I measured the beam pointing sensitivity of all the eight ISS second loop diodes. The naming is the one corresponding to the new slow channel acquisition, we'll check tomorrow the correspondence to the PD1-8 signals. For some diodes, in pitch, I was not able to see any significant modulation

For those not familiar with the meaning of these numbers, the typical values measured for the ISS array before installation were between 1 and 30, depending on the array and on the diode. So the numbers I got are in general quite large.

This beam position on the PD is maybe not the best one, since we are not at the maximum power for all the diodes simultaneously. For each diode, it is possibile to find a beam position that gives maximum power. This also corresponds to undetectable coupling of beam motion to dP/P (at least at this level of excitation). However, the good position is different for each diode, and the non optimal one can have a significant loss of power and large coupling of beam motion to dP/P. The table above is representative of what we normally get.

Here is the procedure in details

1. Move the beam using the picomotors to center the QPD. Dither IM3 with an amplitude of 100 counts at 1 Hz in both pitch and yaw (H1:SUS-IM3_M1_OPTICALIGN_P_EXC and H1:SUS-IM3_M1_OPTICALIGN_Y_EXC), and measure the normalized signal on the QPD. I got 0.14 peak to peak in pitch and 0.22 peak to peak in yaw. The beam size on the QPD should be 250 um. The normalized signal should be given by sqrt(8/pi) * deltaX / w where deltaX is the beam motion on the QPD and w is the beam size. Therefore, 100 cts at 1 Hz of IM3 correspond to 22 um in pitch and 35 um in yaw of beam motion on the QPD. I assume that the motion on the photodiode is the same. 
2. With the dither on, I measured the mean value of each of the photodiode and the peak to peak oscillation at the dither frequency, separately in pitch and yaw. I had to move pitch of 300 counts (70 um beam motion on the diodes) and yaw of 100 counts (35 um of beam motion on the diodes). Larger amplitudes were giving weird signals.
3. From the photodiode signals I computed the dP/P and then the dP/P/x reported above