Owl Calibration 2017

[lah]

This ticket records the results from the Owl wavelength calibration in 2017. This is the first time this has been attempted and the scripts used are still experimental/ not yet reviewed, but a record is still very useful.

Data location: /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength

The process uses the raw data to calculate a transmission spectrum which is then compared to that measured by FSF instruments (a wiki page has been prepared in /users/rsg/arsf/scratch_space/wiki_text_DELETE_AFTER_BECOMES_LIVE/Owl_Wavelength_Calibration.docx).

This ticket also records the first attempt at comparing the radiometric stability of the Owl over a year, with a percentage difference plot included in the data quality report. The process is documented in: /users/rsg/arsf/scratch_space/wiki_text_DELETE_AFTER_BECOMES_LIVE/Owl_Radiometric_Calibration.docx

[lah]

The script used (and developed) is: calc_owl_wavelength_offset.py in the owl subdirectory of libarsfcal

[lah]

Multiple files were collected at different integration times, but there are only 2 complete measurements of the same sample. The commands used were:

calc_owl_wavelength_offset.py -r /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_Ref_Int1_00_2017-04-19_13-38-30/capture/Wavelength_80_Ref_Int1_00_2017-04-19_13-38-30.raw -s /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_PP_Filter_Int1_00_2017-04-19_13-48-37/capture/Wavelength_80_PP_Filter_Int1_00_2017-04-19_13-48-37.raw -c /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_Black_Int1_00_2017-04-19_13-51-16/capture/Wavelength_80_Black_Int1_00_2017-04-19_13-51-16.raw -e -t /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Docs/plasticTruthFSF.txt -o ~/scratch_network/transfer/Ttests/

calc_owl_wavelength_offset.py -r /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_Ref_Int1_10_2017-04-19_14-06-48/capture/Wavelength_80_Ref_Int1_10_2017-04-19_14-06-48.raw -s /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_PPsample_Int1_10_2017-04-19_14-08-46/capture/Wavelength_80_PPsample_Int1_10_2017-04-19_14-08-46.raw -c /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Raw/19.04.17/wavelength/Wavelength_80_Black_Int1_10_2017-04-19_14-10-38/capture/Wavelength_80_Black_Int1_10_2017-04-19_14-10-38.raw -e -t /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Docs/plasticTruthFSF.txt -o ~/scratch_network/transfer/Ttests/

[lah]

The results were:
PP_Filter: rms = 0.024, offset 6.836 nm
PPsample: rms = 0.027, offset 0.610 nm

Given the difference between the 2 measurements the offset measurement is not conclusive.

Manual inspection of the peak values suggests that fityk is correctly identifying the peaks and that the algorithm is working properly, so the difference may just be due to the spectral resolution of the Owl (50 nm).

More measurements should be taken in future for comparison, but with such small values compared to the spectral resolution, the wavelength shift is not currently of major concern.

[lah]

The output files have been stored here: /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Processed/wavelength_transmittance

[lah]

The merge request for the script is ​https://gitlab.rsg.pml.ac.uk/arsf/libarsfcal/merge_requests/13

[lah]

script committed under version 1.0.2

[lah]

Wavelength calibration plot was recreated with a different legend for inclusion in the data quality report.
​https://gitlab.rsg.pml.ac.uk/arsf/libarsfcal/merge_requests/19
​https://gitlab.rsg.pml.ac.uk/arsf/arsfhome/merge_requests/13

Will tag once radiometric comparison has been merged.

[lah]

Radiometric comparison

Plotted results using script under development: ​https://gitlab.rsg.pml.ac.uk/arsf/libarsfcal/merge_requests/17.

Results stored in:/users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Owl_Variation/Cal_comparisons. Includes percentage difference plots for daily, yearly and the internal black body calibration. Also tabulated the median values (to ignore blinker spikes) to examine trends.

Temperature of 30 degree plot used for data quality report as this is closest to real data settings (and struggle to go lower than 30 degrees in the lab).

Example command:
compareOwl.py -o /users/rsg/arsf/arsf_data/2016/misc/2016_cal_feb_cambridge/processedOWL/d3/BBradiance/BBradiance30_2016-02-03_12-38-56/BBradiance30_2016-02-03_12-38-56_calibration.rad -n /users/rsg/arsf/arsf_data/2017/misc/Owl_pre_BB_cal_Feb/Day1/Processed/Blackbody_30_2017-02-06_15-01-08/Blackbody_30_2017-02-06_15-01-08_calibration.rad -c -s /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Owl_Variation/Cal_comparisons/Yearly/Feb3_Feb6_30.png -z 90 90 -l 1

[lah]

Owl Data Quality Report

Discussed the radiometric comparison results with Chris (FSF) and decided to not include these in the data quality report. We should wait until we have established a repeatable method, where the distances between the Owl and the laboratory black body are tightly fixed. This should be achievable in the next calibration using the newly finalised (Apr 2017 calibration) calibration jig. Not using the calibration file results as we can't explain the pattern.

Looking at the differences in the Apr calibration (12th - 19th) using the calibration jig doesn't show a massive improvement. There are still differences between the different bands. Plots are stored: /users/rsg/arsf/arsf_data/2017/misc/Owl_post_BB_cal_Apr/Owl_Variation/Lab_BB_comparisons/apr_comparisons

[lah]

Tagged ​https://gitlab.rsg.pml.ac.uk/arsf/libarsfcal/tags/libarsfcal-1.0.3