I am attaching some plots that I think can be useful to us while doing the CARM offset reduction. When we are using the arm transmitted powers for the CARM error signal, the CARM offset that a particular TR_CARM offset gives us depends on the recycling gain, and the reflected power dip can be used to infer both the power recycling gain and the CARM offset.

These are similar to the plots that Gabriele added in his comment on 43344, but I have expanded the range of arm losses plotted and changed the axis so that we will be able to judge what our PRG is even in the first few steps of the CARM offset reduction while the transmitted power is still low. In this case I am setting the PRC losses to 0 and assigning all losses to the arm cavities, as Gabriele showed this method can't distinguish between arm cavity losses and losses in the PRC.

The first attached plot can be used early in the CARM offset reduction sequence, and includes rather bad power recycling gains.  In our locking attempts last week, we reached an arm transmission 4 times that of a single arm lock, and saw the reflected power at 95% of its off resonance value, and 98% after improving alignment 62054.  This corresponds to a power recycling gain of 4 after the improvement, and worse before.

The second attached plot also shows the arm transmitted power vs reflected power for different power recycling gains, but plotted to show how the late steps of the CARM offset reduction will look when we have good or decent power recycling gains.

The final two plots can be used to estimate what the CARM offset is as a function of the single arm normalized transmission, for different recycling gains. We transition DARM control from ALS to AS45 at a single arm transmission of 4, before the changes in recycling gain make much of a difference to the CARM offset.  As we go to smaller CARM offsets the power recycling gain makes larger differences in the true CARM offset we get for a fixed arm transmission, which is why our locking sequence becomes more sensitive to changes in recycling gain as we get closer to resonance. Stefan wrote some guardian code that adjusts the transmitted arm power at which we transition CARM control to REFL to get around this problem 44075

The code used to make these plots in is a directory I created in Craig's git repo: aligo_commisioning/power_budget