Laser output dropped substantially... diode or driver problem?

[echelon]

Hi guys,

I'm a programmer without much EE experience, and I've got a problem that I might need some major guidance to fix. I do have access to some EE friends, but they don't have much lasing experience.

Last year I bought a projector from djpsych on PL; it's a 30 kpps, 2x red, 1x blue, 1x TTL green projector. That same autumn, both of the red lasers dropped in power to the point where they cannot be seen unless drawing only red onto a small surface area. I'm wondering how complicated and expensive it'll be to repair my reds. How can I diagnose the source of the problem, and what would I order to replace the faulty components?

Also, during some shows the power of the TTL green will drop in half or by three quarters. When I draw different geometry it picks up again. Is this indicative of future problem about to crop up?

Thanks so much!

(If anyone is in the Atlanta area, I might be able to pay for some hands-on help.)

[buffo]

How much power were the reds making before they died? And do you know what wavelength they were?

If they were 660 nm long open-can reds, or the Opnext single-mode 638 nm reds, replacing them is a cinch. If they were the Mitsubishi 638 nm reds it's a little more tricky, since you'll probably need to re-align the secondary correction optics, but either way the job is still well within most people's ability.

As for cost, a bare red diode will run you ~ $12 for 660 nm, or ~$35 for the Opnext, or $60 for the Mitsubishi. Multiply that by the number of diodes in the projector and you have your price.

Can you post a picture of the internals of the projector? That would be helpful.

The issue with your green laser could be a number of different things, but to me it sounds like typical poor blanking performance of a DPSS green. This is rather common among all DPSS lasers, and especially the cheaper TTL-only models. You can try reducing your scan rate to see if the problem gets better. (If it does improve, that's a sign that the problem is related to the blanking performance of the laser itself.)

Unfortunately, there's not much you can do to help this issue, apart from replacing the green laser. (Though if you can run your beam shows at a lower scan rate, that may be good enough for you...)

Adam

[echelon]

I'll post some pictures tonight and get the info. There's a thread on here detailing the build, so I'll dig that up as well.

Thanks so much for the input!

[echelon]

Here's the thread detailing the projector I bought from djpsych last year:

http://www.photonlexicon.com/forums/...B-Build-Thread

Red: 2x 250mW diode extracted from LPC-826 (660nm)

I'll take some pictures later today/tonight of the reds, but I think the thread had great info on the build specs.

Where/whom should I order the replacement diodes from? What kind of equipment will I need to replace them? (I'm a total engineering newbie! Do I need a decent soldering iron? How do I wire it?)

As far as the DPSS green goes, I suppose I'll just put up with it for now.

Thanks so much buffo!

[The_Doctor]

I'll stay on the reds for my tuppence worth.. I pushed those and learned things on my own it took some time for others to accept.

When new, they usually have a clean operating region, but people wanted more power, and usually pushed beyond that. The new reguion of safe current ranges was moderately clean and well-behaved, but puts a diode at risk of early death. The result is almost always the same, catastrophic optical damage of the facet, but the causes may be various. For the damage, imagine that scene from James Bond's 'Die Another Day', where he has a widget in a ring that allows it to force resonance into a plate of glass strong enough to cause total fracture of the whole plane. Writ microscopically small.. With later greens and blue diodes maybe it's not so extreme, wasn't on a blue I tested, but with red is almost always is.

As you're not familiar with tech handling I guess you might want a long maintenance free life out of your diodes rather than push the envelope too much. So when setting power, look at the scattered light cast obliquely across a dark flat surface. That will show up the specularity, the weird brain-defying pattern in the light that moves as you move. If you power a diode too hard you will force the resonance to become dirty, like the singer who is trying to break a wineglass with a soprano voice. Push hard enough, and... So, watch the specularity. Dirty resonance means more harmonics, broader spectral content, and the sharp monochrome specularity visibly deteriorates, looks soft of pastel-shaded. Back off now! That would usually be the last, and only warning a diode will give you. If you're quick the diode will usually survive. But if you don't back off far it's still a matter of time.

Another thing that kills diodes is reflecting enough of their own light back at them, it extends the lasing 'cavity' and the facet now has too much power on both sides, doesn't like it, and gives up the ghost. Even if you run a diode at conservative levels and it withstands this, you might find that it will not after a few hundred hours runtime! Strong reflections in the optics will hasten this early death.

Diodes are cheap now, but buying mounts that make it really easy to replace them likely isn't. So that's maybe what you need to look at next.

EDIT: Tad bit more on that 'regions' thing... A diode for a clean sharp beam is single mode, meaning TEM00, meaning that across its beam section, up/down or across, there is one blob, a 'Gaussian' distribution that has most energy in the centre. If you push these diodes, they 'mode hop' into some more obviously complex pattern, and that's another early sign that they're not happy. People pushed beyond the first mode hop the way Chuck Yaeger pushed through Mach 1 (but with less personal risk) suspecting that things would be quiet again once they got through it. They are, but it's not a safe way to run a diode for a long life. Anything that destabilises it, and it crashes and burns so fast all you see is is no more bright light. Looking at diode power to determine safe region is no good. That's because after the first mode hop as you push current up, power drops off sharply before eventually 'recovering' and becoming even stronger as you continue increasing current. So if you really want a long life, you need two things: don't ever persist at the point where specularity gives way to ugly pastel-shaded light, and for best results don't even go beyond that first mode hop. That can make for a boring life maybe, but if you want to trust your diodes, you have to treat them well.

[Picasso]

You can pop those reds out and replace it with ones from here https://sites.google.com/site/dtrlpf/home/diodes would need to double check your mount size. But from the link of the build pics you got an easy to update machine. You should put a power meter on those red to see what voltage your putting out to them right now. Your driver might have gone bad and the current ones fine.

[The_Doctor]

You should put a power meter on those red to see what voltage your putting out to them right now. Your driver might have gone bad...

Current. Not voltage. A diode's Vf changes very little even with large changes of current, hence current control being vital. Inserting an ammeter into a diode circuit is the best measure, but is a dangerous move for a beginner because electrostatic discharge is likely during handling, and even if laser diodes were good before, there's a big chance they won't be after this test if it's done wrong. Even connecting the meter in ohmeter-mode by accident with voltage reversed can kill a diode.

A better move is to identify the diode driver, find its current sense resistor and read the voltage across that. Still got to be careful (and you'll need a MILLIvolt meter with good accuracy) but it is a lot safer than trying to read current directly. Reading voltage across a diode will tell you next to nothing unless you use a fast scope and see spikes or anything else that really shouldn't be there. Or perhaps if the driver has failed so badly that the voltage has collapsed almost entirely. But in that case there will be no light, and echelon did say he'd seen some. Got to be pretty much full Vf covered for that to happen. I'm betting the current is there too, and that diodes have died.

One way to look at diode output (not always helpful especially if diode outputs are mixed for single beam) is to project obliquely across a flat black surface like I described before. But not for specularity (though look at that too), but for dark bands across the bulk of what ought to look like a long bright ellipse.

[buffo]

OK - Your red diodes are long open-can 660 nm diodes. Easy enough to change (and quite affordable!), but they are also very sensitive to both static and back-reflection. So you'll want to handle them carefully, and avoid accidentally sending the beam directly back to the emitter face.

Here's a link to one of many E-bay listings for the diodes you need. (I don't think DTR sells these anymore - he doesn't have them listed on his site anyway, but they are available from many other sources.)

The Z-bolt heatsinks are easy enough to work with - you just loosen the allen screws on the side and the round 12 mm module will slide right out.

Getting the diode out of the round 12 mm module will be a bit more difficult, but since the diode is already dead you don't really care if you damage it further. Just remove the lens barrel from the front and the extension cover on the back, and then press the diode out from the front (output end) using a drift pin.

Installing the new diodes will need to be done using a grounding strap and an anti-static mat. First you will need to solder lead wires onto the anode and cathode pins for each diode. (It's a good idea to put some heat-shrink tubing over those connections once you finish soldering them.) Then twist the ends of the two wires together to keep the diode shorted. (This helps protect against static.) Now you can press the diode into the 12 mm round mount.

I find that using a C-clamp and the back extension cover of the 12 mm round module works best. Set the diode in place, then turn the extension cover around and center the small hole over the diode pins. Clamp the whole thing together until the diode is pressed firmly into the mount. (The diode should be flush, or even slightly recessed, from the back face of the mount.) Then re-install the forward lens barrel.

Now you can un-twist the lead wires and solder them to the driver output wires. (Make sure you get the polarity right!) Once that is done (and properly insulated with more heat-shrink tubing), you can insert the module back into the Z-bolt heatsink. At this point you will need to apply power to the diodes and rotate them in the heatsinks until you get the polarization correct so that each beam passes through the PBS cube. Then lock down the allen screws that hold the heatsinks together, and finally you can adjust the PBS cube slightly to correct any miss-alignment. (You may need to shim one of the heatsinks slightly if the near-field alignment is also off, but hopefully that won't be necessary.)

Not exactly the easiest repair, but it's not a killer job either. Just make sure you take proper anti-static precautions and you should be fine.

Adam