I would think so, yes. However most bench power supplies can dial in to .1V so it should be fairly close. I have also noticed that when the prose use similar calibration bulbs that they power them up for a set amount of time before taking the measurement. I would think it the case of this bulb a bench mark of somewhere between 1 and 3 minutes would do fine.
The bulbs have a built in constant current driver. They do take a while to temperature-stabilize… like 30 minutes or so. The light output drops quite a bit during the warm-up period.
Guys, I know how a true calibration light is done, but what we try to achieve here is to solve calibration differences between BLF-members that are in the order of tens of percents. An adapted flashlight that is constant within close to 1% is way way way sufficient to straighten that out, and the flashlight that I have send around has the added benefit of convenience, and the option to compare throw measurements (luxmeter calibration) and integration quality of lumen contraptions as well.
Please carry on with the theoretical discussion on how it really realy should be done, but the way we are doing it now also works for what we want. It is nice btw that the MR16 bulb is added to the package, thanks tp! :-)
+ 1 On the thanks to TP for his input and generous donation of the light for the good of the cause for us poor plebs.