In order to check the new Faradays and measure the mode matching of the squeezer beam, we need to generate an IR beam from the OPO in the optics lab.  We are doing this using a REFL PDH lock on the IR beam injected through M2 (normally the seed or CLF), rather than locking using 532nm and controlling the crystal temperature.  The cavity is very undercoupled when injecting through M2 which transmitts 0.1%.

I am using a New Focus 4004 phase modulator, 15mrad/V (Manual), with the reflected beam off the PBS used for IR power control which cleans up the polarization to avoid RFAM (first attachment shows this part of the setup). The modulation is at 60MHz, 3 Vpp is split using ZSC-2-1+ (-3dB), sent to the mixer and attenuated by 10dB, and I am using a ZHL-1A RF amp with 16dB of gain to drive the EOM. This means the EOM is driven with 13.4Vpp, 100mrad peak, and gamma should be about 0.05. 

I am injecting 50mW of IR into the thorlabs fiber, getting 25mW in reflection off of M2, and using a R=98% BS I measure 200uW with a power meter on an 1811 in reflection that we are using for the PDH (second attachment is a photo of this path). The DC output of the 1811 is 190mV, indicating that there is 270uW on the diode.  We are getting 75uW of power in transmission when the OPO is locked. 

The Prometheus that we are using is the ALS spare, which has been modified so that the noise eater can be remote controlled using beckhoff.  This means that the noise eater needs to be turned on by adding a jumper between the two pins on the cable where the noise eater switch normally is, without this there is a lot of noise at 700kHz which adds a fluctuating offset to the PDH signal.  This laser tends to mode hop, so I've moved the temperature around, it started at 26.16C.  I see evidence of mode hopping in the cavity scan at 25.9C and 22.3C, so I've set the laser temperature to 24.1C. 

I believe that there should not be a polarization rotation generated by the EOM, since the beam we are using is the reflection from a PBS.   Since I see offsets on the PDH signal even with the noise eater engaged, I tried to align the EOM to reduce RFAM by moving it in and out of the beam until I could place it in the beam without misaligning the beam into the fiber. This did seem to help reduce the wandering offset, but we still have about -30dBm of RFAM measured by blocking the cavity with an IR card and looking at the 60MHz signal on an RF analyzer.  This level of RFAM fluctuates over 10s of minutes, and is comparable to the level of RF we see from the cavity resonance, so it may cause difficulties for locking the cavity stably enough to measure the Faraday isolation and throughput.

The PDH signal from this set up is roughly 90mV peak to peak, using the 4.8MHz cavity pole and the approximation here  we have roughly a gain of 53MHz/V  (the error signal isn't carefully phased and the sidebands aren't well outside the cavity linewidth).  The PZT gain should be 16.5nm/V, or with the gain of 15V/V of the MDT694 PZT driver 395MHz per V applied to the PZT driver input. 

The OPO is locked using an SR785, with a single pole at 0.3Hz, a gain of 100 and no inversion, the OLG is shown in the blurry photo in the 3rd attachment, it is a simple loop with a ugf at 75Hz and 83 degrees of phase margin .  It seems stable for 10s of minutes yesterday afternoon, but we may encounter difficulties for the Faraday isolation and throughput measurements from the changing offset due to what seems like RFAM.