Any downsides to running 12.6V for DC filaments?

[raiken]

Can anyone think of any downsides to running DC filaments on 12.6V instead of 6.3V, other than the fact that off-the-shelf transformers always come with 6.3VAC windings?

I can think of a few upsides:

(1) The filament path current will be cut in half when running on 12.6V, so there will be less voltage drop in the filament path, and smaller wires or PC traces can be used if desired.

(2) Rectifier efficiency will be improved, because there will be half as much current, so half the power dissipation in the rectifiers - longer life, less heat

(3) Regulator efficiency will be improved (unless you are using a switching regulator), again because of half the current draw - linear regulators dissipate power equal to the in-out voltage differential multiplied by the current

(4) Less filter capacitance will be required for the same amount of ripple voltage

(5) Resistive power losses will be cut by a factor of four, because of half the current draw

(6) Output stages that use four tubes can be wired as two series pairs on opposite sides, so if one filament burns out, two tubes on opposite sides will shut off, leaving you running at half-power on two tubes instead of running on three tubes with unbalanced primary currents.

(7) The 12.6V supply is useful for lots of other things, like relay supplies (12V relays seem to be more plentiful than 6V relays) and 12V MIDI phantom power (use a series diode to drop the extra 0.6V and protect the supply from accidental back-driving if plugged into a device that also sources 12V).

RA

[loudthud]

The downside:

Actually designing the power supply.

What will the dropout line voltage be? 115V, 110V? A 12.6VAC secondary probably won't be enough.

What will the heatsink dissipate when the line voltage is 127VAC?

If the DC current is 3.5A, the RMS transformer current will need to be rated about 7A.

The cold tube heater surge current will be about 2X the rating. You'll need a good current limit to do a soft start, or slowly ramp up the voltage.

Did I leave anything out? Will it still work at 50Hz?

[NateS]

I took the question as 12.6 vs 6.3 assuming DC. Most of the DC heater projects I've seen all just run the preamp on DC and the power tubes still run AC. The two questions are still somewhat orthogonal, but Randall's point about taking the other half offline automatically is a darn good reason to run 12.6V.

[Enzo]

As to wiring output tubes in series, I really think that open heaters in output tubes is too uncommon to make a design focus. Output tubes fail all the time, so sensing a blown tube and disabling its partner might make some sense, but usually it is a blown screen resistor from a shorted element in the tube, not an open heater. On the other hand, it cost little or nothing to include it as a "feature."


Peavey has for quite a few years used DC for preamp tube heaters, and takes it a step further by running them in series. Like on a 5150 or 5150-2, they run four preamp tube heaters in series across the +24 and -24v rails, thus 48v. That way the same 150ma flows through all four, and the +/-24v rails are left for relays and whatever. Three tube amps get a 36v rail to power three. 150ma is not going to tax any wiring or pc trace copper.

Does heater current really need to be regulated? My mains voltage wanders around, but not usually as much as 10%, 5% is more like it, and that is close enough for me and my heaters. I don't think regulator dissipation gets much lower than not having any. And if you need the supplies to be regulated for other things like op amps, that 150ma ought not to be too much of a drain.

That of course does not include power tubes.


And another Peavey example is the Valve King, both the 50 watt with a pair of 6L6 and the 100 watt with four. They ran the 6L6 heaters in series, so they all share the same 0.9A, then they put the three 12AX7 heaters in parallel wired for 6v, then placed that combination in series with the 6L6s. Three 6v 12AX7s makes a 0.9A draw, so they fit the series arrangement. They use different voltage supplies for the two models. 18v for the 50w and 30v for the 100w. I think it is clever, it certainly solves the burning high current 6vAC connections that plague so many amps. The downside of it is if you want to try differnt tubes like EL34 or KT77 or whatever that have different heater current, it upsets the balance. Still, it might suggest something. For example, run four 6L6 heaters in series across 24v.


But your points in favor certainly have merit, so it looks like it is worth pursuing.

[Gingertube]

One big plus, simplicity - a simple fixed 12V 3 terminal regulator with the common terminal going to 0V via a standard diode, gives you 12.65V
Cheers,
Ian

[raiken]

I took the question as 12.6 vs 6.3 assuming DC..

Yes, I guess I wasn't clear enough. The question was not DC vs AC, but assuming the decision is already made to use DC heaters, why not 12.6VDC instead of 6.3VDC. Most of the designs I see are just rectified 6.3VAC, sometimes with a linear LDO regulator.

As for regulation (assuming it is used), my preference would be a switchmode buck design, as all the heatsink, dissipation loss, and dropout issues go away. Of course, there are conducted and radiated EMI and noise issues that have to be addressed through proper design and layout techniques, but that's a whole different issue.

BTW, I have been working on a bunch of new high-gain designs lately, and DC heaters are the only way to go, in my opinion. The noise floor is so much lower, and PCB layout is much easier if you don't have to worry about induced hum. I think I'm going to go ahead and try the 12.6VDC heaters, as I already need a MIDI supply and 12V relay supply.

RA

[pdf64]

Power tube failure can lead to the 2x100 ohm ground referencing resistors on regular heater supplies blowing.
The only way I can see that could happen is if significant high potential B+ current found its way into the heaters and then to ground, via the resistors.
Mitigation for such a failure mode may therefore be beneficial when designing a re-arranged, dc supply, in order to ensure that it is sufficiently robust to survive.
Pete