When I was in college, I worked in a laser lab one summer, for a visiting professor who was doing research on metal-vapor lasers. The first half-dozen times we turned his apparatus on, all it did was make the grad student at the next optical slab over really mad. Because the apparatus was using a big honking RF generator to ionize the metal vapor and prepare it to respond to a pumping laser, and there was enough energy leakage bouncing around the lab to trip the breakers on all the grad student’s power supplies.

Yes, our equipment was grounded. But it wasn’t really grounded. I spent the next couple weeks helping to make perforated-steel enclosures for all our RF stuff, connected every few inches to a the building foundation by copper-braid straps the size of your thumb. I have come to realize since then that this probably wasn’t the best way to get rid of the RF, but at least the breakers stopped tripping.

We went on to do some measurements and found really remarkable results. The metal vapor was showing gains of 30-40% in one pass of pumping light through the tube. This was going to be really big and important. Then summer was over  and it was time for me to go back to classes.

A few months later, I asked the professor who ran the lab what had become of the fantastic metal-vapor laser experiment. Garbage, he told me. It turned out that when you turned the RF generator on, there was still enough RF interference leaking around that it acted as an additional power source for the pumping laser, and all the measurements of increased output had been due to that.

***

So I’ve been working on an off to build a DLP printer using a regular projector with a mercury lamp, and I almost had it going (my test piece was sticking to the bottom of the vat instead of to the stage, as they do). Then I left a vat on the machine for a couple minutes with no image going through, and found that a bunch of the resin had started curing anyway because of stray UV bouncing around. Not good.

Quickly I got some tape and aluminum foil and covered all the light leaks, so that only light from the lens would get through. Improved the image quality quite a bit. And the machine stopped working at all. Instead of exposure times of a few seconds per layer, even exposures of a minute per layer weren’t doing anything. 10 minutes of continuous exposure, nothing.

But wait. I’d been getting images of what was being projected. Images stuck to the floor of the vat, sure, but still little pieces of cured plastic in the right shapes. How could that be? My best guesses are that either the light from the projector managed to piggyback on the UV light leaking from the lamp, or that the projector (remember, this is a couple thousand lumens delivered to an area a few inches across) was heating up the resin where the image was and making it just a little bit more responsive to the UV. Because I took apart the lamp and there’s a nice thick glass UV filter right across the front.

So now I’m at a crossroads — do I redo the optical path so that some UV gets through after all, or do I carefully open up the light leaks again and see if I can make use of the background instead of suppressing it?