sorenb

What makes you do that over and over again? do both phases require adjustment? it might be a clue to look at the CCS's, and whether those are properly connected to the heatsinks ... You might look at the amount of current being supplied by each CCS with and without hum

20B octals 300b UX4 I got inspired from your Octal/UX4 pattern ;o)

Yep, EL34 for the CCS tubes and (EMS 20B or 300b) for the output tubes Here's the 20B version before power switch/volume knob etc. Kevin already posted what is necessary. Using 20Bs is pretty straight forward, 300b is another story and requires a bit of tail resistor tweaking. Not using either, but you probably figured that out looking at the above pic 🐵

Have been tinkering a bit with the XL'ish. Build one based on EMS 20Bs and another based on EMS 300b (picture) A major improvement to the Megatron for sure. Really enjoyable to listen to, toe tapping. Whether 20B or 300b I found those pretty similar sounding in this configuration.

maybe start by checkin PSU rails at the AMP boards entry points, and check that all diodes are properly lit

PM me your address, I have plenty

Sounds like it's not only the +500V having a problem ... Have you measured the drop across the "current setting" resistors for the various 10m90's? Summing up those per rail, might shed some light on the matter ...

The PSU's I have at hand consumes ~25W unloaded Exploding 10m90's indicate massive current draw ... have you checked for any carbonization in the vicinity of those?

@jokerman777 Based on your post's I'd recommend disassembling both angle brackets, making sure all holes are properly deburred, and all the tappings are clean, also clean all semi's. Make sure the screws can go into the tapped holes using only finger force - if not re-tap the hole. Re-assemble, using regular compound (the white stuff). As for the buzz, you stated to have observed a flash in the region of the negative rails, and showed a burn'ed screw from the positive rail, showing that something was cookin here as well, and might have cause excessive current draw from the transformer causing it to buzz

@jokerman777 The thermal compound looks kind of grey'ish ... do you use some exotic compound? The screw needs to be able to slide into the 7721 without any force, otherwise it is too big. Maybe you found the cause of the HV transformer buzzing with "no load" ...

STTH512FP is ready available and fully insulated. If you look through the KGSShv thread, some had troubles with un-insulated rectifier tabs shorting.

@jokerman777looks like you did not use stth512's ?

As I understand, the channel has been working all along, just thrown off by the bal servo. Wonder what the benefit of troubleshooting with the LF353 installed are?

Look at the Uberamp output stage.

.... are D19/D20 the correct value and both oriented correctly (not flipped)? There are some good advice to be found in the posts from @simmconn, make sure you have read those.

Did you test Q24 (J79 in R+ ) ?

To me it seems R+ is not working correctly as it seems to be stucked at ~0V ...does it swing toward B+ at power off? have you tried pull the LF353? does the EL34 heater glow? maybe swap the EL34's in that channel?

Provided the readings are right, I'd look at the right channel LF353 and re-check all the diodes in the right channel servo are properly oriented if everything seems right, pull the right channel LF353 and see what readings you get.

Kevin, wish you a UM66T08

"you should get ~2mA" probably a typo, ~1mA is what I usually get, depending on the DN2540 ...0.5mA might be just fine

I've seen similar. DN2540 leaking usually shows up as the output voltage of the PSU drifting. As far as I remember the supplies breaking down was all wired as negative rails. However, I have not seen it on any GRHV using the cpc1117.

did the SiC survive (someone else’s board with a TO220 DN2540)?

Kerry most likely points to the zeners dynamic impedance which varies with current 1.3mA might operate the zeners too close to the knee

did you check the zener string is actually seeing ground? and the 3k is actually 3k value, and sees source and gate?

did you test whether the css is actually working?