A few other things tonight:

• I added to the PRMI locking states lowering the MICH trigger tresholds, this is something I often do by hand when the alignment is not great to help lock.  I also added some lines to LOCK_DRMI_1F to set them back to the DRMI values. 
• We tried turning off the PRC1 loop to see if that would have an impact on the PRCL/SRCL/MICH noises.  The motivation for this is that we can see coherence of the LSC signals with the POP QPD, which mostly seems to be due to fluctuations in the sum. As expected, this didn't have much impact but there was an improvement in the PRCL noise below 20 Hz.   
• To see if it is plausible that lock point errors in the REFL9I sensor are the cause of our jitter noise problems, I measured a quick transfer function from REFL control to DARM.  By looking at the screenshot, you can see that the transfer function from REFL control to DARM is high enough that if you assume that the entire spectrum of REFL control is due to lock point errors (probably not true) the noise you would have in DARM is almost a factor of 10 above the level of noise in most of the bucket. 
• I remeasured the REFL 9I to DARM transfer function, once several hours into a lock and once in the first few minutes after arriving at nominal low noise (second attachment). The coupling changes by a factor of 3, and has a different slope than the coupling measured by Evan Oct 3rd. The last attachment shows the spectra at the times of these measurements (approximately). 

A little bit of ASC work:  

• The changes to the REFL WFS centering servos I made in 29990 were being undone by the REFL WFS centering state in SIC_GEN states (only used in inital alignment).  I've removed the offending logic. 
• I didn't find any signal for SRM Pit in the REFL 45 Q signals, which is what the ASC design says we should use.  Even with the centering loops not oscillating there is a large fluctuation in all the refl WFS at around 0.2Hz, which could just be the optic motions.  It doesn't seem to be caused by a loop instability of INP1, PRC1, PRC2, or CHARD. 
• I broke the lock trying to use the template we have used in the past for CSOFT P, I started to retune the amplitudes but haven't been able to finish.  
• A new CHARD P measurement for low noise is in the svn.