PDA

View Full Version : How to modulate a laser diode at high speed ?



KoD601
09-23-2010, 06:41
Hi,

I'm new here and looking for help to achieve my current project :)

My goal is to convert a video signal from a PAL/NTSC source into a proper video-projected image via :

- a FPGA
- 2 x 8 bits DAC
- a 20kpps scanner
- a video decoder board
- a 650nm LD (dvd burner extracted)
- a laser diode driver

I don't intent to make a perfect and usable projector but rather a fun prototype displaying image in shades of red :p

My biggest concern is the LD driver (electronic is not really my "thing") for a 720x576@25fps video that means 10,368 Mpixel/s.

I understand that analog modulation is not an option at that speed and I can't find a pre-made board for TTL modulation above a few hundred kHz, knowing that for changing the color of a pixel with TTL I would need many times that speed (10Mhz ON/OFF = 2 colors, 20Mhz = 3 colors, 40Mhz = 5 colors, 80Mhz = 9 colors, etc ...).

This is where it gets tricky (for me at least :eek:) I need to make that driver myself and my first guess was to look for fiber optic LD driver IC as they can modulate much faster (and they involve not much electronic in the circuit) but I'm not sure they suitable for my needs and I'm afraid they won't give enough power.
exemple : http://focus.ti.com/docs/prod/folders/print/onet4201ld.html

Also I get TTL and analog but if anybody could clarify CW, burst and pulsed modulation that would be greatly appreciated.

I'm pretty sure that's doable since a 16x DVD burner LD write data at more than 177 MHz (16X x 1.385.000B/s x 8bits) so I'm wondering if the fiber optic IC is the way to go or if a more conventional circuit like this one : hhttp://www.photonlexicon.com/forums/showthread.php/7802-Fast-modulable-Laser-diode-driver-with-high-side-current-control could be enough ?

I'm really discovering lasers so I may have said a lot of crap (my english is also far from perfect ^^), so please correct me if needed :)

I will post updates on this project as soon as there is something to show.

Solonar
09-23-2010, 07:13
I doubt you will be able to achieve what you want.

Raster is possible, but not really at the framerate and resolution you want.

20k scanners will probably be waaaaaay to slow also.

KoD601
09-23-2010, 07:27
Thanks for the fast answer :)

I known that the cheap scanner set from eBay won't get me really far, but that's all my budget allows.

720x576@25fps is the "best case" scenario, I'm planning to play on the optic angle, the fps and the resolution if necessary to obtain a valid image (even if it's a 10cm square at 10m :p), moreover I could still upgrade the galvo for my fast axis later and I would like my diode driver to be able to follow the speed when that happen.

Solonar
09-23-2010, 10:12
The fiber optic driver only supplies 50ma max, which is not enough to power the diode.

you would need at least 30k scanners( 576 x 25=28800 pps)

I don't see any way you would be able to modulate the diode fast enough to get the pixel density.

The only laser raster output I've seen is low resolution and limited pallet using high end 60k scanners.

Maybe a multi diode array that scans multiple lines at the same time using multiple drivers, thereby dividing the load.

mixedgas
09-23-2010, 14:53
I have modulated diodes to almost 1 Ghz. Around 1 Ghz they start to have interesting intracavity effects. The drivers are nothing magical to get to 10 Mhz for video. Old cathode driver chips for fast CRTs and a booster transistor will get you there.

You need a resonant scanner for the horizontal. I have one I no longer need at 3.89 Khz, PM me, I'll sell it to you at my cost, unless you live in one of the nations I cannot export to.

if you notice, 3.89 x 4 = 15,570 Hz NTSC rate.

This guy got banned here, but he has Youtube videos and finally, after realizing MEMs chips made his "great new secret technology" obsolete, posted some details at laser pointer forums. He filed a piece of steel sheet to resonate at the PAL Horizontal line frequency. So you would use a resonant scanner and a normal galvo. You phase lock your clock off the resonant scanner with a probe beam or other means to pick up the resonance. The probe beam scans a diffraction grating to make the pixel clock, this helps with linearizing the display as the resonant scanner does not produce a ramp waveform but a sinusoid. SLTVM did not linearize, and his stuff still came out decent because the sync board he ripped out of the TV had the "parabolic" correction for the TV yoke and screen curvature nonlinearity.

His downfall was making a A&& of himself here, we have enough engineers and PhDs and skilled tinkers here that we could not stand his blurry pictures and attempts at finding a licensee without disclosing details. While we respect proprietary research here, he blew it with his annoying posts. Which is a shame, because the way he did it could have been mass produced for cheap. It was his masters project, and you are supposed to disclose somewhat if your college is paying the bills.

Good Luck,

Steve

Some Links,


http://laserpointerforums.com/members/sltvm2007/albums/homemade-laser-scanner/

http://www.youtube.com/results?search_query=sltvm2007&aq=f

http://www.youtube.com/results?search_query=sltvm2007&aq=f

http://www.fpga4fun.com/

http://www.knjn.com/


Steve

KoD601
09-24-2010, 03:43
I have modulated diodes to almost 1 Ghz. Around 1 Ghz they start to have interesting intracavity effects. The drivers are nothing magical to get to 10 Mhz for video. Old cathode driver chips for fast CRTs and a booster transistor will get you there.
I checked rapidly and there are cathode drivers running at 80Mhz or more and the chip can control 3 cathode so that's perfect for adding other laser later :)
The booster transistor is used to up the power of the cathode driver to the one needed by the laser if I understand correctly.
If there any schematics for that circuit (or a close one) ? I'm no good with electronic but that's a great starting point.


You need a resonant scanner for the horizontal. I have one I no longer need at 3.89 Khz, PM me, I'll sell it to you at my cost, unless you live in one of the nations I cannot export to.

if you notice, 3.89 x 4 = 15,570 Hz NTSC rate.
I checked for resonant scanner a while ago but couldn't find any, PM sent ;)


This guy got banned here, but he has Youtube videos and finally, after realizing MEMs chips made his "great new secret technology" obsolete, posted some details at laser pointer forums. He filed a piece of steel sheet to resonate at the PAL Horizontal line frequency. So you would use a resonant scanner and a normal galvo. You phase lock your clock off the resonant scanner with a probe beam or other means to pick up the resonance. The probe beam scans a diffraction grating to make the pixel clock, this helps with linearizing the display as the resonant scanner does not produce a ramp waveform but a sinusoid. SLTVM did not linearize, and his stuff still came out decent because the sync board he ripped out of the TV had the "parabolic" correction for the TV yoke and screen curvature nonlinearity.
I saw his video and frankly that was my starting point for my project as it proved it was doable, even if I realized later that he gave no info on how he did it :rolleyes:.
About the linearization I will look for details later as I'm not really there yet.


It was his masters project, and you are supposed to disclose somewhat if your college is paying the bills.
It's also my master project and I am paying the bills myself so I could not disclose it but that would also mean not getting any help or feedback, and I don't want that :D