I get the same negative voltage at all four tubes. I tried a new run and I notis that when I reached appr. 50ma then it all flies to the moon. The current go skyrocketing real fast. Up til 45 - 50 the current rises nice and slow.
A new run. When the bias goes bananas the negative voltage at appr. -62 -63v stays stable. I`m using a less expensive quad set of tubes for this testing.
Check the 100K resistors
Check the -62v bias bias voltage
Check the value of cathode resistor
Check the plate voltage (check for AC component)
If all good, its either a tube or the OPT or the amp is oscillating (check with a scope)
Check the -62v bias bias voltage
Check the value of cathode resistor
Check the plate voltage (check for AC component)
If all good, its either a tube or the OPT or the amp is oscillating (check with a scope)
I mean, are you absolutely sure the bias current is rising? Or does the meter think it is?
Measure the value of Rk (with everything off), then with everything on measure V(Rk), the voltage across the Rk.
I = V(Rk) / Rk.
Probably enough to do one tube to verify that the current really goes up or not.
Jan
I agree with above, don't trust the meter, use your multi-meter.
Make sure all your grounds are good
Make sure all your grounds are good
But dear people. I use one multimeter to check the voltage and the other to check the current. On both tubes on one channel. I`m NOT using the meter in front.
Think I`ll tear the amp apart and stuff another board. With even higher ratings on everything. I really enjoy this Hashimoto design. I have a MF transistor set I can use in the meantime.
I had a problem like you describe with bias being ok and then taking off. It was an oscillation in my case. Is the input shorted when you're doing these tests? Have you tried disconnecting the gNFB loop and seeing if it still happens? Just ideas...
There is a gain control on the input. Set to zero when testing. Not tried to disconnect the NFB.
That's nonsense. Why would the rating of everything be an issue? You have no idea about the problem but want to built another one?
I think I'm out of here.
Jan
Just as the frustration rises, yes this kind of nonsense comes into my mind. I certainly would like to find out what's wrong but then you feel lost in the desert when there is no logic to the problem.
Disconnected the NFB and problem solved. When I installed new tubes 6 months ago I inserted a 680pf + 10K in parrallell to the 47K in the feedback to get a little more sparkel to the treble. Remove this and now its working as it should. Shocking.
You answer: "A normal stereo amp.". That doesn't clarify things. Does each channel has its own sepearte power supply, and its own seperate bias supply, or not?
You answer: "Uses two multimeters, one per channel". But in post #1 you wrote: "Tried to readjust the bias with just the tubes in one channel and got it down to 50ma but now the voltage across the 10ohm cathode resistor measures only 20mv.". Why would you use one of your two multimeters on the channel that didn't have tubes in? It also seems to indicate that when adjusting the bias of one channel, you were only monitoring one of the two KT88's in that channel.
In post #29 you write: "But dear people. I use one multimeter to check the voltage and the other to check the current. On both tubes on one channel.".
Now your two multimeters are used in only one of the two channels? And again it seems to indicate that when adjusting the bias of one channel, you were only monitoring one of the two KT88's in that channel.
Strange case... I would normally suspect the (circuit around the) meter. But who knows, maybe this is more a case of "pulling someone's leg".
I'm out of this thread.
Glad you found it. What you inadvertently did was to increase the gNFB and that caused a phase shift in the HF region which showed like you saw which made your power amp into a power oscillator.
The small pF cal in || to the 47k is not a bad idea. What I would do is remove the 47k, try 33k for stability (same ratio I use). If it's stable, then feed 10kHz square waves into it and pick a shunt cap to make the square waves not ring. Probably around 300pF will do but it's really up to the OPTs and how good they are. I even have a "pair" of Hammond 1650N which one needs twice the compensation cap as the other. Effectively the cap is almost always AOT (adjust on test).
In my case, it's usually 2k||1000pF tying to cathode between 330R and 68R, 68R on the bottom, cap across the 330R.
Will certainly try that. Now the amp is somewhat "dry". And again, it feels like someone in here thought I was pulling someone's leg. I`m from Norway and isn't good at expressing the right description of what the problem was and how I did the measurement.