Last night I put it through its paces with 2.50 volts AC on the filament, and DC with series resistor on the anode. Warmed it up in free air at 100 mA. The following pictures were taken with currents and voltage drops of:
1 mA, 12.5 volts (zoomed in to show the filament glow. No photo of permanent magnet play.)
10 mA, 10.49 volts
100 mA, 10.37 volts
1 ampere, 10.43 volts (incl. the I*R drop in filament, interconnection, and transformer)
For the same PIV rating, 10 kV, those forward voltage values would be not too shabby in a silicon diode.
Most makers of 866-A specified condensed mercury temperature range of 25°C to 60°C (pressure 2 microns to 26 microns).
Some have extension to 70 or even 80°C, with a doubling of maximum current and a greatly reduced PIV.
Datasheets for plain 866 from Hytronic http://www.r-type.org/pdfs/866.pdf and Eitel-McCullough (later Eimac) go down to 10°C. 866-A's have a metal cylinder blocking our view of the filament, allowing greater emissive area at same temperature and power.
At 100 mA, vigorous fanning of the tube envelope eventualy brought the plasma voltage up to 10.9 V.
WIth a paper envelope loosely wrapped around the tube for thermal insulation, I let the voltage get as low as 10.15 V.
The partial pressure of Hg in fluorescent and "neon" lamps must be similar. The lamps contain on the order of 10 torr of noble gas. Anybody know if the rectifiers also use a noble gas? I suppose their glow spectra could give us a clue.