using multiple laser diodes to scan multiple lines with one mirror movement

[Psi]

First off, i'm pretty new to this. If i'm saying something totally obvious sorry.

I was thinking about the problem of using spinning mirrors instead of galvos. Where you get a full 180deg laser arc out of the rotating mirror and you might only want 10deg to project the image on something.

Has anyone heard of multiple lasers being used to project multiple lines with one mirror rotation. So you'd have say 10 laser diodes set up pointing at the mirror at different angles.

The goal being that as the mirror rotated through 180deg you would cycle through sending data to each of the 10 diodes. This would project 10 lines of your image with one mirror revolution. Thus if you had a 4 sided mirror one motor revolution would draw 40 lines. Obviously you'd have a 2nd motor scanning the Y at the same time.

Am i making any sense?

[loopee]

Hi PSI

Check out this link:

http://web.media.mit.edu/~stefanm/TinyProjector/index.html


Troy

[Psi]

Nice!

[mixedgas]

Get on youtube, there are videos of working versions, usually called "Laser clock". Due to limited utility without doing full blown laser video, I urge you to go the vector route.

Obviously the MIT guy never looked in a optics catalog.

Steve

[weartronics]

Has anyone heard of multiple lasers being used to project multiple lines with one mirror rotation. So you'd have say 10 laser diodes set up pointing at the mirror at different angles.

Typically, this is achieved using a multi-faceted mirror (instead of multiple diodes). This way, you can use 10X fewer diodes or have 10X more power for the same cost. You can find this type of motor and mirror assembly pre-built for you inside any laser printer. Here is one example being used in a laser projector:

http://www.photonlexicon.com/forums/...19&postcount=3

[Psi]

I thought ya ran into problems spinning the mirror fast enough to get fluid motion when doing scanning with fully rotating motors instead of galvos.

Hence multiple lasers would reduce the rotational speed required.

[weartronics]

I thought ya ran into problems spinning the mirror fast enough to get fluid motion when doing scanning with fully rotating motors instead of galvos. Hence multiple lasers would reduce the rotational speed required.

In the laser printer, these mirrors are spinning at several thousand RPM. In my projector, I had to *reduce* the rotational speed by about 100 times. I can't estimate what rotational speed might be required for your projector, because your design is not clear to me.

[buffo]

If you want to do raster laser graphics, then yeah, you need to spin the X mirror really fast, since even NTSC resolution needs ~15.7Khz on the horizontal. So if you have 10 faces on your mirror, to get to 15.7 Khz your rotational speed is 1570 revolutions per *second*, or 94,200 RPM!

In reality, you only need 10 to 12 degrees of optical scan angle, so most solutions go with 30 to 36 faces on the spinning mirror wheel, which yields a more manageable speed of between 26100 and 31,400 RPM. And remember, that's just for 640x480.

Doing HD video (1920 x 1080, lets say) in raster-scanned laser would be an application where multiple beams might make sense, if only to keep the speed of the spinning mirror for the horizontal scan down to something reasonable. But I don't know if anyone is doing this yet.

Adam

[Stoney3K]

If you want to do raster laser graphics, then yeah, you need to spin the X mirror really fast, since even NTSC resolution needs ~15.7Khz on the horizontal. So if you have 10 faces on your mirror, to get to 15.7 Khz your rotational speed is 1570 revolutions per *second*, or 94,200 RPM!

Using galvo's on the X axis wouldn't be a problem, would it? It would require a scan rate of about, er... exactly, 15734kpps. (Personally, I prefer the 15625Hz European flavor though...)

The Y galvo would be running at the same speed, but it would step through the field lines, where the X galvo would be running a continuous sawtooth. You might run into problems with the X galvo being too slow on large scan angles, but any 30K or more kit should be able to do it.

The RGB signals would be fed by, well, RGB image data . There might be some hitches here when you're running off frame or at the end of the scanlines, but nothing that can't be figured out.

When you're in an adventurous mood, you might be able to rip out a sync generator from an old TV. That would already put out the right waveforms, including blanking, the only difference would be a laser scanning instead of an electron beam. Hey, you could even recycle the tuner!

I'd be more concerned about the right spot diameter, though. It might not really look like real video but more like a large grid of points.

[buffo]

Using galvo's on the X axis wouldn't be a problem, would it? It would require a scan rate of about, er... exactly, 15734kpps. (Personally, I prefer the 15625Hz European flavor though...)

No, more like 188.4 Kpps, or 188400 pps. Remember that point speed is not the same thing as bandwidth. (Unless you want your picture to be just 1 point wide...)

30,000 pps yields roughly 2500 Hz on a small step signal. So you could probably scan at 3 to 4 degrees at 15,700 Hz if your scanners could hit 188Kpps. But that is an absolutely insane speed. (Also note that the "K" means thousand... Your example of 15734Kpps is 3 orders of magnitude too high, not that it wouldn't be great if we could ever move a mirror that fast.)

The Y galvo would be running at the same speed

No, it wouldn't. There's no need to run the Y galvo any faster than 60 Hz, and that's only if you wanted to interlace. Otherwise, you could run it at 30 Hz and do a progressive scan of an entire frame, rather than lacing two fields together. So you could use a cheap galvo for the Y, but you still need something exotic for the X. Either an insanely fast scanner (which hasn't been invented yet), or a spinning, multi-faceted mirror-wheel.

You might run into problems with the X galvo being too slow on large scan angles, but any 30K or more kit should be able to do it.

Um... No. Not even close. The *top* speed for scanners is 60K. That's still 3 times too slow. (The goal is 188K, remember.) If you're running 30K speeds, then you are 6 times too slow. Which means that your resolution will suffer.

This is why most raster images you see in laser shows with standard scanners are limited to around 80 lines of resolution. They just can't move the mirror any faster.

Now, if you have a spinning polygonal mirror, then you can run the speed *way* up, and true raster laser television at NTSC (or PAL) resolutions becomes possible. However, it's tricky to get the timing synchronized between the horizontal scan and the vertical scan. (You've got to have very accurate position feedback on the spinning wheel.) Also, you have to use all direct modulated lasers or a PCAOM, since otherwise you'll never get the color modulation to respond fast enough.

The spinning wheel trick has been done on a large scale by several of the big laser show companies to demonstrate true laser television, but as for the hobyist/enthusiast crowd, I'm not aware of any successful projects. Of course, the MEMS scanners in this little device might change everything, if they can be scaled up to fit a larger diameter beam and if the cost stays down... (See discussion here.)

Adam