QUBIC

Minutes:

• Sky dips with varying azimuth (Steve):
  - plot done with b.plot_timelin(TES=96, plot_Tbath=True, plot_azel=True, plot_1K=True)
  - the impact of azimuth on data is explained by Earth magnetic field
  - it might have an effect on elevation
  - weird correlation between azimuth change and T1K change

• PCA analysis (Noah):
  - hysteresis effect seen, could be created by some metal parts? near the squids? they are not seen on all TES
  - check if the effect is seen on TES by groups of 4 (they are physically stacked in the squids)
  - or same by groups of 32 (there are 4 boxes of squids stacked on top of each other with 32 squids in each)
  - might use a neural network to predict the azimuth-related noise and then substract it from data
  - discuss with Damien, Manuel, Michel about squids

• hyperparametrisation (Tom):
  - there is a bug in the fitting code, with walker diverging in some cases
  - it seems to be a bug, one that is quite complicated to find because it works fine sometimes
  - the more Planck bands we put in, the worst it gets
  - with QUBIC only, it works sometimes and fail some others