I'd guess it uses a spinning mirror. http://www.youtube.com/watch?v=9FiR2rGv7i4

From the article, they acknowledged that this would be unsuitable for cargoes sensitive to high G-forces. They suggested that rocket fuel would be a suitable payload.

I'm less than certain. Intersecting with the Moon would only require changes in ellipticity, whereas causing the satellite to become unbound would require a considerable increase in the satellite's kinetic energy.

That's a neat thought, but how are you planning on communicating with your experiment once it gets out into deep space?

There's a nice bit of software called PUFF that was written at Caltech. It is available as an MS-DOS binary, which you can run on XP or Mac through DOSBox. Also, on the PUFF website they report that the source code comes with the program, and some have had success in compiling it for Linux. Unfortunately, you can't buy the software directly. Some textbooks come packaged with it, though. I can recommend the Rutledge book as a nice overview of lab electronics.