hybrid ga-20/5e3 odd preamp tube voltages?

[ALBATROS1234]

i have been working on this as a harp amp for a friend.at first i was going to try to do 1 channel of a gibson ga-20 with the 6sj7 preamp tube but i kept getting red plates on the power tubes and it only sounded good for a few minutes until the tubes started to red plate. so i decided to splice a 5e3 fender pi and power section with the ga-20 6sj7 . i am using 12 volt tubes actually with a separate filament trans and a big fat step up control trans which gives 230vac out with 120vin. i get 326vdc from the power supply rail.

so the problem is that it has more breakup than i want for a harp amp even though it sounds pretty decent for guitar with a thick creamy crunch. it also seems to sustain in an odd way. when i first strike a chord it sounds good,punchy and crunchy but it seems like it doesnt sustain as well as it should and slightly warbles as the volume decreases. so i checked the voltages.and the 1st preamp tube is really low in voltages. 8vdc on the plate and 26vdc on the screen.

12SJ7- plate 8vdc  screen 26vdc
12SL7- plate 2 used as amp stage-154vdc plate 1 used as pi-135vdc
12AQ5s plates 326vdc and screens 312vdc

i have checked my wiring and it seems right but perhaps i have to tweek some values and i am hoping someone has some advice to give. this is my first push pull amp and 1st one using a pentode in v1.

[VMS]

not sure if amp is missing it but the drawing is missing ground on first stage

[ALBATROS1234]

i forgot to draw it ,the 10 meg is grounded definitely. as are the kathode resistor and the .02 cap which connects to the screen.just fixed the pic.

[ALBATROS1234]

what is just dawning on me is perhaps i need to add a node to the ps? both the 12SJ7 and 12SL7 are run off of the 3rd node . but that is how its seems to be done usually. the ga-20 has the SJ7 and SL7 run off the same node as does the 5e3 albeit the 2 12ax7s. i am grasping at staws i suppose.

[ALBATROS1234]

it must be a bias issue on the 12SJ7 because i have 315vdc before the resistors that lead to the plate and screen. i am not sure if i should go up or down with the 10meg grid bias resistor to get the voltages up on the plate/screen.

[ALBATROS1234]

i just alligator clipped another 10meg in parallel to the 12SJ7 grid bias resistor figuring it would make it 5meg. it made a slight improvement in tone but the voltages on the plate and screen are the same. im confused i thought this would have been a big change.

[VMS]

You have probably done this already but check the plate resistor

[ALBATROS1234]

yeah, i checked its value before i installed it as per my routine. i try to be really meticulous in such things. perhaps i need to change the plate resistor to a lower value. the champ 5c1 has a 5 meg grid bias resistor and has a 2 meg where i have a 1meg on the screen and a 250k instead of 470k on the plate. maybe if i set up like the champ it will be better. perhaps pad the input with a resistor too.

[sluckey]

The balance of that cathodyne phase inverter is totally screwed up. Replace that 220K plate resistor with a 56K resistor (same as cathode). Any better?

Put another tube in the V1 hole. Any better?

[VMS]

Also check that you have 0 VDC on volume pot wiper. (maybe faulty coupling cap)


You can also try to cathode bias the first stage, like is done in ga-9 amp

[VMS]

Oh and do what sluckey says, he is much more successful debugging amps than i am 

[ALBATROS1234]

ok i didnt realize the cathode and plate loads had to match, the ga-20 had a 220k on plate but it used both halves of the triode.

[ALBATROS1234]

what about the 1.5k btw the 56k and kathode? i dont get that,does it not unbalance the 2 56k loads?

[sluckey]

Visit the Valve Wizard's website an read about cathodyne phase inverters.

[PRR]

> what about the 1.5k btw the 56k and kathode? i dont get that,does it not unbalance the 2 56k loads?

Much less than 56K versus 220K.

Yes, 57.5K<>56K but the 3% error seems to be utterly negligible, or perhaps "toneful".

[ALBATROS1234]

well i fixed the 56k resistor and used the 6SJ7 values from the 5A3 champ and now it actually sounds pretty damn sweet. it still has some hum i gotta work out but i am almost there. most complicated build to date.soon  i want to try and build a low power marshall type. likely late 60s to mid 70s but want to find a pair of pentodes that are low dissipation , perhaps a pair of nos rca 6K6gts with plate diss at 8.5 per i could get a 15 watter going.also i do like 6f6gt quite a bit and read great stuff about the 6f6g .they sound different to a 6v6 or 6l6 perhaps because they are pentodes not beam tubes. i think marshall i think pentode power tubes .anyway thanks to all who responded.