It occurs to me that another way to do this without a collimator lens is to photograph the fringes in a well defined setup. Knowing the water thickness accurately and the distance to the camera you could calculate how many rings of power you would expect to see then using a fringe analysis program like Fringe XP subtract them out to get the true wavefront. You would need to make a fixture sort of like a spherometer but with a needle tip to measure the water thickness exactly. It might be worth contacting Dave Rowe to see what it would take to do this. You could then use a light box like Dave built. What do you think?