EL84 Cathode bias push-pull oscillating

[kanurys]

Hi there,

I'm working on an old Silvertone with two EL84's in push pull, sharing a single cathode bias resistor. When the output is increased to 3+ watts, 1KHz sine wave input, There is a big high frequency oscillation on the lower half of the wave-form. The oscillation stays there even if I swap the tubes. I've tried a known working phase inverter 12ax7 to rule that out. Any thoughts?

This is a Silvertone 73, I'll attach a schematic of the power section.

An externally hosted image should be here but it was not working when we last tested it.

 

[kanurys]

The oscillation seems to go away when I disconnect the negative feedback circuit from the output transformer secondary winding. Do you think this points to a capacitor issue?

 

[kodabmx]

Could be bad coupling caps. hard to tell without actually pulling them out of the circuit and testing them. By that time you might as well replace them.

 

[kanurys]

Well, I increased the 47pF capacitor to 147pF that goes from the phase inverter input grid to ground. That attenuated the oscillation.

 

[kanurys]

Could be bad coupling caps. hard to tell without actually pulling them out of the circuit and testing them. By that time you might as well replace them.

Thanks for the suggestion. I tested the caps and they were all good. The 12K negative feedback resistor measured about 1K. I replaced that and it only helped the oscillation a little bit.

 

[Osvaldo de Banfield]

Try also a snubber across the entire primary of the output traffo. Try with a resistor of 1.2 times the impedance plate to plate, and a cap of about 4700pF @ 1KV, Decrease the cap or increase it until no more oscillation appears.

 

[kodabmx]

Well, I increased the 47pF capacitor to 147pF that goes from the phase inverter input grid to ground. That attenuated the oscillation.

If you have a scope, you can check if this change has effected the HF response.

Too large a cap tends to round off the nose of a square wave quite a lot

 

[kanurys]

Try also a snubber across the entire primary of the output traffo. Try with a resistor of 1.2 times the impedance plate to plate, and a cap of about 4700pF @ 1KV, Decrease the cap or increase it until no more oscillation appears.

Thanks for the suggestion. There is already an appropriate snubber across the primary. This is a 50kHz oscillation. It seems to be completely gone with a little bit more capacitance at the input to the phase inverter. If the problem persists I'll try to snubber cap, as well.

 

[Osvaldo de Banfield]

I can see in the pic, a cap of about 50pF (?) from first plate to gound. This cap places a HF rolloff that ensures reducing the gain in high frequencies. Please, check if this cap is OK, and eventually add a small cap (33pF) across it, to decrease still more the HF cut off frequency.

 

[kanurys]

I can see in the pic, a cap of about 50pF (?) from first plate to gound. This cap places a HF rolloff that ensures reducing the gain in high frequencies. Please, check if this cap is OK, and eventually add a small cap (33pF) across it, to decrease still more the HF cut off frequency.

Thank you. That is exactly what fixed it. A larger value than 33pF did the trick. The speaker for this old console has no tweeter, so high frequency performance is not an issue. The existing 47pF cap was good, but not big enough for the variance in other parts of the circuit.

 

[Osvaldo de Banfield]

Congratulations.