Due to Sebastians work today, the Xarm is locking more stably
(for a few minutes at a time, rather than seconds).  We still have a problem with mode hopping to the 01 mode. 

I tuned the demod phase by watching the transfer function from the common excitation on the PDH locking board to the Q phase signal and trying to minimize it.  I think this was somewhat sucesfull, although it was made more difficult by the cavity often loosing lock.  I changed the demod phase from 88deg to 110 deg.  There is clearly a signal in the Q phase of our locking signal, which I couldn't really make go away by changin the demod phase.  It is clearly in phase with the power fluctuations of the transmitted light, and has coherence with the pitch oplevs at 0.5 Hz. (plots attached)

So we still have a pitch problem, even though we have the ISIs working at about their nominal performance, and the initial alingment servos are running.