We got a new signal generator in at work, so I thought I'd subject the Slow variant I picked up in a previous order to some tests. Check out the results.

Of the scenarios I mentioned, 1 and 3 applied - and a 5th reared its head: errant program behavior. Did not expect that one :)

SFE - or somebody else who has one - would have to test it. It's not generally included in a datasheet because basically.. you're not supposed to PWM them :)

There's several things that could happen, depending on how fast you PWM it / duty cycle.

1. It runs as usual (high enough frequency + high enough duty cycle)
2. It doesn't run at all (too low frequency + too low duty cycle)
3. It burns dimly at its initial color (frequency*duty cycle high enough to turn it on, not high enough to keep it 'in program')
4. It runs as usual, but dimly (fringe case - never seen it.. would have to discharge enough for the LEDs to dim/turn off, but not far enough that it goes out of program).

Typically, you're going to hit #3. If you hit #1 there's no point in PWM'ing it to begin with. #2 is obviously useless. #4 might be the desired effect.

Would be pretty neat to test, though - and maybe make for a nice primer on flashing / auto colorcycling LEDs by SFE? Who do we bribe to make it happen? :)