8X Diode Murder fund

Must have been a lot of work Igor. :gj:

I am surprised at the differences in output power from highest to lowest. I guess I was just hoping for more consistency with the 8X's. But it does look like most are pretty closely bunched, and that there are only a couple of higher efficiency ones, and a couple of lower efficiency ones...

With those kind of power readings, the graphs must have been done using my 405-G-1 glass lens correct?

very nice Igor. Were those all tested with the same exact lens? You used the 405-g-1 right? But did you use the same 405-g-1 for each diode?

great job, thank you!

..so you were right, they all are perfectly linear, no kinks, no knees or similar! yay!

manuel

jayrob said:

Must have been a lot of work Igor. :gj:

I am surprised at the huge difference in output power. I guess I was just hoping for more consistency with the 8X's. But it does look like most are pretty closely bunched, and that there are only a couple of higher efficiency ones, and a couple of lower efficiency ones...

With those kind of power readings, the graphs must have been done using my 405-G-1 glass lens correct?

Click to expand...

If it was 11 6x's, it would be a lot of work. Here i was five times more careful on top... I made the diode analyzer to make my life easier, but i still don't really like doing so many plots....

As for the differences, the first two diodes i tested showed a large difference, but then again, both were good diodes.. At that time i guesstimated that the two represent the maximum and the minimum efficiency and that the rest will fall in between, because #2 overlapped FrancoRob's 203BK spot on, and from the other collected plots, i saw it was a very high efficiency one before, while the second came in at the lower end of the data i had available...

Luckily only one diode came out lower, and the two lowest ones are being murdered anyway.
I was surprised that one came out higher tho!


Interestingly, in the P/I plots, the lowest seems closer to the group, while the highest seems to be separate.
But once their Vf is taken into account the opposite becomes apparent. The highest is actually a part of the group, while the lowest stands appart from the rest.

Many diodes are also very tightly packed together around a certain average.. So tight in fact, that you can't even see half of those, except where they peek out a bit randomly...

I'll re-do the plots with thinner lines tomorrow evening, and also calculate the slope efficiencies for each diode.


Overall, i'm happy that no diode pushed the two "winners" out of their spots.. I was very worried while testing the last three, but luckily for nothing...
And while the efficiencies vary, if these were reject sleds, there would be a whole Low Efficiency group around or even mostly under diode #3, judging by GGWs!

Anyway, there is some work left to be done on the plots, but i can safelly start torturing the first one on the cycler now. :yh:


And yes, all diodes were tested with the 405-G-1 lens. Somehow putting plastic lenses into 8x's just doesn't feel right.... And then variations would be influenced by the wavelengths...



Bryce: The only (properly working - not counting the shitload of useless prototypes) glass lens i have on hand right now is one 405-G-1...

The diodes were all tested in one and the same module and lens.


I'll take photos of the testing module and heatsink and the entire testing setup (i'll use the same module/heatsink combo for the Cycler experiment), i wanted to take some photos now, but i can barelly keep my eyes open.

Tomorrow i have a lot of work, so i badly need to get some sleep, but i'll post pics of the setup and more explanations as soon as i hook it all up...
Right now i can barelly even think straight..

Hey IgorT...

wow... nice work on plotting those 8X graphs...
can't wait to see the testing module..heatsink and entire
testing setup pics...

Thanks for all your efforts...


Jerry

Thanks Igor!! you are working very hard to plot and torture all of these diodes! The results are looking good so far...nice and linear...
You deserve a rest now....:gj:

IgorT said:

Ok, all diodes plotted ... one came out high, the other low. I guesstimated that the rest will fall in between. Well, all except for two did - #3 and #9...

Click to expand...

Great effort, Igor :beer:
Looks like almost 100mW between the "winner" and the freak ("Beauty & the beast", lol)...

@ IgorT: imho, is not important what lens you use, as far as you use always the same lens and focusing for all the tests.

Being sincere, i personally was more oriented in using no lenses at all, measuring the bare output of the diodes themselves, cause at near-UV range, also glass may have some problems (mean, except specific optic-grade crystals and glass mixes, some other glasses can contain small parts of fluorescent impurities, but suppose that these lenses are planned for this specific use, and exempt from this problem, and this may also be the cause of the better power gain, not just the AR treatment) .....

But i suppose that, except building a specific setup, or use a thin heat filter in front of the sensor, there's no way for measure the real output from the diode, without affect the measure with the heat produced from the diode itself (at least, if i try to measure the power in this way, my meters reads more power than the real, just cause the diode is too much close to the sensor)

Anyway, great work ..... seem the #9 is a bit "freaky" in output, and the #3 is the lowest one, the others seems inside the usual +/- 10% that the manufacturers usually declare for their products ..... The #1 is a bit low, too, but if i read correctly the PO/PI plot, it also "eats" a little less power than the others (also if the difference is minimal) ..... if you continue this way, then maybe we can write a datasheet, too

:beer:

Krutz said:

great job, thank you!

..so you were right, they all are perfectly linear, no kinks, no knees or similar! yay!

manuel

Click to expand...

Yes, starting from 100mW CW / 200mW Pulsed, kinks seem to be gone with 405nm's for good!

BUT... As mentioned before, there is at least one case where a kink developed later through degradation with an 8x - Hemlock Mike reported it on a diode that used to do 400mW+, still starts there but drops rapidly with it's beam splitting into two...

The good news is, he says he was among the first to make a 400mW+ 8x laser, meaning it lasted quite a while before this happened only question is how much it was used in that time... I also don't know what the efficiency of his diode was (is his diode among the plots Franco collected?)..


In any case, this means i need to be careful when testing these 8x's, make sure i replot the diode DURING the test several times, perhaps every four hours (4h of run-time = 8h cycling time), so as not to miss this if it happens, and of course to record degradation....

Now that the diodes are already plotted, testing for degradation will be quick and easy. I just need to measure Ith, 40mA and 280mA, after letting the diode cool down for a bit....
- Degradation would show in an increase in Ith and lower powers at previously tested steps - 405nm's supposedly degrade by Ith climbing, while the slope efficiency remains the same - the plot just shifts to the right.
- A kink would show in powers up to a certain step being the same, but at the higher steps they would drop drastically.


Otherwise, while testing i decided to go to 280mA with all diodes, so as not to have to extrapolate four steps for every diode.. The accuracy would be reduced - especially in the Po/Pe plots....


I kinda doubt anyone is interested in much less than 280mA with 8x's, so i thought it wouldn't hurt to go that far for a quick plot (only takes seconds to check the numbers and quickly turn the current back down). Going to 360mA really would be too much, before we even know how long they last.. The last two steps (320 & 360mA) are what i'm extrapolating...

Power vs. current is easy to extrapolate, since it behaves pretty much linearly (well, it does drop a little with every next step, but it's easy to predict). But i wanted to compare diodes according to their actual efficiency. And for that i need their Vf.


Extrapolating the Vf is the hard part.. Basically i would have to find a pattern of how it behaves, and then continue it... But since with two diodes i actually went to 360mA, i have their Vf plots, and could adjust the others so they have the same curvature.
But even better than that is using the Efficiency Plot (% of Pe), because it drops somewhat linearly after the peak, and if i have Po's (which are easy to extrapolate) it only depends on Vf... That way i can adjust the Vf until the efficiency drop matches that of other diodes, and make approximate but pretty accurate Po/Pe plots up to 360mA without having to push the diodes there...




Only question left is, when i hook it all up to the cycler, what current do i set the first diode to? The lowest efficiency diode only produces 336mW at 280mA.

At 300mA it would do just over 360mW.... We said current specific testing and 300mA for the first one.

But since it stands quite far appart from the group, i was wondering if 280mA might give better results, closer to those of an average diode at 300-320mA perhaps....


This is of course a guess based on the assumption, that efficiency equals, or rather is closly related to diode "health"..

I don't have a lot of evidence for this, i have not killed enough GGWs to proove it, while PHRs were not consistent enough.. In one experiment a low efficiency GGW was the first to die at 200mA (only a few hours), while the normal or high efficiency ones "out in the wild" all still live...

For now it's just a feeling... But it would make sense...




Otherwise, here is a photo of the testing setup i use for plotting..

The big heatsink was a present from Tallaxo, attached to it is the head piece of a v3 Heatsink/Module.
This particular module is slightly different, in that it allows me to press a diode in, where it is grabbed firmly, for good thermal contact, but also allows me to safelly remove it afterwards.

The heatsink is big enough, that i can make a full plot, and when i go back to check the previous steps, they still show the same power - no drop from heat.. During plotting the heatsink is on a plastic stand i made to hold it at the height of my LPM's thermopile opening.

I wanted to add a third display to the Diode Analyzer for temperature, already put the module in the enclosure, but it's getting tight on the panels (i'd have to move the switches to mount a third display). Besides, while it would tell me the temperature, it would not respond nearly as fast as the Vf does.


Basically, when i'm plotting a diode, i use the Vf display not only to write down the voltage data, but also to watch for any temperature drift... If the Vf would start dropping, i would know the diode is warming up and the plot might be affected....

The Vf changes MUCH faster, than the temperature would, besides, the temperature probe would measure the heatsink, not the diode, which means there would be a considerable lag.... The Vf directly corresponds to the die temperature on the other hand.

Because of the size of the heatsink, the Vf stays put during plotting, but if something was wrong (if the thermal transfer was imperfect at some point) it would immediatelly become apparent in Vf.


Anyway, this is the heatsink i am going to use for the Cycler test, except with another module head, which is meant for permanent installation...



Otherwise, i have to "calibrate" the sensor now. It has to register powers over a wide range, but at the same time not respond if the diode goes LED (or 5~mW like the F03's - just in case), so it needs a "callibrated diffuser", partially so as not to fry the photo transistor (already killed one with my 6x :yabbem, and to stop it from registering below certain powers...

I miss my big test heat sink :cryyy:

Damned thieves:gun:

Peace,
dave

impressive setup!

i cant add much to the question where to set the first "candidate". i know i will surely push mine higher than that anyway, i still wish for a classIV laser.. ;-)

well, do we want to produce recommendations for the average diode, or for the worst case diode? if we say "put your diode to xx mA", should it be safe for 99% of all users, with low(er) output, or safe only for 90% of the users, but with higher output?
..simply assuming that the one-from-13 "low" diode ratio is any representative..
(i dont think its very clear what i mean..?)

manuel

Krutz said:

impressive setup!

i cant add much to the question where to set the first "candidate". i know i will surely push mine higher than that anyway, i still wish for a classIV laser.. ;-)

well, do we want to produce recommendations for the average diode, or for the worst case diode? if we say "put your diode to xx mA", should it be safe for 99% of all users, with low(er) output, or safe only for 90% of the users, but with higher output?
..simply assuming that the one-from-13 "low" diode ratio is any representative..
(i dont think its very clear what i mean..?)

manuel

Click to expand...

I don't understand what you mean here. Didn't we decide on 300mA as the current for the first murder?

Traveller said:

Great effort, Igor :beer:
Looks like almost 100mW between the "winner" and the freak ("Beauty & the beast", lol)...

Click to expand...

Wow, you are right!

Looking at the P/I Plots, three diodes stand appart from the group. The "freak" (#9), and the two murder candidates (#3 & #1).. The rest of the diodes form a tightly knit bundle, which varies just under 40mW at 280mA..

But the difference between the freak and the first "winner" is a whopping 91mW at 280mA! In the Po/Pe plots, the difference is even a tiny bit higher (at 1.6W in)!


Otherwise, it's interesting, how three diodes stand appart from the group in the P/I Plots, but in the Po/Pe plots where their Vf is factored in, only one really stands appart - the first winner.

The second winner is practically overlapping with another diode in the Po/Pe plots... At that point i started wondering which diode should get the silver, but then i decided that since diode #1 needs more current to reach the same power and is a hair lower in the Po/Pe plots too, it should remain in it's rightful place (i'm sure it feels very honored)...

bryce007 said:

I don't understand what you mean here. Didn't we decide on 300mA as the current for the first murder?

Click to expand...

He is answering to something i was wondering about...


Basically, before i started testing the lot, i only tested the first two.
They showed certain differences, but nothing where i would consider a diode to be "bad"... I was hoping the rest would all fall in between (and luckily almost all did)...

But now, suddenly one diode fell short of that by quite a bit... The rest all seem huddled around a certain average in a group.

Groups like that have a certain meaning... It's the diodes telling us they are all good and normal. Except for one, which is an "outsider"...


Since that one diode is considerably lower, i was wondering, if setting it to 280mA instead of 300mA would yield end results closer to those of an average diode at 300-320mA. Killing this one at 300mA will only result in the minimum the weakest diode survives 300mA..

But what do i base that on? Nothing really, except a gut feeling. I don't even have a calculation handy in this case.


Well, i did consider setting it to a current where it would produce as much heat as an average 8x at 300mA, instead of 300mA but other than that....


But before that we all seemed to agree on 300mA and besides i did say i want to test the minimums, so i probably shouldn't complicate... :angel:

I should just start the test quickly (and pray it doesn't end too quickly)...



P.S. The thing that worries me is, that if we had bad luck and stumbled on an actual weak diode by accident and it dies too soon, we might be forced to repeat the 300mA test on the second one...

But this is just me, worrying to much as usual..

daguin said:

I miss my big test heat sink

Damned thieves

Click to expand...

Wait, what?!? What happened? :thinking:

I know Tallaxo made you such a heatsink first, but.... Thieves?!?

Igor;

I suggested a 8X "safe operating level" of 240 ma for >100 hour lifetime.
(link below in signature).

Dr. Lava's testing showed diode failure @ 42% over the "safe operating level" I calculated.

So, with that background I am guessing failure at ~346ma on your first 8X diode.

I would start @ 300 ma.

LarryDFW

I've been running my 8x on 300ma and it doesn't get hot at all. I'm buying another host... the 18650 one and I want to increase it to at least 390ma. The beam isn't really all that visible in its current state but it burns like hell and damnation itself. It even burns without being completely focused. The beam on my dilda is brighter than it is.

It doesn't even get warm for me. I have it in the side button host right now.
Again... no heat.