Sheila, Jenne, Stefan

TOday we changed the calculation of TR CARM in the LSC model.  The model had a saturator before the square root, when we first turn on TR CARM we can be at low enough arm powers that we are below the lower limit of the saturator, (1e-4) when the powers increase above that the loop can get a large transient.  To smooth this out, we added a second path that has a saturator with an upper limit of 1e-4, subtracts a fixed offset  of 1e-4 and applies a fixed gain of 50 and is summed with the original path.  

This means that for values of the arm transmission above 1e4, we have the same signal as before, and for values below that we have a signal that is just the sum of the arm powers.  This should not cause the kinds of transients we were seeing last night.