Some experiments with the blue LD's

[Joe Farina]

I was surprised how many sellers are now offering these diodes on eBay, at least 5. Prices seem to be dropping, the least expensive is $54.99 (free shipping). To see these, type in "445nm" in the search field on eBay.

The reasons are pretty straightforward. In the +/- $700 beamer the 24 diodes are in a heatsink block that can be removed in less than 30 seconds if you've done it once. You don't even have to open the beamer itself. The diodes are connected to each other by a ribbon pcb, takes 30 seconds to remove. Diodes are not glued or bolted down, just stuck in the heatsink with heatsink compound. If you don't mind the hassle of packing, dealing with things going wrong in the mail etc. and don't mind the chance of your diodes ending up in a 14 year old's 1W handheld 'pointer'... it's actually a pretty good business.

I see. So, if someone gets $60 per diode, 24 times 60 is $1,440. Subtracting the cost of $700 yields $740 profit, or about $31 per diode. Not too bad for going through the hassle of selling and mailing. Maybe those projectors have some other parts too.

[dave battin]

Thats correct Joe! It was better when they were selling for $75-$100 each! I did two projectors and quickly made my money back and still have an extra projector. there are various lenses, front surface mirrors, dicloric beamsplitters,and a red LED Phat light in each of these projectors.......... aside from the DLP module and PCboard!

by the way i did a test last night to see how deep a transmission hologram i could make(w/one beam), and i got an image at 20 inchs deep!

[Joe Farina]

Wow, 20 inches is really impressive Dave. I've just bought a diode to get my feet wet, maybe I can work on some of the less demanding problems like collimation Thanks a lot for the planned tests with DCG.

[wler]

Wow, 20 inches is really impressive Dave. I've just bought a diode to get my feet wet, maybe I can work on some of the less demanding problems like collimation Thanks a lot for the planned tests with DCG.

Actually not so much single mode operation but beam quality is more the practical problem with those diodes, especially if one desires an even illumination. It is almost impossible to get the beam through a spatial filter without great losses, for example. A way to ameliorate this would perhaps be to split the beam at a dark band with a knife-edge mirror and so use the otherwise wasted part as object beam for example.

So in order to make good use of the diode, some things need to be specifically tailored for it. The very least is a pair of anamorphic prisms or suitable cylinder lenses to get the beam in a better shape to begin with.

[Joe Farina]

Wow, 20 inches is really impressive Dave. I've just bought a diode to get my feet wet, maybe I can work on some of the less demanding problems like collimation Thanks a lot for the planned tests with DCG.

Actually not so much single mode operation but beam quality is more the practical problem with those diodes, especially if one desires an even illumination. It is almost impossible to get the beam through a spatial filter without great losses, for example. A way to ameliorate this would perhaps be to split the beam at a dark band with a knife-edge mirror and so use the otherwise wasted part as object beam for example.

So in order to make good use of the diode, some things need to be specifically tailored for it. The very least is a pair of anamorphic prisms or suitable cylinder lenses to get the beam in a better shape to begin with.

Yes, it will need to be worked on. What I should have said was less demanding on "me," and not in general

Hopefully some others will do tests on these diodes, and post the results here, so we can all benefit. New approaches and experiences from a variety of people might be crucial for success.

[dave battin]

yes it seems the hard part will be to get a good beam profile, my uncoated asphere assemblies do a pretty god job of focusing a spot, this is the area to attempt a spatial filtering. I suspect these diodes can be easily tricked into producing an infinite coherent lenght laser, as i can easily "wink" this laser diode (on and off) when doing an interfrometer test (via retro reflections)...............................and that is a favorable sign

this a wall shot at 24 feet from the laser.


when all is said and done ill only be using the center portion of the beam profile anyway, what is that like 20%..... im still happy

[Thieu]

@Ahmet: The video you asked for is up at my blog: http://piepklein.blogspot.com/2010/07/s ... tem00.html
This is with the higher amount of feedback though.

About the uncapping, I've been practicing it in the watchmakers lathe on some old printer diodes because I wanted to experiment with installing some kind of FAC on the blue ones. I use a very small slitsaw so the diode itself doesn't have to spin fast. It works pretty well:

uncapped diode

uncapped diode 2.JPG (16.74 KiB) Viewed 116 times

I'm not sure how much dust and other environmental influences will be a problem. There will be at least some dust because of the sawing and my house isn't exactly a class 1 cleanroom It may be managable by recapping them or something after installing the FAC. I have no idea if gently clearing the die from dust with a high purity solvent is an option. Any comments are welcome.

[Joe Farina]

Sorry I don't know what FAC is. Regarding uncapping, my first reaction would be to avoid it when possible. About 10 years ago I uncapped some red diodes using the throlabs can-opener, and the diodes "seemed" rather unstable. At the time, I attributed this to reduced thermal stability because they were exposed to the open air, but this was just a guess. I would also be concerned about contaminating the face of the chip, or possibly damaging the diode if using a non-optimal method of removing the cap. I would suggest re-capping the diode (maybe with some other kind of cap or covering) after making your modifications. Then it wouldn't be exposed to the ambient air.

By the way, superb job with the microscope slide feedback. After re-reading your posts and blog, I'm starting to get the picture. I was wondering, were you able to measure output power at ~300mA?

[Thieu]

FAC = fast axis collimator

The beam coming out of a bare laser diode is nearly always non-symmetric. It's spreading much quicker in one direction than in the other. The direction in which it's spreading the fastest is (unsurprisingly) called the fast axis. It's beneficial to collimate the beam along this axis as soon as possible, because the fact that it's spreading so quickly means that optics to do so further along the path of the beam will have to be much bigger. So often the FAC is a tiny cylindrical lens very close to the output facet of the diode.

Unfortunately I don't have an optical power meter, so I can't tell the optical power. But the slope efficiency of these diodes is 1-2 mW / mA, so with a threshold of 220 mA, they'll give out 80-160 mW at 300 mA.

[Joe Farina]

Thanks for the explanation.

the slope efficiency of these diodes is 1-2 mW / mA, so with a threshold of 220 mA, they'll give out 80-160 mW at 300 mA.

Wow, that's a lot of power. It just remains to be seen how long it remains stable over time/exposure. The good thing about extra power is that exposure times can be kept down.

[Thieu]

Just found out that the ferric system at least responds to this wavelength as well. I was able to visibly bleach a small part of a ferric chloride sensitized plate in a couple of seconds with the laser very close to the plate. Don't know if that means a lot though, cause almost anything seems to bleach fast at this wavelength and intensity. But the fact that it does respond is hopeful I guess..

[a_k]

Thanks a lot for the video, Thieu. So it seems the feedback has a visible impact on the transverse modes. The interesting question (for me) is if there is a correlation between the SLM operation regions with feedback and the transverse modes. That would mean comparing the spectrograph results and a sychronous projection of the beam (like in the last video).

About dust and exposed die: I've been using several open can LDs in an Aixiz housing from which the collimator was frequently removed and in a semi-open setup for about a year and did not have dust landing on the front facet so far.

Martin and me have been testing ferric compounds quite extensively and there is a very high sensibility to 405nm (about 1mJ/cm2). For 445nm there should be considerable sensitivity as well.

Ahmet

[wler]

Re the new video: I believe, based on the literature on supermodes, that the "mode pattern" is due to coherent superposition of supemodes and so can be interpreted as interference pattern. The less modes are present, the better the contrast becomes; in a sense the diode acts as its own spectrum analyzer. I was pondering to make use of this for detecting single mode behavior: just read out a CCD array and maximize the AC component of the signal (highest fringe contrast). But it seems that simple noise detection suffices for the same purpose.

On related matters: especially an ECDL has its own built-in spectrum analyzer, ie the grating. When studying the beam profile after a few meters, one can see with the naked eye if extra modes are present, as they give some extra stripes, and one can clearly see when the modes lock or unlock against the grating.

[Thieu]

@Ahmet
I'm not sure if I already wrote that somewhere, but there is a correlation between what I see on the spectrograph and on a y-axis scan of the beam. I see some longitudinal modes on the spectrograph disappear when I place the entrance slit at the center of the y-axis in the beam. It would be nice to change the setup so that the beam would just fill up the slit in the vertical direction so that you'd get a two dimensional map with a spectrum for every spot in the vertical direction of the beam. With my current MacGyver'ish tape&glue setup that wouldn't be practical nor safe, but I may be able to set it up at work where I got access to a lot more optics and space.

Because space is so limited where I live now, peaceful coexistence of holographic plates and dinner plates is the only option in the kitchen at the moment so dichromate is a bad idea. That's why I'm particularly interested in FEG again. The last thing I remember is you were using ferric ammonium oxalate and developed with H2O2, is that where you both settled on or did you explore other salts/processing after that?

[wler]

It would be nice to change the setup so that the beam would just fill up the slit in the vertical direction so that you'd get a two dimensional map with a spectrum for every spot in the vertical direction of the beam.

Such a correlation between location and spectrum has been described in the very interesting article: "Supermodes in Broad Ridge (Al,In)GaN Laser Diodes" which seems to apply well to the blue diodes. (I can't post it due to copyright, but can send a copy for private purposes). There are 2d plots which show how the spectrum changes along the beam profile.

This applies to multimode operation, for true single mode operation the spectrum stays the same due to locking of the supermodes. An issue is of course whether it is really true SLM operation in a given situation, or whether other modes hide at different locations of the beam; ie, how one can identify true SLM operation if one samples the beam at a small region only (which is what most spectrum analyzers do to some extent).

I didnt find this problem serious, as SLM operation seems pretty robust once it is there, but some more time needs to be spent on this. In ECDL there is the phenomenon, though, that at higher powers the spectrum tends to have smaller side peaks, let's say at 10% level or less, and the composition of those side peaks seems to vary over the beam. The bad thing is that the spacing of those can be very narrow due to the external resonator, and it is difficult to resolve with a grating spectrometer; this calls for an scanning FP interferometer.