Twin Reverb Issue

[tdlunsfo]

I have a weird one. Late 70s UL Twin Reverb came in with a blown fuse and an open 22uF electrolytic. I recapped the amp, replaced the screen resistors, tubes were tested with no shorts (output tubes are newish Tung Sols). I fired the amp up, gave it a 20 count, flipped it to play mode and I get this loud hum and another blown fuse. Recheck my work and find no issues. This time I brought it up slow on the variac and it starts fine and plays great. Turned it off and on a few times and no issues. I come in this morning, fire it up again and the heavy current draw is back. This time I am able to reduce my variac to keep the fuse alive and around 1.5A I hear a pop and the heavy current draw dissapears. Almost like a part is failing under high startup voltage. Anyone ever experience something like this? I subbed in new tubes and it still does it. I lifted the coupling caps to the 6L6 grids and same thing. Rectifier and bias diodes test fine.
Fender-Twin-Reverb-SF-135-Schematic.pdf

[The Dude]

Do you have the overcurrent problem with no output tubes installed? I'm wondering if you have a bias supply problem. I'd remove the output tubes and check voltages on the sockets- specifically bias voltage.

[tdlunsfo]

When it pulls heavy current it pulls the whole B+ down so it's hard to tell if it is bias or not. The last time I got over current, I was able to two pull a set of tubes but no difference, of course when I pulled the remaining tubes the overcurrent dissapeared.

[The Dude]

Well yes, if you lose bias, tubes will over-conduct and B+ will drop. Best to just check bias supply. When you recapped the amp, did you install the bias filter cap correctly (plus side to ground)?

[tdlunsfo]

Bias cap is fine. I can start the amp slowly on the variac and it will play. I can see the bias voltage..it's there but dropping along with the other voltages. I also tried adjusting the bias balance pot while pulling heavy current but no change.

[g1]

See if bias voltage drops at grids when power tubes removed.

[tdlunsfo]

Bias voltage is stable with tubes removed but then again it will also not pull heavy current with tubes removed.

[g1]

Then I would try some other power tubes. Or remove the ones you earlier called 'the two remaining tubes' and try with the other 2.

[tdlunsfo]

One of the issues is getting this thing to mess up consistently. Just to show you how weird this is, here are are a couple of pics. The meters on the side are looking at plate current. The Fluke is looking at B+. The left meter on my variac (blue box) is AC current and the right is AC voltage. The first pic is a high current event with the line voltage reduced to around 80v for it to stabilize and keep the fuse from blowing. I then drop my variac down to around 20v, I hear a pop and ramp it back up to full voltage to get the 2nd pic where all of a sudden all is well. I've tested these tubes in my hickok three times... no shorts. Then again I am pretty sure my tester doesn't put this kind of plate voltage on the tubes (~500v). If I ramp the line voltage up in it's high current state, one set of tubes nosedive and the other skyrockets while B+ dives. It doesn't seem to matter which of the tung sols are in which position or which of the quad are installed. Its almost like these particular TungSols cant handle the plate voltage. I have a spare quad of JJs in it now, I'll report back if the issue continues.

[Helmholtz]

What kind of bias probes are you using as your meters run in current/mA mode?

[Tom Phillips]

Possibly one or more of the tubes in your Tung Sol quad has an intermittent short or is going into thermal runaway. If the JJ set performs OK then you can decide to just go with the JJs or spend the time to sort out the bad Tung Sol by process of elimination. That can be worthwhile when it is your amp and you want to cull out the bad tube(s). However, if you are working on the clock then it's often not economically practical.

[tdlunsfo]

Bias probes are the simple inline plate current type. One is made by eurotubes and the other a homemade copy.

[Helmholtz]

Bias probes are the simple inline plate current type. One is made by eurotubes and the other a homemade copy.

Are you sure the probes want the meter set the mA and not to mV?

The probes I use insert a 1R resistor at the cathode and require the meter set to mV range.
Principle is reading the voltage drop across the current sense resistor and divide the voltage reading by 1 Ohm to get the current.

Using a current meter only makes sense if the probe disconnects the tube from the circuit and connecting the ammeter closes the circuit.
In other words the tube won't work without the meter connected.

If your probes actually "measure" plate and not cathode current like mine, the probe leads need to be connected to the plate.
Some amps don't like this and oscillate causing wrong current readings.
Also don't like that such probes put full plate voltage on both meter leads. If the amp oscillates plate voltage may exceed 2kV peak.
To prevent oscillation pull the PI tube during power tube measurements.

[tdlunsfo]

Positive. I've been using these probes for years. These route the plate through the meter to calculate the current. There are no resistors, and to be honest, somewhat dangerous with B+ on your meter leads but I like their simplicity.

[Helmholtz]

Nothing more simple and safe than measuring cathode current.via a sense resistor.
Hardly any handheld DMM guarantees safety with 2kV wrt ground.

Be aware of potentially wrong results caused by oscillation with your probes.