Amp cuts out when driven, HELP!

[Firestorm]

Make the grid leaks and coupling caps smaller. But the problem is too much signal. Reduce gain.

[WRC34]

Ok, but which resistors would be considered the "grid leaks"? (where in the circuit would they appear?)

Is there any other way to reduce gain in general between the two triodes in V1?

[JazzGuitarGimp]

Yes, they are grid to ground. 1M should be fine for all preamp stages. If there is another path to ground (say the grid ties to the wiper of a pot, and the bottom leg of the pot is grounded), then that is sufficient for that stage.

But if you've got signal coming into the grid through a coupling cap, and no ground reference on the gridd, you will get the symptom you are describing.

Good luck!

[WRC34]

Killer, thanks JazzGuitarGimp! I am guessing that the first triode of 6SL7 V1 (the one that is connected to the input jack) achieves this 'grid leak' from either the 1M resistor that is connected to the legs of the input jack (and then ground) or perhaps through the volume pot? If so, then I would just need to add a 1M resistor from grid to ground on the "second" half of the 6SL7.

I'll get it set up tonight, put it to the test tomorrow then report the results.

[tictac]

sounds like parasitic oscillation to me.... tricky to troubleshoot

[Firestorm]

sounds like parasitic oscillation to me.... tricky to troubleshoot

Can't tell without a pic. I'll stick with the blocking distortion diagnosis till then.

[WRC34]

Firestorm & JazzGuitarGimp - you guys were right. I installed a 1M resistor between the grid of the second half of the first 6SL7 last night. This morning I tried not to get too excited since if this didn't work then it probably meant I did have a parasitic oscillation and the amp would need to be disassembled and rebuilt. BUT when I fired it up - pure glory!

I'm calling this amp "Yeah BUDDY!" because that's how I felt when it worked properly today.

The cabinet is one that I ordered from the eBay seller 'sorrycharly' and was damaged in transit. It is no longer suitable for sale since it's all chewed up, but I decided to sand it down and finish it anyways.

Inside the PTP wiring might look like a prime candidate for a parasitic oscillation, but I have built many amps like this (this I think is the 9th one) without ay issues before. I have only begun venturing into the 'cascaded' gain stages territory so this kind of hiccup is part of my learning process.

Thanks again everyone!

[Firestorm]

More Jazz than me. Without a grid to ground reference tubes just run away . I've never heard what that sounds like. Sometimes the grid load is a pot, in which case a high value "safety" resistor protects against the wiper going open.

[JazzGuitarGimp]

Good work, WRC!

[Firestorm]

I hate point to point (well, no actually I hate working on them. Too many layers). But hats off to you for tackling it and solving the problem.

Like I said, Jazz called this one. A tube permanently overdriven (through lack of grid to ground reference) is one thing. Blocking distortion is the other side of the same coin, a cap that can't discharge fast enough (because of too high resistance to ground). The usual solution is to make the cap smaller, thus changing the time constant. Might attenuate lows. You can drop the load, too, and sacrifice some gain. Which you make up somewhere else. And then you start over.

Damn you Isaac Newton. (That was a Calculus complaint.)

[WRC34]

Ha! Firestorm, ptp has been a big challenge for me to keep tidy. I swapped components so many times trying to diagnose this issue and to do so indeed required much un-soldering of components and layers to get at what I wanted to change. ugh.

Thanks again for everyone's help and also thanks to Littlewyan for being the first one to reply and get the conversation rolling.

[Firestorm]

It's the only way they made them up through the '40s. Wired one "layer" at a time. Also the reason for resistor color codes (and I guess banded and dot coded caps). You could discern the colors under incandescent light.

Still hate them. The trend persisted. I have a '66 Thunderbolt on my bench now, getting ready for sale. It had an unused triode so of course someone f'ed with it. Also under filtered. Fenders from ten years(!) earlier were better thought out. Good you got it going.

Now, build a proper "parts panel" for the next.