How to build an Elitzur–Vaidman bomb tester? I would like to personally recreate the Elitzur–Vaidman bomb tester experiment. Is that generally possible with maybe only a few thousand dollars to invest or maybe even for a few hundred? Any recommendations where to buy the parts?
The experiment is based on the overlapping of the waves of the photon, so one has to align the components with a precision which is much higher that the wavelength of light, correct? So when I use a 532nm laser I have to precisely align the elements of the experiment to multiples of 50nm precision to get meaningful results, correct? 
How was the original experiment set up? 
 A: Emilio is correct that the crystals and photon detectors are probably the most expensive elements. You will probably need precision optomechanics to mount and align all the beams too, like the kind they sell at Thorlabs (www.thorlabs.com), which can add up to a substantial cost for a hobbyist- almost certainly in the thousands rather than hundreds. The good news is that the laser diode itself can be cheap (thanks, Blu-ray!), and the FPGA or similar device to handle the photon counting shouldn't be too bad either.
In general, I think development of a simple, affordable setup for this would be a sort of novel thing that one could publish in a suitable journal, along the lines of : http://aapt.scitation.org/doi/10.1119/1.3116803 (paywall, sorry). It is the kind of thing that educators, and maybe other hobbyists too, are actively thinking about and working out.
The material in the above paper and at (http://people.whitman.edu/~beckmk/QM/) might be useful to you in working out some of the basic things you would need. The pictures on that site can also give you an idea of what this kind of experiment usually looks like, if you are unfamiliar. It might be easier (although at additional cost) to try to recreate the kind of Bell violation test that they perform as a stepping stone to a Vaidman bomb experiment, which is similar but a bit more complex. 
One clarification : spatial alignment doesn't generally need to be comparable to the wavelength of light (thankfully). The relevant size is just the spatial mode of the beam, or in other words the size of the laser spot. So that's not so bad- at least, if you have standard optical mounts and some patience. I would guess that more challenging than alignment precision is stability, since you need to make some kind of interferometer where the path length does need to be stable over time to nanometers.
Good luck!
