Twin Reverb 2 with 20Hz oscillation at the Phase Inverter

[christarak]

I am hoping you can help with trying to stop this 20Hz oscillation at the Phase Inverter of a Twin Reverb 2.

I am having difficulty getting more than 60watts RMS at the speaker terminal (22.5VAC RMS using Fluke 87 V into 8ohm resistive dummy load). The amp came in for buzzing, minimal clean headroom. I changed the original caps and the power tubes. The buzzing has gone, but the clean channel is clipping at less than 60w.

All the DC supply voltages seem to be spot on. I was tracing the 100 mV p-p sine wave through the clean channel until I came across this huge 20Hz oscillation at both plates of the PI. The supply at the other side of their 82K and 100K plate resistors looks clean. The strange thing is that just touching the pin 1 plate of the PI with the Fluke sent it into oscillation. Removing the probe did not stop the oscillation. I am hoping that with this oscillation gone, I may get more power from the amp. With a 1kHz 100mV sine wave at the input, the sine wave is superimposed on the oscillation resulting in a very ugly signal. I have tried several different 12AT7 and a 12AU7 as PI. I have chopsticked around and can't find any dodgy joints.

Here are the voltages around the PI. All volumes set to 0, Reverb at 0, all tones at 5, signal generator removed.
Pin AC (oscilloscope) VDC
1 49v p-p @ 20Hz 172v
2 38.7v
3 4.5v p-p @ 20Hz 46.9 v
4 Filament 6.3VAC
5 Filament 6.3VAC
6 48 p-p @ 20Hz 205v
7 39v
8 4.4v p-p @ 20Hz 46.9v
9 Filament 6.3VAC

I am on 240VAC 50Hz mains power.

Thank you

Twin Reverb II.pdf

[Mick Bailey]

A low-frequency oscillation suggests to me something involving the larger value caps and my first check is to ensure the supply nodes are being decoupled correctly. Did the amp oscillate before you replaced the filter caps? Check the solder joints with your DMM (from the cap leads to their destination) just to be sure and check that nothing came adrift.

[cerrem]

Check for a open resistor in the PI ....
Had a very similar situation a few years ago...
It wound up being a open resistor in the tail circuit of the PI .... I believe it was one of the grid to cathode resistors..
Either way...also as was previously suggested to make sure you have proper de-coupling ....

[pdf64]

It’s a good idea to check all 0V connections too; these are often common to several stages, so if they go bad, feedback loops are then created, in just the same way as on the HT side with decoupling issues.
It’s easy to overlook 0V issues so bear in mind that those connections still carry HT current.

[christarak]

A low-frequency oscillation suggests to me something involving the larger value caps and my first check is to ensure the supply nodes are being decoupled correctly. Did the amp oscillate before you replaced the filter caps? Check the solder joints with your DMM (from the cap leads to their destination) just to be sure and check that nothing came adrift.

Will do, Mick. I wish I knew if the 20Hz oscillation existed before I replaced the main filter caps, or if I created the problem. I'll check those joints and get back to you.

[christarak]

Check for a open resistor in the PI ....
Had a very similar situation a few years ago...
It wound up being a open resistor in the tail circuit of the PI .... I believe it was one of the grid to cathode resistors..
Either way...also as was previously suggested to make sure you have proper de-coupling ....

Thanks Cerrem. I'll let you know what I find today. From the advice so far, I am starting to suspect I did something wrong with the filter cap replacement, but I'm definitely checking the resistors all around the PI.

[christarak]

It’s a good idea to check all 0V connections too; these are often common to several stages, so if they go bad, feedback loops are then created, in just the same way as on the HT side with decoupling issues.
It’s easy to overlook 0V issues so bear in mind that those connections still carry HT current.

You are right, pdf64. I have overlooked the 0V connections. I've got a good check list now. :-) Thank you.

[christarak]

OH the shame. I am almost too embarrassed to say what the problem is, but maybe there is one other person who may benefit from my humiliation. I performed the tests recommended and the oscillation was still there. I thought I'd try the amp in the cabinet. No oscillation. Back on the bench with 8 ohm dummy load. Oscillation back! I checked the Fender speakers:a resistive load of 6.5 ohm. Good. I checked my 8 ohm dummy load: Open circuit!! The earth lead had broken off in the speaker lead. Thank God the O/T didn't fry. My lesson: double check the impedence of whatever I plug into the speaker socket. Do not assume

So now its back to see why the Normal channel is putting out 10V RMS at the speaker terminal compared to the Overdrive channels 27V RMS. I'll report back.