Making glass frit by hand is labor intensive. A machine should do it. This is perhaps the simplest machine which will do the job:

In want of steel balls, cut-up lengths of rebar should work too. That steel is very hard. (If anybody can get the original I'd be much obliged. I'm storing a copy of this on my server in case it gets deleted.)

I might postulate a cyclone separator to get the right sized grains under the laser for sintering. It adds complexity but Hackaday readers are already familiar with those. In industry the combination ball mill and cyclone separator is commonly used in for example coal-fired power plants, which burn powdered coal.

[ED 23.08.2015 - Removed competition-specific and import-related speculation.]

It's a 405nm, violet 10mW laser. I ordered two. After a bunch of research I determined that I need at least 10^2 voltages on the piezo discs to control a CO2 laser like I imagined. The short 405nm wavelength will let me use the 30Vmax voltage regulators to demonstrate destructive interference and also characterize the voltage/amplitude response of the piezo discs through the same mechanism.

I'll probably aim for a maximum voltage of 10^3. 1kV. This high voltage, low current has me researching CCFL tube drivers as I expect the volume they're made in to make their components cheap. Turns out that Jim Williams spent 10 years working on these things, to the point he was fed up with them. I'd very much like to stick one of his designs verbatim on an Arduino shield so I've been getting deeper, much deeper into electronics than I have been before. Current mirrors, totem poles, Darlingtons, the different breeds of TTL and the protocols I2C and SPI... Fortunately I have found a teacher who seems to know what he's talking about, presenting simple things as simple as they are. Too bad he's no longer taking phone calls...

http://cds.linear.com/docs/en/application-note/an25fa.pdf

http://cds.linear.com/docs/en/application-note/an65fa.pdf

http://cds.linear.com/docs/en/application-note/an70.pdf

http://cds.linear.com/docs/en/application-note/an118fa.pdf

Perhaps I can just refer to his documentation for this part of the project? Hmmm!