STIX field of view
For spectroscopy, the requirement here is that at least we include any pair of fully-illuminated non-adjacent pixels in a given row. This ensures that the result is not dependent on source morphology.
For imaging, the requirement here's that at least 2 adjacent stripes be fully illuminated in a given row AND at least 1 row be fully illuminated for these pixels. Partially-illuminated pixels don't count. With 4 pixels in a row, it is business as usual; with only 3 adjacent pixels, ambiguities can be cleanly eliminated using redundancies in the data; with only 2 adjacent pixels, we can still do Yohkoh-style imaging.
The criteria below apply to all windows, but do not include the effects of internal grid shadowing, which are different for each grid. There are separate limits for x and y coordinates (y is parallel to the long axis of the pixels), both of which must be satisfied.
- Use of all pixels in a row for imaging or spectroscopy: |y| < 0.518 degrees
- Use of only one row for imaging or spectroscopy: |y| < 0.960 degrees.
- Full sensitivity imaging or spectroscopy (four pixels in a row): |x| < 0.614 degrees
- 3-pixel/row imaging or 2 pixel/row spectroscopy: |x| < 0.844 degrees.
- 2-pixel/row imaging: |x| < 1.056 degrees.
Note that until the next perihelion, full sensitivity can be used for flare studies. The limits are relevant for the Crab, however.
One can gain a little spectroscopy s/n outside these limits but at the cost of enhanced systematic errors. Also, one can gain some imaging insights with as only 1 pixel in a row, but one cannot determine visibilities.
Once the trajectory of the Crab is known, these limits can be translated into time boundaries for imaging or spectroscopy under the assumptions above. Note that the limits do not take into account the finite size of the source, or the effects of misalignments.
When other effects are folded in, my goal is to come up with a self-contained IDL module (similar in function to that used on RHESSI) that returns a mask of pixels to use for a given purpose for a specified xy offset, and/or a normalization factor that could be used (at ones peril) for partially-illuminated pixels.
Finally, a corollary of all this: Ignoring cross-subcollimator trajectories, an upper limit on x,y for the illumination of any part of any pixel in any detector is: |x| < = 1.48 degrees AND |y|< 1.40 degrees.
Author: Gordon Hurford (firstname.lastname@example.org)