8X Diode Murder fund

rtyrty100 said:

Ill donate 10$ if the second diode still needs to be payed for, but I dunno if it does, cuz all the edits on the main post confuse me :undecided:
Let me know what I can do

Click to expand...

Thanks for the offer. We paid for (and IgorT has) two 8X didoes to test (and a 12X).

Peace,
dave

Very well then. I will sit on my cheeks whilst I wait for another fund to come around!

Personally, I don't want to ever have to replace a diode unless I am upgrading it, which means it needs to last at least 50-100 hours. And I don't observe duty cycles, so I run my lasers for 10-20 minutes occasionally which tends to go through hours pretty quickly.

My point is, not everyone thinks it's useless to do any testing over 20 hours. These tests are not just to judge how long an 8x will last, but also to learn more about the nature of these diodes(e.g. how and why they die). If you change the test in the middle you will not learn any definitive information after the change, unless that is the specific test you are performing.

daguin said:

IgorT has already stated that he would like to watch it until it dies to track the degradation. Changing anything now will NOT tell us ANYTHING about how a "new" diode will act in a laser.

If it refuses to die, we can award it to a contributor or something. Maybe IgorT can build a laser and sell it at a discount to buy another diode to test.

Peace,
dave

Click to expand...

I will agree with IgorT then but I feel this diode has over 200+ hours in it.

Dave, how will we know if the diode "refuses to die"? 200 hours, 250, 300?

Changing the current now to something "ridicules" like 400ma and then we get 20 hours addition will tell us we can maybe expect at least 20 hours from a new 8x at 400m would it not?

sportcoupe said:

Dave, how will we know if the diode "refuses to die"? 200 hours, 250, 300?
Changing the current now to something "ridicules" like 400ma and then we get 20 hours addition will tell us we can maybe expect at least 20 hours from a new 8x at 400m would it not?

Click to expand...

Refusing to die is a relative concept. If we get so tired of watching it that we give up before it does, the we move on.

Living for 20 more hours at a high current AFTER it burning at a lower current for an extended time means nothing when considering what to do with a "new" diode. We have no idea what (if any) changes may take place at lower currents. Will a "new" diode react the same under these conditions than if hit initially with 400mA?

Take the S03 debacle for instance. I was able to push the first one well above 400mA when I turned it up 5mA at a time. However, not one other diode survived being hit initially with 300mA. Why? IDK. The conditions were different so the outcome was different.

Peace,
dave

Yeah... I think we will learn the most by just continuing to watch how it degrades as the experiment continues with the same settings.

It seems that we've sentenced it to death, and it continues to degrade. But we have already seen some interesting changes lately...

well, i agree we leave the setup as it is to watch it die, no matter if it takes another 10 or 100 hours. but i agree as well that we got the important answers for most of us - will it survive long enough?
so why not declare this first experiment over, leave it running at that current anyway, and setup the second diode?
not sure if enerything is ready for this, second heatsink for example.. dont want to push you, igor.

..and i would turn the second diode a good bit higher, this time. 360mA maybe? several other pointers use this. and if the weakest diode survives that long already..

manuel

Igor is building the second cycler position now. He was looking at making two more positions but needed another part.

Peace,
dave

The diode reached 123h+ over the weekend, uploading the latest degradation plots..



123h P/I Plot:



123h Efficiency Plot:



The graphs now only show the plots that are evenly spread appart in time...


Diode Stats:
Ith = 36mA (no change here)
Po @ 300mA = 317mW (89.04% of initial)
Slope Efficiency = 1.204mW / mA
Actual Efficiency = 17.95% Peak / 17.72 Avg.


The diode dropped under 90% of initial power quite a bit sooner than predicted, so the degradation clearly is speeding up. Altho it did speed up once before, but then slowed back down again.


Only thing i still don't understand is why it didn't show a difference from 90 to 100h, while a big difference is noticable after an additional 20h.

Something similar happened once before (don't remember at which point).


One interesting thing to notice is, that the first 30h caused pretty much the same power drop, as the 90h after that.



This diode most likelly won't last much longer. If it dies in under 150h it would answer the original question - what current can these diodes survive for 100-150h (of course it will vary from one to another and we can't really afford to kill enough to get to the averages).


One thing is for sure at this point - the 8x's are WAY tougher than i ever thought possible!

I feared 300mA might proove to be too much - i was worried i would have to give everyone bad news, but in this case i couldn't possibly be happier prooving myself wrong! :evil:



So we are definitelly pushing forward with the next one, but i think it would be best to chose a current, where the results of the two experiments are still possible to correlate in some way.


Ideally we would find a power where they last approximatelly 50h, which is still somewhat reliable for those willing to take the extra risk.

But if we go too high and get 20h or less as a result, that would not be good, because at those currents survivability will vary anywhere from 1-20h - the higher you go, the less the results of one diode will be comparable to another (some can pop instantly, others can last a surprising amount of time - let's just hope this is not what happened here - another reason i don't want to go too far with the second one, if done right, it could reaffirm the results of the first one, as well as provide answers about higher currents - how fast the life drops, as current is increased).


Does anyone know which currents they start dying at unpredictably. I heard there were several losses at 360mA+?

As i said, the higher the current, the more the results will vary from one diode to another - if we go too high the diode could die prematurelly OR it could last due to pure chance and provide a fake sense of safety at that current.


In my oppinion, it would be best to test 330mA next.. That would allow comparing what a difference in lifetime the higher current caused, perhaps correlate the results of the two in some way, but everyone who donated has a word in this of course...


Hopefully we come to an agreement about the second current, which will provide information useful to everyone.




P.S. There is one question that really bugs the hell out of me. How long a high efficiency diode would survive the same current as this low efficiency one.

I'm almost considering sacrificing one of the two 8x's i got for myself. I'm thinking of offering one of my personal lasers for sale, and if it gets a buyer, i think i'll kill one 8x diode out of my own pocket, to confirm the first current, (or the second one, IF the second current should seem "reliable" enough).

I badly want to see how much of a "health indicator" efficiency really is with these diodes!

Attack of the disappearing attachments.

As far as a current setting for the next diode, my uneducated guess would probably be about a 10% increase.

I'm with you Igor... 330mA's sounds good!

Why not start with a higher mA like 365ma on a new diode? afterall this is a murder, right?

Im guessing because they want an accurate base murder, and you have teh start somewhere! The more time it takes the for diode to crash, the more detailed information they have on its progress (or degradation).
Also they need a second diode to create somewhat of a constant of the first diode, so that these details can be confirmed. Once that base information is attained and confirmed, and the sharp details are aquired, then its time for a desolative murder :eg: So yeah I do agree with you on the 365ma or so for the THIRD or after diodes! But whatever the experts decide to do...is obviously better than what our weakling minds can fathom

And "technically" any mA is a murder, just obviously one murder quicker than the other

I like 330 for the second 8X, and maybe 380 for the 12X...

Igor;

I don't think you need to sacrifice your hi-efficiency diode.

Look at this graph from Nichia on many 320mw 12X BR diodes:
The graph on the right shows very little aging differences in BR diodes with various efficiencies,
even @ 80° C at 260ma Pulse Power.

With only an 11% power drop at >123 hrs., this 1st diode should last about 200 hrs or so.

That assumption, leads me to think a 15% increase in current would be a good level for the next 8X test,
which would be 345 ma.

This is 172% of rated peak current (200ma).

LarryDFW

P.S. I also found this quote from Sharp:

Optical power of 250 mW will be required by 2008 to record at 8X speeds on two- layered disks. By 2009, 350-mW power will be needed for 12x recording on two-layered disks.

Click to expand...

(I think Nichia, Sony & Sharp finally settled on 320mW for 12X this year)

Uhm ..... one point in favour, and one point against .....

Making the second 8X test similar to the first, give us more repeated data for "consistency" in the plots that we already have, thanks to IgorT, but don't give you data at higher current ..... doing it at higher current give us different plots in different conditions (that is a good thing), but don't let you know if the data that we have, coming from a single exemplar, are common to all the others, or coming from "freaky" diodes, and not repetitive .....

Just need IgorT to decide what is better, have different plots, so more quantity of data and accept the risk that they are not common, or have less quantity of data, but more "sure" ?


Edit: what also about the suggestion from some users, to shorten the on-off cycles ? ..... like, making them 30 seconds, as companies do ? ..... or keeping them 60 seconds ? ..... this, about the fact that, usually, when they refer to pulsed power at 50% duty cycle, they refer about high frequencies of usage ..... as far as i know, none of them tested the diodes for more than 30 seconds on-30 seconds off, cause also so, it's considered almost continuous use (quasi-CW), for the degradation effects of the chip (usually the duty cycle is referred about 1 second time, with 0,5 seconds on - 0,5 seconds off, not more)