3x 445nm LD PIV Plot

I can't see any reason why you shouldn't be able to get at least 1.8W.. sometimes you'll get some small losses. If you were to raise your current slightly I'd guarantee you'd see 2W..

If it were me though, I'd use that kit with a PBS on the second flex mount instead of a mirror. If you're only combining 2 diodes a PBS is far superior.

Meatball said:

This is the kit laserman532 made for me.

Laserman532 445 Dual Knife Edge Beam Combiner KIT

You think I can get 2W+?

I'll bet I can. My driver is set for 1.25A max for each diode..

I'm going to use Dr Lavas cylindrical lens kit to fix divergence, and after the expected 3% losses, your graph is telling me I might get something as high as 2.5W! Your results seem to be very consistent with lots of others' results!

Click to expand...

I see, i've seen setups like that with an even larger number of diodes than in the pic ElectroFreak posted IIRC. But i didn't know they were called that.

I do believe you should be able to get ~2W, whichever way you go, since the diodes appear to be relativelly consistent and two at 1.25A would generate ~1.3W each, if not slightly more.

But as ElectroFreak says, you might not need such a complex setup for it since it's only two diodes in your case...


At first i thought you were planning to use one diode driven to obscene currents, that's why i mentioned a TEC rig, cos there is a noticable drop-off in power at higher currents, which might turn into a proper knee if you went even further (above 1.2A where i stopped testing).

Igor, I've found beyond a doubt that these diodes LOVE the cold. You will get an improvement from cooling it down..

Here is a video showing it.
http://dl.dropbox.com/u/3175365/Video 1.wmv

Now this diode has been kinda zombified. but it does still put out a lot of light.. albeit not as much as it should.

500ma = 100mw
700ma = 200mw
1A = 400mw
1.5A = 600mw
1.6A = 700mw
1.7A = 800mw ish.

Kenom said:

Igor, I've found beyond a doubt that these diodes LOVE the cold. You will get an improvement from cooling it down..

Click to expand...

I believe so myself, because of the accelerating drop in efficiency after the peak.

Here is an Efficiency Plot for the same three diodes. Efficiency is plotted against Pe (electrical power going IN) and shows the percentage of how much of the electrical power comes out as light:




I believe proper cooling (= with constant temperature) would flatten the curve after the peak somewhat.


Otherwise, when i started testing these diodes i expected the PI curves to have many severe bends due to being multi-mode, but they came out relativelly straight for this type of diode... Still, cooling would probably straighten them out some towards the end of the PIV plots in the first post...

Kenom said:

Now this diode has been kinda zombified. but it does still put out a lot of light.. albeit not as much as it should.

Click to expand...

This is one thing that worries me if i should attempt a torture test on them. Do they often go zombie or do they mostly just die completelly, like any self respecting diode should?

Problem is, if one went zombie, my setup would not detect a problem and might keep triggering the counter, since diodes of the undead variety can still sometimes pass for alive, when in reality they are anything but.


Also, is there a difference in the shape of the output now? Like that 12x that not only went zombie, but also started projecting this evil smiley face... :crackup:

lasersbee said:

Yes... The XJ-A130 is the Casio 2000 Lumen Projector... where the
XJ-A140 is the Casio 2500 Lumen Projector..

BTW... thanks for the Graph.... I would also like to know which
model of projector that your LDs came from...

I tested one of each and there were differences that I saw..

Click to expand...

I see.

But the thing is, from Cassio's point of view it is MUCH cheaper to buy a bunch of same diodes and power them at a lower current in the lower rated projector.
So it would be unusual if they were different unless the A130 projector was made first, and by the time of the A140 better diodes became available.


Were the differences much bigger than the ones depicted in the PIV plot above?

I believe these diodes came from an A140, but i need to ask to be sure.


P.S. The two links you posted don't seem to work for me - i get thrown at the outer-most forum layout with both...

I asked about the projector model, and i discovered these are actually A130 projector diodes - leftovers from after people didn't want to buy A130's anymore in favor of A140's... I'll edit the first post accordingly.


The PIV data shows them to be exceptionally good, all their electro-optical characteristics fall into the range described in the Nichia datasheet that is guesstimated to belong to them. I kinda find it hard to imagine the A140's surpassing these, except for the occasional freak or weakling as found in every diode model.

I will probably just buy a couple of the A140 type and plot them for comparison, but i suspect they will fall very nicelly in the range - simply because it makes much more sense for Cassio to buy a bunch of one diode model and set them differently as per projector requirements, than buying two smaller bunches of two different diode models, simply for finantial reasons....

IgorT said:

I see.

But the thing is, from Cassio's point of view it is MUCH cheaper to buy a bunch of same diodes and power them at a lower current in the lower rated projector.
So it would be unusual if they were different unless the A130 projector was made first, and by the time of the A140 better diodes became available.


Were the differences much bigger than the ones depicted in the PIV plot above?

I believe these diodes came from an A140, but i need to ask to be sure.


P.S. The two links you posted don't seem to work for me - i get thrown at the outer-most forum layout with both...

Click to expand...

Sorry about the links above... I just fixed them...

I understand what you are saying about buying one type of
LD in larger quantities to satisfy the A130 and A140 projector
needs (for quantity pricing) and just turning down the LD power
on the A130s..

It has also been suggested that perhaps there is only one type
of 445nm LD being used and they are binning them to use the
higher output/efficiency ones in the A140 projectors... IDK


Jerry

lasersbee said:

Sorry about the links above... I just fixed them...

I understand what you are saying about buying one type of
LD in larger quantities to satisfy the A130 and A140 projector
needs (for quantity pricing) and just turning down the LD power
on the A130s..

It has also been suggested that perhaps there is only one type
of 445nm LD being used and they are binning them to use the
higher output/efficiency ones in the A140 projectors... IDK


Jerry

Click to expand...

Yes it's possible they cherry pick them for efficiency but to be honest such an activity is A LOT of work (i know from personal experience as i've been doing it since PHR times)...


Otherwise the data you posted there is really extensive, enough so i could plot your two diodes.

Would it be OK if i plot them and add them to my multiplot above?

I believe it would help see if there is any difference at all besides the usual efficiency variation.

IgorT said:

Yes it's possible they cherry pick them for efficiency but to be honest such an activity is A LOT of work (i know from personal experience as i've been doing it since PHR times)...

Click to expand...

Binning isn't a lot of work for diode manufacturers, since they already have to test every single diode they make anyway. The test labs are often as big as, or bigger than, the manufacturing facilities.

Since they have to test every single diode anyway, usually multiple times, it's very easy for them to bin diodes along the way, maybe even multiple times. It's standard procedure for semiconductor devices of all kinds to be binned.

IgorT said:

Yes it's possible they cherry pick them for efficiency but to be honest such an activity is A LOT of work (i know from personal experience as i've been doing it since PHR times)...


Otherwise the data you posted there is really extensive, enough so i could plot your two diodes.

Would it be OK if i plot them and add them to my multiplot above?

I believe it would help see if there is any difference at all besides the usual efficiency variation.

Click to expand...

I was reffering to the LD manufacturer perhaps Binning the LDs not
Casio...

The data on my Posts are Public... of course you can use them...:beer:

BTW... if you read after my Data posts... a Member posted
some graphs of my data...


Jerry

IgorT... Welcome back!

pullbangdead said:

Binning isn't a lot of work for diode manufacturers, since they already have to test every single diode they make anyway. The test labs are often as big as, or bigger than, the manufacturing facilities.

Since they have to test every single diode anyway, usually multiple times, it's very easy for them to bin diodes along the way, maybe even multiple times. It's standard procedure for semiconductor devices of all kinds to be binned.

Click to expand...

I realize they (diode manufacturers) have to test every single diode to make sure it fits into the ranges described by their datasheets.

But i haven't thought of the possibility of them actually offering these diodes separately by their efficiency... I was thinking more along the lines of Cassio (or the end user) testing them by themselves... If they can actually simply ask the manufacturer for a specific efficiency range it's a whole other story.


Otherwise as a number of my own tests would seem to indicate, higher efficiency diodes not only deliver more power at the same current, but also live longer while doing so.

I could imagine the diode manufacturers selling the higher efficiency diodes as a higher power model based on that too...


However when it comes to the difference between A130's and A140, i still think they are likelly one and the same.
I'm just now finishing including Laserbee's data into my plots and the result is quite interesting. Posting in a bit...

lasersbee said:

I was reffering to the LD manufacturer perhaps Binning the LDs not
Casio...

The data on my Posts are Public... of course you can use them...:beer:

BTW... if you read after my Data posts... a Member posted
some graphs of my data...

Click to expand...

Yes, i was thinking about the end user (Cassio) testing their diodes individually. If they can simply ask for higher efficiency when buying that's something else entirelly. Haven't thought about that option.


I haven't stumbled upon the graphs but the idea is combining the two of yours into the three i made to see if they form a neat group or if any stand out.

I'll be done in a sec.

Ok, here's what happens if i include your data in the multi-plot:




I used pinkish colors to make your two diodes stand out.

As you can see, your "A140 #1" has almost the same efficiency as my "A130 #02"...
Your "A130 #4" does indeed have a MUCH lower efficiency and stands appart from the group. If they are all one and the same diode model, i'd classify it as a "weakling".

But since the A140 is so close to the main A130 group, i see that as further evidence (but not necessarily proof) they are indeed one and the same model.
Unfortunatelly 5 diodes is still too small a group to have a statistical significance, but for now i see no reason to think they are different.


I do however understand why you felt they were of a different model - the difference between the two of yours really is significant!

Another weird thing is, that your A130 has an absurdly high Vf in comparison to all the others!
That means that in terms of absolute efficiency it would stand even further appart from the rest of the group (much lower than now)!
Unfortunatelly i can't easily plot this one into an efficiency plot as the Vf data is not available for all steps (unless i make up the inbetween numbers so they fit into a neat curve). Maybe tomorrow...


P.S. IF my three diodes were actually A140's i would agree the A130's might be a different diode model or at least lower efficiency diodes of the same type. But so far every single one of them (even the weakest) still falls into the ranges defined by that Nichia datasheet, which specifies their slope efficiency as 0.8-1.8W/A

P.P.S. The high end of the specified slope efficiency gives me hope i will discover a freak if i go through enough of them... Anything close to 1.8W per 1A would be absolutelly amazing! :evil:

jmgallego said:

IgorT... Welcome back!

Click to expand...

Seconded... welcome to the 445nm party, IgorT!

Incidentally, Igor, do you think that you can still say "no duty cycle" about these beasts in your hosts? I'm planning a build of my own in a substantially larger host (a PGL-III-C), and may boost my target power if 1W is still reasonable in builds the size of yours.

Canuke said:

Seconded... welcome to the 445nm party, IgorT!

Incidentally, Igor, do you think that you can still say "no duty cycle" about these beasts in your hosts? I'm planning a build of my own in a substantially larger host (a PGL-III-C), and may boost my target power if 1W is still reasonable in builds the size of yours.

Click to expand...

Thanks! :yh:


That is a good question. I haven't finished testing it just yet, i also plan to put the module into an aluminum host of the same model to compare thermal dissipation into air - the alu one is black so it might even radiate more heat away instead of mainly cooling from convection (mirror polished stainless steel can't radiate as much).


As for the duty-cycle question, i'll put it like that. As long as the head of the laser is not too hot to touch i am not worried. In the projector these diodes operated at 70°C if i'm not mistaken.

And so far the laser easily took any kind of my (ab)use (and i used it quite extensivelly for long periods of constant ON).

It did get very hot quite fast - much hotter than anything else i've ever built obviously. But if i touched the head with my lips it still wasn't too hot to touch. (unfortunatelly i can't use my IR thermometer on stainless steel...)


What i'm trying to say is that i haven't (yet?) imposed any limits on it's duty cycle and still haven't noticed a problem.
But the final answer will come from a Po over Time graph and a comparison between the SS and alu hosts powering the same module.


For now i am pretty certain of the following:
- My previous builds including a couple of 12x's i made for members of the forum managed to radiate the heat away fast enough that they could even be left ON for half an hour if not even permanently without damage to the diode.
- While the Stainless Steel 445nm did not frighten me with heat yet i do not believe it could be left ON for half an hour straight. It does create 3.3W of heat (4.5W in minus 1.2W out)! But as long as i can hold it in my hands without getting burned i am not worrried about the diode.


I also put thermal paste where the module is squeezed between the lip of the head and the middle part of the host body, to lead as much heat down into the handle as possible.

Like i said, for now i used it without any restrictions and haven't noticed a problem even tho it does get very hot (in human terms).


The only real problem i am having now is the extremelly rapid battery drainage. The voltage quickly falls below the minimum required for the driver to keep the current steady. But even this doesn't bother me too much as i always have a couple of batteries in my charger.


Obviously a larger 18650 battery host will keep it cooler with a larger heatsink, but the main benefit there is that you don't have to constantly replace the battery.


Basically i believe it is safe to use a 14500 host, as long as you don't run it past the point where the head is too hot to hold, but the power stability will be lower than in a 18650 host for obvious reasons.


I'll definitelly post my findings in a separate thread, once i do more extensive tests.


Otherwise i am getting 1.190W out of it at 1A, not just 1W. You can definitelly go further than that in a larger host as long as the driver can deliver that current (i have plans of using two drivers in parallel in an 18650 host if it turns out that 1.5A is still "safe")...





VERY IMPORTANT FOR ACCURATE CURRENT SETTINGS ON THE MicroBoost DRIVER:
If using a MicroBoost driver close to it's limits one HAS to do it with the dummy's Vf AS CLOSE AS POSSIBLE to the actual Vf of the diode (or a slightly lower dummy Vf just to be sure!) while powering the driver from a PSU set to 4.2V!

This is because at such HUGE currents, the driver can and will drop in current slightly as the battery voltage decreases (or as the voltage it has to put out increases)....

That means it's possible that if you set the driver to 1A on a dummy which has a Vf higher than your LD, the current through the LD could end up being higher than it was set to on the dummy! That's because you could have set the actual current higher than what you measure on the dummy, but this actual higher current setting would be easier for the driver to reach, once it was on the LD with it's lower Vf! It could end up pumping the lower Vf diode with the current you actually set it to - the "dropping out" (at higher dummy Vf) could cause that you unknowingly set the actual current considerably higher than you think, but you'd only see 1A for example, because that's all the driver could give out at the higher dummy forward voltage.

I do hope i explained this clearly enough. :yabbem: