8X Diode Murder fund

[davidgdg]

Hi Igor

If this diode continues to do well, will you still set the next one at 300ma or might you use a bit more?

 

[IgorT]

But keep in mind that, usually, the definition of "end of life" from manufacturers, is normally referred to the "useful life for the planned use" that they do ..... and that this is normally referred about when the power go 30% less that the tag power ..... (and also that we are over-use and oer-abuse those poor diodes )

Good thing that for us, the "useful life" ends when the diode dies, eh?




Our little diode already made it through 25h+ of ON-time!

That's over 50h of torture (+ breaks, there have to be breaks so it doesn't develope tolerance)..

This is already way better than my worst fears. They are tough diodes, that's for sure...

 

[IgorT]

Kewl! Yea if they are a reasonable price, I'll probably take like 5 of them. Keep me posted

I think they were 3 or 5 USD each. Or maybe EURO.. Not sure.. It's just an MCU that when turned ON records the time it's ON..

The reader cost a little more, but still not enough to say it was expensive.

I just don't know how far the project ever went. The ones i have are prototypes. I wish i could have gotten more.

 

[IgorT]

Hi Igor

If this diode continues to do well, will you still set the next one at 300ma or might you use a bit more?

We want to squeeze as much info out of the two as possible.
In an ideal world, i'd want to kill 20 at 300mA to establish some averages of how long they will live.
And then kill another 20 at a higher current and another 20 at an even higher current, and so on, to create a chart of how long they will live at different currents...


But we "only" have two at our disposal (altho that's great, the original idea was one diode!). If this diode continues to do well, and really impresses us, we might want to test how they do even higher up.
Altho i think going to 360mA might not be too good, perhaps something in between, so that it can still be somewhat "comparable"....


It's still open for discussion. We will have to decide by the time the first diode gives up...



Otherwise i just measured it's power at 300mA.. I didn't re-plot it completelly, i just measured what it puts out at room temperature at the tested current...



Po @ 300mA is 338mW after 25h+, maybe a bit more, the meter couldn't decide between 338 and 339mW.


So while there is some additional degradation, it again seems very little over 5h. Just like there was little difference from 10-15, but a noticable difference from 10-20h.

So i guess it wants to go to 30h and further... I should just let it.



I stopped the cycler yesterday as the counter reached 1500+ counts. I was too tired to continue working or to replot it, but didn't want to miss anything.
Now that i measured it, i realized i should have just left it cycling on.


At the risk of missing something i think i should stop trying to check it at such regular intervals, so we get more ON-time done faster....
Re-plotting it every 5h was not part of the original plan anyway, and i should have just done what i said i would yesterday - only replot it once a day.


A lot of time gets wasted when waiting for it to cool off. Even more if i can't replot it immediatelly.
The torture chamber would really benefit from a constant temperature rig...

I was considering making it completelly computer controlled, but unless it had a constant temperature TEC setup, even the computer would have to wait for it to cool off before re-plotting...

 

[IgorT]

We are reaching the 30h mark!!!

I have to admit, i did not expect this. Not at 300mA with the weakest diode!

I am very impressed by these 8x's! :yh:



EDIT: We have reached the 30 hours of ON-time! :bowdown:


Two pics... :yh

The diode deserves a short break... :bowdown:

 

[bryce007]

schweet!

 

[CHP]

Ok, the 20h degradation plots...


There was no obvious drop from 10 to 15 hours, but 5 additional hours of ON time left their mark... The difference does not seem larger than that from 10-15, even tho it's the result of 10h of ON-time (20h of cycling)..


Anyway, attached are the plots, the diode already prooved it is tough, we shall see how much longer it can take this kind of abuse, but i'm already impressed... :yh:



Other data:
Ith is at 36mA
Po @ 300mA is 339mW
Slope Efficiency dropped to 1.275mW/mA



P.S. I think in one of the previous reports i measured 37mA for Ith, but Ith is kinda hard to determine. I usually simplify it by recording the value where the power reaches 1mW, even tho lasing actually starts sooner...
Measuring Ith is made a little harder by the fact that spontaneous emissions also register on the LPM, so i try to dial in just under the lasing threshold before zeroing in my meter, and then raise the current until i see 1mA...

There are some errors always involved, and they are the most apparent at such low numbers. Higher up the numbers become much more accurate.


P.P.S. This is it for today, i have a bunch of work to do, i will measure power after another 5h of ON-time, but unless it's drastically different (indicative of a kink), i will not replot the diode until 10 more hours of ON-time...

I plotted relative output power (power@time/initial power) data at 300 mA versus hours in a log/log format. If you extrapolate the curve, at 10,000 hours the power will be at ~90% of the initial power.

 

[IgorT]

30h Degradation Plots​

Diode stats:
Ith = 36mA
Po @300mA = 334.5mW (= 93.96% of initial power - 364mW)
Slope Efficiency = 1.264mW/mA
Actual efficiency = 19.14% Peak / 18.88% Avg.


I calculated the average efficiency from the data points where the efficiency flattens out - starting at just under 600mW El. Power (120mA and up)..




What i find interesting is, that the degradation seems to be slowing down!

The first five hours left a noticable mark on the diode...
The second five hours caused the most degradation.
But then it took TEN additional hours to cause LESS degradation than the second five hours did!
And now the last ten hours of ON-time caused as little additional degradation as the first five. :thinking:


I mean, fact is, that the more the diode degrades, the lower it's output power - the lower the optical flux at the die, which is what causes it to degrade.. But at the same time, the imperfections (or in this case damage) should be growing and accumulating.

So while the optical flux is a little lower, it has more imperfections to put "pressure" on..

I was expecting the changes to become more and more noticable with time... I thought this was confirmed, when i saw the second 5h causing more degradation than the first... But then, five more hours didn't even cause a measurable difference, now it takes 10h to cause noticable degradation....


I also did some tests to see if any kinks might be developing, by allowing the diode to warm up at different currents (i removed it from the big heatsink, leaving it in the small one only), but it behaved normally - it would heat up, the Vf would drop and so would the power.

As soon as i reattached it to the big heatsink, the power returned to normal. I tried this at the current it is being cycled at first, then at higher (320mA) and lower currents (280 & 250), but nothing happened, it behaved as any normal healthy diode would..


The only thing changing is it's efficiency, which started out at 20.43% peak, 20.15% average, and dropped to 19.14% peak, 18.88% average..
Or in slope efficiency, 1.339mW/mA dropped to 1.264mW/mA..



I'm starting to wonder what the second diode will be like, and if there is a chance it would actually degrade faster due to a higher optical flux.
My assumption was, that the lower efficiency diode would degrade faster, and at the same power, i'm sure it would. But... No, i'm not gonna make any guesses right now, we should just wait and see.

The first diode surprised me in a positive way so far. If the second does the same, i'll be very happy!



From now on, i'll just measure the warm power during cycling after 10 more hours of ON-time, to see if there is a rapid drop indicative of a kink.

If not, since Ith is clearly not changing (the differences i noted initially were errors, since they go back and forth), all i need to make a degradation plot is the diode's power at 300mA with the diode cooled down...

I'll only re-plot it if i notice strange behavior - power rapidly dropping with heat and the spot splitting in two.

 

[IgorT]

I plotted relative output power (power@time/initial power) data at 300 mA versus hours in a log/log format. If you extrapolate the curve, at 10,000 hours the power will be at ~90% of the initial power.

Can you show the plot you're talking about?

The diode is already at 94% of it's initial power after 30h... If it goes on like this (and survives a drop to 90%) it will be at around 50 or 60h.

 

[o0ICE0o]

Im waiting for 8xs to be PHR cheap.
It may take a while, but the 22x reds are only $10...
who knows?

....now to wait.

EDIT: are we there yet?

 

[IgorT]

You might be waiting for a long time, i'm afraid. GGWs appeared as sleds approximatelly 10 months ago. And so far they are still holding their price. Well, it dropped a little recently, but it's still several times the PHR price...

Hopefully our Chinese "friends" can get 8x sleds, since they have their "foot in the door" with LG, perhaps even Pioneer. If they can get them, they'll buy them by the truckload for a few dollars per sled, and then lie to us about availability to keep the price up (as usual).


In any case, i've been watching this diode for almost a week now and i can't believe how tough it is! In my oppinion, they are definitelly worth their price even if we have to rip them out of drives. It's amazing how much power they can handle...


This diode is like that Duracell bunny commercial. It just keeps going, and going, and going....


I can't believe i'm gonna say this, but we're most likelly testing a higher current with the next one! :evil:

Funny thing is, when i suggested this experiment, i did so because i believed they were being pushed too high. Now i actually think we can squeeze more out of them (at the cost of more lifetime of course)!

Of course we won't know for sure until the second one is tested. So far everything is hanging on the premise, that efficiency is directly related to diode health....



P.S. 35h ON-time passed a while ago, still no kinks in sight, nothing new to report...
Except that i'm starting to wish i could speed up the process, but the diode has to cool off in between.

One thing i've noticed is, that it's reaching slightly higher average temperatures now, with the efficiency having dropped quite a bit, but that's only normal...

 

[Flaminpyro]

I's the Enegizer Bunny not duracell, duracell has a cow :crackup:

You might be waiting for a long time, i'm afraid. GGWs appeared as sleds approximatelly 10 months ago. And so far they are still holding their price. Well, it dropped a little recently, but it's still several times the PHR price...

Hopefully our Chinese "friends" can get 8x sleds, since they have their "foot in the door" with LG, perhaps even Pioneer. If they can get them, they'll buy them by the truckload for a few dollars per sled, and then lie to us about availability to keep the price up (as usual).


In any case, i've been watching this diode for almost a week now and i can't believe how tough it is! In my oppinion, they are definitelly worth their price even if we have to rip them out of drives. It's amazing how much power they can handle...


This diode is like that Duracell bunny commercial. It just keeps going, and going, and going....


I can't believe i'm gonna say this, but we're most likelly testing a higher current with the next one! :evil:

Funny thing is, when i suggested this experiment, i did so because i believed they were being pushed too high. Now i actually think we can squeeze more out of them (at the cost of more lifetime of course)!

Of course we won't know for sure until the second one is tested. So far everything is hanging on the premise, that efficiency is directly related to diode health....



P.S. 35h ON-time passed a while ago, still no kinks in sight, nothing new to report...
Except that i'm starting to wish i could speed up the process, but the diode has to cool off in between.

One thing i've noticed is, that it's reaching slightly higher average temperatures now, with the efficiency having dropped quite a bit, but that's only normal...

 

[HIMNL9]

I's the Enegizer Bunny not duracell, duracell has a cow :crackup:

In the USA, maybe ..... here in Europe, Duracell have a bunny as mascotte

LOL

 

[o0ICE0o]

In the USA, maybe ..... here in Europe, Duracell have a bunny as mascotte

LOL

American.

...Back on topic now.

 

[CHP]

Can you show the plot you're talking about?

The diode is already at 94% of it's initial power after 30h... If it goes on like this (and survives a drop to 90%) it will be at around 50 or 60h.

I meant minutes not hours. My fault. Here is the graph including the 30 hour test data with a projected line to 90%. The projected line crosses the 90% mark at about 10,000 minutes.

 

[LarryDFW]

CHP;

Your projection is for 166 hours if the diode fails @ 10% power loss.

According to the manufacturers test methods, the diode should actually last even longer.

The diode is now requiring 6% more current for the same power.

I think it will easily surpass the 100 hour lifetime that Igor was hoping for.

Not bad at all, since we are overdriving it by 36% of maximum pulse power current (300/220).

Only the testing will tell for sure.

LarryDFW