search

HERA-Team / hera_cal

Library for HERA data reduction, including redundant calibration, absolute calibration, and LST-binning.
MIT License
11 stars 8 forks source link

Investigate high variance in a few autocorrealtions #513

Open jsdillon opened 5 years ago

jsdillon commented 5 years ago

In looking at how much autocorrelations vary from day to day, we found a few antennas with markedly higher variance than others. This is a bit suspicious and should be investigated. In particular, I noticed (70, 'Jxx') was poorly behaved in zen.grp1.of1.LST.1.31552.HH.OCRSL.uvh5 when looking at the ratio of the STD over days to the LST-binned autos. See below:

They all show some spectral structure, but not a ton. Here the worst antennas (median ratio > 2%) are bolded and included in the legend.

image

I'm not sure if excluding 70x is the right answer... it might be an LST-dependent effect. It also might only affect the even or odd days. There's a lot to look into. @nkern curious if you have any thoughts?

_Originally posted by @jsdillon in https://github.com/HERA-Team/hera_cal/issues/490#issuecomment-515629422_

jsdillon commented 5 years ago

Discussed 8/21/19. Will need to be revisited for the Tier 3 paper.