8X Diode Murder fund

[sportcoupe]

I let the diodes run all day yesterday and through the night, and they reached 160h on Murder Candidate #1 (LG 8x #3) and....

20h on Murder Candidate #2 (LG 8x #1)! :yh:


20h explains how they can be set to 360mA and then last a while...

Excellent, especially 8x #2 at 360ma. I predict it'll last much longer too because I have an estimated 20 hours on my 8x at 400ma (BDR-203). These 8x diodes seem tough.

 

[IgorT]

Excellent, especially 8x #2 at 360ma. I predict it'll last much longer too because I have an estimated 20 hours on my 8x at 400ma (BDR-203). These 8x diodes seem tough.

Murder Candidate #2 is at 348mA (i was aiming for 345mA, came out a little higher).

I meant if they survive ~350mA so far, they will also have a chance at 10mA more.


Unfortunatelly i can't deal with usage estimates, because i have noticed that many people, including myself, drastically overestimate their own use, when it comes to time.

So much in fact, that someone claimed to have used their 4x by hand for more time then i cycled mine in only twice the number of days. And in that cycling test the duty cycle was not even 50% but closer to 90 (90% of time ON, 10% of the time OFF)....


So while i believe you used your laser a lot, i can't be sure it's anywhere near 20h, especially since you might also say you haven't noticed any degradation in the form of a power drop.



Don't get me wrong, we all do it, that's why i say i have no idea how many hours my personal 8x has on it until i dissect it and take the time recorder out.



EDIT: Altho it is obviously true that some diodes will degrade slower than others. We've seen (and still are seeing) the two low efficiency diodes degrade, i am seeing a much higher efficiency diode degrade rapidly at a reasonably low current of 320mA, due to it's higher optical flux at the die. Another high efficiency diode will be tougher and possibly survive the same current for as long as the first low efficiency one, due to being "tougher".

And unfortunatelly it seems there are no external indicators of this elusive "toughness". I thought efficiency was it due to some experience with GGWs, but we concluded that some there may have been prototpe diodes...

When i take the time recorder out of my personal 8x i may very well conclude that a high efficiency diode can die faster than a low efficiency one at the same current....

 

[IgorT]

Here is the money shot...

Murder Candidate #2 at 20h (and #1 at 160h):

An attempted beamshot:

20h PI Degradation Plot (Murder Candidate #2):

Diode Stats:
Ith = 33.5mA
Slope Efficiency = 1.319mW / mA
Absolute Efficiency = 21.09%
Po @ 348mA = 416mW (91.73% of initial)


Now this is really interesting. The last 10h did very little damage. Barelly noticable at all. Degradation has almost slowed down to a halt.

On the PI plot the difference seems approximatelly as much as the one from 6 to 10h.


This diode seems to just wanna keep going.


Looks like the top 8-10% is the fastest to go, after that the diode almost seems to "stabilize" a bit. :thinking:



I'm trying to get more hours under the hood as quickly as possible so i'm leaving it till 30h ON-Time for the next testing, as for the first murder candidate, i am testing it with longer intervals, since there the degradation has already slowed down quite a bit.



In any case, the second diode took 20h to degrade to the point for which the first required 100h. It's a big difference, but 20h is not a low age for a diode at this current!

I'm quite impressed already, if the diode makes it till 50h, i'll consider it reliable.

 

[pullbangdead]

I admittedly haven't been following the details and ins&outs of this testing process, but this behavior is how lasers age. The aging rate decreases as the laser "burns in". Here's an image from an OSRAM paper published last year on nitride laser aging (these were actually 450nm diodes, so there are some differences vs. the 405nm laser discussed here). But the point remains, this is typical laser diode behavior.

 

[IgorT]

Nearing 30h ON-Time with the second diode @348mA! :yh:

The old diode has crossed 10.000 cycles by now @300mA, waiting for 180h...


For those using this info to decide on a current setting, please keep in mind that power matters too, just as much as current.

Just because one low efficiency 8x survives 30h or more at 348mA, doesn't mean all higher efficiency ones will too (nor that they won't).

Some higher efficiency 8x's could degrade themselves faster at that current due to their higher optical flux. Others again can live longer, if they are tougher.
Also, this doesn't mean that all lower efficiency diodes will survive higher currents longer...


Unfortunatelly, there is no way to tell how tough any particular diode is.


For example, LG 8x #1 started out at approx the same power as my personal 8x, altho my personal 8x is at a lower current, and they both seem to degrade with approx the same speed during the first hours, possibly because of the similar power.





Otherwise i dug up some degradation info on an older 60mW 405nm diode...


This first chart shows how the tested diodes need more and more curent to maintain the same power as time elapses:

The second chart is a lifetime prediction chart. Every blue rectangle represents a "dead" diode and they were declared dead when Iop becomes 1.3 times that of initial current (the diode needs 30% more current than at the start to reach the same power):

Both charts show the murder of 15 60mW CW diodes similar (or identical?) to PHRs, but powered at their rated power.



As you can see in the first chart, every single one of them degrades even at their rated power and this is noticable even during the early hours. In fact there is a small bend in the lines of the first chart showing the current increase, indicating that degradation there might also be fastest at the start.



This is why i'm confused, when i see so many claims of no degradation with time at the high powers we are using.



One other thing the first chart shows is, that the diodes took 1200h to reach approx 10% of degradation.
The second chart shows that it then took a total of 5200h for them to become declared "dead", or required 30% more current to maintain the same power on average (MTTF).

On average those diodes lived a total of 4.33 times as long as it took to reach the first 10% of degradation. Of course that doesn't mean anything for our diodes driven to much higher powers - they might not even reach the level of degradation where it would take 30% more current to maintain initial power.




In manufacturer's tests, the diodes are always driven at a constant power, in this case 60mW, while we drive them at a constant current.

The constant current part could further explain why degradation slows down so much in our case, while it seems almost linear in the first chart - our diodes drop in power as they degrade, the lower optical flux degrades them at a slower rate...



4 more cycles till 30h re-plotting. :yh:
EDIT: Diodes succesfully reached 30 and 170h respectivelly. Cooling off for re-plotting.




EDIT:

... this behavior is how lasers age. The aging rate decreases as the laser "burns in". Here's an image from an OSRAM paper published last year on nitride laser aging (these were actually 450nm diodes, so there are some differences vs. the 405nm laser discussed here). But the point remains, this is typical laser diode behavior.

I'm assuming those tests took place at a constant power?

If it's typical there, where the current is constantly increasing as efficiency drops in order to maintain a power, it makes much more sense here, where the power drops as the diode degrades, while the current stays the same.

 

[sportcoupe]

This is why i'm confused, when i see so many claims of no degradation with time at the high powers we are using.

Oh I am sure my 8x has degraded, I just have not bothered to check at all. My 8x has went through three increases in ma's so far when I got tired of the mw output and wanted more. That and my lack of plotting blinds me of any actual degradation. I sure enjoy 500mw of 405nm light though for sure.

 

[IgorT]

30h Murder Candidate #2 PI Plot:

Diode Stats:
- Ith = 34mA
- Slope Efficiency = 1.294mW / mA
- Absolute Efficiency = 20.58% Avg.
- Po @348mA = 408.7mW (90.11% of initial Po)



It took 30h of ON-Time for the diode to lose just under 10% of it's power. This is where the first diode was at approx 110h.

If this diode makes it to 50 or so hours, the first one may very well survive as much as 300h.



Going on 40h and 180 for the old diode. :yh:

 

[IgorT]

Oh I am sure my 8x has degraded, I just have not bothered to check at all. My 8x has went through three increases in ma's so far when I got tired of the mw output and wanted more. That and my lack of plotting blinds me of any actual degradation. I sure enjoy 500mw of 405nm light though for sure.

Changing the current is a different story, you lose track then. But i was referring to other cases. Besides it doesn't take replotting, just measuring a laser's power.

Only case of reported 8x degradation i know of was when Hemlock Mike posted about his diode dropping into a kink.


How often do you guys check your laser's power?
I'm sure you don't just measure it once it's made, but i check it quite often to see how it's doing.. Definitelly a couple of times per week together with an IR temperature measurement of the laser head (so if the power is lower than expected i know why).


EDIT: I'm gonna try to replicate my original 8x power measurement by force cooling the laser down to 17°C. If it gets close to 437-440mW it will make me very happy.

EDIT 2: Nope, i'm lucky to get 430mW when the laser feels freezing cold. I guess i really can't put it down to temperature anymore...

 

[Slipstream]

I'm also sure mine has degraded but because I've adjusted the current multiple times I don't know where the degradation and how much it happened.

 

[pullbangdead]

EDIT:



I'm assuming those tests took place at a constant power?

If it's typical there, where the current is constantly increasing as efficiency drops in order to maintain a power, it makes much more sense here, where the power drops as the diode degrades, while the current stays the same.

No, those charts I gave are power plotted at a constant current. Papers in science/industry do it both ways, power at a constant current OR current at a constant power. They both apply, I see it both ways in papers.

 

[IgorT]

No, those charts I gave are power plotted at a constant current. Papers in science/industry do it both ways, power at a constant current OR current at a constant power. They both apply, I see it both ways in papers.

Cool, thanks for the info. Usually i assume constant power with data like that cos of drive manufacturers..



Murder Candidate #2 nearing 40 hours!
The old diode is almost at 180h.

Time for a double re-plot in a few..

 

[sportcoupe]

Murder Candidate #2 nearing 40 hours!
The old diode is almost at 180h.

Fantastic!

 

[IgorT]

Both diodes survived till the last point (40 and 180h) and both did something completelly unexpected in terms of degradation, but not in a bad way. :thinking:

I plotted them both, but unfortunatelly i will have to post the data in the morning, cos i have a bunch of work to finish, in the meanwhile i will let the second Murder Candidate push onto 50h, which it now actually has a good chance of reaching.

50 hours is actually probably good enough for a year of low to average use!


With cooling off and re-plotting (and other) interruptions i am getting approximatelly 10 hours of ON-Time (20h of cycling) done per day, so the diode should reach the 50h mark tomorrow if all goes well.

Looking good so far. Very good in fact. :yh:

Fantastic!

Indeed. From this point on i'm not worried about the second diode anymore.

For those using this info to decide on a current/power setting, it is probably safer to say that a certain power is reliable for these diodes, rather than the current they were tested at.

In this case it would mean, that diodes set to 360mW will live over 180 hours (on average?), and diodes set to 450mW will live over 40 hours.
If a diode can reach these powers at lower currents, it's hopefully better for them.


What i'm getting at is, that a high efficiency diode might not survive 350mA as long as the second test 8x, due to the higher pressure it puts on itself, but a lot is up to chance here.

 

[ossumguywill]

Both diodes survived till the last point (40 and 180h) and both did something completelly unexpected in terms of degradation, but not in a bad way. :thinking:

I plotted them both, but unfortunatelly i will have to post the data in the morning, cos i have a bunch of work to finish

sounds cool... did they suddenly jump up in efficiency or something? That would be interesting, after "aging" your 8x it suddenly turns into a superdiode

will

 

[IgorT]

Here's the double re-plot data from last night...


First our second Murder Candidate (8x #1) at 40 hours:

Diode Stats:
Ith = 33mA
Slope Efficiency = 1.306mW / mA (1.294mW/mA @30h)
Absolue Efficiency = 20.71% (20.58% @30h)
Po @ 348mA = 413mW = 91.97% of initial Po (408.7mW = 90.11% of initial @30h)


This is the weird part i mentioned last night. Strangelly, the efficiency has actually gone UP! And i triple checked and made sure the temperature was the same as during the last tests... :thinking:

The diode has aged 10 hours longer, and yet it produces more power than before at the same current. Clearly more damage must have accumulated, but it's not obvious from outside, unless this is how it is showing.

In the plot it's the green line between the previous two, actually where i expected the red to show, only that at that point degradation seemed to have sped up a bit. Now it's not just slowing down but reversing?!?


I just hope this is not a bad sign...




Next, our old Murder Candidate (8x #3) at 180 hours:

Diode Stats:
Ith = 36mA
Slope Efficiency = 1.267mW / mA (1.183mW/mA @150h)
Absolue Efficiency = 17.80% (17.29% @150h)
Po @ 348mA = 315.5mW = 88.62% of initial Po (312.5mW = 87.78% of initial @150h)


The even stranger part is, that the same thing has happened with the old diode over the last 30 hours of it's ON-Time. It's efficiency and power have also gone up by a bit.

In the plot it's again the green line between the blue and the red, it's actually touching the blue line slightly, meaning the efficiency went almost back to the point where it was at 123 hours...



In both cases i was going back and forth on the diode analyzer several times until the temperatures (and Vf's) were as close as, or spot on with where they were in the previous plots, but the results remained the same, i couldn't put it off to lower temperatures (i actually had to warm the diodes up a bit, i allowed them to cool off a bit too far).



Last night i was worried the diodes were saying they are giving up in their own strange way, however only 180 cycles to go and they will be at 50 and 190 hours respectivelly. :yh:

 

[IgorT]

And indeed we have 50 and 190 hours. I don't know what i was worried about before.


Plots will be up in a bit.