I had a look at the refernce cavity today to see where we losing our gain. Initially I tweaked the alignment increasing it by approximately 10%, the 01-mode is now much dimmer on the monitor (I didn't do a scanning alignment, and tried to kept the cavity locked).

Also, I rotated the QWP just infront of the cavity to maximise the power on the RF LSC diode. In the beam just after the PBS, which taps the beam for the RF LSC diode, is a ~90/10 beam splitter (unlabelled), directing 90% of the beam to a CCD camera! Due to a lack of looking for a smaller splitting ratio beam splitter I left it inplace, so we now geting ~4 mW onto the RF LSC diode (of the 34.4 mW available in the return beam).

I dropped lock, so I could measure the power. Into the cavity is 46 mW, reflected is (out of the PBS) 34 mW. I measured the power incident on the transmitted diode but forgot to measure the total power transmitted! There is 4 mW on the RF LSC diode (reading 169 mV at the DC output). Once locked the power drops to 40 mV out of the DC output.

The UGF is 287 kHz, with the folowing Interface Settings

It runs with Common at 850 and Fast at 700 (UGF 315 kHz), but it becomes more sensitive to bumping and floor motion and starts to oscillate.

You can recover from the oscillation by reducing the Common gain to below 160, and bring it back up.

The relock of the RefCav is at a slightly different mode, asI tweaked the coarse tuning down wards to find the 00-mode. There is 210 micro Watts on the transmitted diode, reading ~828 mV (it has a gain setting of 20 dB).