Hmm, a 10us response time is well within photodiodes range (normally nanoseconds), but of course as you pointed out, large area photodiodes are hard to find - not to mention expensive!. Then there phototransistors. Huge gain, but probably maybe a bit slow.
I was also thinking of modulation, until I saw the required response time, sure it would increase the sensitivity no end since you coudl filter out ambient light and all other frequencies except yours, but for 10us...you would have to modulate at a very high frequency. Then again, the delay *should* be fixed, so for measurement purposes this won't be a problem, just do it in post processing. All depends on the shortest period you're measuring.
I think others have covered the lens idea pretty well. You must remember that a convex lens will focus a large area, to a small one. This means if you move your beam half way across your lens, or at an angle (attempting to align), say 4cm, the focal point will only move a few mm. Coupled with a a pseudo parabolic reflector, the type you see used in flashlights, I wouldn't have thought you would have a problem.
I just thought....a cone shaped reflector, long and thin, with a convex lens at one end, and your photodiode at the other. Should sort out the alignment problem, with minimal light loss.
For a 'small' detector, nulike the cone idea above, instead of a specular reflector (like a mirror, shiney) use a lambertian reflector....white paper. Again, sacrificing sensitivity for alignment simplicity. So you wouldn't diffuse the beam as it enters, firstly it gets focused more or less on the detector, with the surrounding white reflector picking up any misalignment and scattering it around the inside of the reciever. A few carefully placed photodiodes around the box (not in the same place) might give a half decent signal?
You still have to deal with natural light though. Ambient light would be DC, but also anything that moves will produce a changing signal. Perhaps in that case we can assume the laser 'swamps' the reciever *when* it is aligned. So looking for a large sudden change would just be a case of using a comparator, with hysteresis. - Like a dataslicer, the kind used in RF recievers to detect changing signals, whilst 'tracking' a slow moving DC component (changing ambient light in your case).
Hm, I don't have much experience with this. Another simple solution would be to widen your laser
Or to help you align one of those reflector strips that you see on cyclists. They light up nicely even when a laser isn't 'directly' on them.
I hope that gives you some ideas, I'm just waffling on. If its uselses ignore it, but if you want some dirty diagrams, I'd be happy to knock up something in mspaint.
Blueteeth