Craig, Georgia

This afternoon we could not make it through ENGAGE_DRMI_ASC without (1) the verbal alarms saying that IX had saturated, and (2) a 0.5 Hz oscillation ringing up when the beamsplitter oplev servo was turned off.

This first problem was baffling since nothing should be feeding back to IX at this early stage in the lock acquisition sequence. We had a look at the ITMX top mass (M0) and reaction chain (R0), and found some strange behaviour on the reaction chain OSEMs. Before we had begun working on alignment this afternoon, while the interferometer was just sitting trying to lock the arms in green, the sensor input of the SD (transverse) OSEM had plummeted, and the noise on the other OSEMs (particularly LF and RT) had increased. First attachment shows inputs to the R0 OSEM input filters at this time. Second attachment shows M0 and R0 master out signals (bottom 3 rows are R0) which go to the coil drivers.

I had a look at the M0 and R0 channels for the other test masses and didn't see anything similar.

We tried adding positive and negative offsets in the test filters, which adds an offset to the coil driver and should be seen in the sensor. We expected that if the we were close to the edge of the shadow sensor, an offset in one direction would be visible in the sensor. Our offsets were not seen in the SD sensor (though there was some coupling to the other sensors) as shown in the third attachment. Does this suggest that the light source for this particular OSEM has died? Or are we missing something?

We also noted that when turning off the damping feedback to this sensor the noise in the other sensors was reduced (visible early in the time series in the third attachment). We have left this feedback off, and managed to reach lownoise_length_control without a problem.

(We never solved the 0.5 Hz oscillation problem when engaging DRMI ASC. We had resolved let the IFO lose lock at this step and work on single bounce RF9 modulation problems, when for no apparent reason we succeeded in locking, got up to lownoise_length_control, and lost lock due to a 4.7 Hz ETMX L2/L3 oscillation that caused many locklosses yesterday.)