Ronnie
I see a lot of cheap and expensive microscopes, and I notice that expensive is not always better for protein crystallization.
Almost the most important thing is that illumination comes from one particular direction. Often this means that the light source is small and far from the sample stage. What does not work well is to have a large light source (eg multiple LEDs, large white screen, mirror or sintered transparent sheet) that is close to the sample - even with the best optics in the world, you won't see crystals well.
Dark ground (see only scattered light) and ordinary transmission mode can both work well - good to have both if possible.
Good luck,
Patrick