What's the term for this?

Probably, but try it with modern high efficiency leds first. You will find that they (and likewise laser diodes)have very little margin for extra current. They are very different from the old style of led where you could push them like that.

There seem to be enough parts out there that are specifically designed for pulse operation. They don't have a fancy term for it.

RL

You can't. LEDs can be abused in this manner, it certainly shortens their life, but you can go to large extremes of pulsed power and get away with it.

The optical power being reflected back and forth in a laser cavity is HUGE, and most diode lasers, operated at their rated current, are close to the top of their ability to withstand this power. If you exceed the rated power by even a fairly small margin, even for a few us, the laser will mechanically self-destruct. Some lasers probablt can handle 15% more power, some really good ones might even do more. But, I suspect that if you go to 50% extra power, you will find that 99% of your lasers will fail instantly.

Jon

Right, so it's basically a non-starter. Don't exceed the maximum continuous current rating even if you're only pulsing it - and for longevity, buy the best diode you can afford from an established manufacturer. Glad we got that sorted out!

There are parts spec'ed for the app.

5min Googling got vl905p75 as an example.

RL

Well, it's not clear what you mean ("Pulsing" is what you're doing, after all.)

Pulsed-mode rating?

Whether it works at all depends on what gets damaged when you drive more current through it. With resistors it's all about what gets hot, how quickly, and how quickly it can bleed out to the outside world. It's mostly the same for power transistors and IGBTs. Devices are often marketed with pulsed-power ratings, and the data sheets have little time/ power charts to help you figure out what size parts you really need.

HOWEVER, my understanding with LEDs and lasers is that one of the significant damage modes is aluminum migration -- the Very Nontechnical explanation I was given is that all the electrons whacking into aluminum atoms makes them move to spots in the crystal where they don't belong, and your LED stops being an LED (or your laser stops being a laser). If that failure mode is, indeed, the limiting factor, then you're talking about time constants faster than any reasonable pulses you can generate, and there's probably no pulsed-rating joy to be had.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

The laser diode I use (spec sheet tacked to my wall) lists Absolute max ratings...

CW power = 80 mW Pulse power = 85 mW. (1 us pulse/ 50 % duty cycle.) That's the light output power.

George H.

40 amps?? It?s a standard 5mm package. I can?t imagine those leads sufficient for 40A pulses, let alone the bond wires...

Buy this book, it will tell you the answer

--

John Devereux

Inducing this kind of controlled current for less than 200nSec will also be interesting. With a forward voltage in the +30V range, you wonder what might be more easily controlled...

...and with an average power level of

Why not? Energy is the integral of power. It takes time for things to heat, even if it is just milliseconds.

--

Rick

So for "my" low power LD (sanyo 7140, 785nm 80mW) I think I measured an output coupling of ~20%, (power in retro reflected beam off a second cavity grating..(ECDL for those playing along at home.) I had to estimate grating efficiency so the numbers are a bit fuzzy.) Which means (at maximum power) about 5 X 80, 400mW of optical power bouncing around inside. What's the major failure? Heating at the surface?

During ESD testing I bashed these with sparks for a piezo ignitor, nice big pulses of optical power and then a higher threshold current. (from a "failed" design. :^)

George H.

You need to include at least a fragment of the quoted test, on no one knows who you are talking to.

GH

Dielectric breakdown at the surface, where there's much less restoring forc e on the atoms.

Also probably the axial component of the E field (which is far from zero d ue to the highish NA at the surface). Maxwell's equations lead to continuit y of tangential E and perpendicular D, as you know, so there's a giant disc ontinuity in the perpendicular E field at the facet. (D = epsilon E, and epsilon goes from like 11 down to 1.0 at the surface.)

Experimentally, you can blow the facet right off the diode, leading to a ve ry sad looking beam with a lot of fringes and not a lot of output power.

Cheers

Phil Hobbs

Good one!

It can be microseconds in this context.

--
Bill Sloman, Sydney

Are you using a threaded reader? Oh... no, I see it?s through Google Groups.

Get a stand-alone app and follow the threads. It?s then a much more cohesive discussion.

Not really. Google Groups does a better job of presenting content. A purely threaded presentation sort of works, but I prefer Google Groups to what I get out of eternal September. And the Google search facilities are a whole lot more powerful, if you can find them.

--
Bill Sloman, Sydney

It's essentially just a super-condensed summary of the all the advice in this thread.