Looking at datasheets like the SFH4641 from Osram, I noticed the pulse handling capacity diagram, and although I hadn't really thought about it before, I don't know how they measure that. Niavely, they could hook up LEDs under a bunch of conditions: different pulse length, duty cycle, and current, and look at degradation over say 500 hours. But to do that Osram would have to have 10 leds per condition for nearly 1k conditions, with maybe another 3x for uncertainty. But that seems like either a huge rig to measure all those LEDs, or a long time, all for a diagram. Does anyone know how they actually do this? Are there short cuts I am not aware of?
3\$\begingroup\$ You're maybe thinking that this would be measured by hand. In industry that's too unreliable and takes too much time. So they automate it. You'd need a programmable current source and some setup to measure the light output. When that's all hooked up to a PC a test engineer writes some programs to do all sorts of tests. Also a type of robot can be used to change the LED after measurements are done. All this is "standard practice" for almost any mass produced electronic component. \$\endgroup\$– BimpelrekkieSep 3, 2020 at 19:06
3\$\begingroup\$ For "lifetime" tests that would take too much time the measurements can be done in extreme conditions (high temperature and/or humidity) to determine how long the device would last. Also there are formulas that can be used to make an educated guess of the "real life" lifetime from the data resulting from the accelerated tests. \$\endgroup\$– BimpelrekkieSep 3, 2020 at 19:10
1\$\begingroup\$ You should make your two comments an answer. \$\endgroup\$– MadHatterSep 4, 2020 at 0:09