12AU7 as a power tube red plating

[kent]

Hello gang.

It has been literally years since my last post but I am in a pickle on my first new amp in just as long. I have done a design based on a 12AU7 for a 1 watt output using this setup:
http://www.mojotone.com/site/specs/M...plications.pdf

With this transformer impedance:
http://www.mojotone.com/transformers/MOJO-1580-OT.pdf

After doing initial wiring and getting my preamp section happy I have remade the circuit board to clean it up. Preamp section works fine but when I first turned it on I noticed the power tube red plating in a few seconds. It is possible I had this issue before and did not notice it but that seems unlikely because I played through the amp during testing on the first board for long periods with no issues. I have since checked the wiring multiple times and can find no deviation from the schematic above. My plate voltage is 312V - just 3 volts below the shown value so within a doable tolerance. When measuring the voltage on the grid and cathode I got a voltage that just kept creeping up. I assume the grid should be at 0 with no signal but it was at a couple of volts or more and rising. The Cathode was at 13 volts or more and rising. I wasn't going to keep it on at that point because the sucker was glowing red hot. This is with or without the preceding phase splitter tube in the socket. I have also swapped power tubes with the same result.

My preamp skills for troubleshooting far outweigh my power section skills. Does anyone have a suggestion on where to start breaking things down and looking? Or Maybe a reference I can look at on the design end that's not a typical pentode setup?

[eschertron]

From your description, I'd say that the grid leak is not keeping the grid referenced to ground.
13vdc bias aught to be OK for the 12AU7 (I have a design that uses one, and I bias about 10..11vdc SE with around 280vdc B+). It's the 'creeping up' voltage on the cathode that makes me think thermal runaway, i.e. the grid isn't doing its job.

[kent]

So I will check again my grid leak connections to ground.

As a related question. I would assume since this is not a single ended setup that I would calculate bias voltage differently. By that I mean the cathode resistor is shared between the tube halves, the plate resistance would be 11500 ohms as half the transformer primary etc. Am I off or over thinking this? Plus it wouldn't be biased "in the middle" to get the biggest gain before clipping, correct?

[eschertron]

All right, I'm looking at the docs you posted, this is interesting.
First, it seems as though mojotone wants you to put double the load impedance on the selected tap. Have you done this? If not, that could cause overdissipating the tubes.
Second, you could go with a bigger cathode resistor, but at some point the cold bias will affect the sound. Entirely your call. If the amp is overheating by design, then you'd need to 'fix' it. But I'm not sure that is what is happening, yet.

My SE amp has both triodes in parallel, so the cathode resistance is comparable, feeding both amplifying tubes. As far as OT primary impedance and plate load lines, each plate does 'see' half the plate-plate resistance until the opposite tube goes into cutoff, then sees 1/4 the resistance... I tried to hand draw a curve once, but lost my nerve. There are web page simulators that make short work of that part of the design.

Using a standard 12AU7 SE load calculation, the 11k5 line looks a lot like the push-pull graph, but the dissipation is a little higher, and of course the output power goes way down
Here's what I'm using Interactive Valve Data Sheets to get my 'first pass' design results.

[kent]

I thought I would close out the thread with a resolution.

Eschertron, you were right in having me check the grid leak resistors again. I built my amp with turrets on a PCB Aiken style. Turns out that one of the pads for the turrets had some kind of coating that I did not get off when I cleaned the board so the solder did not adhere the turret to the pad. After reseating and soldering the turret I tested it with the "backup" 12au7 and it was red plating still. However I noticed on closer inspection that it was the opposite side of the tube this time compared with the original tube. Popped in the first one and voila - it worked now! Turns out the "backup" was a bad tube which gave me a double whammy problem with the same symptoms. Made it much harder to troubleshoot. I'm up and rocking now - thanks for the help (and the link)!