8X Diode Murder fund

IgorT said:

This diode has it's mind set... It won't give up any time soon!

It just reached 90h this moment. Waiting for it to cool off for re-plotting.. :yh:

Click to expand...

wow that's awesome!

IgorT said:

This diode has it's mind set... It won't give up any time soon!

It just reached 90h this moment. Waiting for it to cool off for re-plotting.. :yh:

Click to expand...

Excellent. Now if an 8x can survive 90+ hours at 300ma I would think that it could take 400ma for at least 10 hours, more then enough for an average use pointer.

sportcoupe said:

Excellent. Now if an 8x can survive 90+ hours at 300ma I would think that it could take 400ma for at least 10 hours, more then enough for an average use pointer.

Click to expand...

Before you jump too high, remember that we have had several die at >350mA with "normal" use already.

Peace,
dave

daguin said:

And this is what we consider to be the worst diode of the bunch

Peace,
dave

Click to expand...

I am starting to wonder if i should change my definitions of "good diode" and "bad diode"... :crackup:
With GGWs i could always rely on efficiency to answer this question, here i'm starting to wonder... :thinking:

That's why i would so much like to test another diode of a higher efficiency at the same current (i have a theory that would be absolutelly horrible if correct :undecided!
But unfortunatelly we "only" have two diodes to kill (not that i'm complaining - i was expecting one when i proposed this), and we have to push forward, and test higher currents...


I just think it would be a good idea if this higher current would not be that much higher, that the two experiments would be completelly impossible to corelate...
That's why i believe somewhere in the "middle" might be best for the second one.


I will of course listen to, and consider testing ideas and proposals from everyone, but the idea has to make sense and be well explained (- as in what, why, what do you hope it will tell us, and so forth (and keep in mind, it's almost a $200 USD question!))....

Krutz said:

but but but!
but it should be possible to find out how diodes degrade at their rated output? from the manufacturer, datasheets, pressreleases? i expected its simply stated in datasheets?
we dont have to identify unknown diodes. we simply try & error until we know at what output they degrade like 1% every 100 hours? (if thats what the other datasheets say about rated output)

sure, if we cant find out what the "rated degradion" is like, we are out of luck again.

perhaps the datasheet wouldnt help too much anyway? we want to know at what output the diodes survive many hours, not at what rated output they survive several thousand hours..

manuel

Click to expand...

This information IS known.. Altho not exactly in this form...
It's declared as diode life at rated power, and it's usually declared as 6000h for disk writer diodes...

But obviously that doesn't mean every diode will last 6000h if used at the rated power!

6000h is just MTTF - Mean Time to Failure... It's the time it will take for half of the diodes either to die AND/OR degrade to the point, where they are considered useless (the latter varies slightly according to the manufacturer, but usually it's when it takes 130% of the initial current to reach the initial power).


Now try to imagine getting a diode rating using this data in reverse!

You'd have to kill a huge number of diodes at one power, to see how long they last on average. Then adjust the power and kill again, till you reach a power, where they die / become "useless" at 6000h - again on average...

Even with the cheapest diodes this would be a VERY expensive experiment! Never mind the time it would take....


But if you then rounded the result down to a value that would make sense, you might be close to the diode's actual power rating.



On the other hand, this is preciselly what diode manufacturers have to do, when they create a new diode / improove their diode manufacturing process..

They have a certain goal they want to reach - a certain power rating...
Then they have to start killing diodes at that power, to see if they can survive it long enough on average to get that rating...
If they die too soon, they can rate them for a lower power, but have to improove the manufacturing process if they want to make a higher power diode... And they know they reached their goal, when their diodes die/become useless at around the predetermined number of hours on average..

Even then, some will die MUCH sooner, others will live WAY longer than average. And yet they are all rated for the same power.
The power rating doesn't hold a magical special meaning, it's just a number the manufacturer was comfortable to rate their diode at....


When you read a datasheet, especially the extremelly rare detailed ones, you see, that they don't guarantee every diode will survive that number of hours.
In fact, they don't even guarantee that an individual diode will survive any number of hours at all, for that matter! Just that most will!
Every datasheet i've seen has a disclaimer, that any diode can die at any time, even when used within it's ratings....



Every diode is different. Every diode will degrade at a different rate.


If this was an ideal world, we'd have to kill 100 diodes at 300mA, then 100 at 320mA, another 100 at 340mA and another 100 at 360mA.

Then we could use the results to create a plot of how long diodes will live at different high currents - curves, representing the average, minimum and maximum lifetimes one can expect at each current...

Now wouldn't that be something?


But with only two diodes at our disposal, we have to pray, that the number of hours the first one survived so far means, the 8x's are really tough and not, that i just sentenced the toughest 8x ever to death... :angel:

We will know more when the second test diode reaches a certain number of hours.. :shhh:



But when it comes to the power ratings of the diodes. What's wrong with the charts Larry dug up?

Before, we were only guesstimating that GGWs are 100-105mW CW diodes, now the charts confirmed that, as well as told us, that 8x's are 120-125mW and 12x's 150-160mW CW diodes...

Uhm ..... just a weird though, and can be wrong, but ..... if, as Larry said somewhere, those diodes are rated for 100mW CW / 200mW pulsed, AND if the MTTF declared from manufacturer is 6000 hours at rated power ..... if it last as example 150 hours at 50% duty cycle (also if, being corrects, our use is just "technically" a 50% duty cycle, cause the declared one, usually, is based on an intermittence of max 1 second on / 1 second off, and any longer interval can be considered quasi-CW, for the damage effect), this mean that the degradation can be exponential ..... so, maybe, 400mW can give you not more than a pair of minutes of light .....

Ofcourse, this is only a free consideration, cause i don't have precise data ..... is just as actually the tendence resemble, not a precise measurement.

(and we don't have the some-hundred of diodes to burn, as IgorT said is necessary for get these data ..... )

But, at least, we now know that there's a good possibility to use them for some decent time at THIS current, that is anyway more than that what we was knowing before ..... and data still accumulates

Patience, patience .....

The worst is that we can't make another test at less current because it would take us months... xD

Igor;

This is a graph from Nichia on 320mw 12X BR diodes:
The graph on the right shows very little aging differences in BR diodes with various efficiencies.

LarryDFW

The test diode has passed 90+ hours........3 days ago.

I'm absolutely positive it's well over 100 hours by now. :bowdown:

edit: I had said the diode broke 90 hours and some people misread that to mean the diode was broke. I changed "broke" to "has passed".

Oh.. you said "broke" and I said, "WAT".

You confused me for a second. The diode is NOT broken, it alive an kicking! It should be at about 120 some hours now that I think about it...

Or, maybe, we don't have any other news, cause the diode decided to take his vengeance, and set IgorT torture chamber in fire ..... J/K :crackup: :crackup: :crackup:

it was mentioned some time ago that a torture test that should be done is to turn the diode on and off repeatedly for a good long time. Since we use our lasers for short periods of time and not for long, it's the on/off cycle that ends up killing them fast. (I think it was climbak that suggested this)

I would love to see 5 sec on / 5 sec off repeat cycle. This would truly be a torture test.
However, I did not contribute to this test so havn't got a say.

I think that is what he is doing Ken

Does that mean that pulsed operation is the worst form of powering on and off?

so... ETA on completion? or did i read over and miss something?

Kenom said:

it was mentioned some time ago that a torture test that should be done is to turn the diode on and off repeatedly for a good long time. Since we use our lasers for short periods of time and not for long, it's the on/off cycle that ends up killing them fast. (I think it was climbak that suggested this)

I would love to see 5 sec on / 5 sec off repeat cycle. This would truly be a torture test.
However, I did not contribute to this test so havn't got a say.

Click to expand...

IIRC The "cycler" is set for a 50% duty cycle (60 on; 60 off)

Peace,
dave