I'm getting 34v bias from the curves, at 100v P-K and 65mA. But an rp of 300 ohms. It's hard to read the Tung-Sol curves though.
You can run the 12AU7 with 1.2v bias or so, and at 2mA plate current it will have 50v at the plate with +/-25v peak output. Not enough to drive the above, but plenty for a 70v/45mA operating point.
You can easily have direct coupling, or a high-power operating point, or a C4S plate load; you can probably get any two of those with the Crack power supply, but not all three!
I was looking at the WE 421A tube. It thought they were equivalents. It's clear now that they aren't. I'm looking now at the 6AS7G Tung-Sol datasheet to make sure we're on the same page.
And you're right, of course, 100v p-k, 34v bias and 300 ohm rp. Good for 1.8 watts and 1.35% THD. Not bad! Not direct coupled but I could use a C4S load on the plate (yes, Tom, the plate of the 6AS7). At 65mA I'd better use big heat sinks. Would the heat sinks of the Speedball upgrade be enough? Maybe not since the speedball upgrade is for 29mA.
And at 70v/45mA I get 25v bias. With a grid voltage at 50 the cathode would be at 25v and the plate at 95v. I get about 800mW of output, 7% THD but rp goes up to about 700 ohm.
So I guess I'll cap coupled at 100v 65mA and use the original 12Au7 parameters and a big heatsink on the 6AS7G.