This morning I did the usual noiise injections (logfile 1223376492), in the figures there the usual OLTFs and the noise budget. The PRCL gain has been reduced by ~ 10% in LOW_NOISE_3.
The main difference with the previous measurement is the MICH loop: I had to add a whole new 300 us delay in order to fit correctly the phase, and the strange behaviour between 40 Hz and 80 Hz is still present and a bit worse. To be investigated.