From: dstamp@watserv1.waterloo.edu (Dave Stampe-Psy+Eng) Subject: Re: Integrated laser arrays for Eyephones Date: Thu, 31 Oct 1991 05:35:25 GMT Message-ID: <1991Oct31.053525.20923@watserv1.waterloo.edu> Organization: University of Waterloo 70353.3056@CompuServe.COM (Christopher Fry) writes: >I wrote the original report on Microlasers. >>Leaving aside the fringing effects of coherent light ... >Please don't leave them aside, I'm interested in the problem. If we've got >a very narrow laser going directly into the eye and it hits a few rods/cones >which are not hit by another laser, or, for now lets say, any other light, >what's the problem? Basically, it's that coherent light forms interference patterns when scattered by defects in the eye's optics. With incoherent light this just creates blur, but with laser light you get very visible rays and sparkles. Ever looked at the laser line in a supermarket checkout scanner? >>the technology is overkill for eyephones. >Great. We've finally got a technology that's overkill. Generally its easier >to scale down something that's overkill than something that's underkill like >LCD's and CRTs. I don't see that, myself. It's not proven that laser arrays will be any better for eyephones, and they CERTAINLY will be more costly and take a longer time to get to market. >>- No yield was mentioned. I suspect that the yield per laser is pretty >>low (99.95% is _low_) which means the yield for a 100x100 array is nearly >>zero. > >By "yield" I presume you're using the traditional "everything's gotta work" >definition that we need for conventional chips. The more pixels we've got on a >display, the less percentage of them we need to work in order to have a good >picture. Glasses and car windshields work pretty well when dirty. I'd prefer >to have all the pixels there, but 100% simply aren't necessary. > >Also these are regular arrays. Small arrays will naturally have higher >100%-good devices. If we can make "mosaics" of these smaller chips to build up >our whole display, then having to create any one large chip with all of its >lasers working becomes unnecessary. Making such mosaics may be hard to do >without line gaps in the display, though I saw one article in the IEEE tech >article that discussed cleving such arrays of lasers precisely between rows of >lasers using some kind of etching. Didn't work with LCD's, CRT's or plasma panels when they were small and expensive, so it probably won't work with laser arrays. >>Here's some caveats: >>- Reliability isn't what it seems. If EACH laser in a 100x100 array has a >>MTBF of 100K hours, one laser will die every 10 hours. >The specs I've seen indicate that lifetime for one of these microlasers is >about 10K hours. I presume that when you initially run the array, not many >fail, but when you start getting close to the average lifetime, its time to >throw out the whole array. Depends on the rate of failure curve. Some devices are like that: don't know about microlasers. >>To get colors (most semiconductor lasers work in red or infrared wavelengths) >Yep, as I mentioned, making a visible-frequency microlaser still can't be >done. >>you could use a frequency-doubler crystal deposited on the chip. >Can you tell me more about frequency-doubler crystals? This might fix the >problem, though perhaps they're inefficient and just compound our heat >problems elucidated above These are nonlinear optical devices that are basically tiny lasers themselves. They need a high level of energy (another laser is perfect) to translate 2 or 3 low-frequency photons into 1 high-frequency photon. Not sure about their efficiency, though. >>SO: Why not use LEDs? They're cheap, reliable etc. I guess you'd have to >>deposit different chemicals for the different colors (GaAs for red, >>GaAsP for green, and InGaAlP for blue) but it's doable. >I don't know what the beam-spread on the micro-lasers is, but if its not much, >then all of the light output can be directed to a specific spot on the retina. >Won't this help us save power AND lenses? Well... Power isn't really the problem here. The power of light at the retina is measured in femtowatts (10e-12). The control thing is the only possible reason for using lasers, and beamspread actually helps make your optics a bit easier to design. If you want to cut down beasperad, use a colimmator plate. Lasers would be nice for use with holographic optics, but modern thick-film holograms work pretty well with incoherent mono- chromatic light. And high-density Fresnels might beat them both in the cost area. >Demonstrated frequency performance for LED's is clearly better. >Are they more efficient? >Can you make them as small as these microlasers? >Can large arrays of them be fabricated with known technologies like >IC's use? [I'd be interested in seeing articles on Mega-LED chips.] Laser diodes are just large, VERY carefully made LEDs. >If the answers to all of the above are "yes", they are a contender! > >>Or, if you want a proven solution, a group announced last year that they >>had built a flourescent TV on a chip! Just like that flourescent blue >>display on your VCR, it has a thin filament in front of the chip which gives >>off electrons. The chips switchs 24V to small dots of phosphor on the chip, >>which can glow with ANY color you please... > >Sounds interesting. Got a reference? > >>Still, at present, CRT or LCD displays are really the only options >>available at any reasonable cost. >True enough. The operative phrase there is "at present". One of the things >that excites me about the microlasers (and the above-mentioned "flourescent >TV" is that they are essentially chips. As you know, chips in quantity are >cheap and get cheaper per device regularly. The article was in a 1990 issue of SID (soc.information display) journal. Unfortunately I don't have the exact journal name (the Society puts out several journals) or the date. But it's a small white-and blue paperback size journal. Hope this helps. -------------------------------------------------------------------------- | My life is Hardware, | | | my destiny is Software, | Dave Stampe | | my CPU is Wetware... | | | Anybody got a SDB I can borrow? | dstamp@watserv1.uwaterloo.ca | __________________________________________________________________________