mosfet trem blowing PS resistors...

[CascoSieg]

My trem circuit is informed by several I've seen here (TK Imperial, Fender Vibro-champ SS, others...) and is intended to wiggle the cathode bias of a driver stage into a SE power stage. I know little about how SS stuff is supposed to work, but wanted to save space and was curious.  I mocked it up successfully on a breadboard with a test circuit of the final amp design, and planned to make tweaks to the trem in the final build... which is where I'm at now. The amp works fine without the source follower of the trem connected.
When I connect the drain of the source follower MOSFET to B+ it blows the PS resistor for that node - seems like something is just plain shorting out, but I've checked and checked and tested resistors everywhere, and just can't find anything that strays from the schematic, or is likely to be shorting ... except maybe the MOSFETs themselves (?).  Is there any way to tell by measuring their resistances in-circuit, whether they are good (undamaged)?
Is there anything about the schematic that looks like it would cause a big short? Maybe I just overlooked something or just poorly interpreted the various designs I've drawn from?
Thanks for any and all input!

[sluckey]

I think the Drain is connected to the metal mounting tab on the back of the MOSFET. And it appears that you screwed the mounting tab to the chassis without any isolating bushings. Therefore, direct short.

If the drain is connected to the mounting tab you must isolate the tab from chassis.

[CascoSieg]

Hmm.... It is isolated using the rubber pads and screw gasket. I get same behavior when I unscrew and lift the MOSFETs into the air, breaking all contacts with chassis.
Drain to chassis resistance measures 67R in situations (screwed in and suspended), and similar for the other MOSFET: 130R.
?

[CascoSieg]

I went back to my notes from the mock-up and, big surprise, I did not mock up this precise thing - it was the same circuit without the CF (source follower) part of it. I'll reconfigure things and start again from the mock-up prototype, like I thought I had...
Thanks!

[CascoSieg]

Update and Questions:
So, I got some new IRF820s and better mounting hardware, and reconfigured things as a single MOSFET oscillator, and got that working... but it's weak and thumpy.  So I added the follower, and modded a few things to get it to reach nearly full depth without thumping, resulting in the new trem schematic, shown at bottom of the full schematic attached.
The audio coming through the trem'ed driver triode sounds good, but at about 70% to 100% full trem depth I'm still getting clicking sounds in synch with the trem frequency, at a constant volume regardless of the MV setting on that channel. Quiet enough to disappear at higher playing volumes, but just audible when guitar is muted or with quiet playing.
The trem signal/wave itself looks nice and smooth.
The voltage at the driver cathode registers 2.4v to 2.9v with trem at full depth, and 1.3v at zero depth, so it seems the trem is shifting the bias in addition to modulating it. I *think* I used a lower value R on this triode specifically to compensate for that upward voltage shift I noticed in the circuit model. When I subbed in a 800 ohm resistor for the 1.2k that is there, that diminished the click (and everything) but didn't remove it.
10 second audio/video sample here: https://photos.app.goo.gl/9q4BKon57kELyXc99
Am I fighting a trem design (cathode wiggler) that is just going to have this problem, or do we think there is a solution to the click noise somewhere?
If I can't solve it I'll try to implement a version of the Trem-o-nator, but early design scribblings made me wonder whether it would load down the already pretty low signal too much... and I really like the sound of the bias trem. 

[Williamblake]

So you already tried a cap to keep the DC bias unaltered?

[tubenit]

CascoSieg,
That's a really interesting and innovative amp design!       I like it! 

- FET input
- paralleled V1 with dial-able cathode cap values
- James tone stack
- switchable OD gain stage with cathode follower
- very interesting dual single ended power amp with a master volume for both power amp tubes
- mosfet tremolo

Way to go thinking outside the box! Very remarkable approach.  THANK you for sharing it!  While some of those features can be found on other amps, I've not seen an amp topology that incorporated all those in one amp design.  Bravo! 

With respect, Tubenit

[CascoSieg]

@WilliamBlake: No I haven't tried that - would I put that on the follower just before the depth pot? What size cap would you suggest starting with?
@Tubenit: Thanks very much! High praise coming from you.   I wouldn't yet call this build a success, but it's a work in progress. 

[Williamblake]

I'd start with 100V 1mF or similar where it says Tr.

[CascoSieg]

I've been playing with different cap values between the trem and cathode, as well as resistor values in various places in the trem circuit, and diodes in places based on suggestions I've read here and elsewhere. WilliamBlake's 1u suggestion was pretty much right-on in terms of providing just enough attenuation of the modulated signal but not too much - larger (2.2u is the next value I have) doesn't suppress the clicking and smaller .47u really limits the depth.  I'm still hearing a low thumping at my desired trem depths even though the clicking is gone, and it just seems like, with this design, noise is a just a function of trem depth and can't really be separated out on its own. I think this is the best I can do with this design unless anyone has other suggestions.

I may try the Trem-o-nator next.

Thanks!

[Williamblake]

As i understood it, clicking results from the oscillation not being a real sine wave or from the wiggled tube going too far out of bias at one point in the cycle, asking for different improvements in the circuit. Begging to be corrected.

[tubeswell]

Apart from increasing the 47k at the source follower (to decrease the maximum intensity), the way your FETs are laid out (with all long lead wires in parallel close to each other) could be a source of ‘clicking’ as intensity is increased. (The TO220 package is designed to be mounted in a PCB with the excess lead length clipped off.) A ‘fix’ for that could be a 100R on the Source Pin of the SF ( like a stopper), or not.

[CascoSieg]

Ah! Hadn't considered the close parallel wire issue here - I'll try to separate them a bit when bolt things down. I ordered what I thought were 3-pin sockets so I could avoid soldering directly to the MOSFETs, and got something unexpected but workable. Will post final pic when I'm there. WilliamBlake's comment about bias made me wonder how a different tube might work, so I tried a 12AY7 and that solved the noise problem... but of course isn't driving the power tubes nearly as hard. So I tried a 5751, and that did the trick!  : nearly full depth (meaning silence at the lowest point) oscillation, with just the slightest hint of thump at full depth on the knob, and no clicks! Now I need to go back and see whether my Rl, Rk, and Ck are near where they should be, for the 5751, but it seems plenty loud.
(and I was about to give up - sheesh!).
Thanks all!
[EDIT]  Oh, and the 47k is actually now 82k on the MOSFET source follower - still 47k on the OD CF.

[CascoSieg]

Pics as currently implemented, with IRF820s wiggling the cathode of the 5751 driver.
Still planning to add little heat dissipaters (that came with the TO-220 insulation kit) to the MOSFETs, when I have it on the bench next. But they work Ok for now just sticking up in midair... made it through a couple of 2 hour rehearsals so far.