The AM Forum

what's the thought on running DC on directly heated cathodes?  I'm contemplating a quad of 813s as a leenyar but don't have the filament transformers.  I acquired, last summer 2 high current DC power supplies that each can easily supply 25 amps up to 18v on the cheap.  I was thinking of powering the filaments with these vice filament transformers.

In normal operation, you ground the center tap of the filament transformer to allow current to flow, and it flows evenly from the entire filament.
With DC, only one side can be grounde and might cause problems.
There may be other issues with RF into the power supplies also.

10 volt center tapped transformers are not expensive and small, Hammond makes nice ones.
I had one in my 4x 813 amp.
It runs 4x 4x150's now, they are 6 volt tubes but they work fine at 5 volts.
(its likely 5.5 volts at the socket)

813s take 10 volts at 5 amps, at least the ones I have.

You need a proper filament xfmr.  10 volts at 20 amps.  I think I have one,  big one, I think it's 13 volts at probably 20+ amps.  Stepped 240 volt primary.

Fred

Yes, Hammond makes a nice 10 vct 20 amp one, or maybe its 25 amps.
Fist sized.

http://www.rfparts.com/165v10.html

Didn't know Hammond still made xfmrs like that, good to know.

Fred

I just went through their catalog, I didn't see anything bigger than 10 vct, 10 amps.

Another place that has them:
https://www.fairradio.com/catalog.php?mode=viewitem&item=7132

You can also variac a 12 volt transformer and use it for whatever you want.

I have a ton of low voltage high current xfmrs.  I used a twin secondary 12 volts @ 8 amps each to make a xfmr for a single 304TL.  I put both secondaries in parallel, located a CT on one of the windings and with a variac I can get 10 volts at 12.5 amps to light one 304.  Works perfect, I measured the voltage from the HB ct to the ends and it was exactly the same voltage.  The CT was right at the edge of the windings where i could solder on a ct lead without damaging the insulation.

As for the OP, I guess there's no reason why he couldn't use two 10 amp xfmrs to light the 4 813s.

Fred

Salvage an old microwave oven transformer, cut the secondary winding off and wind some #12 wire in its place.  I think it works out to be 1 turn per volt.
If you use one from a 750 watt or so oven, it will carry a pair of 813s.  I have an amplifier running 813's and that is where the filament transformer came from.
Don't forget to make the CT connection.

The 813 runs on DC in the ART-13. In some book it says to periodically reverse the polarity on high power filaments operated on DC so as to even the usage of the filament emission material. A latching relay based flip flop would be statistical enough do so automatically each time the set was powered on.

I think DC is better than AC. Sure handles one hum source and sometimes allows a direct connection to ground. Bypass whatever your RF path is on both sides, whether to ground or perhaps the input of a GG hookup.

Some very early directly heated tube circuits, they actually counted on the voltage drop across the filament to do the biasing of the stage.

In some cases DC results in much shorter filament life although it probably isn't a problem with the high current 813 filaments.  DC operation apparently can result in localized hot spots causing rapid consumption of available filament life.  In any case finding the proper filament transformer isn't difficult and avoids concerns of a complex regulated DC supply where it isn't needed.

yep, looked at the Hammond transformers.  A bit out of my price range.  But it sounds as if a transformer is the only way to go.  Didn't think about the center tap.  OK have to rethink.

Here is a cheap one on eBay: 161579683457   Ends shortly.

Many others as well.

Use a pot or Variac in the pri to get 10V.

73DG

A pot would be also a nice current limiter at turn on.

DC is frequently used in audio amps where one wants to eliminate hum.

However, the "low side" has low voltage, like zero at the bottom, so it does not light up quite the same
as the high side. The recommendation to flip the fils makes sense in this context.

I seem to think that one could lift the ground side up with a suitable resistor and not have the zero volt condition? Of course one is dumping some voltage that way.

People also have been known to run the DC using a "current regulator" as opposed to a voltage regulator.

                                     _-_-

How about a DC supply with +5 and -5 attached to either side of the filament?