[extracted from email from JDM on 2009Nov05] You could do this in person in the lab, if you want to turn knobs by HAND and watch results on your laptop using ircam. but otherwise, you should be in the control room with thep icomotors on. i recall that maybe the motor for the slit array has partially failed.. where it doesn't home any more. this means you need to index it.. which I don't even know if thats possible in the software?? we may have a small problem here. probably you power cycle the system and just move everything relative. and you can home the prism mover by hand Ithink. or just don't bother homing the slit array but will have to do things relative to 0 position? the way this works is that the slit is not adjustable in focus, so this means we have to move the Doublet (the lenses just before the prism) to bring the SLIT into focus. I've attached an image of the slits . I think we are unable to reach all positions with the stage, but can read the 4 positions on the right hand side (please check by eye if this is right.. it should be easy to see which side of the slide has the grid of holes). The MIRC pixels are 40 microns, so if we ant to have the BEST spectral resolution we need to use the 40 micron slit. However, this may lose more light than we want to use, in which case you can try the 80 micron slit. For this test in order not to lose much light, you could center up on the 80 micron slit if you want. [ideally, we should do this on the 40 micron slit first and then measure how much light we lose and then if its too high, then go to the 80.. but, if weon't have time, we can do this the opposite way -- that is, lets do the 80 micron slit and if we are happywith the throughput, we can try to go thinner at a later date. After the slit is in focus (and not tilted -- if its tilted, there should be a goniometer than can rotate the slit mask to make it straight -- this may need done, since we replaced the slit holder as part of photometric channel upgrade), then we move the slit out of the way and then look at the line of light from the cylindrical lens. we then focus this onto the same set of pixels as tightly as psosible -- getting as much light in peak row as possible (this can be inspected using IRCam and looking at the central row and the 2 neighboring rows. There is a focus pico for the cylindrical lens. Last time I was there, the focus of the system was kinda messed up by the constant jiggling of the mirc system as part of xchan upgrade. Ideaally we can get 90% of light (even 95%) in one ROW.. but I couldn't do this well ... maybe 80%-85%. [if this is true then you see that probably we can live with a 40 micron slit and not have to use a 80 micron slit). Once we have the line of light focused using the cylindinder (and re-focused the photometric outputs) , all you have to do is NOTE the peak counts and then move the slit to hit the same row. and then note the peak counts with the slit and see how much light is lost when you move the slit in. **Remember that updown on the detector is really sideways outside the dewar (90 degree rotation).. A. Setup 1. get white light source on beams 3,4 2. do fiber explorer. 3. take prism OUT 4. reduce flux by moving fibers so that the system does not saturate B. focus slit. 1. move in the slit you like (ideally the 40 micron slit, but 80 micron slit could work if we lose too much light) 2. you will need to make it so you can SEE the slit. this could be done by having a little light on the slit so that the slit is dark. OR probably better is go to Kband, because the slit will be warm compared to reflective gold. The lens is supposed to work at H and K so the focus at K should b eclose to the right place. 3. check tilt of slit too. adjust goniometer. C. focus cylindrical lens. 1. move out slit array 2. focus line of light from white light into a central line a much as possible by moving cyl lens focus. (go back to H band) 3. check how well the lines from B3 and B4 overlap and are focused. We had some issues where B3 and B4 don't hit quite the same line. if this happenes, we must talk! 4. hopefully 80-90 of the light lands in one row for both B3, B4.. Write down the counts D. focus photometric output 1. you will need to also refocus the photometric channel outputs since you mooved the doublet in step B. 2. I'm not sure that Xiao wired up the picomotors, so you will need to do this by hand. 3. you should see two spots from B3 and B4. These should have most all the light in one pixel.. not 100% but like 80%. you will need to focus these spots and make sure they arei n the same ROW as the light coming from the cylindrical lenses. 4. to focus the spots you move the vgroove holder forward-backward towards/away from dewar window. please look at the optics (including the fold mirror) so you can understand the geometry before moving to many motors. 5. you can move the spots up/down by moving of the other screws. **Remember that updown on the detector is really sideways outside the dewar (90 degree rotation).. E. check throughput 1. move slit back in to the exact same row as light from cylindircal lens 2. align it carefully to maximize the light through the lsit 3. compare with how much you measured in part C and repor this 'loss' hopefully 95% or so of the peak counts still gets throughs.