I spent the morning playing the picomotor game, to improve as much as possible the beam centering on the ISS photodiodes. The procedure is the same explained in 14211: I moved picomotor #8 and then used a python script to recenter the beam on the QPD using picomotor #1. My figure of merit was the sum of the power read by all photodiodes. I improved the total power by something like 0.5%, which isn't impressive, but at least now I know that I'm very close to a maximum. For future reference, the total sum of all PD signals, in counts, as measured by the IOP-PSL0_MADC1_EPICS_CH24-31, is about 40050.

After this, I injected a 1 Hz line on IM3 pitch and then yaw, and looked at the variation of the PD powers. I moved a bit the beam with both picomotors, to reduce the oscillation that was clearly visible only in PD2 and PD4. Now I'm close to be dominated by the two omega oscillation in both diodes. The optimal position does not correspond to a exactly centered QPD: PIT = -0.06, YAW = 0.35. Note that PIT and YAW for the QPD are reversed with respect to PIT and YAW for IM3!

Then, I took some (medium) high resolution spectra of all PDs, to estimate the coupling of beam motion to RIN. I used a IM3 pitch excitation at 1 Hz, with an amplitude of 300 cts. This gives me a signal on the QPD of 0.67 1/rHz, corresponding to 105 um/rHz. I used the same frequency for yaw, but with an amplitude of 200 cts, corresponding to 0.80 1/rHz on the QPD or 125 um/rHz. Spectra of all signals are shown in the attached figures.

Here are the results:

In general, those numbers are not too much different from the best we could get in the past. The only exception is PD4, which looks quite bad.