Beacon

Overall Reasoning
//Judit email of 8 Sep, summarising previous discussion://  First thing to decide is whether we want to use the beacon for slow WFS in the lab, or not.

If YES:
 * 1) For a given diode laser + fiber set [not yet selected], what percentage of that laser light is needed to calibrate, and not to blind, the fast WFS. Roughly the rest goes toward the lab, i.e. slow WFS.
 * 2) Is it easy to interlace the slow WFS corrections with the observation? How long does it take a correction? How often do we want to do that? There has to be a shutter for the beacon. In this case, the synchronization with science data collection is more demanding, but probably can be done without much difficulty. Right?
 * 3) First surface: same simple edge filter as beam samplers.
 * 4) Second surface: The substrate wedge would take care of beacon ghosts, so I think AR on the second surface in the VIS is not necessary. Am I right? Or is there something else to consider?We want AR for the IR to maximize transmission. Right?

If NO:
 * 1) What OD is enough blocking at the beacon laser wavelength toward the fast WFS (in transmission) not to hurt it? [Almost the same question as 1a, but in this case we could add extra blocking filter before laser injection into the fiber, if necessary and the selected laser + fiber arrangement allows that.]
 * 2) Shutter or we could turn it off remotely?
 * 3) First surface: How wide of a band does the slow WFS need to operate, versus how long does it take to do a correction? Where would that band be in the spectrum?
 * 4) Second surface: We want AR for the IR to maximize transmission. Right? **MJI: Yes... within limits. The second-surface reflection should be less than e.g. 10% throughout the visible. A single-layer AR coating for the IR may be OK.**


 * MJI: If NO for bluer stars and during system on-sky calibration and YES at other times: All questions from YES plus (4) from NO.**

System Design, Layout and Cost
Laszlo email of 7 Sep:

//On the other hand if John is right and the lab seeing is fast and we want to correct that, a laser beacon could provide much faster WFS. These two approaches are not mutually exclusive, of course. Initially we will probably use starlight because the beacon won't be ready. Laser diodes are available from about 400 nm to 2000 nm for under $1k and the most commons are less that $40. With driver and optics the beacon could be as little as $1.5-$2k. //

MJI notes that the most expensive part is probably a collimator to get a high-quality wavefront with 125mm beam.