Regulated heater reference

[Quince]

I've this regulator for the heaters of four EL34 tubes. The reason of two regulators on top of one another is the lower portion of voltage lost in rectifiers than if I had only 6 V for output.

My question is, since the cathodes sit around -350 V, how do I reference the heaters to that so that the heater to cathode potential rating is not exceeded, without frying the regulator's low voltage components? The thing is, a relay delays the HV so when the heaters come on initially, there's no voltage applied to the tube's cathode or plate, and the HV supply is not even on (the relay is before the rectifiers since relays have much lower DC ratings).

[SY]

If you draw an actual schematic, it will be easier to answer your question. I got a headache looking at that! You needn't throw in every wire resistance.

[Wavebourn]

I would connect filaments in series instead of using 2 regulators. Also, I don't understand why do you need a regulated supply for filaments of power tubes.

[FastEddy]

" ... I don't understand why do you need a regulated supply for filaments of power tubes. ..."

'cause most heaters have an AC source, which injects an added noise component into the tube where the action is ... sometimes this AC noise can leak into the audio amplification resulting in, well, extra noise on the audio sgnal ... Sy?

[Quince]

I don't think the details matter. Here's an abstracted version.

http://img482.imageshack.us/img482/8840/hregqb9.png

[Quince]

The matching of filaments is not that great, so in series doesn't work well. Now, the mismatch would seem to be just as bad in parallel except for the fact that filaments are not simple resistors but their resistance varies with temperature. For EL34 tubes I find in parallel works better. However, I can't have all four in parallel since then I get twice the rectifier loss and I don't have enough voltage overhead with this transformer.

In any case, this still doesn't address my original question.

[Quince]

These tubes are driving electrostatic headphones, and I want the lowest possible noise floor. I've even shielded the power supply separately from the amp.

[SY]

Originally posted by Quince
I don't think the details matter. Here's an abstracted version.

http://img482.imageshack.us/img482/8840/hregqb9.png

The details absolutely do matter. There are two considerations- noise injection and heater-to-cathode ratings. Without an actual schematic, it's difficult-to-impossible to lay out the correct way of setting up the heaters or to determine how many heater supplies you actually need.

FE: Generally, you don't have to regulate power tube heaters; noise is usually not an issue and when it is, it's almost always common mode, which takes a different means to prevent. The exception is direct-coupled circuits, where any drift of output tubes off their nominal operating points can fry something.

[Salas]

Just use a common mode choke.