inrush current kills tubesocket pin 9 [Conn organ conversion amp]

[eroometep]

i'm guessing this enourmous transformer was powering about a "1.5 gallon jug worth" of 12ax7's filaments associated with black and white organ keys. And now "overcurrent's" the few remaining tubes [of the now just a simple 20watt or so power amp], which look like a pair of pushpull "EL84's or so", a couple preamp tubes and a "gas-glow" regulator valve, [A-02 ?] known for it's 70v drop.
An output tube sockets pin 9 shows me lightning, the scorched insulator and copper of pin 9 ripple with dazzling plasma very soon after startup, I assume the socket temperature needs a second or two to ramp up before before the lightning initiates socket destruction.
What hack could I use to have startup and run without the filament "over-supply" from the [Ginourmous Xformer] ?
Is it possible to get the Xformer filament supply to run a fraction of tubes for which it was designed ?
Hypotheses include huge resistor, huge resistor that heats up and changes R, adding filaments to load the supply, Lm317's, Separate supply transformer...perhaps you can imagine something which is quick, light, small, reliable and free [lol]... like it's not actually there but reduces filament destruction from the now [after conversion reduced the #of filaments by 1/6th or more] seriously overrated Xformer filament tap. Thanks for reading, the amp sounded really great though.

[Leo_Gnardo]

If you choose a new, separate filament transformer be aware they don't cost all that much. And they don't take up a whole lot of space. For an amp that has a couple EL84 and a few more 12AX7, all you need is 3 amps or so at 6.3V.

There may be a leakage path at that first tube socket where the filament leads from the power transformer connect.

Why did this happen in the first place? First, consider that on startup, all those cold filaments look like a short circuit to the transformer so there's an enormous current jolt maybe 6x more than the average running current, give or take. In the old organ, with all those tubes, that's a ship load of current! Perhaps the connection at that first tube wasn't quite perfect right from the factory. Or the heat from those turn on jolts eventually compromised the solder joint. Eventually the heat from an "ohmic connection" one that isn't a near perfect conductor but rather a resistor, chars the nearby socket into carbon and that carbon conducts filament current to ground with a welding display. Either way here's your opportunity to improve the situation. Hope you get your project working, without spectacular fireworks on power-up.

[nickb]

You did say pin nine of the EL84s right?

That is the screen grid. Seems to me that it's arcing. Replace the arcing sockets.

The only issue I see with the lightly loaded transformer is all the voltages will be a little higher than usual. I would not expect that to be a problem.

[eroometep]

Thank you for the experienced response. Instead of hoping for a quick workaround-fix I'll press in a more appropriately sized filament xformer, and swap out the failed socket. I tried hardwiring right to the tube pin hoping that path would be a more attractive current path, but the path through the socket insulation remained viable with continued toasty plasma effect. I'll be ''rebooting'' the filament supply...

[nickb]

Thank you for the experienced response. Instead of hoping for a quick workaround-fix I'll press in a more appropriately sized filament xformer, and swap out the failed socket. I tried hardwiring right to the tube pin hoping that path would be a more attractive current path, but the path through the socket insulation remained viable with continued toasty plasma effect. I'll be ''rebooting'' the filament supply...

Why? Based on your description the filaments have nothing whatsoever to do with the problem. The filaments are on pins 4 and 5 but your problem is on pin 9.

[Enzo]

Maybe he is assuming pin 9 is part of heater as on a 12AX7. It isn't.

[eroometep]

http://tdsl.duncanamps.com/show.php?des=12AB5
PUsh socket was brown wafer style, Pull socket black fancy style, LMto believe the pull socket had been replaced...new socket is in, amp plays perfect again ! Should probably be louder, one side of the 12ax7 is unused ! ! I've got the 100k and a .022uf...looking for a fat 1k5...thank you for helping me sort this out @ !