pulse drivers

Indeed... if you now drive a diode at 200 mA CW and want 20% duty cycle, you'd have to run a full amp through it. Regardless of frequency, i'm affraid it will not survive that, since some failure mechanisms can be rather quick - catastrophic optical failure for example.

With LED's the limitation is mostly thermal, as long as you manage to keep the die cool, they will tolerate great peak currents. Also, with LED's forward voltage and wavelength depend on current much more than with laser diodes. You can see this in green power leds: If you run them at high current but with low duty cycle, the color goes somewhat into the cyan... even if the brightness isn't (much) better compared to running them CW on the average power/current of the pulsed configuration.

Laser diodes need reflective end faces on the die to lase, and those are failure points leds lack... so its the light itself that kills the laser diode, whereas a led will only die from running to hot internally.

Benm said:

You can see this in green power leds: If you run them at high current but with low duty cycle, the color goes somewhat into the cyan... even if the brightness isn't (much) better compared to running them CW on the average power/current of the pulsed configuration.

Click to expand...

Agree with everything you said.

Just adding that this effect, the large blue-shift, is a cool materials science problem, and in research labs can be largely averted. It goes back to the polar nature of GaN, and non-polar and semi-polar planes can help this effect a lot. In the future, it's likely that blue and green LEDs grown on non-polar or semi-polar planes can greatly lessen this effect, but the substrates necessary aren't currently anywhere near economical (making this change currently would mean going from a 2-inch wafer to <10mm wafers, not exactly good for your economies of scale).

Is it that big a problem in practical application with leds? I suppose in most applications you could just PWM them to have brightness control without any color shifting.

Perhaps it becomes problematic when you need analog intensity control, for video projectors, megatron displays and such. I always figured that would be solved in software for systems that complex.

In a way is unfortunate that this effect doesnt work for laser diodes - it would be very nice to be able to shift a red laser up or bluray down in frequency by tinkering with voltage provided you only keep dissipation within acceptable limits.

Invisible lasers in an accessible price range. Now on the high energy side of the spectrum, wouldn't that be just great...

:undecided:

Depends on what you consider invisible really. If the limit is 400 nm exactly, some bluray diodes may already fall into that category when run at cold temperatures (perhaps using TEC). In reality things near the border of the 400-800 nm range are often visible to some degree, so its hard to set a limit.

I'd consider bluray hardly visible really. With enough power you can still see the beams, but if you compare it to 532nm at equal power is almost nothing.

Benm, I like your idea of a comparison test, possibly multiple tests at different pulse frequencies, is there anyone with the technical know how, parts and time to pull off such a test?
-nick

I suppose i could do it, in the sense that i can build CW and pulse drivers, and measure laser power thermally (not very accurately, but good enough to tune two lasers to equal average power).

The problem is however in the 'proof'. The difference in brightness is in the eye of the beholder here, so it might make a biggere difference for you than it would for me.

The best way to test is is probably at a meeting or something like that, where you have a bunch of people. Just set up the beams next to eachother and make people point out which, if any, appears brighter. Obviously they should be unaware of which one is pulsed, though i suspect that might be difficult at lower frequencies.

well if it was worth doing on a large scale the difference in brightness would have to be apparent to everybody not just me.