Windy days are bad for locking, but good for other side activites. So I sat down to write a script that automatically tune the IMC aungular offsets to reduce to minimum the coupling of jitter to intensity noise. It basically follows the same procedure I performed manually earlier today. The script is attached to the elog entry. You can sae it and run it from terminal. it will open up a plot showing you the progress of the tuning. Run the script with the IMC locked, no IFO and ISS second loop open!

Here is what the script does

1. Turn on excitations on the PSL PZT at 101 Hz (pitch) and 153 Hz (yaw). The excitation is removed when the script exits (even if you kill it with CTRL+C)
2. Open up a window that shows you the progress of the tuning. There are three panels:
   1. Error signals for the tuning: real part of the transfer function of ISS_SECONDLOOP_SUM14 over excitations, both pitch and yaw. if the tuning is working, these two traces will be driven to zero
   2. Monitor of the band-limited RMS of ISS_SECONDLOOP_SUM14  in the 200-400 Hz region. this is where most of the PSL periscope noise sits, so if everything goes well you'll see this trace dropping down close to zero (typical values today when the alignment was good were 0.01-0.02 a.u.)
   3. Monitor of the IMC WFS offsets. The tuning acts on IMC-DOF_2_P_OFFSET and IMC-DOF_1_YAW_OFFSET, since those seem to be the most important offsets
3. After an intiial wait of 30 seconds, the script will start servoing the error signals to zero by changing the offsets. Basically, it reduces to minimum the coupling of the two PZT angular lines to the power transmited by the IMC (as seen by the ISS array)

The script has a built in check of the IMC power, and it doesn't change any offset if the IMC is not locked. There is no check on the convergence of the error signals, but the script terminates automatically after 10 minutes, which is typically more than enough to find the right offsets. However, you can terminate the script with CTRL+C when you're satisfied withe the results. The excitations will be automatically removed.

I tested the script several times with random starting offsets, and it worked all the times.

The first attached plots show a comparison of IMC transmitted RIN before and after the script was run. The other plots show some examples of the plots generated by the script, showing the tuning in action.