Bypass caps in choke PS

[john dozier]

The main resevoirs of my tube monoblocks have 2-220uf caps before and 2-220uf caps after the choke. Should I bypass these caps and if so what value would you use. Thanks

[simon7000]

I would put 470 PF on the input and 10 UF on the output for a first try.

[john dozier]

The 2-220ufs are in parallel in both cases. Would you bypass only one of each pair or each cap? Thanks

[astouffer]

Smaller value film caps like 1uF or 0.1uF will clean up the HF noise that the large caps don't filter.

[Vinylsavor]

Hi!

I don't recommend bypassing. I know my opinion is controversial since bypassing is a common practice. But bypassing can also have negative effects. If your large caps are so bad that they need bypassing they probably have large lead inductances. These can create a parallel resonant circuit with the small bypass cap, depending on the value. That will have the exact opposite effect at the resonant frequency as intended. I bet a lot of the 'improvement' heard from bypassing is from resonant effects.

Use caps which are good enough to begin with. Elcaps improved during the last decades. Modern ones don't really need bypassing.

Best regards

Thomas

[john dozier]

I share these concerns. The problem aside from physical size, is that there are very few manufacturers of 500v caps. I know of no "shoot outs" testing these very few available caps for SQ. That is the only reason I have considered bypasses. Thanks

[trobbins]

John,

Some people like the idea of bypassing to sustain a low supply impedance at higher frequencies - as seen from the amp circuitry. To do this correctly really requires the electrolytic capacitor datasheet impedance curve or relevant info that allows an understanding of where ESL kicks in. It then also needs a bypass cap with a sympathetic impedance curve that achieves its impedance minima at the time the main cap's impedance starts increasing.

Some people like the idea of using the bypass cap to filter diode noise. Filtering diode noise in a valve rectifier HT circuit is better done with filtering cap across secondary winding. The high frequencies involved, and the long lead lengths to a bypass cap on the main filter are not a good filtering method. The issue of rectifier noise stems from HT winding inductive energy, and focusing on the source of the noise is of most benefit.

Ciao, Tim

[simon7000]

I find an RF filter as soon as possible is useful. For high voltage circuits such as this that usually means it is vacuum tubes (As stated in the first post.) Since these do not have the same problems as semiconductors with EMI (RFI) the filtering is not as important. So 470 PF should get rid of enough. There are interesting issued with small caps at high frequencies. Their leads have enough inductance so that they self resonate! This limits their HF performance. The electrolytic capacitors should have enough ESR that the internal L will not be an issue with such a small cap.

The 10 uF on the output is based on personal preference. I find 300 uF film capacitors make a difference on solid state Class A amplifiers driving 8 ohm-ish loads. So I scale this value down as I expect you will be using an output transformer to provide a higher impedance load to the power supply. I expect the 10 uF to be as close to the output as possible and grounded in a way not to induce currents into other circuit paths.