Aha... But why ground via the chassis?
To be specific here, let's distinguish between circuit '0V', 'chassis/shield', and 'safety earth'.
The stock implementation (from Alctron) is trying to balance between noise immunity and ground resistance: the stock circuit joins circuit 0V and chassis/shield (right at the 7-pin connector, and again at the microphones output connector) in order to minimize ground resistance. This is the way it's set up stock from Alctron.
The "best" implementation (conceptually) is to ensure that circuit '0V' only returns to the PSU via the 7-pin cable pin 7: the microphone's pin 1 should be bonded to the shell of the microphone body (via it's retention bolt), and chassis should be bonded to safety earth at the single point where the safety earth attaches to the chassis. The PSU's XLR's pin1 and shell(s) should be bonded directly to the chassis where they exit the PSU. Then, circuit 0V should return to the earth point so that the chassis and circuit 0V don't drift too far apart (which is better for noise).
However for a given microphone in a given environment, having the shield return circuit 0V (by tying chassis and pin7 of the PSU connector) can actually be less noisy for a host of reasons: for example, return impedance of 0V can be much less when you factor in the cable shield and body shield go in parallel with the 0V ground wire. However with DC heaters, I never noticed much difference either way.
I wired a temporary ground to the circuit to measure bias, filament and B+. It seems to fire up nicely. I just have one question. What's the spec for B+ in this circuit (referring to the v1.4 PSU). In my build B+ sits at 200 Vdc, unloaded.
What is a 'temporary ground'?
Unloaded, it could be anything. Earlier in the thread we describe using test resistors inserted directly into the output 7-pin XLR connector: use a 180K for B+, and 20 Ohms, 2W for the heater. This will allow you to trim for 120V for B+ and 6.3VDC for the heaters without using the microphone at all.