n" cap across the first filtercap - 220k bleed resistor is doing.
http://www.mhuss.com/Plexi6V6/Plexi6V6.gif

Thanks Joe.

[Tonewood]

I think of it as being in parallel with the first big filter cap. Maybe to pass to ground hi frq that the big cap may not do a good job with.

[J M Fahey]

^^^^^^ what he said.
Millions of amps out there without that extra cap working perfectly, but it's a nice touch and definitely does not hurt.

I mean, unless a heated discussion starts about the "tone improvements" involving cap colour, paper vs wax vs plastic, etc.

[JoeK]

Thanks for the replys. I did find a few more threads here that went into theories about using 1/100 or 1/1000 value of electrolytic cap and also theories that the cap will filter hash noise from noisy diodes. I'm using UF4007 so I guess that negates the need for that. Shall build the amp and if it fails to impress I'll know where to snip it out I actually built the amp before as an experiment in a rough chassis without the choke etc etc and it was great.

[R.G.]

The reality is that no capacitor stays a capacitor. As frequency rises, a capacitor's impedance gets smaller in the same proportion that frequency rises. So if a cap has a reactance of 10 ohms at 100Hz, it's 1 ohm at 1000Hz, 0.1 ohm at 10k and so on.

But every capacitor also has minor imperfections. They have some internal series resistance and some internal inductance. While these are small, they are there. The resistance stays whatever it is at all frequencies, and the inductive impedance rises directly as frequency rises.

At some frequency, the capacitive impedance drops to near the same value as the internal series resistance, and from there on the resistor dominates. The capacitive part of the impedance doesn't matter any more, even if it goes to nearly zero.

Also the inductive part of the impedance rises steadily with frequency, so that at some frequency, the inductive impedance gets bigger than the internal series resistance, and the "capacitor" now behaves as an inductor at all higher frequencies.

The way to fix this is to note that the bigger the capacitor, the bigger the equivalent series resistance and equivalent series inductance. So you go smaller on the capacitance, resistance, and inductance by paralleling the big honking filter cap with a smaller cap that doesn't store as much energy at low frequencies like DC and low audio, but which also does not reach the "turnaround point" frequency so low. It lets you have effective bypass caps at higher frequencies, and may be useful in keeping the power supply impedance down for bypass and decoupling.

The rules of thumb about 1/10, 1/100, 1/1000 etc. of the capacitor being bypassed are pure raw guesswork. Electrolytic capacitors vary widely in their ESR and ESL for any given capacitance. So basing a secondary bypass cap on the capacitor value of a filter cap is - well, let's just say "optimistic". To do this properly you have to measure the ESR and ESL, or run an impedance versus frequency plot on the filter cap. No one will ever do that though; so a rule of thumb is as good as anything else, I guess. But intense discussions about the right value of bypass capacitor bypass capacitor (note that this term is technically correct, not a typo ) are futile.

Secondary bypass caps are intended to keep power supply impedance low at high audio frequencies and above, and are particularly useful at RF. They may be useful in suppressing the squarks of RF ringing caused by slam-off ordinary diodes in the power supply. They may be useful in suppressing squealing and RF oscillation in an amplifier. Or they may not be needed at all. Horses for courses.