The polar lissajous generator:
On the right is the ps: three AA's wired in series.
Between that and the two motors is the laser diode: the green circuit board has the LD current regulator and old power switch. I stripped the guts from a laser pointer I got from Pearle; I like these because they're inexpensive but the body and collimation lens are machined brass. They're very easy to cross-wire and bypass the on/off button. This is the same laser head I used to transmit audio across the room, by hooking it to an impedance-matching transformer. They're just good for experimentation.
The LD is mounted by press-fitting the brass collimating lens into an aluminum chunk, that's bolted to an L-bracket that screws to the baseboard. I slotted the L-bracket with a hacksaw so the height of the LD was adjustable. It also points upwards slightly so the image is thrown above the generator; this prevents loss of part of the image from the side of the generator or the table on which it sits.
The pair of motors, cheap 5v jobs from RatShack, are bolted down with 3/4" copper tubing clamps. On the shafts they have steel bushings with centered holes that press-fit the motor shafts; on the ends of the bushings I double-stick-taped fragments of FS mirror.
The proto board has a 556 timer driving (through resistors) two MOSFETs that handle the motor switching loads. The 556 is wired as a simple PWM, using a single cap to ground and a pot, seen on the far left, from power, to wiper connecting to the 556, to cap.
The controls are two carbon-film 5K pots (values of the cap and resistor almost don't matter, since all we care about is off/on ratio, not speed) and a power switch.
There's an extraneous trimmer pot when I thought that I needed to carefully adjust the power to the laser, and an electrolytic across power/ground. Oh and lots of duct tape.
This is against my fiberglass shower wall. It's a triplet with precession because of the exposure time:
And another, this one a quartet with precession:
With both of them you can see that the motors do not share an optic center. I'll have to shim them to try and get the rotation points collimated.
The lesson for today, by the way, is: don't skimp on optics. The initial version used some small FS mirrors I tore out of tape cartridges, used in the copy-protection beampath. They were teh suck. The image was fuzzier than Bon Jovi's head. The new version uses a chopped up nice FS I got from American Surplus & Science. It's about 20x better.
Total cost: about $15, although I did scavenge a lot of the parts from my junkpile. The FETs are IRF640's. The caps are mylar except the power/ground.
Next: clean it up, use a real board rather than a proto, and put on a third axis to get really amazing swirly designs. I also need a more stable controller schematic; eventually I'll use a PWM off a PIC. Probably having more massy mirror mountings would stabilize the motors considerably, and it's not like this needs rapid response time.