Are diodes killed by mW or mA?

[Eudaimonium]

Good discussion you guys got there,
Well one correction, it's not LPC815 that was driven that high it was the new LPC822 which has proven to be exeptionally hard to kill, tuning it down from overcurrent simply gets you the laser back...
PullBangDead, care to explain those new diodes that appear to be Chuck Norris of diodes?

Also, here is a link to wiki:
Laser diode - Wikipedia, the free encyclopedia
That might help some new guys understand the diode better.

 

[seoguy]

Well one correction, it's not LPC815 that was driven that high...

No correction needed, that's not what I said. Please re-read my post.

 

[Eudaimonium]

No correction needed, that's not what I said. Please re-read my post.

Many apologies, somehow I flew past that comma, so I misunderstood.
Sorryz :yabbem:

 

[pullbangdead]

Are you talking about this curve?

I completely missed that whole discussion of a new diode.

In short, that is 100% perfectly normal and expected for any and every laser diode. The rollover is 100% caused by heat, and that rollover doesn't happen when pulsing the diode (as pulsing has the effect of removing heating from the equation, not a perfect assumption but good enough). It's not seen in other reds because they die before they get to the point where it rolls over. It's not seen in violet or blue diodes because the material they're made out of (GaN-based materials) is *much* less sensitive to temperature, and they die WAY before they reach the heat/power dissipation required to see the full rollover. The rollover is not intentional, it's not a design feature, and would be completely avoided if at all possible, but that's just what happens when you're running CW and heating a laser (by running it CW, just to clarify) and it hasn't died from other factors yet.

Current, heat, and light all kill, but you can tell which one is doing it by how they behave. COD happens suddenly and all at once when you reach a certain power level, and will happen whether you're running CW or pulsed. Current kills whether you're running CW or pulsed, but generally happens as a slower degradation (and you can test more to confirm it's current-caused degradation). Heat causes this roll-over in performance, and doesn't kill when pulsing, so you can differentiate between heat-caused deaths and other deaths by pulsing the diode. It's more complicated than that, naturally, but that's the idea of how to differentiate between causes of death. Light can cause slow degradation I suppose, but you can still do work to isolate what is causing your diodes to degrade, even if it's just a slow degradation.

----------------------------------------

ETA: Alright I read the whole other thread now, and what I wrote above still stands. The one "pulsed" test on the other thread was 50% duty at 1kHz...that's not CW, but it's not pure pulsed either for the purposes of laser diodes. 50% duty at 1kHz is still enough to produce significant heating. Notice he said the rollover does happen at a higher current than at CW, which makes sense because you are getting LESS heating at that duty cycle, but you are still getting heating. Pulse fast enough to remove the heating (which is basically the working definition of "pulsed" in laser terms, fast enough to remove all effects of heating), and you'll get rid of the rollover completely.

 

[laser_freak]

Thanks guys for all the useful information!

:thanks:

 

[Hemlock_Mike]

PBD ---

Your points are likely correct but with heating causing this "rollover", why does the power follow backwards immediately as I reduce the current? Can the die cool that quickly?

HMike

 

[pullbangdead]

PBD ---

Your points are likely correct but with heating causing this "rollover", why does the power follow backwards immediately as I reduce the current? Can the die cool that quickly?

HMike

Yes. The junction temperature reaches a quasi-steady-state very quickly. By the time you look at it at any given temperature it has reached a quasi-steady-state, and as long as the diode wasn't degraded at the higher temperature, it'll return to the same output characteristics as it had before as soon as it reaches the same junction temperature again.

If you had the equipment to control junction temperature while running the diode pulsed (as in truly pulsed, where heating due to operation is zero and junction temperature is affected only by the controlled temperature that you set), you could plot the typical perfect, straight lines straight to every point on your curve as you changed the temperature. It would look something like the attached image. As you increased the temperature and retook your measurements at each higher temperature, the threshold moves out, and the slope will likely decrease as well. The equations are complicated, but that's the general idea. Each temperature would have its own perfect L-I curve in such an experiment.

This is a rudimentary, badly drawn estimate by me, but it gets the point across. At each current level, a quasi-equilibrium is reached very quickly, and the temperature stays the same while at that current.

 

[LASER66]

Are you talking about this curve?

I completely missed that whole discussion of a new diode.

In short, that is 100% perfectly normal and expected for any and every laser diode. The rollover is 100% caused by heat, and that rollover doesn't happen when pulsing the diode (as pulsing has the effect of removing heating from the equation, not a perfect assumption but good enough). It's not seen in other reds because they die before they get to the point where it rolls over. It's not seen in violet or blue diodes because the material they're made out of (GaN-based materials) is *much* less sensitive to temperature, and they die WAY before they reach the heat/power dissipation required to see the full rollover. The rollover is not intentional, it's not a design feature, and would be completely avoided if at all possible, but that's just what happens when you're running CW and heating a laser (by running it CW, just to clarify) and it hasn't died from other factors yet.

Current, heat, and light all kill, but you can tell which one is doing it by how they behave. COD happens suddenly and all at once when you reach a certain power level, and will happen whether you're running CW or pulsed. Current kills whether you're running CW or pulsed, but generally happens as a slower degradation (and you can test more to confirm it's current-caused degradation). Heat causes this roll-over in performance, and doesn't kill when pulsing, so you can differentiate between heat-caused deaths and other deaths by pulsing the diode. It's more complicated than that, naturally, but that's the idea of how to differentiate between causes of death. Light can cause slow degradation I suppose, but you can still do work to isolate what is causing your diodes to degrade, even if it's just a slow degradation.

----------------------------------------

ETA: Alright I read the whole other thread now, and what I wrote above still stands. The one "pulsed" test on the other thread was 50% duty at 1kHz...that's not CW, but it's not pure pulsed either for the purposes of laser diodes. 50% duty at 1kHz is still enough to produce significant heating. Notice he said the rollover does happen at a higher current than at CW, which makes sense because you are getting LESS heating at that duty cycle, but you are still getting heating. Pulse fast enough to remove the heating (which is basically the working definition of "pulsed" in laser terms, fast enough to remove all effects of heating), and you'll get rid of the rollover completely.

No, the rollover is not designed- in, just an artifact we get to observe, as stated. As I posted on Mike's Thread was the fact that what IS designed- in is what is one point discussed here: Light kills. And what diode manufacturer's have done, is reduce optical absorption at the facet, and surrounding material, to reduce heat at the emitter. What this ends up doing is in effect, allowing us the get to the point where we can observe to rollover, like discusses.
Before, we would have seen COD failures before getting to that point.

-Will

 

[pullbangdead]

No, the rollover is not designed- in, just an artifact we get to observe, as stated. As I posted on Mike's Thread was the fact that what IS designed- in is what is one point discussed here: Light kills. And what diode manufacturer's have done, is reduce optical absorption at the facet, and surrounding material, to reduce heat at the emitter. What this ends up doing is in effect, allowing us the get to the point where we can observe to rollover, like discusses.
Before, we would have seen COD failures before getting to that point.

-Will

Glad we're all on the same page. :beer: