We had a problem with the ISS tonight, which was probably caused by the second loop. We haven't taken the time to look into it but have disabled the second loop by turning off the input H1:PSL-ISS_SECONDLOOP_ENABLE, which made first loop stop railing. Attached is a screenshot.
One problem was found and fixed. Feed forward of AC coupling was suspected at first but turned out to be OK, but I adjusted it anyway.
1. Slow offset servo was super aggressive.
The main problem was that the output matrix element for diffraction part of the slow offset servo H1:PSL-ISS_SECONDLOOP_REFERENCE_IN_MTRX_1_4 was set on Oct/23 to be a huge number (0.5). Behavior of the main ISS part of the 2nd loop is dependent on the offset servo's behavior at and below its UGF, and you don't want to set it super aggressive or you might suffer when there's a slight change in the main ISS part.
This loop is there to bring the diffraction back to a pre-defined number (-1*H1:PSL-ISS_SECONDLOOP_REFERENCE_DFR_CAL_OFFSET) VERY slowly, but the gain of 0.5 is like the UGF of 0.1Hz-ish and that's really too much. Even when I reduced this from 0.5 to 0.15 there was a clear gain peaking. I set it to 0.05.
(But this number was 5e-5 before, which is overly small.)
Also, I set H1:PSL-ISS_SECONDLOOP_AC_COUPLING_OFFSET to 2 so the diffraction becomes about 2.3 with AC coupling ON but without slow offset servo.
Finally, 10Hz pole of this path was off (H1:PSL-ISS_THIRDLOOP_POLE) and I switched it ON because we're not really doing anything at around 1Hz and faster.
2. Feed forward path of AC coupling.
Since PSL commissioning/maintenance, ISS 1st loop diode power level dropped by about 15% though the IMC in and transmission only dropped by 10% or so. This is OK, but I was annoyed that H1:PSL-ISS_PD_NORM_OUT was about 0.84 because if the power decreases by another 5% and it will hit the lower limit of 0.8.
I changed FM2 of H1:PSL-ISS_PD_NORM so the output becomes close to 1.
We enabled the ISS second loop again and we found that it started ringing up again. At -120s I switched it off, guardian switched it back on again around -80. We disabled it again.
That's because my changes were reverted back by something. Especially H1:PSL-ISS_SECONDLOOP_REFERENCE_IN_MTRX_1_4 was back to 0.5 again.