Matamp GT120 eating preamp tubes

[Helmholtz]

About 3 years ago MEF member nickb reported on early tube failures in the DCCF position of a newer Marshall amp.
He did a number of tests and IIRC the conclusion was that the issue could be attributed to high grid potential at the CF triode before sufficient heater emission after turn-on.
Chinese ECC83s did not reveal the issue (maybe due to faster filament heat-up?).

Maybe someone is able to find the thread?

[Mick Bailey]

Maybe this is what you are thinking of; https://music-electronics-forum.com/...-tweaks/47283-

[Helmholtz]

Yes, thanks a lot!

[Purgasound]

That's a great thread. Thanks for posting the link.

I did have a feeling it was something similar to what was happening in that thread. The GT120 does not have a Standby at all so when it's switched on there is full B+ on the preamp plates until the tube warms up.

[Chuck H]

Since the PI tube is the one that was failing I'll just point out that though it's not a CF it is a DC coupled cathodyne. Not sure if this could suffer the same failure mode in the above linked thread but it seems possible. At the very least the bias tends to run hot in these circuits when only 12ax7's are used. The bias does seem hot, but not unreasonable in GT120 circuit at around half a volt (negative at the grid simulated in LTspice). But if you reduce circuit impedance for the cathodyne to 47k for the plate and cathode resistors this could be improved to about .9 volts negative at the grid. This shouldn't make any noteworthy difference in gain. If you do this you could also increase the coupling caps to from 47n to 100n to correct the LF knee. Though you probably wouldn't notice it. At any rate, and Helmholtz will correct me if I'm chasing nothing here, I'm hopeful it might help to reduce circuit impedance and improve the bias condition for the cathodyne.

[Helmholtz]

I did have a feeling it was something similar to what was happening in that thread.

That's what I thought (and what made me remember the old thread) when seeing the DC coupling with V2. This essentially results in the same cold start conditions as with the DCCF in the Marshall.

[Helmholtz]

Since the PI tube is the one that was failing I'll just point out that though it's not a CF it is a DC coupled cathodyne.g here, I'm hopeful it might help to reduce circuit impedance and improve the bias condition for the cathodyne.

As said above the DC coupled cathodyne will see the same cold start conditions as the DCCf in the Marshall, as with cold heaters there's no voltage drop at the plate resistor.


If you lower plate and cathode resistors to 47k tube current must double as plate and cathode voltages are fixed by the DC coupling. To me 2mA means hotter bias than 1mA.
What would be the benefit?
Actually I think it will lower available grid drive.
While the lower value resistors will give a somewhat lower source impedance at the plate output (cathode impedance is low anyway) this will not noticeably change the LF corner frequency with the coupling caps, which is determined by coupling capacitance and the bias feed resistors (being the load).

[Purgasound]

I figured I would post a follow up to this thread since I don't think there is any problem with the amp in question...

Low and behold the customer brings me another dead amp. No fuses blown so I followed my instinct and replaced the first stage preamp tube and the amp roared to life. I called my customer and asked him how he was running these and his response was that he was diming the Monitor Output from a Kemper Stage Profiler unit into the front of these amps. That sends a +4dB line level signal and I have no idea what the actual signal level and voltage are coming out of that thing when it's maxed out. That'll surely fry an input driver that's designed to see instrument level signals.

There's a lot I can say on that subject but I'll leave it be. Neither repair was expensive but they did have to pay more than once to learn this lesson.

[pdf64]

Bear in mind that any input mixer / grid stopper resistance will limit grid current at the input stage. And the signal level into the 2nd or 3rd stages of any valve preamp, when turned up, will probably be higher than +4dB.

[Chuck H]

Bear in mind that any input mixer / grid stopper resistance will limit grid current at the input stage. And the signal level into the 2nd or 3rd stages of any valve preamp, when turned up, will probably be higher than +4dB.

I was thinking the same thing. It's not uncommon for guitarists to run over a volt into a tube amp using boost pedals and dirt boxes. But a +4 dB would be something greater than two volts (?) and from a source that can supply a 600 ohm load, right? So, depending on the stringency of that source as it relates to typical audio circuits what happens when you run it into a 1 Meg load? It could be even higher. Could that stress the bias on the first triode? I wouldn't think so because cascade tube stages typically take tens of voltages for the purpose of overdrive/clipping. So this is still a mystery to me. Why this amp should be suffering.

[Greg Robinson]

I agree with the others, that's not an explanation. 12AX7 can take 10's of volts on the grid, even 100+, so long as it has an appropriate grid stopper.
And wasn't this amp killing the heaters? High input signals definitely won't do that.