Comments on Relative Gains


Instrumental polarization completely degenerate with relative gain
Therefore a non-issue for a pair diff experiment

We measure rel gains every 30 minutes
Fluctuating errors average down (T sometimes leaks to +Q, sometimes to -Q)

Even systematic errors average down due to observing at different angles
In a sim with zero intrinsic EE/BB and fore/aft gain flat at 1.03 EE/BB power 1e-4 of TT power