Busy week. Lots of work.

So, Here I am at Salt Lake city airport, on my way to APO to do the 2016 Summer shakedown calibrations.
Last week has been quite busy. Made a lot of progress with the SoS ELG problem. The new improvement is that we are now using the clean wavelengths in the z-band to estimate the S/N of the red cameras. I implemented this in the SoS and looked at the S/N for the three plates in the SGC (Plate numbers : 8763,8767 and 8768) and three plates in the NGC (Plate numbers : 8957,8954 and 8958). The S/N ratios does improve as compared to the normal procedure of using the i-band wavelengths. Remember, when we used the whole z-band wavelengths, the S/N values had actually worsened. The improvement in S/N with clean z-band is promising. But, not enough.

I presented my results in the eBOSS pipeline telecon. Stephen probably was sceptical about the new approach I tried, i.e, fitting the standard stars and using the coefficients to estimate the S/N at some fiducial magnitude. The main concern was that, in the SGC plates , normal LRG/QSO plates a discrepancy between R1 and R2 cameras SN2 values. The cause was traced out to the ADR effects affecting the two spectrograph differently. But, the new approach SN2 values in R1 and R2 are almost same. In other words, the new approach is not sensitive to the ADR discrepancy that we saw in the normal LRG/QSO plates. This seems to be a valid suspicion. But, the clean z-band approach has the ADR discrepancy. So, it was agreed commonly that the clean z-band approach is a more correct solution. I proposed to run Julian's routine on the clean z-band approach. Julian's routine doesn't fix the slope of the fitting. Instead, it uses all targets with in the broad magnitude range 16-24 (Old range was 21-22). This will ensure that there is a smooth transition between poisson dominated noise regime to sky dominated noise regime. So, we can make use of the standard stars in the fitting. This should definitely improve the SN2 values. Fingers crossed.