I've been setting up the cross-sampling test of the Hartmann sensor, Right now I'm waiting on a 50/50 BS so I'm improvising with a BS for 1064nm.
The output from the SLED (green-beam @ 980nm) is around 420uW (the beam completely falls on the power meter.) There are a couple of irises shortly afterwards that cut out a lot of the power - apparently down to 77uW (but the beam is larger than the detection area of the power meter at this point - by ~50%). The BS is not very efficient on reflection and cuts down the power to 27uW (overfilled power meter). The measurement of 39uW is near a focus and the power meter captures the whole beam. There is a PBS cube that is splitting the beam unequally between s- and p-polarizations (I think this is due to uneven reflections for s- and p-polarizations from the 1064nm BS). The beam is retro-reflected back to the HWS where about 0.95uW makes it to the detector.
There is a 1mW 633nm laser diode that is used to align the optical axis. There are two irises that are used to match the optical axis of the laser diode and the SLED output.