I've been out of town for the Fourth (yes, the cabin is still standing!) so I'll try to answer several questions.
Running a 2A3 in a Stereomour set up for the 45 will run the tubes at very close to the same operating point - 35mA and 275v plate to cathode. The lower plate resistance of the 2A3 would normally make me want a slightly higher load impedance; the 4K standard is 0.8 times my target. But my choice of load impedance is pretty conservative; many commercially-available amps use an even lower impedance. In fact, the Paramour II upgrade runs at 0.75 of my target. In both cases, the large plate choke inductance minimizes tube distortion at the lowest frequencies, compensating for the slight distortion increase of the low impedance. Then lower impedance will also raise the efficiency. I have not done a full set of measurements yet, so I will only guess at 2 to 2.5 watts for the 2A3 in 45 mode. At a guess, the power and distortion differences will be no more and possibly less audible than the intrinsic differences between the tubes.
Calculated uptimum values for the parafeed cap are 2.5uF for 2A3, 5uF for 45 or 2A3 in 45 mode. The range from half to twice the optimum is considered suitable, which is why the stock 3.3uF is used for either tube. A further consideration is that the smaller capacitor protects the tube from distorting due to low load impedance at infrasonic frequencies, which is of course more likely to be triggered with a lower-powered amp - hence a slightly small parafeed cap with the 45 is part of the plan. Experimentation with these values is quite reasonable, because of the interactions possible with real-world speakers.
(Parenthetically, I have posted a different but similarly exotic reason for the 3.3uF in Paramour I and II where that value is larger than the theoretical optimum. The cynical will have noted that all these ideas lead in the end to the same cap value of 3.3uF :^) I maintain nevertheless that this is just a fortunate coincidence. Really and honestly, I do.)
I still do not know what to say about the EML tubes. They do not like certain kinds of startup transients - EML is quite straightforward about this on their web site, and they have been helpful in identifying and diagnosing problems. In most cases, a cap-coupled circuit such as Stereomour is considered safe with those tubes, as long as the time constant is short enough. Many people with many different cap-coupled amps are very happy with these tubes. Our time constant is 25mSec. I would say that is "normal" while a few more exotic amps have longer time constants, up to 5 times as long. I do not have enough detailed information to do a design that is "known to be safe" - I kind of suspect nobody does. That critical time constant probably depends on several other variables, which is probably why nobody will specify a specific universal safe value. I am not going to purchase ten pairs of EML tubes in order to do destructive testing to discover those limits! So, while I think it very unlikely there will be any problems, I won't make any promises and Bottlehead won't cover your expenses.
The PC board in Stereomour will be the same, and can be configured for, a startup time delay - but the parts are not included and I still have some cautions about the EML tubes with those boards. Even with the startup delay, the amp needs to stay off for a long time once turned off - I am saying 30 minutes at this point - in order for the slow start to re-arm itself.
Hope all that is helpful.
