gate resistor burning up

[sbirkenstock]

Hi everybody,

I got a Marshall 6100.
It uses 5k6 gate resistors for the EL34 on the board.
Two of them created some smoke (by coincidence while I was maintaining some noise pots...)
Both look burned.

First time it happened, I was not quite sure which exact part was smoking, so I turned the amp off real fast.
Next time I turned it on, it was playing fine.
Around the area where the smoke came from, there is also the 2k2 5 Watt resistor (power supply to the PI).
I was suspecting this at first. The resistor measures around 5k and then it raises on my Fluke 179 Multimeter.
Like to 6K in 20 seconds or so...
So not right either.

Any suggestion what the cause might be?
When this happened, the Amp was set to low power=2 output tubes instead of 4.
Both resistors are on the same phase though...
They are both connected to the 82K "output" of the phase inverter.
With one EL34 switched off, only one resistor should be "active"?

Well, maybe the resistor suffered way earlier when the amp was set to high power = four EL34...


Any suggestions?

cheers,

Stephan

[R.G.]

P = V * I
If the resistors burned, there is too much voltage across them, and current through them - which is two ways of stating the same thing.

There are capacitors in series with the gate resistors from the PI, and the resistors go directly into the GRIDs of the output tubes. If the capacitors feeding the grid resistors are not shorted, the only way to get enough current through the grid resistors is to use high frequency AC. One possibility is RF oscillation.

But the most likely scenario is either a shorted cap from the PI, or something in the output tubes that caused a shorted cap, or a current path (short or other low resistance path) around the caps to dump DC onto the grid resistors. This would pull the tubes' grids high and dump current into the grids. It would also cause high current in the affected tube and some amount of tube and/or power supply destruction. Changing the now-damaged tube would leave the problem with the cap and/or DC path.

Check those caps, and if they're good, look for high frequency oscillation. Then go look for the collateral damage to the bias supply, tube sockets, and power supply.

There's a whole host of other possibilities, but these are the first things I'd check.