I have realigned the primary beam from scratch; the spots are now centered fairly exquisitely on each of the cavity mirrors, as well as on all of the pre-cavity optics downstream of the input optics that Koji adjusted on Wednesday. I also aligned CW/CCW splitting PBS better.
I had somewhat of a difficult time getting everything to lock again very stably, but it appears to stay locked indefinitely now. Some notes:
- Regarding the RFAM, I wanted to verify that it was still minimal after I got things working again. I reduced the experiment input power to the point where the DC level on the REFL PD was ~100 mV, as in Koji's post. The RFAM had increased from -85 dBm to about -75 dBm, but this was still the minimum level achievable by rotating the HWP before the EOM. Though it has gotten worse slightly, the large RFAM swings we were seeing before appear to have gone, and the cavity stays locked, so it looks like Koji fixed it. The DC error offset on lock is ~20 mV in INPUT_MON, and so it's < 1 mV in the actual error signal.
- A somewhat more troubling issue is that, for some reason, the DC signal from the PD appears to oscillate at ~2 MHz for incident power levels corresponding to lower than a few hundred mV. At 100 mV, there is a peak-to-peak swing of almost 50 mV. This exists with and without the LO on.
- I had trouble keeping the loop under control at low frequency, so I increased the gain on the MEDM filter module for the thermal loop from 0.05 to 0.1, which it accepted without becoming unstable.
- I was still seeing some weird effects wherein the TRANS DC level on lock would hop up and then lose lock, or generally switch between levels. I thought this might have to do with some sideband coupling, so I increased the modulation frequency back to the original value of 33 MHz, which is fine now that we are using fast diodes again.
- Realign the AOM and extract the 1st-order double-passed beam
- Direct this to the cavity and maximize TEM00 coupling
- Reinstall second faraday isolator and reconfigure CW REFL readout
(side note: we still need a modematching telescope for the CW beam)
After this, I can go back to the other list I made concerning loop measurements and beyond..