Check platescale over mission using measured occulter diameter

[jburkepile]

Use the occulter radius size from the routine kcor_find_image.pro to compute an image platescale. Record the value of the image platescale in the FITS header keyword IMAGESCL. Image platescales will be compared with the original determination of the instrument platescale derived from the K-Cor effective focal length calculated by Dennis Gallagher, and the lab measurements of the K-Cor occulter diameters.

Tasks

• [x] IMAGESCL keyword
• [x] add IMAGESCL keyword to database
• [x] make mission plot of IMAGESCL in end-of-day
• [x] add RCAM_SCL and TCAM_SCL FITS keywords ("distortion corrected image scale for camera {0,1}")
• [x] add rcam_image_scale and tcam_image_scale kcor_eng table fields
• [x] plot each camera separately for mission plot plus a difference plot
• [x] make daily plot of IMAGESCL in end-of-day; use different color or symbol for each camera
• [x] annotate RCAM/TCAM in plots
• [x] create CSV file for daily values, columns: filename, IMAGESCL, RCAM_SCL, TCAM_SCL, and RCAM_SCL - TCAM_SCL

[jburkepile]

The effective focal length of K-Cor Objective lens, using data from JenOptiks and Corning, is 2018.5 +/- 0.2 mm.

K-Cor occulter diameters in mm (+/- 0.00254 nm - estimate from Dennis email):

• Old Winter Occulter (OC-1018.9") = 19.941 mm
• New Winter Occulter starting Jan 12, 2021 (OC-1017.0") = 19.905 mm NOTE: (19.905 is by DFS-avg of 6 measurements) company that fabricated it measured 19.944mm
• Spring/fall Occulter (OC-1006.9") = 19.705 mm (ranged from 19.702 to 19.708)
• Summer Occulter (OC-991.6") = 19.480 mm

Using the focal length and occulter sizes (ala UCoMP):

magnification = distortion-corrected occulter_radius_in_pixels / (occulter_diameter_in_mm / 2.0)
platescale = 206264.8 / magnification / focal_length

[mgalloy]

I have notes about a OC-1018.9" occulter. Was this never used?

[jburkepile]

Mike: Thank you for pointing that out. We swapped out the 'old' winter occulter of 1018.9" on Jan 12, 2021 for the new tapered winter occulter. I had forgotten about that. I edited my info above to include both the old and new occulters. Thanks!

[mgalloy]

I am running code that calculates IMAGESCL on 20221007. I am getting results about 0.034 arcs/pixel less than the current plate scale (5.643 arcsec/pixel).

level2$ for f in *_kcor_l2.fts; do echo $f; fitscat -r $f | grep IMAGESCL; done
20221007_171731_kcor_l2.fts
IMAGESCL=               5.6105 / [arcsec/pixel] image scale for this file
20221007_171746_kcor_l2.fts
IMAGESCL=               5.6100 / [arcsec/pixel] image scale for this file
20221007_171801_kcor_l2.fts
IMAGESCL=               5.6098 / [arcsec/pixel] image scale for this file
20221007_171816_kcor_l2.fts

[jburkepile]

Not surprising. I have been spot-checking days and getting platescles as much as 0.06 arcsec/pix different - always less than the 5.643 arcsec/pix I got back in 2013-early 2014. Using the larger difference of 0.06 arcsec/pixel as a worse case, we are off by 0.03 solar radii at the outer edge of the field-of-view (the error is much smaller closer to the occulter). That is not awful. The smallest error bar I use for measuring CME positions in the best images is 0.03 solar radii. Many measurements are twice that.

Nevertheless, the platescale is off and needs to be recomputed. I will set up a separate ticket for that. Yesterday, I asked Ben to retrieve 2 of the K-Cor occulters next time he is at MLSO. I want him to mail those back and have the IG measure them again.

It is also possible that a change in distortion correction is playing a role.

So glad you are going to save this info in the headers.

[mgalloy]

Here is the mission plot (only 1 day has been processed with IMAGESCL values).

[jburkepile]

Thanks very much for doing this. Just to clarify the above plot: the red line is the platescale (5.643). Is the black vertical line near the end of the mission the image scale? Are you plotting the distortion corrected occulter radius for each camera separately?

[mgalloy]

The black vertical line is comprised of the image scale values for a single day. These image scale values are computed from the average of the two distortion corrected radii. I could store separate IMAGESCL values in the FITS header and database, and then produce two image scale plots.

[mgalloy]

Here is the daily plot for 2013-11-13:

Mission plot is not too valuable until we reprocess:

[mgalloy]

The CSV file looks like:

p$ cat 20131113.kcor.daily.image_scale.csv
2013-11-13 18:52:01, 5.644, 5.649, 5.639, 0.010
2013-11-13 18:52:16, 5.644, 5.649, 5.639, 0.010
2013-11-13 18:52:31, 5.643, 5.649, 5.638, 0.011
2013-11-13 18:52:46, 5.643, 5.648, 5.637, 0.011
2013-11-13 18:53:01, 5.642, 5.647, 5.637, 0.010
2013-11-13 19:01:19, 5.637, 5.642, 5.632, 0.010
2013-11-13 19:01:34, 5.637, 5.642, 5.632, 0.010
2013-11-13 19:37:16, 5.646, 5.652, 5.640, 0.012
2013-11-13 19:37:31, 5.644, 5.649, 5.638, 0.011
2013-11-13 19:37:46, 5.643, 5.649, 5.638, 0.011
2013-11-13 19:38:01, 5.643, 5.649, 5.637, 0.012
2013-11-13 19:38:17, 5.643, 5.649, 5.637, 0.012
2013-11-13 19:38:32, 5.642, 5.648, 5.636, 0.012
...

[mgalloy]

Updated daily image scale plots:

Updated mission image scale plots:

[mgalloy]

Mission image scale from the 2013 reprocessing: