Serious Issue, Filament Smoke

[ChrisM]

About a month and a half ago I had a working Deluxe Reverb. At the time I was working on reducing hum and tweaking. Anyways one night I decide to redo the input jacks, re solder some grounds on the ground buss bar and shorten the choke's leads.

I turn the amp on and let it warm up, I go to my bench to wait....I turn around and there is smoke coming out the back. Filamnet winding (CT and all) completely charred black, insulation melted away. Now I thought it was my power tubes at the time for somereason, dont ask.

Fast forward to yesterday...Finally installed a new PT and brand new power tubes and fired it up with the amp out of the cab but hooked it up to a cab using a speaker cable. After no issues while warming up I flipped standby...Got sound, hit a few open strings just to make sure. Ok powered down, amp probably only on for about 3min total, 40sec of which was when the amp was fully on.

So today I put the amp in its cab as I wanted to see what the noise was like....I also wanted to check the bias. Anyways within 30sec of warming up I saw the faintest trace of smoke and immediately shut it off. When I took the amp out of the chassis I found the filament winding bubbled and all melty (clearly on the track of the last PT).

So basically power transformers aren't cheap for a 16 such as myself and I am getting super frustrated, super quick.... Such a piss off, I have built three amps prior to this one and they have been just great.

Any ideas?

Thanks...Chris

[bnwitt]

Well obviously you are pulling more current in the filament winding than it is rated for. I would check the heater wiring and all of its connections throughout the amp. Look for other wires touching the heater connections, little whiskers maybe. Disconnect the PT filament wires so you don't have that resistance path on the circuit and then ohm out the whole heater circuit with the tubes in. You have a pilot lamp, 6 dual triodes and 2 6V6's sitting in parrallel on the heater circuit. At 0.3 amps per dual triode heater that is 1.8 amps total on them and at 0.45 per 6V6GT thats 0.9 The total without the lamp (you can pull it and check it separately) is 2.7 amps. With that load, resistance should be about 2.3 ohms unless something is shorted. It its alot lower keep looking.

[Phil_S]

It appears you've got a short on the filament string. Can you post a picture?

I wonder, no basis for this, did you install an artificial CT with two 100 ohm resistors? If you did, you can't have that and the CT.

A more remote possibility is a tube has shorted.

First I would go over the filament wiring very, very carefully. Use your meter to check for continuity, first end to end on each leg and a then between each leg (socket to socket). Remember to check the pilot light. Lift the CT when you do this. Then check each tube socket pin to ground, and the circuit should be open. Look for a whisker that is touching something it shouldn't. Use a magnifying glass.

Then, when satisfied, with no tubes installed, you might be brave enough to power it up and take some voltage readings. Keep the CT lifted.

Stop at this point and post your findings. A picture would be a good thing to post.

Since you caught it early, you may have saved the winding. A bit of heat shrink over that messy insulation will probably do the job for one more go round.

If you blow the winding again, the low budget solution is a Radio Shack filament transformer.

[Structo]

Another good reason to use a current limiting device when powering an amp up for the first time or after major surgery.

I don't own a variac so I use a light bulb limiter.

Pretty cheap insurance to show if you are having excessive current draw.

[ChrisM]

All good thoughts guys, thanks. No I am not using a artificial centre tap.

I do have the pilot lamp touching the chassis but no voltage is present on the part that touches.

Doesn't it seem odd though how the amp has worked fine in the past and had these circumstances where things just failed? It happened with both PTs.

Anyway to test to see if a tube is shorted? I am guessing check for continuity between some heater pins?

Also why the hell isn't my fuse saving me from this mess?

Also does the bias circuit operate when the amp is in standby? Could my issue be there?

[bnwitt]

Well the fact that you did some internal work leads one to believe that something could have been bumped or moved. Is this a reissue or a handwired unit? Can you provide a chassis gut picture for us? Your main fuse is too large to save you on a heater winding short. If this is a handwired unit, disconnect the PT filament wires from the pilot lamp assembly. Pull the lamp out of the socket, check for ohms (not continuity) at the pilot lamps two terminals where the heater wires leave for the tube array

[ChrisM]

Will do, Ill also get some pics up.

[Alexo]

If the filament string was fine in standby mode, it seems unlikely that there is a short there.

What is happening seems consistent with what would happen if you exceeded the cathode-heater reference limits on a tube, like if you had a cathode follower or a pi with 100 volts or more on the cathode while your heater center tap is referenced to 0 volts. I can't imagine this happening unless you redesigned your amp somewhere, but it might be possible that a shorted preamp tube *could*, *maybe* be involved....

[ChrisM]

I am going to add a inline fuse... Would I want a 3A slo-blow?

[rfgordon]

It is possible for a power tube to short plate-to-filament. I had a KT88 do that in a build this year. It smoked the 100 ohm resistors in the artificial CT.

I put the DMM across the pins for filament and plate and got 1k5 ohms resistance! Yikes! That made a very happy path to ground for the 486 volts on the plate!

Check your tubes in like manner. The only pins that should show any conductivity at all are the filaments.

[drz400]

I am going to add a inline fuse... Would I want a 3A slo-blow?

I worked on a Candian Marshall that used in line fuses for the heaters
The resistance of the Slo Blow caused the fuse to get hot and unsolder itself from the PC board. Removed them and everything was fine.