Project

General

Profile

Anom-R5

Anom-R5: Image Health,Quick Reduce, or other mechanism recognizes and reports major image anomalies that require immediate corrective action:
a. loss of data on 1 or more amplifiers,
b. dome occulting beam (i.e. large zeropoint change on otherwise clear night?)
c. guiding errors > 0.5'' or other guider failure.
d. PSF FWHM too big compared to site DIMM: FWHM^2 > DIMM^2+x^2 for x=0.60''
e. If DIMM is not in operation, other indicators of poor focus: indications from donuts, sudden change in PSF FWHM.
f. Sudden change of sky brightness (e.g. lights on in dome), other than twilight.

Prerequisites:
All calibration images of the given night (since they are probably required by the first reduction tools), i.e. master flat, master bias. In addition:
a) one raw science image taken with all 62 chips
b) another raw science image taken of the same part of the sky right after this with a 20% shorter exposure time; we need to then fake the exposure time to again be the regular one, so the raw counts are just 20% lower than expected
c) an image taken with detuned guider, such that stars are offset by 0.5'' between beginning and end of exposure; an image taken when the telescope is still moving
d) this can be simulated by faking the DIMM seeing; we should also test it with a slightly de-focused image
e) a de-focused image with donut data
f) another raw science image taken of the same part of the sky as (a), but turning the lights on in dome

Procedures:
Feed sets a-f of faulty data to Image Health / Quick Reduce. Test whether flags are raised.

Code required:
Image Health / Quick Reduce

Verdict:
Flags should be raised in all 6 cases and information useful for the observer to fix the respective issue should be provided.

Consequences:
If requirement is not met, contact developers of existing tool with feature requests / bug report and re-test after changes have been made. Iterate until test is passed.

Note:
I mailed Angelo, Marcelle and Josiah ; ; for feedback, since they're uploaders of IH/QR documents on the docdb.