8X Diode Murder fund

[IgorT]

42 hours 19 minutes...op:

 

[IgorT]

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

LOL

:crackup:

Cow? :crackup:

 

[bryce007]

42 hours 19 minutes...op:

I read this post in the email, and I thought it said that it popped at 42 hours and 19 minutes. That scared me, then now I see the smiley, lol.

 

[jayrob]

:crackup:

Cow? :crackup:

That is funny...

I've always liked Energizer much better than Duracow. :crackup:

Hey Igor, that is over 42 hrs. of 'on time' correct? (not half of 42)

 

[IgorT]

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.

I thought you might have been thinking in minutes, yea...


I guess it's possible it would reach ~166h... It's been quite tough so far.

But that requires that degradation keeps slowing down at the current rate - that's what the plot is based on...


At this moment the diode already made it one quarter that far. I need to re-plot it to see what's happening, if degradation is stil slowing down....


If it actually gets that far, i would even dare test 360mA next.. Altho it might be better to keep the current somewhere where the two experiments would be easier to corelate...

 

[IgorT]

That is funny...

I've always liked Energizer much better than Duracow. :crackup:

To me it seems like it's one and the same company... I mean, seeing how the logo under the letters is the same? :thinking:

Hey Igor, that is over 42 hrs. of 'on time' correct? (not half of 42)

Over 42h ON-time, yes... Cycling time is twice as much, experiment time even longer, due to the interruptions i made for cooling off and re-plotting (it'll be one week tomorrow).

This is also why i reduced re-plotting to once per day - so we would get more ON-time done faster...

 

[jayrob]

It's going to be interesting with the next one huh?

And the 12X too!

Thanks for all your work Igor...

 

[IgorT]

I read this post in the email, and I thought it said that it popped at 42 hours and 19 minutes. That scared me, then now I see the smiley, lol.

Hehe, i was thinking of how to make sure no one would misunderstand it...

Didn't think as far as email tho...


But even IF it were to pop in 42h as long as it is true that efficiency is a sign of diode "health", i'd be happy with the results! With normal use that's already a year if not more.

I think i was overestimating average pointer use myself as well.. Of course, regardless of anything, the longer it lasts, the better...

It's going to be interesting with the next one huh?

And the 12X too!

Indeed! In fact i think we should start deciding where to test the second 8x at soon...

We could test very high currents (like 360mA), but that might be impossible to relate to the first experiment.
We could test closer to the original current, but i imagine we also want answers about how long they can survive high powers...

I definitelly want to go higher, only question is where. Perhaps somewhere in between would be best. Like 320-330mA or so..


I need to check my data. Yes, 8x #1 would reach 410mW+ at 330mA.. I originally tested it at just over 400mW at 320mA. 400mW is definitelly a BIG step for BluRays in my oppinion!


12x's on the other hand might turn out to be "reliable" (in hobby pointer terms) at 500mW+...

 

[IgorT]

42h Degradation Plots​

These are the latest degradation plots:

Diode stats:
Ith = 36mA
Po @ 300mA = 332mW (93.26% of starting power)
Slope Efficiency = 1.254mW/mA
Actual Efficiency = 19.16% Peak / 18.87% Avg.


One thing is clear - degradation IS actually slowing down - and not just a little either! CHP's prediction may actually come true!

I would like to point out one thing... If you look at the P/I Plot, you will see that it now takes 320mA to reach the power the diode started out with at 300mA... That means it takes 6.67% more current, to reach the same power....


Strangelly, actual efficiency barelly shows a change, in fact the peak efficiency seems higher, but this could be in part due to the fact that the Vf was lower from a slightly higher temperature (this is why i wish i had a constant temperature rig)..


It is really strange to see degradation slowing down like this.

It's as if the lower power of the diode - the result of degradation - causes a lower "pressure" on the die...

But the difference is small! Only 6.74% less power than at the start! And since the damage is supposed to be growing, i expected degradation to speed up, not slow down... Somehow i find it hard to believe, that the 6.74% less power would account for such a big difference in degradation speed..


This makes me wonder what higher efficiency diodes will behave like.

I don't want to guess, but i have to say i'm really anxious to test the next one. It's efficiency was a lot higher, in fact it was one of the "normal" ones!



There are two things we need to decide on:
- What current do we test the second murder candidate at?
- Do we wait for the first one to die, or do i expand the "Torture Chamber" and start the second experiment ASAP?

 

[Crazy Jay]

IgorT lets look at possible options.

What if you were to continue increasing the current on the one you're doing now?
Would destroying it be more beneficial than putting it aside and testing it later after a certain amount of shelf life

What I mean is what would be more productive?

 

[davidgdg]

Dear Igor

Since there is only one remaining diode to test, we need to get as much useful information out of it as possible.

So rather than having a fixed current, I suggest that the next test be done starting at 300ma but with an X ma increase in current for each Y hours of on-time.

The X then needs to be small enough to give a decent spread whilst large enough to get up to high power levels before we all get old. Something like 20ma increments?

The Y needs to be long enough to provide a bare minimum of useful lifetime. Say 10 hours which would give a year's worth of useage at 2 minutes a day.

So the protocol could then be:

0-10 hours @ 300ma
10-20 hours @ 320ma
20-30 hours @ 340ma
etc etc

X and Y above are just suggestions, but as a general protocol I think this could be a good way forward.

Best regards

David

 

[HIMNL9]

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.

Hi, sorry, but isn't that chart logarithmic ?

Usually, it's not correct, make prevision curves using a linear path on a logarithmic chart ..... not with so few reference points ..... at least, as far as i remember



@IgorT: sorry, is just my impression, or also the degradation rate is slowing down ? ..... i mean, it's not degrading in linear way, as probably expected ..... perhaps you've found a diode that don't want to die, regardless what you do :wtf:

 

[daguin]

You do not change the conditions in an experiment UNLESS you have decided it is over. I do not believe that it is over. All I see is impatience rearing its head. Our original "goal" was to see if it makes it to 100 hours at this current on the cyler. If it makes it to 100 hours, THEN we can decide if we want to alter the conditions.

However, even if we do alter the conditions then, we will not know what it would have acted like if it had been at that higher current all along. We will simply see what it can do AFTER being run at this current for 100 hours.

If it makes it to 100 hours, it should be retired. It can no longer tell us anything about how a new diode will act. Put it into a host in which the diode can be changed fairly easily and play with it. It has done its job.

Peace,
dave

 

[CHP]

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

I wasn't assuming failure at 90% (10% loss). Just a figure or merit for comparison. I didn't know what the manufacturer's lifetime failure criteria was (20% loss, 30% loss, etc.).

 

[Deleted member 8382]

if it lasts 100 hours I vote for pushing it until it pops to see how far can they be pushed for short periods. Even if it's a freak diode it's degraded now, so we could get reasonable data...

OR, we could also make a lottery and donate it to the donators...

 

[Grix]

if it lasts 100 hours i vote for pushing it until it pops to see how far can they be pushed for short periods. Even if it's a freak diode it's degraded now, so we could get reasonable data...

Or, we could also make a lottery and donate it to the donators... :d

Kill! Kill! Kill!