Mark B. (posted under Arnaud's name by mistake)

In response to a question by Dennis and Norna, I set out to confirm that the magnets for the L1 and L2 OSEMs on the ITMy and ETMy were installed with the proper checkerboard pattern called out in E1000617. The sensing at these levels has had some attention from Szymon but the actuation less so.

I set up some DTT jobs that applied 0-6 Hz uniform excitation to the EXC test points in the COILOUTF filter blocks (H1:SUS-ITMY_L1_COILOUTF_UL_EXC etc) and looked at the phase, magnitude and coherence to the IN1 test points of the OSEMINF blocks (H1:SUS-ITMY_L_OSEMINF_UL_IN1_DQ etc):

/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ITMY/SAGL1/Data/20130301-ITMY_L1_SignCheck.xml

/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ITMY/SAGL2/Data/20130301-ITMY_L2_SignCheck.xml

/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1ETMY/SAGL1/Data/20130301-ETMY_L1_SignCheck.xml

/ligo/svncommon/SusSVN/sus/trunk/QUAD/H1/ETMY/SAGL2/Data/20130301-ETMY_L2_SignCheck.xml

The four OSEMs at each level were done in parallel. To get good coherence given the weakness of the lower actuators I had to use large excitation levels and increase the WD actuation thresholds temporarily from 25000 to 50000. Damping was off, but preliminary runs showed that it didn't make much difference.

When I found a good combination of settings, the phase below about 0.3 Hz was near either 0 or 180 degrees and tracked the sign of the gain field in the COILOUTF block.

For both ITMY and ETMY the L1 signs to give 0 degrees phase at low frequency (i.e., actuation and sensing with the same sign in the spring-constant dominated regime) were UL=LR=-1, LL=UR=+1. This is consistent with E1000617 with N=+1.

For both ITMY and ETMY the L2 signs to give 0 degrees phase at low frequency were UL=LR=+1, LL=UR=-1. This is consistent with E1000617 with S=+1. Apparently the AOSEMs are wound the opposite direction from the BOSEMs, or there is some similar consistent sign swap. This is _not_ the way the gain had been set - all L2 OSEMs had been configured as pullers rather than pushers as intended. I left the corrected settings in place and updated the safe.snap files.

The only case where the conclusion is marginal is for ETMY:L2:UR, where the OSEM is indpendently suspected to be badly aligned, so the response is weak and the coherence is not as good. We hope to align this OSEM better when the ETMY is reinstalled in the H1 tank.

This study does not constrain the second magnet in each flag assembly, which is supposed to be installed oppositely so that each flag assembly has zero net dipole.