Is my Gibson Atlas blowing up my tubes or am I just unlucky?

[edenexter]

Again I'm not clear, sorry. I have a 10k resistor + a 25K trim pot acting as my adjustable bias. Thanks for the tips!

[Rhodesplyr]

To echo and amplify, pun intended, the advice of others above--what if you removed your bias mod, put the bias circuit back to stock form, and then tested it to see if it still fried tubes? That would be much more informative than testing the power supply voltage with no tubes installed. In general, you can't get any useful information at all about tube operating points if you have no tubes installed. The fact that this is your only mod and that you're having problems is a very large red flag to me.

If you're unwilling to try that, could you at least tell us exactly how you have the trim pot wired? I can think of at least two ways you might have hooked it up. And is the trim pot of sufficient wattage? Most small ones won't take very much power; it's not what they were designed for.

We also might start thinking about leaky bias supply capacitors.

[edenexter]

I took a while off from this project, but I'm back for to fix this now. Thanks a lot for all the previous comments, they helped quite a bit.

I installed a new bias control at R34 using proper components (22k resistor + 20k multi turn trimpot) wired as a variable resistor. I set it to give my maximum negative voltage on pin 5, which is around -56.5v.
I have a new pair of Sovtek 6L6WXT+ installed and just took some measurements.

Plugged in the 8 ohm speaker and turned on the amp to let it warm up.

Measuring the plate current at 458V. And 38mV on pin8.

38ma * 458v = 17.4 watts. So I'd say I'm in a safe range now. Amp sounds pretty okay but distorts quickly as I up the volume to about halfway.

Now here's another newb question: I had the multimeter still connected to pin 8 and on fairly loud strums the voltage reads over 100mV! Everybody talks about what the reading should be when idle, but what are safe/normal readings while in use?

[MWJB]

Idle current reading is set lower than max dissipation, because in a fixed bias design (& even in a cooler cathode biased design) plate current may rise significantly when a signal is applied...it is to be expected. Frankly I'm surprised you're only seeing 100mV...you can see twice that!

The only thing you can do is set the idle current, you should be fine as you are, even if you set a little cooler (15W?) you will still see a massive rise in signal current...if the signal plate current bothers you, then just stop looking at it! ;-)

Seriously though, you're not the first to have this reaction, it can be a shock.

[edenexter]

Haha! I just don't want to burn up anymore tubes, so I'm extra careful. For a while I was a little scared of this thing, but I'm getting over that.