Is it practical to collimate a (cheap) laser source to less than 1mm diameter? I am working on a research project that needs to project a small UV laser dot at a 3D object. Because of the geometry of the sculpture, the projection distance varies between 500mm to 1000mm as the dot scans. I would like this dot diameter to be somewhere between 0.2 to 0.5mm.
I believe by collimating a laser beam and telescope it to a small diameter (0.2 to 0.5mm), I can avoid focusing the beam to an ever changing distance (500 to 1000mm). Is it realistic to create such a narrow beam from a cheap diode laser source? I plan to purchase a 20X and 40X microscope objective for changing the diameter of the beam, with a pin hole in the middle that acts as a spatial filter.
Money is a limiting factor and I currently have a 405nm laser source 500mW that look like this: 
 A: A 0.2mm UV beam can easily be kept narrow up to a distance of 1000mm (a meter).
The Wikipedia Gaussian beam page tells you pretty much everything you need to know to think about these kinds of issues. For example, two useful parameters are the Rayleigh distance $z_R$ and beam divergence half angle $eta$:
$$z_R = \frac{\pi\,\mu^2}{2\,\lambda}$$
where $\mu$ is the $1/e^2$ beam diameter at the focus / laser output and 
$$\eta = \frac{2\,\lambda}{\pi\,\mu}$$
is the vertex half angle of the cone defined by the $1/e^2$ diameter of the asymptotic farfield. 
The Rayleigh distance is the distance the beam must propagate from the focus (laser output) to expand to twice its diameter. For your parameters, assuming a wavelength of 400nm, I make the Rayleigh distance to be 150 meters! So, if your beam is anything approaching a Gaussian beam, you are all set to go.
As in the comments, sometimes cheap lasers can have a great deal of beam aberration, particularly astigmatism. Depending on your application, a way around this is to use two infinite conjugate objectives to focus your laser onto a subresolvable pinhole and recollimated the output. This will give you a perfect, aberration free beam, but you will lose power: the more aberrated the beam, the more you lose. You may have enough power, or you could compensate by choosing a higher power laser. Be mindful of laser safety considerations. You calculate the pinhole size needed as follows. The numerical aperture of the focussing beam is:
$$\eta = \frac{\mu_L}{2\,f}$$
where $f$ is the focal length of the objective and $\mu_L$ the $1/e^2$ laser beam diameter. Don't be distracted by the numerical aperture printed on the objective: the laser sets how full the objective is, so the effective focal length is the pertinent specification. Now, given the numerical aperture, the resolvable diameter is:
$$\mu = \frac{2\,\lambda}{\pi\,\eta} = \frac{4\,\lambda\,f}{\pi\,\mu_L}$$
Choose a pinhole of about 2/3 of this diameter, and use the arrangement I have described to purify the beam.
