How bad is grounding one side of AC filament ?

[vladn]

I've been thinking of using filament windings (12V) for a second purpose of generating negative bias supply for output tubes using a voltage doubler. But this requires asymmetric grounding of one side of the winding. Is there any kind of isolation between filament and cathode or there is a parasitic diode between the two ?

There are two ways to make such a doubler - a simple one with one side of the winding DC grounded and a bit more complex one where I can shift the DC on the winding arbitrary but it still requires AC coupling of one side of the filament winding to the ground.

Intuitively I think that capacitive coupling from filament to cathode should be small so asymmetric AC feed should not be a big deal. However I have no clue how bad is the parasitic diode so I'd appreciate some feedback.

[JohnAtwood]

If your filament supply is driving only tubes with high-level signal (output tubes, driver tube) you should be OK. Unbalanced AC on the input tube heater may very well give hum, depending on the tube. Tubes that were specifically designed for low hum, such as the ECC83, 7025, 5879, EF86, 12AY7, etc. will probably be good. More general purpose types, such as the 6SN7, are more susceptible to hum. In this case, trying different brands of the same tube may be needed.

Another technique to reduce heater hum is to cancel it. You need to inject a very small amount of the AC heater voltage into the amp at the correct phase to cancel the hum. This isn't recommended for general-purpose users or people who want the amp to be completely "plug and play", but if it is your own design and you don't mind tweaking a pot, this actually works quite well.

- John Atwood

[Eli Duttman]

Vlad,

The last circuit shown here resolves all the issues, but it's not inexpensive to execute. Low noise/low forward drop Schottky diodes are in order. Metallized film caps. are needed in the "arms" that connect to the trafo. 'Lytics are fine in the central "spine".

BTW, the schematics on that page are for positive rails. Don't forget to "stand things on their heads".

[nigelwright7557]

I designed a pre amp with an ECC83 and had quite a few problems with hum.

I found that the HT smoothing had to be very good.
After much playing around with the heater cct in the end i found a DC drive with one side grounded worked best.

I was using a lot of gain in the pre amp for an electric guitar soi any hum was amplified a lot through the 2 stages of amplification.

[vladn]

I guess I have to do some experiments. My curiosity is to the nature of that hum - is it coming mostly from a parasite diode or parasite capacitance. I think it's the former but not sure. If it is mostly from diode than DC bias should solve it regardless of how symmetrical is the AC feed.

BTW this is for a power amp front end not a preamp so it is not an extreme case.