Georgia, Craig

- Lost lock twice in CARM_TO_ANALOG, once while trying to increase the Sum node REFLAIR9I gain to 23 dB, and once when trying to engage the CARM Servo compensation boost.  Managed to get through by slowing the increase of the REFLAIR9I gain, and increasing the final gain on sum node to 28 dB.  Made changes to guardian, now able to successfully request PREP_ASC_FOR_FULL_IFO (made it through directly twice).

- Increased DARM offset to 9e-5.  Had to do this to lock the OMC to be over 11 mA on the DCPDs.  Still not very robust, our last lockloss came while we were trying to relock the OMC by hand after it had locked on a 45 MHz sideband.

- Tried to get to DC READOUT and failed once.  We measured the TF between LSC-OMC_DC_OUT and ASC-AS_A_RF45_Q_SUM_OUT and got a ratio of 3.36e7 at 100 Hz with 150 degrees of phase.  The input matrix element for ASAIR_A_RF45_Q to DARM was -1e-6, so we put 33.6 into the OMC DC matrix element and tried to make the transition.  It is extremely unclear what ASAIR_A_RF45_Q actually is.  Georgia and I looked in the matlab models and in h1lsc we found that LSC-AS_A_RF45_I/Q_SUM goes to ASAIR_A_RF45_I/Q.  Not clear yet if ASC-AS_A_RF45_Q_SUM_OUT is the same as LSC-AS_A_RF45_Q_SUM.

- ALS_DIFF is intermittently noisy when we lock ALS.  We did a quick test where we locked only ALS_DIFF while ALS_COMM floated about.  We discovered that DARM becomes more quiet when the COMM PLL unlocked.  Plot 1 shows a striptool where DARM is green and COMM PLL control voltage is blue.  It's unclear to us right now why the COMM PLL should have any effect on DIFF noise.  

- ALS_YARM is struggling to reacquire upon locklosses.  Georgia keeps restoring ETMY and TMSY to their OSEM witnesses in order to lock.

- PRM yaw is bad when locking DRMI.

- PR2 has bad scattering.  Pic below.