Self designed tube amp problem

[Manowar1985]

Hi everyone.I really need help with an issue.I have designed an OPS. It is a pair of el84's in push pull that works at fixed bias.It is switchable between pentode and ultra linear mode.Everything seems to work fine when the outputs(the anodes) of the differential driver stage are not connected to the inputs(the coupling caps) of the OPS.When I connect the driver stage with the OPS the power tubes are getting very hot(because the vgk is going very positive).The vgk of the 2 tubes are not the same but both are positive..Such thing does not happen when the driver stage is not connected.In detail when the driver is connected I measure -14 Volts at the grid leaks of the 2 output valves at the other end of each grid leak before the grid stopper I measure.+5.85V and +400mV each.At the grids I measure +2.4V and -50mV.I must say that I did not direct couple the 2 stages.Also the same thing occurs when the preamp tube is removed but the preamp circuit is still connected.I have ckecked the circuit and it looks fine Any help will be appreciated.Thank you

[Helmholtz]

Schematic?

What does OPS mean? Output stage?
Positive Vgk may have already damaged you power tubes. Are your coupling caps good?
If you're using global NFB, have you checked for HF oscillation with a scope?

[Manowar1985]

yes it means output stage..I am not using any global NFB.How do I check a cap?All I did is to check with an ohmmeter if the 2 pins of the capacitor touch each other internally

[Helmholtz]

How do I check a cap?All I did is to check with an ohmmeter if the 2 pins of the capacitor touch each other internally

Not sure I understand.

I'll describe a method to check for coupling cap leakage. But please first post a schematic of PI and power stage to avoid misinterpretation.

[Manowar1985]

I attach the schematic

[Helmholtz]

I attach the schematic

Ok,
1) It seems you're using polarized electrolytic coupling caps. The polarity in your schematic is wrong. Reverse polarized ecaps will inevitably leak and get damaged.
2) Coupling caps need to be rated for at least 250VDC.
3) 1µF seems much too high. The corner frequency would be 0.7Hz. I recommend to use 0.1µF foil caps, rated for 400VDC or more. This would result in a corner frequency of 7Hz, still much lower than a typical OT can transfer.

[Manowar1985]

ok I will switch to non polarized 100n but why is the polarity wrong

[Helmholtz]

ok I will switch to non polarized 100n but why is the polarity wrong

Because PI plates are positive and power tube grids are negative.
In case of doubt use your DCV meter to determine polarity between the points where the cap is to be connected.

[Manowar1985]

Can you tell me a method to check a cap for leakage just to know how to test caps.Thank you

[Helmholtz]

Can you tell me a method to check a cap for leakage just to know how to test caps.Thank you

With coupling caps, connect one end to the high voltage (plate) side. Measure the voltage between the open end of the cap and ground with your DCV meter.
Should read below 0.2V.
This assumes a meter resistance of 1M or higher.

[J M Fahey]

Curious about why you use 8k2 as E88CC plate loads and 100 ohm as cathode bias.

[Mick Bailey]

Electrolytic coupling caps are something you don't come across with tube amps and apart from the 1uf caps I'm not sure what your reasons are for using 330uf and 290uf caps - what's the thinking there?

Edit- just taken another look and realized these are nF.

[Manowar1985]

Mick Bailey it's 220nf and 330nf.

JM Fahey It's 8.2k because at that point Rl is almost 2ra. We have anode current 9.45 mA each and Vgk=-2V at that point. The cathode resistor is Rk=2/(2*9.45)=2/18.9=105.8 Ω


Helmholtz you are a saver.I used 150 nF at 400volt non polarized for coupling and it worked fine so the leakage of the caps was the problem


Anyway thank you all for your help. I will proceed to adding some more stages to be able to apply global feedback .I will add 2 μ followers and 1 cathode follower with active load.What are your suggestions.

Thanks again.

[Manowar1985]

Well folks there has been a significant improvent to my design but when I add more preamp stages(μ follower or/and cathode follower with active load) before the phase splitter there is weird stuff going on. In my oscilloscope I see very distorted waveform sometimes turning into a triangular wave especially in higher frequencies and there is so much noise and oscillation.I dont know why.Without the preamp the things are not as expected(it distorts at smaller waveforms than it should) but at least there is no oscillation or ringing or significant noise.Can you tell any opinions.Thanks

[pdf64]

Schematic?
Is the HT supply for each stage decoupled from the others? eg dropper resistor and decoupling cap.