FPGA to drive 50 RGV lasers scanners

Hi everyone,

I am building an art installation for Burning Man with a budget of about $3000. It's gonna be a laser system with 50 RGV lasers all placed in a row. Each laser points to a 45° mirror attached to a stepper motor rotating in the same direction, which makes it a one-axis scanner able to scan a 360° plan (see schema). All the plans will be parallel, spaced by a foot. The installation will be seen from very far away in the desert by night (there is a lot of dust in the air), and I will do animations with the lasers like a wave will the 50 lasers in a row.

I am thinking about using a FPGA (Spartan-3AN) to control the system (from the DMX output of a computer), and build a custom driver that will power the 50 laser scanner units.

I have already started a thread about the RGV part (http://laserpointerforums.com/f40/cheapest-rgv-laser-55606.html) and I want to focus here more on the driver/mechanics.

I am a beginner with laser systems, and am asking gratefully the community for help about the following:
1) What type of stepper motor should I use for this case of one-axis scanner (knowing that 1.8° step is OK for me)? What are the advantages/drawbacks of uni-polar vs bi-polar?
2) Does it seem realistic to use a FPGA to control 50 lasers in angle position and RGV color? That would make about 400 outputs from my driver
3) Is DMX a good idea as input? Any issue with using a FPGA with it?
4) What are the biggest pitfalls I am facing?

Thanks a lot,

That cost is going to go WAY up 3,000. a driver to drive 50 stepper motors is going to cost a ton. why not just master and slave they all move the same way or do you need everyone to be independent.
maybe have 5 sets of 10 linked up so you only 5 stepper drivers. but 50 that alone will cost you an arm and a leg.

Unfortunately I want the 50 lasers to be independent. I will build the driver board myself, for few hundred bucks in components I guess. The driver board will just be here to deliver power to the different leads depending on the FPGA output (where all the driver logic will be). Does that make sense?

yes but what a task and you will want to find 4 wire surplus stepper motors.

I have found stepper motors between $5 and $10 on the web like Stepper Motor (Small - Low Torque) [RKI-1128] - $4.69 : Robokits World, Easy to Use, Versatile Robotics Kits... or Lin Engineering 4018X-07-01 bipolar stepper motor | Alltronics, but I don't know if they are good for what I need (never played with steppers before).
I would guess the paramaters are torque, step precision and max speed. I think the torque should never be a problem to simply move a very tiny mirror, the precision of 1.8° is fine with me, and I have no idea for the max speed.
Do you have experience with stepper motors? What do you think?

i do and 1.8 is good but if you are just drawing a circle that does not matter much
3.1 VDC STEPPER MOTOR | AllElectronics.com
these are the cheapest i have gotten them from.
but ebay sometimes there are people with huge lots of this stuff for a cheap price.

Are you using a custom board for the FPGA or a Dev Kit?

Do you have any experience in programming an FPGA? And how do you plan to get 50 RGV lasers within a $3000 budget?

jib77 said:

Are you using a custom board for the FPGA or a Dev Kit?

Click to expand...

I am using the a Spartan 3AN starter kit, and I may build a custom board to go on top of the starter kit.

Bluefan said:

Do you have any experience in programming an FPGA? And how do you plan to get 50 RGV lasers within a $3000 budget?

Click to expand...

I have played a bit with my Spartan 3 starter kit for another project and am so slightly familiar with the programming of FPGAs.

For the 50 lasers, I think they will more likely be 10 for now :yh: , and I am evaluating if http://laserpointerforums.com/f51/t...ser-using-white-fusion-kit-sightfx-41942.html is the solution I need. But I am *so* surprised that nobody sells RGV modules. I guess blue/violet lasers have become common only very recently and we need to wait about one more year to see RGV modules come on the market.

My experience is with Altera's CycloneIII ... starter kits have a lot of those I/O lines used for ram, flash, and other hardware goodies. Even though they claim a high I/O count you may not have exclusive use for all of them. For 50 you would need a custom board, but for 10 it may be fine.

jib77 said:

My experience is with Altera's CycloneIII ... starter kits have a lot of those I/O lines used for ram, flash, and other hardware goodies. Even though they claim a high I/O count you may not have exclusive use for all of them. For 50 you would need a custom board, but for 10 it may be fine.

Click to expand...

You are right, so I will try to start small scale to make it easy. Also I think I have a choice to make between:
a) creating a big custom board, plugged to the FPGA board, that will power all the 10 laser + motor systems at once
and
b) creating 10 small boards powering each one laser + motor system and controlled directly by the output of the FPGA board

What are your thoughts about those two options?

You might think about using a PWM LED driver to control your stepper motors. LED drivers such as TI's TLC5940 are serially controlled (3-inputs, and chainable), easy to service (DIP chip in a socket), and provide 16 outputs with 12-bit PWM control. This may not work with the stepper motors, but you can easily just hook them up to h-bridges or other motor controls, and use other forms of feedback.

$3000 might be enough to fund the basic needs of your project, but be prepared to spend twice that to get everything working and right, supplying the power needs, transportation, etc. not to mention the cost of labor. Tuning 50 RGV lasers to form beams will be a major chore. Sometimes the neatest effects are made with randomized or patterned graphics, not individually controlled mechanisms.

Before you invest in large amounts of anything, make your prototype with a single laser (don't even bother with multiple colors for now). Also make sure you have your signal lines in good order. One project I was on had major problems with signals being degraded to the point that it affected the graphics. We only discovered the problem after most of the hardware had been built and sealed. The workaround for us was resetting the hardware periodically to clear out these glitches. It cost us a lot of extra money to find the right solution for that problem.

Good luck!

EDIT:

A good source for FPGA components and boards, etc. is Digilent. If you use one of their boards I would get a serial Pmod and make your own simulator using Processing. Having a simulator was invaluable in getting one of our projects working properly and predictably before the hardware was fully built.

I would create 1 board for 10 lasers ... this board would have a small CPLD with bridge logic to interface to the FPGA. This way you can always add boards of 10 and all it takes to interface to the FPGA is the lines needed to communicate to each bridge. Also careful using FPGA pins to directly drive anything ... they can only drive/sync about 12mA max. We use N-Chan MOSFET's to drive LEDs or anything else requiring larger currents.