OT Shunt biasing question...

[EFK]

Here's an issue I have not seen addressed previously and it had me scratching my head yesterday.

The way I like to bias via this method is to take the meter internal impedance right out of the equation, so (1) I measure unloaded resistance and then (2) measure DC current, and do the math.

Now here's my question: the OT internal resistance changes, like a pickup, depending upon whether it is warm or dead cold. So which number does one use? My JTM45 OT unloaded resistance measures 20-30 ohms higher after it's been warmed up for about an hour, and generally you want to bias the tubes after they've warmed up and settled down and while the amp is hot and running; so, does it make more sense to use the resistance value obtained from a warm transformer?

[R.G.]

Here's an issue I have not seen addressed previously and it had me scratching my head yesterday.

The way I like to bias via this method is to take the meter internal impedance right out of the equation, so (1) I measure unloaded resistance and then (2) measure DC current, and do the math.

Now here's my question: the OT internal resistance changes, like a pickup, depending upon whether it is warm or dead cold. So which number does one use? My JTM45 OT unloaded resistance measures 20-30 ohms higher after it's been warmed up for about an hour, and generally you want to bias the tubes after they've warmed up and settled down and while the amp is hot and running; so, does it make more sense to use the resistance value obtained from a warm transformer?

The thermal coefficient of resistance of copper is 0.0039 per degree C. For a 10C rise, its resistance increases by a factor of 1.039, 3.9% higher. For a 100C rise, it's 39% higher than its 20C value. The problem then is - how hot is the copper wire inside when the transformer is "warm". The resistance of the lead wires is, in fact, how transformer internal temperature rise is measured.

The simplistic answer is that you have to measure resistance at the temperature you're going to be operating to get a modestly correct answer. The detailed answer is more complicated.

First, if the resistance changes 20-30 ohms, how much of the base (cold) reading is that? 1%? 10%?

Then, how accurate is your meter in (a) reading resistance and (b) internal resistance measuring current? A 1% meter is pretty good, and that "1%" may be 1% real accuracy (which is what you hope for) or 1% repeatability. And then there is the issue of how cross-accurate your meter is; that is, how are the accuracy numbers measuring resistance, then measuring either voltage or current. The errors may cancel or accumulate.

On top of all that, how accurate do you really need a bias adjustment to be, and then all the secondary and tertiary issues with bias adjustment, like how safe is it to do the measurement?

My personal take on the shunt method is that it's fast, only modestly accurate based on the issues you bring up, and others, and slightly dangerous to do. I vastly prefer the measuring resistor in the cathode method because of these issues. YMMV.