Gibson GA-95 RVT

[Travst]

Corrections submitted above... I forgot about the difference in tubes.

[Firestorm]

It looks like V4a (pins 1, 2, 3 ) is not grounded. The 277V plate reading seems to be the B+ from that node (you can verify that: see what voltage you get at the anode of C29A).

Find R28 (I can't see it in the pics). It should be a 3K3 bypassed by a 1uF (unless you changed it). One end should connect to V4 pin 3; one end should go to ground. With the amp off and discharged check: resistance to ground from the grounded side (should be 0); resistance to Pin 3 from the "hot" side (should be 0), resistance of R28 itself.

Somehow the ground connection is bad, which is why you get no current flowing in V4a, which is why you get no voltage drop. This may be intermittant. It could be bad solder joints, but it might be a broken cloth-covered lead. There is no grounding bus in this amp, per se. Just that middle row of eyelet/cups which mostly get connected to ground points by wires. The wires are probably solid-core and they do break.

Please confirm that the tube in V4 is a 12AU7. I am more concerned with the plate voltages in V4 than the cathode voltages. At least one of the two schematics is dead wrong there. The Apollo version shows V4 as a 12AX7 with and shows V4b developing 4.5V across a 1K5 cathode resistor with a 100K plate load; V1a gets only 1.5V across a 1K5 with a 100K plate R from the same B+ node. 6EU7 and 12AX7 are identical tubes except for pinout and internal capacitances, so both cannot be true. We have to take some schematic voltages with a grain of salt.

[Phil_S]

FWIW, my Minuteman preamp is nearly identical and uses a 6EU7 for V1 and a 12AU7 for V4. I tried a 12AX7 and it wasn't right to my ears. It didn't make the amp appreciably louder either. I'm attaching my drawing of the schematic that is glued to the cab - I was unable to get a good photo. This is not the commonly published schematic, but the "other" GA20-RVT. I have crossed paths with others who have my version of the amp, so I know it was not a production error. I believe this is an accurate reproduction with the following exceptions: I had no symbols for the following: 5-pin plug for the f/s, incandescent-LDR optoisolator, reverb tank, correct looking transformers - no transformer with 2 secondaries so filament circuit is completely omitted, the on/off switches with three positions (on-off-on), shielded cable (f/s). All of these are improvised and you shouldn't have any trouble "getting it."

It is possible that a component is missing. I believe there is no C15. This disturbs me (not enough to lose any sleep) but I have never found it.

Because my preamp is so similar, I'm also attaching my drawing of the board and terminal strip. I wouldn't be surprised if it is a very near match.

I had forgotten how useful these were and, until now, it didn't occur to me they might apply. I made both of these quite some time ago to help me when I was working on the amp. They proved their worth many times over.

Big differences between the Minuteman and the Apollo: cathode bias EL84's vs fixed bias 6L6.

[Travst]

Find R28 (I can't see it in the pics). It should be a 3K3 bypassed by a 1uF (unless you changed it). One end should connect to V4 pin 3; one end should go to ground. With the amp off and discharged check: resistance to ground from the grounded side (should be 0); resistance to Pin 3 from the "hot" side (should be 0), resistance of R28 itself.

R28 hides under two caps.

resistance to ground from the grounded side = 0
resistance to Pin3 from the hot side = 0
resistance of R28 = 3.4K

Now... there was a 12AX7 in V4. The top of the amp is marked in Sharpie with that, and there was a 12AX7 in it all this time. I picked up on the difference when I tested tubes, but had forgotten about it. I placed a 12AU7 in V4 and took these measurements:

V1-4 1v
V1-6 150v
V1-8 2.6v
V1-9 1.2v

V4-1 112v
V4-6 86v
V4-7 167v
V1-9 1.2v

Volume was level on both channels, but was still at the 5-8w level.

Dumb mistake on that tube. New chart is attached.

[Firestorm]

You excel voltage chart doesn't match what you've written here. You also say you've got 86V on V4 pin 4. Pin 4 is a heater. Recheck that. You're looking for the pin 3 voltage. You've got 3K3 (measures 3K4) on one V4 cathode. What cathode resistor is on the other triode of V4?

Changing the tube in V4 seems to have changed your voltages on the problem triode. It's not likely the tube. Probably the sockets are full of crap. If you have a can of contact cleaner, it might be good to give each preamp socket a spray and insert, remove, insert, remove the tube in each socket. The old in-out in-out. Might be some of why the voltages change each time you measure them.

[Travst]

Sorry Fire, the worksheet is correct. I'll clean the tube sockets again.

[Firestorm]

I don't think the numbers for V4 pins 6, 7 and 8 are right on the worksheet. Even the "schematic" numbers don't jibe. After you get sockets cleaned up, how about a fresh set of measurements for V1 and V4. Then we'll think about why so little power is coming out of the PA. Phil raised the possibility that the interphase transformer may be dicey. Could be. Wish they'd never used those things ...

[Travst]

I've attached a new worksheet done after cleaning the sockets. Note the voltages for V4 are back up, particularly at Pin6. After seeing this, I reseated the tube twice and got the same readings. I can't understand all the fluctuations in voltage.

[Firestorm]

Dirty pins and sockets contribute to all sorts of things. The voltages seem a bit more "expected." What is the resistance of R12? Could you measure across the resistor itself and also pin 8 to ground to verify that it is the same. There is a lot of variability in tubes, so nothing ever matches the schematic. + or - 20% is the norm.

[Travst]

Dirty pins and sockets contribute to all sorts of things. The voltages seem a bit more "expected." What is the resistance of R12? Could you measure across the resistor itself and also pin 8 to ground to verify that it is the same. There is a lot of variability in tubes, so nothing ever matches the schematic. + or - 20% is the norm.

I lifted the leg on R12 and it measures 3.52k and is the same as the measure at Pin8.

[Firestorm]

If you have another 12AU7, I'd be curious whether the cathode voltage on V4 pin 8 comes up above 2.2. It's odd that the plate volts and cathode volts would be low. With more current, the plate goes lower and the cathode goes higher. Now with clean sockets, maybe short the pin 7 to ground and measure plate and cathode one more time.

[Travst]

If you have another 12AU7, I'd be curious whether the cathode voltage on V4 pin 8 comes up above 2.2. It's odd that the plate volts and cathode volts would be low. With more current, the plate goes lower and the cathode goes higher. Now with clean sockets, maybe short the pin 7 to ground and measure plate and cathode one more time.

With pin7 of V4 grounded, pin6 is 153v and pin8 is 7.6v. I don't have another 12AU7 to test with.

[Firestorm]

That implies more current flowing in the cathode and less current flowing in the plate. Shorting pin 7 is essentially the same as turning the volume pot all the way down (assuming the pot reaches all the way to ground) so those results are not what I expected. Tube may be bad, but that's a failure mode I'm unfamiliar with. You could try a 12AT7 if you have one, or just go back to the 12AX7. We've seen that a properly working 12AU7 triode gets cathode and plate voltages in the territory (other triode, though).

Have we ever lifted the cold side of C4 to check for DC?

[Travst]

I have a 12AT7 and inserted it in V4.

P1 - 192v
P3 - 4.5v
P6 - 196v
P8 - 4.5v

lifting the grounded side of C4... getting about 50mv DC with the amp on.

[Firestorm]

50mV is nothing to worry about. The voltages with the 12AT7 look right and both halves behave the same. For curiousity's sake, what is the node voltage on the anode of C29A? So you will eventually need a new 12AU7 for that position, but Phil's experience is that 12AX7 would work. When you had a 12AX7 in V4, was it a new or known good one?