Post by: klugesmith on November 18, 2022, 10:05:29 AM
I have some replacement lamps, with special V-shaped filaments. A sad reminder of going-out-of-business sale at our local everything-photographic store. Note strange shape of glass envelope.
.
.
The other day I fired one up and projected its arrowhead shape using a small lens.
As with any optical projection, the brightness on screen depends on the source luminance, the magnification, and the lens speed (f/ ratio). Luminance is the apparent brightness of a distributed source, for example what photographers measure with a spot meter. A traditional unit in USA is the footlambert; SI unit is the candela per square meter, also called the nit. Brightness of TV, computer, and phone screens can be hundreds of nits.
For this bulb, as with tungsten-lamped slide projectors and movie projectors familiar to we boomers, the limit is on order of 10 meganits according to literature. (Solar disk is in the low giganits. Slightly higher viewed from surface of Mars. If a cheap laser diode has uncollimated brightness of 1 Cd from emitting area of 3 square microns, that's 3E+11 nits! )
Here's a SWAG of luminance for the V-shaped filament in picture. Say 5 mm length of 0.05 mm straight tungsten wire, for luminous area of 1/4 square mm. If the luminance were 10 meganits, the uncollimated lamp brightness would be 2.5 Cd on axis. Feels about right. [edit] I guess that when properly measured, the wire thickness and lamp candlepower will both be higher.
Post by: klugesmith on November 19, 2022, 03:56:32 AM
The basic unit, candela, was originally defined as the directional brightness of 1/60 cm^2 of blackbody at the melting temperature of platinum. That corresponds to blackbody luminance of 600 kcd/m^2 at around 1770 °C. For comparison, here are some familiar sights according to wikipedia:
That standard could be realized with much better reproducibility than standard flames or incandescent ribbons.
Still very challenging to get absolute accuracy better than 1%, since the light varies by more than 4th power of temperature.
Interesting report here: https://technology.matthey.com/article/30/2/84-95/
Discussing work at US National Bureau of Standards in 1930: "Some 185 g of platinum in the crucible surrounds the radiator tube and the whole is heated by a high-frequency induction heater which can be precisely regulated. In use the container and its contents are initially heated above the melting point of platinum, they are then allowed to cool very slowly through the period of solidification of platinum while a small central hole in the lid of the crucible acts as the source of light."