Why is my LOC diode broken?

[Jimmymcjimthejim]

Edit: After a long and friendly conversation with BobBoyce over Laserchat, it seems the problem was identified. Key points are pasted from the chat log:
(12:23:35 AM) BobBoyce: the oc could just have indium splash on it causing the output malfunction
if you have a reflow station and a good magnifier, it is possible to use a small fine point and reseat the die while reflowing at about 110C
but it can even be seen in the unpowered photo the dark strip of the die side is exposed even when viewed from above
(12:15:07 AM) BobBoyce: if you heat a LD enough the indium will let loose
(12:16:09 AM) BobBoyce: it really does look like mechanical shock damage
(12:14:15 AM) BobBoyce: ok, the optical coating damage would be from too much drive current
(12:11:21 AM) BobBoyce: but the output end of the die appears to be damaged as well
(12:10:17 AM) BobBoyce: in the above die shot you can see the pump energy leaking out along the side of the die, looks as if the die has broken loose from mechanical shock
(12:08:27 AM) BobBoyce: looks to be output optical coating damage Jimmy


I'm making a build for a friend, and I have a LPC-815 LOC diode hooked up to a "Groove Driver" at 330mA. The build is all finished, but the diode isn't working properly (it never was working since it was first powered at 330mA, driver tested). Instead of a nice round projection, the laser diode is emitting a odd triangle projection, and is very dim. So I buy another LPC-815 sled, extract the diode, hook up the driver, and the same thing. What could possibly cause this? Is it my extracting procedure? Below are some macro pictures and a picture of this odd triangle projection. Note that instead of the light being emitted from the tip of the diode, it is emitted from the center.

Normal diode (below lasing threshold) on the left, odd diode on the right. Both LPC-815

This next picture is not mine, but a nice picture of a lasing LOC diode.

I know I'm an experienced member, but this is just something that baffles me. 2 bad diodes in a row? Not right..

Thanks,
Jimmymcjimthejim

 

[CompWiz]

Actually, it's not lasing in that picture, I had it tuned just below threshold. It's lasing in this one though: http://dl.getdropbox.com/u/1547250/LOC_07.JPG

Your diode is dead, though I'm not sure what would cause that. I'd guess ESD, because the ones I've killed with over current and back reflection looked like mine but dimmer.

 

[Jimmymcjimthejim]

It can't be ESD. I was wearing an Anti-Static wriststrap with a grouded soldering iron. Plus here in Florida the air is moist so there is not a lot of static electricity. The possibility is there but I doubt it. I think I have a diode that died of ESD (when I wasn't wearing my wrist strap), I'll see how it lases.

 

[TheMonk]

Did you use a power switch to power it up before completing the build?

I had killed many diodes testing them without a power switch.:banghead:

 

[millirad]

It's broken because it hasn't been replaced with a new one. At ~$10.00 a sled it isn't even worth it to figure out. I usually provide a dead LD with a new LD and newly built/tested DDL and go from there. Most of the time I'm back up and running. Good Luck friend!

 

[Gryphon]

That first pic looks like LOC's that i have killed in the past but that was from driving them at over 500mA, you on the other hand are driving them well below the trusted 420mA for this diode. Strange..... try a different driver or maybe alter the way you harvest the diode or install the diode in the module to try and reduce any physical stress on the die. I doubt you could get two dead diodes in a row for no reason :thinking:

 

[Benm]

I guess the only way to really find out, is to see when the problem occurs. Next time you get a sled, wire the diode to a driver before harvesting, and see if it comes in working condition. Then snap off the metal diode holder, and check again. The metal will still provide heatsinking for a short test.

Tricky part is extracting the diode from the holder - i'm not sure how you do that, but i think its the most risky step.

Testing the bare diode is a bit tricky too since it lacks any heatsinking. I'd recommed holding it gently whith some pliers from the sides, and testing it with low, but above treshold current (say 100 mA).

I suppose doing these few simple tests is a lot easier than speculating about what goes wrong, and will provide proof in the process.

 

[Mohrenberg]

what is your technique for harvesting since you think that may be it?

i use a triangular file and file down the thinnest part of the the heatsink on both sides, then i grip it with pliers and use a.

when it comes to the driver, i've never used a groove driver, as rkcstr is about the same price and smaller. so i won't be any help in that area.

 

[Jimmymcjimthejim]

Did you use a power switch to power it up before completing the build?

I had killed many diodes testing them without a power switch.:banghead:

Yes, I use a switch, and take out the batteries before I solder everything in. I also short the capacitor.

It's broken because it hasn't been replaced with a new one. At ~$10.00 a sled it isn't even worth it to figure out. I usually provide a dead LD with a new LD and newly built/tested DDL and go from there. Most of the time I'm back up and running. Good Luck friend!

I want to figure out so I can prevent the problem in the future.

That first pic looks like LOC's that i have killed in the past but that was from driving them at over 500mA, you on the other hand are driving them well below the trusted 420mA for this diode. Strange..... try a different driver or maybe alter the way you harvest the diode or install the diode in the module to try and reduce any physical stress on the die. I doubt you could get two dead diodes in a row for no reason :thinking:

When the diode was first powered up it was inside an Aixiz module in a larger aluminum heatsink, with heatsink paste. I doubt that it was my fault that the diodes were broken because I have experience, and I use many precautions from situations that have happened in the past.

I guess the only way to really find out, is to see when the problem occurs. Next time you get a sled, wire the diode to a driver before harvesting, and see if it comes in working condition. Then snap off the metal diode holder, and check again. The metal will still provide heatsinking for a short test.

Tricky part is extracting the diode from the holder - i'm not sure how you do that, but i think its the most risky step.

Testing the bare diode is a bit tricky too since it lacks any heatsinking. I'd recommed holding it gently whith some pliers from the sides, and testing it with low, but above treshold current (say 100 mA).

I suppose doing these few simple tests is a lot easier than speculating about what goes wrong, and will provide proof in the process.

Thanks for the advice Benm, I will definitely do that. The only reason why the diode was bare when powered up in the pictures.. was so I could take the pictures.

what is your technique for harvesting since you think that may be it?

i use a triangular file and file down the thinnest part of the the heatsink on both sides, then i grip it with pliers and use a.

when it comes to the driver, i've never used a groove driver, as rkcstr is about the same price and smaller. so i won't be any help in that area.

My technique for harvesting is the two pliers method, but instead of making the heatsink break, I bend it back and forth (one or two times), and wiggle the diode out. This way there are no small metal chunks or anything. But as BobBoyce suggested, part of the heatsink could be touching part of the die.
The driver I use uses the same schematic as the Groove driver, only with SMD parts. So not only is the voltage drop lower, the size of the driver is smaller than the RKCSTR. (Mine is 15mm X 7mm, and you can directly solder the output to the pins of the diode, if wanted)

Thanks for all the advice guys, hopefully this will not happen with the next one. :yh:

 

[Benm]

I couldn't use the twist/wiggle technique on the LPC's, the metal just doesnt bend, rather pieces will snap off whereever gripping it. I use a metal saw to cut right to the diode on one of the thin sides, and then use two pliers to break open the metal around the diode. Before sawing, i very gently put a piece of electric tape over the business end of the open can - to prevent saw dust/filings getting in there.

It doesnt really matter though, if you can test that the diode is working before harvest but dead after, you've at least found the problem.

Another good precaution could be using a vaccuum cleaner instead of tape or anything else protective - just sucks away any metal filings before they go anywhere else. Obviously the whole thing should sit securely in a vice not to get sucked up itself.

 

[Jimmymcjimthejim]

Benm, the 2 pliers technique for the LPC diodes is a little more difficult than on the PHR sleds. You need to grip firmly and twist with a little more force. Once the heatsink bends even a tiny bit, the diode will wiggle right out.

1200 posts!

 

[Mohrenberg]

Jimmy, i got a LPC sled in the mail today, harvested it installed it in a module and power it up. instant LED, not even a flash of red....
and the "LED" instead of shining out the front, seems to be shining out the back of the crystal and onto the diode case.....
i'm going to take a shot in the dark that you bought your from HTD also? assuming that you did, i'm wondering if maybe his prices are cheaper than drew's because these are QC rejects?

 

[Jimmymcjimthejim]

Yes, I bought from HTD. I have the same problem as you, the diode shines out the back, instant LED. This is exactly as I thought was going on, I knew it was unlikely to have two broken LPC diodes in a row, with the exact same problem. Even BobBoyce's analysis surprised me, as I take many precautions when extracting these diodes.. :undecided:

 

[Mohrenberg]

yeah i was very suprised when all i saw was an LED, that almost NEVER happens with me...i'm just going to assume we did nothing wrong

 

[sk8er4514]

The lpc815 sleds I buy are new and appear to be new and never used / tested as QC rejects. they are the same lpc815 sleds you will get from anyone. as far as the strange triangle projection I'm not sure. I've only seen circle lasing, or led when I mess up harvesting as it is very easy to damage the open can die.

I suggest not bending the heatsink back and forth as that sounds like a sure way to cause damage.

the open cans are delicate creatures, you must be very careful.

good luck,
thanks,
Kendall

 

[Jimmymcjimthejim]

I suggest not bending the heatsink back and forth as that sounds like a sure way to cause damage.

In this picture, I used the method as I described. You can hardly tell the heatsink was bent at all.

Kendall, what do you think the problem is? This is too coincidential to have three diodes from the same source have the same problem.
Jimmymcjimthejim