How was the first integrating sphere calibrated? I understand it requires a known luminous flux source but how did they obtain that without a calibriated IS...
Also, for those of you with DIY IS, how a big difference have you noticed in measured values based on the ambient temp? So let's say, you last measured a light at ambient temp of 70F/21C, what do you get at 80F/26.7C or 60F/15.6C?
Thx!,
Tim
It was based on calculating the inner surface area of the sphere along with calculating how much light is absorbed.
the rooster ...
Without knowing the "quantity" of the light, how would they know how much was absorbed?
@Jack - LOL!! You know, all these years of hearing that question asked and that is the first time I've ever heard that answer... good one! =o)
Good point, They probably could test it with a lux meter and a piece of it.
Good question.
Maybe the sun ?
The lumen may have been originally calibrated from the candela which was standardized in the pre-electrical times by an actual candle. From Wikipedia;
One candlepower was the light produced by a pure spermaceti candle weighing one sixth of a pound and burning at a rate of 120 grains per hour.
So that can be how the first integrating sphere was calibrated. That way when they start putting other light sources and comparing they had a known physical standard.
They may have been able to use just the light itself to heat the inside of a black box, and somehow worked it out from there.
The lumen is based on the candela which is based on the Watt which is based on the Joule which is based on the Newton. In practice here I think the Joule is just defined as 1 amp over 1 ohm for 1 sec and they probably use a laser. IOW, the units cancel out and they just shoot a laser or led diode of given power into the sphere/target and see how much it registers.
Ok, so here’s the details: they’re using a detector based standard (ie shoot laser into sphere and standard is the response on a reference detector design).
The download at that page is busted but I have the doc. They use a silicon photodiode of given size (or rather, a hole of precise size over the diode), a photopic vision filter and control for temp.
Or they can do what the Chinese Lumenati do… lube up a couple of fingers and start pulling numbers out of their asses… 
Ah but the question was how would you have calibrated the newly invented photo sensors when they were invented? How would you be able to translate joules into light sensors. Lasers were invented far later than vacuum tube based photo sensors. So back then to be able to measure lumens with a vacuum tube based system you’d have to have a standard reference which I think was the spermaceti candle.
The standard they used before lasers/diodes was black body platinum/gold emitter. And the candle or whatever before that. However, you can see that the arbitrary number they used (683) is a nice around approximation chosen for continuity from the old method more than anything else.
chicken
Thx guys, lots of great (and funny) answers and very informative!