I have measured the sensing matrix at a variety of PRCL offset values.

During this each measurement, I also took a 20 second average of the POP 2f signals and the ASDC signal:

All of this data was taken during a single lock stretch.
If / when I do this again, I want to go out to larger offsets. I won't take as many points, but I do want to see how far I can go before I lose lock, and what the phase separation looks like at larger offset values (this time, I stopped at +700 counts which is about 0.7nm, to start checking the negative values. MC has been unhappy, so I wasn't able to take very many negative offset values.)
I conclude that these sensing matrix measurements do see changes in the phase separation with PRCL length offset (what we saw / said yesterday), but that they do not line up with Q's simulation from this afternoon in elog 9671.
The simulation says that we shouldn't be seeing large phase changes until we get out to several nanometers, however the measurement is showing that we get large phase chnages with picometer scale offsets. Yesterday, Rana and I said that the offsets due to RAM were small (of order picometer), and that they were therefore likely not important (elog 9668). However, now it seems that the RAM is causing significant length offsets which then cause poor MICH/PRCL phase separation.
To Do List:
* Confirm MIST simulation with Optickle.
* Look at sensing matrix data pre-lockins (in the raw sensors).
* Check that there is no clipping anywhere in the REFL path (at least out of vacuum), and that the beam is sufficiently small on all 4 REFL diodes.
* Calculate the new PRC g-factor with the new length. |