LD blowing up fun

I had some fun yesterday. I blew up about 4 laser diodes. I still didn't figure out why, so I decided to make a post about it. Maybe one of you has a suggestion.

But first my 'project'. I'm building a 100mW 658nm laser module.
I use rohm RLD65PZB5 diode, a die4drive (http://www.die4laser.com/dvd-rec/Die4Drive.htm) LD driver and some collimating optics.

http://users.telenet.be/jeroen.derycke1/coll.JPG
Laser diode glued to the collimator.

The two wires coming from the mount are for TEC, which I did not figure out how to use for the moment.

So, what happens: I connect the LD driver to the didoe, I power up and eveything works... for some minutes. Then whoops, an expensive laser pointer diode. Putting out a milliwatt or so.

I drive the diode at 146mA (well below the max 200mA).
I 'm wired to gound the whole time.
I'm NOT cooling the thing down, but when I put my finger on the diode when it just died, there is no apperent heat .

So I just don't know why they die.

Before killing some others, I need some idea's because it's getting a bit expensive... ($36 on diodes killed so far).

By the way, should I seperatly ground the collimator assembly and the eventual housing of the module?
Should I isolate the LD from the assembly?

With the size of the diode I expect the heat will disipate within an extremely short time after the diode blew.
I think you need to keep it cool all the time so it doesnt blow.

Jim

Quote:

---

Originally Posted by JimBo

With the size of the diode I expect the heat will disipate within an extremely short time after the diode blew.
I think you need to keep it cool all the time so it doesnt blow.

Jim

---

Any suggestions how? I've been thinking on drilling a 5.6mm hole in a cooling plate and fix it with that white stuff they use for thermically attach a processor to its cooling device. Thus using passive cooling.

Hi Jeejeedr

Fun... :0 Hehe :)

I think I hold that record!!!

Your mount should have worked fairly good as far as heatsinking is concerned.
If you do use some silicone grease, be sure it is a tiny amount..
too much can hurt it as well.

Sounds like static got to them to me.
Or, They were spiked by your supply.

Be sure you and the devices are grounded during handling..
A good work area may help as well..I am a believer of grounded Anti-static mats .
Also, Keep the diode wires shorted until mounted and you are ready for connections.

I have also tried these Rohm diodes out..at one time they were going to be used in a product, But the failure rate was too high.

Oh, And Robins driver, Be carefull with that..It can still overcurrent the diode
if you put more than 5 volts on the mod input.
Also, That circuit needs just a bit more work as it may spike the diode by overshoot in the pulses..
The diode is floating above ground...be careful not to ground the mount...it will draw full current unless your diode is isolated.

Good luck with your project.

Quote:

---

Originally Posted by marconi

Hi Jeejeedr

Fun... :0 Hehe :)

I think I hold that record!!!

---

How many? :lol:

Quote:

---

Originally Posted by marconi

Oh, And Robins driver, Be carefull with that..It can still overcurrent the diode
if you put more than 5 volts on the mod input.

---

I had 5.15V, maybe a bit too high. :?
I'll put an 7805 between my power supply and the driver.

Quote:

---

Originally Posted by marconi

Also, That circuit needs just a bit more work as it may spike the diode by overshoot in the pulses..

---

I have an updated version. He made some modifications lately.

Quote:

---

Originally Posted by marconi

The diode is floating above ground...be careful not to ground the mount...it will draw full current unless your diode is isolated.

---

Ah, I think there could be an issue here.
I think the diode was touching the housing and I touched the housing, and I was grounded... 8)

Quote:

---

Originally Posted by marconi

Good luck with your project.

---

Thanx, I post things when I have some better results.

Oh,
I also noticed that there is a peltier in my housing.
a 4mm hole in the middle and about 10mmx14.5 (14.7?)mm

Anyone an idea how to find the voltage and maximum curernt to drive this thing, or how to identify the manufacturer (nothing printed on the damn thing).
?

Hehe..
I could buy a nice used car for all the diodes I've done-in

Yep...Ive seen the new board...I think its still is missing a cap.
Maybe it just might have been the test circuit I made up to try it
but I had to add a .22 mf cap across the diode for starters.

Regulated supply is a must but I was talking about the modulation input..
Since most lasershow boards only put out up to 4.99volts
this usually isnt a problem, but on the bench running a generator or
just a supply to run it analog or TTL ..that is the problem.
To get around the problem would involve using another op-amp
to isolate the input overvoltage to just the op-amps limits are..(the rail).

Sounds like it may not be the case of the mount touching but maybe
static..its hard to say without seeing your hookup arrangement.
You might want to add a high value resistor like 22k up to 1meg ohm..
to the base of the mount to the board ground.
That will eliminate any leakage or hand/static effects ..mostly.
I dont remember that circuit going into oscillation but that op-amp is a very high freq one..so its possible.
wire dressing and use of sheilded wire might help too...altho its rare to
pick-up rf or static that way.

Best way to test the peltier is on a heatsink..with a small plate on top of the device..Proper polarity will be the small plate will get cold.
Use a digital thermometer and watch for temp difference ratio.
if you get at least a 30deg c difference I think its usable..
Dont go too far as to freeze the diode or get condensation on the window...hi.
Sounds like the one you have may run at a few volts up to an amp.
Start slow with a volt or two and see what happens..

Seeing that the peltier is isolating the diode from the mount..
I think I would get the TEC operational first so to cool the diode.

Sounds like your going to have some fun learning :)

Nice mount, where did you get that from?

The Peltier can easily take 1A or more, and if you limit the current to that, not much can happen (except if you interrupt the feedback loop). However be careful not to cool too much while driving the diode at the limit, because the max allowable current decreases with lower temperatures.
On the other hand, unless you really need a highly stable laser output, there is not much reason to TEC stabilize the diode anyway, it's not worth the trouble.

I am developing a TEC stabilized laser module with the same diode (no surprise in view of that ebay offer ;-) for holography applications, and so far just one died due to a surge in the power line. I regularly drive them to 175mA and they put out >110mW after collimator and so far none died while running it hard. 145mA should be completely safe and I guess there is indeed something wrong with your driver, there may be spikes or perhaps oscillations if it is not properly designed; a scope may come handy here.

I'll have another try later on this week with these changes:
I'm going to hook up a scope to the driver circuit to see how it reacts in time.
I'm goind to stabilise the input power before feeding it to the driver.
I'll isolate the diode from the collimator housing. (I'll mount it on the peltier)
I'm goind to add that resistor as suggested by marconi.
I'll try to hook-up the peltier. However, using a thermystor and stabilise temp at about 25°C will be a challenge.

About the case, yes a nice thing. Collimates well too. I ripped it from an industrial barcode scanner. I make baggage sortation software for a living so we use a lot of SICK and other barcode scanners. (too bad they're only 10mW)

some more pics of the collimator, peltier and barcode scanner donor.


http://users.telenet.be/jeroen.derycke1/collhousing.JPG
Collimator

http://users.telenet.be/jeroen.derycke1/peltier.JPG
peltier

http://users.telenet.be/jeroen.derycke1/scanner.JPG
complete barcode donor

Hi, regarding multiple losses of diodes: lucerne (lucernelights.com) from the laserfreak forum (link to thread) buit a driver shutting off the diode power in case you short it out. No more killing of innocent diodes! :-)

Send him an email if you are interested in further details (he certainly speaks english).

Cheers
Christoph

Today I killed a diode while trying to gauge the focal length of an AR coated lens I found, one of nine I never used due to longer FL and too wide a beam...

The lens was 20 feet downrange on a very well collimated red beam 2mm wide at that point. I was (foolishly) using a shiny cardboard surface I picked up as target without thinking first. When the focal point was found, at that instant enough light returned back along the 20 foot path to destroy the cavity on the other side of the room.

I'm still trying to work out whether this is cool, or depressing...

Is there such a thing as an 8th wave plate? I have this vague idea that if I could rotate the polar plane by 45°, the return journey adding another 45° might make sure that reflected light can't do this. Is there a cheap, efficient way to make it so retroreflection is not fatal?

To rule out spikes try running on batteries. Feed the battery power into your driver board and see if this improves the kill rate. Traditonally LDs were anode to the case so if you did ground it the worst that would happen would be that the PSU would be shorted out - low risk to the diode. The trend with DVDRW diodes at least is to make the case the cathode. This means you must be more careful as most drivers (all that ive seen) still control the cathode end of the diode. So if you do earth the case you will kill the diode in a blink. I think this also makes it a little more likely to take a beating from static during operation.

good luck and keep us posted

Rob

My problem had nothing to do with static, or spikes. I solved those problems. I just posted about a very different vulnerability, to retroreflection even under remote circumstances. The vulnerability is extreme, so much so that I'd have thought that a DVD surface might reflect enough light to cause failure, let alone using diodes in unintended ways.

Actually you hit the nail on the head, what you need is just a quarter-wave plate combined with a polarized splitter...

If you just want to play, here's a link to how to (literally) "roll" your own plate... (It's an excerpt from a December '77 Scientific American)

http://www.sas.org/E-Bulletin/2003-0...esAS/body.html

[added] I'm feeling insomniac so I drew a diagram...
http://photonlexicon.com/albums/album44/Backscatter.gif

Very cool.
Would this protect a bulk output if the beam was mixed polarization after beam-splitter recombination? I'd hate to think what it would cost if I had to do this to protect each diode independantly...

If the source beam was based on recombination with a splitter, it seems the light could get from one diode, to the facet of the other, then return via second mirror bounce to destroy the source cavity. A long shot (literally) but judging by what I've seen, it could happen.

Another thought: Might the returning light be unable to destroy the cavity even without the polarised beam splitter? As it's returning in the other plane, I'm wondering if it might prevent it from forming an extended cavity at all, which I think was the reason the facet was destroyed (the energy density rising even across the metres long space, enough to destroy the OC facet of the diode).

As far as I can tell, the only lasers diodes I've ever seen that needed protection were the reds->IR... bloth blue and green are pumped so they're just peachy... I suppose the nichias might need it, but I haven't really gotten the time to play with one... So practically speaking, you only need one... If your laser is low power enough, you might be able to get away with an absorbing polarizing lens which are much cheaper than a PBS.

This will not work for non-polarized beams... They DO make wonderful premade prisms which do the above and more, but they cost arm+leg...

In terms of damage, it's because the driver is current controlled and has no optical feedback. The problem is the density of light in the cavity. The return beam doesn't have to reinforce to cause damage because it's already powerful.

Actually it's not the PBS cost that bothers me now, so much as the HWP. >:)
(We seriously need a devil smiley imho..)

I looked up a half wave plate in Edmund Optics, and the cost is something no-one should view unless comfortably seated, and planning to spend someone else's money.

Most times in the past when I asked this question, the advice was always 'avoid back reflection'. Seems about right. My plan is to spend time and effort making sure that fast and easy diode replacement is possible, and to lay in enough stock for that.

Maybe a very stupid idea, but something that may improve the effect without the dramatic cost: either a beam splitter cube with only the faces anti reflection coated where the beams enter (so where it should leave in normal operation :) ) or a glass plate where the diode faced side is anti reflection coated and the other side is highly reflective.

This would reflect partly the reflected beams that would otherwise enter the cavity.

That's actually a good idea, but because light will usually travel back on the path it came on, if relected back, you can't exploit that idea without rotating the polar plane, and that's the difficult and expensive bit. A polarising beam splitter will do the bit you're thinking of, if that rotation is done first, by 45°. (The return trip turns it a further 45°, totalling 90, so the splitter deflects it instead of passing it). It's also sensitive to wavelength, and as diode wavelengths vary with heat, and between diodes, the task might be impossible, or at least inefficient.

One thing I've wondered about is whether reflection itself at certain angles can rotate a polar plane, but I don't think it can. I don't know enough about optics to be sure what effect reflection has on polarised light, other than varying degrees of reflection if the light is strongly polarised already.