I restarted the squeezer electronics and did some checks. While turning on stuff in the FC room I clumsily dropped an oscilloscope on the floor. It still works but has some 45 kHz 4 mV noise when unplugged. Well it doesn't get used for critical measurements anyway so maybe ok. It has the yellow sticker "RM alignment feedback". I tested it using GRMC DEMOD (which was moved now to DDS2 DAC0) in both input channels and it gives basically the same result as the other oscilloscope.
SHG
The SHG was very misaligned, which was a bit strange considering that it usually maintains its alignment for quite a long time. It started out at about 40% mode matching. I used the steering mirrors and XY lens mount and realigned it to (1960 mV, 280 mV, 80 mV -> 84.5%), which is a little bit less than the reference alignment from 2024-12-25 (1840 mV, 200 mV, 140 mV -> 86.8%). I tried also the waveplate to correct the polarization but it had no effect. I also checked the input IR which reads 805 mW on the stick sensor vs 800 mW on 2024-12-25. I locked it and optimized the green output power to 296 mW (temperature controller 3.191) before the green FI.
GRMC
I checkced the GRMC alignment, it was ok, basically where I left it, so I didn't tweak it. I checked the error signal and optimized the demodulation phase to i-phase, then locked. I could see that it stably outputs 25.3 mW (316 mV on oscilloscope) of green power, checked at the OPO steering mirror 1 using the ring sensor, which is stable at the third decimal place on the power meter. I tweaked the servo gain to bring it to the threshold of oscillation (1.1) then reduced it (0.75). Also at one point the power meter randomly spiked to 44 mW, I'm not sure if this was due to the actual green power or just the power meter electronics being weird.
As presented in the research meeting there was an issue with GRMC unlock caused by the demodulation phase seemingly flipping from i-phase to q-phase. By switching channels to DDS2 DAC0, I can see that the problem is fixed. GRMC can maintain lock for > 90 min with the same output power and power stability (at least as seen on the power meter). In other words, the previously used channel DDS2 DAC2 is drifting in phase with respect to other channels. There is a bit of an annoying problem where when the GRMC/MZ assembly is manually unlocked it can sometimes kick one of the PZTs along the green path and requires the MZ to be realigned.
The PDH signal has about 100 mV DC offset as well which is a bit strange. But the error signal is 2.08 Vpk so it doesn't cause any immediately apparent problems.
IRMC
Mode matching was checked (1210 mV, 72 mV -> 94.4%) and was good. PDH signal is the same size as last time I checked (400 mVpk). I locked and saw it transmits 2.30 mV, same as normal.
Homodyne
Homodyne balance was quite bad, about -17 mV. I realigned it to zero, making sure to check the lenses in front of the homodyne PDs to confirm that there is no clipping of the beam. I checked the homodyne spectrum which gives the usual ~ -132 dBVrms/rtHz at high frequencies > 5 kHz. It seems to be about -131.8 dBVrms/rtHz at low frequencies, I feel it is maybe a bit noisy but I guess the squeezing measurement will determine the real performance.
to do: PLLs, OPO, nonlinear gain, squeezing