Details
-
Type:
Improvement
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: ts_aos
-
Labels:
-
Story Points:3
-
Epic Link:
-
Team:Telescope and Site
Description
Update the data/lsst segmentation.txt gain and read noise values and focalplanelayout.txt geometry (x,y position, overscan,prescan), sensor names, to make them consistent with those in obs_lsst for lsst.obs.lsst.LsstCam ( eg. https://github.com/lsst/obs_lsst/blob/master/policy/lsstCam/R00.yaml ).
Currently, the values in data/lsst are consistent with PhosimMapper, but DM-28556 creates a need to move to LsstCamMapper. To allow ISR, the calibration files (made with makeGainImages.py https://github.com/lsst-dm/phosim_utils/blob/master/bin.src/makeGainImages.py for images simulated for `LsstCam` instrument also need to use geometry and gain values consistent with LsstCam.
Attachments
Issue Links
Activity
Now need DM-28553 to be closed (updated lsstCam gain/readNoise values) so that we avoid the 0 gain (which causes issues when making fake flats consisting of 1/gain per amplifier).
Adding the updated focalplanelayout.txt to ts_wep as part of DM-28330
This update looks good to me!
Summary of changes to data/lsst , as in https://github.com/lsst-ts/phosim_syseng4/pull/6 :
Updating `data/lsst` `segmentation.txt` and `focalplanelayout.txt` with geometry and CCD properties appropriate for lsstCam from obs_lsst . Now a lsstCam focal plane contains a mix of ITL and E2V chips, using most up-to-date gain / read noise values from DM-28553 (https://github.com/lsst/obs_lsst/tree/tickets/DM-28553/policy/lsstCam ) .
*Summary of changes for `focalplanelayout.txt`*
- Column 0 : the mapping of names of corner and guide sensors is as follows (only few for example):
| phosim | lsstCam |
|---|---|
| R04_S20_C0 | R04_SW1 |
| R04_S20_C1 | R04_SW0 |
| R04_S21 | R04_SG1 |
| R04_S10 | R04_SG0 |
- Columns 1,2 : x,y -position [microns] updated using `detector.getBBox()`, `bbox.centerY`, `bbox.centerX` (phosim is rotated 90 degrees from lsstCam)
- Columns 4,5: number of x,y pixels updated with `xpx = bbox.getHeight()`, `ypx=bbox.getWidth()`
- Columns 15,16: x,y shift set to 0,0 since correct positions are provided with Columns 1,2 (most substantially this is noticeable for the corner sensors)
- Column 19: updated defocalShift from +/- 1000 to +/- 1500 micrometers
*Summary of changes for `segmentation.txt`:*
- short lines: update raft name, number of x,y pixels as above
- long lines:
- Columns 1,2,3,4: [xlo, xhi ,ylo, yhi] in px per amplifier ; for ITL these are unchanged, for E2V we have eg. instead `[0 1999 0 508]` --> `[0 2001 0 511]`
- Columns 5,6 serial read / parallel read : for ITL serial read is always 1, and parallel read is 1 or -1 ; for E2V, serial read is -1 when parallel read is -1 (Robert Lupton, priv.comm.)
- Column 7 : gain value
- Column 11 : read noise
- Column 15,16,17,17: prescan / overscan values, preserving phosim's rotation of 90 degrees to the left wrt lsstCam . This means that the following was adopted (and tested):
| phosim name | lsstCam name | amplifier property | bbox property |
|---|---|---|---|
| serial overscan | parallel prescan | – | – |
| serial prescan | parallel overscan | amp.getRawParallelOverscanBBox() | bbox.getHeight() |
| parallel overscan | serial overscan | amp.getRawSerialOverscanBBox() | bbox.getWidth() |
| parallel prescan | serial prescan | amp.getRawSerialPrescanBBox() | bbox.getWidth() |
This translates to
| row meaning \ phosim name | parallel prescan | serial overscan | serial prescan | parallel overscan |
|---|---|---|---|---|
| phosim values | 3 | 0 | 10 | 10 |
| lsstCam ITL | 3 | 0 | 48 | 32 |
| lsstCam E2V | 10 | 0 | 46 | 54 |
What was called comcam above is actually R22 of lsstCam (E2V chip). The ISR works properly. Integrated into the AOS loop with
DM-28556(changes to phosim_utils) andDM-28330, (changes to ts_wep). Testing the convergence of the loop with new segmentation / focalplanelayout values for R22.