EXES lookout angle

The X in this image from CD013/B_12_02_28/b170.fits shows the position of the center of the window. EXES is in echelle-side, imaging mode, filter=7.4um, decker=slit=OPEN, ctdelay=10.

The "X" is mode of strips of masking tape oriented up-down and left-right with respect to the floor. (The mounting screws on the window holder were used as a guide to getting the angle right.)

The center of the X is (567,472), +/- 5 pixels on each coordinate.

The center of the illuminated region (defined by the SDF on the right, and probably LN2 baffles and window edge on the left)  is 515,488.

So we would like to move the EXES outlooking angle +52 pixels in x and -16 pixels in y.

To get the orientation and scale on the window, we placed a screw in front of the window (with the bare MUX), backlit by a quartz lamp at the end of the TAAS.

As viewed from the front, the screw was coming in from the right, level to the ground. The orientation of the ellipsoid is shown with the overlaid box. The long axis represents the side with the mounting screw plates.

The screw pitch is clearly visible and gives us the plate scale on the window-- averaging 54 pixels for each thread, which is physically 0.113".

So in terms of magnitude of deflection, we need to move the ellipsoid by sqrt(52^2+16^2)=54 pixels (coincidently, the same as a screw thread). The window to ellipsoid distance is 1200mm, so this angle is 0.00239 rad. This is on the order of 0.011", for the 4"-4.5" ellipsoid. We currently have about 0.025" of shim on the right side of the ellipsoid (as pictured above) and 0.005" on the top.

It was confusing to establish the orientation, so I'm going to try to get someone else's opinion. 

Update: I think this is the real orientation:

 Therefore, the operation will be primarily removing 0.01" of shim from the right side of the ellipsoid.