Looking for flexure

At the IRTF, we tested EXES for flexure by going to a telescope position that mimicked mounting on SOFIA.  We then did 6" nods along the slit direction, although we were using the decker position with a pinhole arc.

ND difference showing night sky through offset pinholes.  Emission lines are shifted from spectrum to spectrum because of their azimuthal offsets at the slit plane.  

A cut along column 400 of the above image showing the values through the 5 pinholes (not the pinhole near row 600 wasn't looking at sky so the the ND difference was very small.  This is to show the data number level on the sky.

By looking at the difference between the two sides of the nod, can look for any change in the location of the pinholes on the detector that might indicate flexure within EXES.  By looking closely at one of the pinhole spectra and shifting the column cut by +/- 1 pixel, we find that the nod difference shows no evidence for flexure.

A random nod difference along column 400 near one pinhole spectrum showing the difference (white), the difference with +/- 1 pixel shifts to one side of the nod (green/blue), and the column DN before difference (red).

Made similar observations at +/- 1 hour to simulate SOFIA telescope rotations of elevation = 25 and elevation = 55 (on SOFIA).  Looking at the difference in pinhole positions for the different hour angles, find change of <3 pixels between "elevation = 25" and "elevation = 55".  At the focal plane, this is <75 micron.  It is a smooth change at the positions examined with half coming between HA = +1 and HA = 0, and half coming between HA = 0 and HA = -1.

Looking at the position of the night sky lines, did not notice any shift with hour angle.  This suggests the echelle grating is the most likely source of the 3 pixel shift.  If this is true, the angular change in the grating to give a 3 pixel shift would be the linear shift / focal length or 75 microns / 1 m = 15 arcseconds or 0.004 degrees.