Hot Rod Deluxe Bias Ramping up ... all by itself.

[christarak]

My brain hurts and I’m hoping you can help.

I have a customer’s Hot Rod Deluxe III in for repair. I replaced the out of balance power tubes and biased them at 30mA each at a plate voltage of 435VDC, and bias voltage of -54v. Over the course of 10 mins I watched the bias current ramp up from 30mA to 50mA. The bias voltage and plate voltage both decreased in line with the increases bias current…but I didn’t touch the bias. I resoldered all the power tube contacts and measured screen grid resistors at 470 ohm each. The 2 notorious 470ohm 5W (R78 & 79) resistors were at an eye-watering 120C (250F) but still metered OK at 470ohm.
Any suggestions as to why the bias current creeps upwards to the inevitable cooking of the 6L6's.

Here are some clues:
With no 6L6's installed everything is rock-solid with Vplate (Vp) =458V and Vbias (Vb) = -55.96

With new 6L6's installed:

Minutes ImA (@V4) ImA (@V5) Vp Vb
0 29.9 28.3 434.4 -54.55
+3mins 30.8 33.6 429.4 -53.62
+6 35.4 37.5 426.4 -52.98
+9 47.3 46.3 421.6 -51.95

Any suggestions on the cause (or inspirational thoughts of relaxation) are very welcomed.

From one of your Aussie comrades
Chris
Sorry, I am not sure why the table did not format properly.

[The Dude]

Was this the original problem before you replaced the tubes? Or, maybe you "installed" the problem? I have to wonder if you have a bad tube or tubes. "New" isn't always "good" in tube land.

[christarak]

Thanks Dude. I really appreciate your quick reply.
The original problem was "amp over-heating. Burning smell". I will put in the original valves and check again. I will also try another new, matched set and try again.
Much appreciated

[The Dude]

If that doesn't do it, I'd try firing the amp up with the output tubes removed and monitor the bias supply. Maybe it's taking a dump after warmup. Also, look for ripple on the bias supply.

[Enzo]

In any case, that declining bias voltage will certainly cause a growing current.

[christarak]

HI Dude,
I put in the original valves to remind myself. Same issue. Over 10 minutes, V4 ramped from 29.7mA to 37.0mA, and V5 went from 18.1 to 23.2mA. Vplate dropped from 431.7 to 427.9. Vbias only small drop from -51.73 to -51.39.

[christarak]

Thanks Dude,

I tried the same test with both power tubes removed and both Vp and Vbias were rock solid. I haven't yet checked ripple on Vp or Vbias.

[christarak]

Thanks Enzo.
I am not sure about the chicken and egg scenario. Is the dropping Vb causing the current to increase, or is the increasing current causing the bias to drop, and the Vplate to drop.
Hmmm...on second thoughts, tube current shouldn't really effect the Vbias. I may try scoping the Vbias to see if there are any clues there.
I really appreciate your input.

[christarak]

On the scope, the ripple seems pretty small (20mV p to p) at 50Hz (Aussie mains freq). Vbias pretty stable at around -53V with no tubes installed. I'm stumped, fellas. I will try a different set of new tubes and report back.

[The Dude]

Another thought is possible leaking coupling cap or caps from PI to output tubes. Are you checking bias voltage directly on the grids or further back in the bias supply?

[christarak]

I'm checking the bias at C38 just after the bias pot. I have just found something that will hopefully give you more than the pain its given me. R76 should have roughly -53VDC on the single diode rectifier side. My R76 has -106VDC with the expected -53VDC on the other side! The poor resistor is copping a huge voltage drop and the C43 capacitor is over its 100v rating.
But, here is why sparks are flying behind my eyeballs: the rectifier diode CR15 has the same -106VDC on the anode. The cathode has the expected 39VAC but also has has -53VDC on the cathode as well (which is where I expect the -53-53=-106VDC is coming).
My brain hurts

[Enzo]

Rising tube current does not have the effect of reducing bias voltage. You must check the bias voltage right at the tube socket, as that is where leaking coupling caps would have the most effect.

What meter are you using? SOme less sophisticated meters can be confused by AC with DC on it. Your bias tap and rectifiier seem to be making a good negative suppy, so whatever SEEMS wrong with the diode and stuff is likely not.

Maybe you already said this: Have you measured the individual tube currents? Or are you taking the test point reading and dividing by two? You might have only one tube running away.

Another simple test, pick one tube, install it in one output socket and fire up. Does it start to run away? The test point should read half what normal is. Now move that one tube to the other socket, same test. By running only one tube at time, we find if one socket or both have the problem.

[christarak]

Rising tube current does not have the effect of reducing bias voltage. You must check the bias voltage right at the tube socket, as that is where leaking coupling caps would have the most effect.

What meter are you using? SOme less sophisticated meters can be confused by AC with DC on it. Your bias tap and rectifiier seem to be making a good negative suppy, so whatever SEEMS wrong with the diode and stuff is likely not.

Maybe you already said this: Have you measured the individual tube currents? Or are you taking the test point reading and dividing by two? You might have only one tube running away.

Another simple test, pick one tube, install it in one output socket and fire up. Does it start to run away? The test point should read half what normal is. Now move that one tube to the other socket, same test. By running only one tube at time, we find if one socket or both have the problem.

Thank you Enzo,

I read another of your posts where you suggested testing one valve at a time, so I had already tried that. Sadly, the current runaway was reproduced on both individual valves.

I am measuring the individual bias currents using my Fluke 87, but was using a fake-Fluke to measure the plate and bias voltages.

I hope you can think of some other things to try. What do you think about the very hot 5w R78 & 79resistors? They share the same 39VAC tap as the bias circuit. Coincidence?

[Enzo]

Hot resistors in zener supply circuits are normal.

[christarak]

Yep. The 16v zeners are both OK, so if everything is going well the 5w resistors should dissipate about 3w. I've seen worse brown spots under R78 and R79 on other amps, but I'm getting to the "clutching at straws" stage. I have spent so many hours on this HR Deluxe.