Ampeg G-15 Gemini ll

[Glenn63]

Been away from my bench for a few years and I'm working on this Ampeg Gemini ll for a buddy.
After a catastrophic sounding breakdown I offered to give it a go.
I've replaced two leaking filter caps and 6L6's.
Verified that voltages look normal while plugged into a current limiter (bulb) and dummy load.
Connected to speaker and wall voltage there is a very loud hum, not effected by volumes.
I began pulling valves and found it quit only after pulling phase inverter. 7199 converted to 6U8a.
Installed new 6u8a and hum remains.
All voltages at PI look normal
Scope shows a 250mv anomaly not present at the PI grid but present at all points from plate 1 to grid stoppers at 6L6s.
I tried replacing all three .1 couplers in the stage to no effect.
Any ideas out there?

[Jon Snell]

What frequency is the hum?
Does grounding the pentode grid stop it?
Does disconnecting the triode grid coupling capacitor stop it?

[Glenn63]

Jon, no changes from grounding grid one or disconnecting .1 from grid 2.
Did you also mean disconnecting dc voltage from grid 2?
Frequency is 125 Hz, if I did the math correctly.
Been a while.

[Jon Snell]

What I meant by frequency was 60HZ or 120HZ.
So it would appear to be 120HZ = twice your mains power frequency, coming from the rectifier stage in the power supply, so check the other smoothing capacitors, they are all very old and may be very high ESR and low capacity.
My reasoning is; As the 6U8 is installed there may be poor hum rejection, when conducting, due to poor smoothing of supply C allowing any current drain (your 6U8 when hot) to amplify the ripple.

[Glenn63]

I as well suspected a filter capacitor, which is located in the cap can and filters the C supply- also to the reverb driver tube. I checked for ripple there at the reverb driver but found none.
I'll go with your idea and try bypassing that can filter with a new and report back.. Thanks.

[Jon Snell]

Was capacitor smoothing supply C, a fix?

[Chuck H]

And don't forget the bias supply filter.

[Glenn63]

'Yes' on the bias filter.
No change after changing A, B, and C supply filters.
Checked filter grounds again and read .5ohm to the power cord ground, which should be the meter resistance. I pulled the PI again and observe only a small ripple at the triode plate resistor 47k. Pulled it, checked it, replaced it. No change when tube installed. This is taxing my patience! Lol...
Got to be something simple I'm missing.
Any chance it could be a NFB issue?

[Helmholtz]

No change after changing A, B, and C supply filters.
Any chance it could be a NFB issue?

Did you verify that both power tubes draw similar idle currents? Either using "bias probes" or adding 1R cathode resistors. Any effect of the balance pot?

Regarding NFB, just disconnect and see what happens.

And please post a complete schematic.

[g1]

Schematic attached (sorry, quality not great).

[Helmholtz]

Found a clear one here:
https://ampeg.com/support/files/Sche...Schematics.pdf

[g1]

Looks like the heaters and hum pot are referenced to the bias voltage. Make sure that pot is ok.

[Glenn63]

Ok... G1 had the correct schematic and Helmholtz had an incorrect one.
Obviously, my troubleshooting technique is lacking and I am still missing something.
Using a variac at 115v:
Supply voltages look close enough to spec.
Balance pot is in working order, centered.
Bias was at -54v, adjusted to -40
Replace 5.6k power supply resistor to 10k (closest metal oxide 1w I had) to lower 'C' when at wall voltage.
Placed 6L6 #1 on bias probe and main fuse blew.
Reflow power tube and PI tube sockets. Just in case.

I think this amp deserves a better filter section.
The cap can contains 40,40,10. I think this must not be stock and the C' supply was on the 10mf lug.
I am installing a parallel 22mf 500v filter at A'.
Supplies B' F' and C' will be 47mf 500v filtered with all grounds to a local chassis lug.
Also installing 1ohm cathode resistors.
After completing these changes, I'll return with results.

Thanks for recommendations
.

[g1]

Ok... G1 had the correct schematic and Helmholtz had an incorrect one.

The one Helmholtz posted has 2 pages. The second page has a cleaner version of the one I posted.

The bias is shown as -50V, not -40.

[Glenn63]

Ah, I didn't look for a second page.

BIAS: Yes, I see the bias is marked as -50v but doesn't that sound hot considering everything is running hotter on wall voltage?
After adjustment it actually tested at -46v on wall voltage.

Almost done installing filters and cathode resistors so I can read the current.
Thinking I'll slowly ramp up to 115vac for first bias and cathode current testing.