UCoMP instrument efficiency by wavelength.

[bberkeyU]

We need to characterize the UCoMP hardware vs. wavelength. Currently, the pipeline assumes the solar spectrum is flat, and the pre-filters have no spectral structures. However, the real sun has a deep absorption line on the disk that makes it into the flats, and our filter has some structures. We need to discuss how to teach the pipeline how to calibrate these issues.

To get things started here, here are some plots of the relevant transmission profiles and real L0 flat data.

Above plots produced from: https://github.com/NCAR/ucomp-studies/blob/master/transmission/UCoMP_compare_data_to_synthetic_profiles.ipynb

@detoma @mgalloy @jburkepile

[bberkeyU]

I think I correctly convolved the atlas and prefilter together. The only area for possible improvement is being a little more thoughtful about mapping the wavelength between the two profiles. We may want to upsample or interpolate if there isn't a perfect agreement between the reported wavelengths. For my analysis, I just chose the closest.

I need clarification on properly combining the whole lyot filter bandpass since we get a different (shifted) profile for each wavelength tuning that must be separately convolved with the atlas and prefilter.

I produced an on and off-band profile for each wavelength tuning. I then convolved the profiles with the combined atlas/prefilter transmission provided and summed the total transmission across the whole lyot filter profile. I then normalized this with a lyot filter profile not convolved with the atlas and profiler. Performing this at all points with a 0.05nm spacing between 1070.7 and 1078.7 nm, I got the "Convolved atlas, prefilter, and lyot filter tuning transmission profiles vs recipe tuning" plot seen above.