Daniel, Terry, Nutsinee
We temporary locked LO loop (with PSL LO) using additive offset on TTFSS just so we can figure out why is it so noisy (slow path still output to OPO PZT, fast goes to TTFSS AO). We saw 3MHz leaking out of PFD IMON and it also shows up in LO common mode board fast control signal. Some filters need fixing.
LO loop OLG transfer function shows UFG at 63kHz (without 4Hz/400Hz slow path compensation). Transfer function at fast path shows crossover ~700Hz (with 4Hz/400Hz). Within a factor of 2 of what we calculated.
With the original Pomona box (alog44446) we would've needed a gain of 36dB on the fast path for a reasonable cross over. Daniel doodly clipped a 220Ohm resister in parallel to the existing 1kHz to give us 14dB less gain in the slow output to PZT. This allow fast gain to be within +32dB slide bar range.
One of the issues we discovered was that the TTFSS was modified to have a pole/zero and more gain in the sensing path. This was done to make the OPO look more like a reference cavity with a 70kHz pole. However, this reduced the AO path gain by 20. To get this gain back we need to move the compensation to the next OpAmp, after the AO is added; see LIGO-E1800283.
The second issue is with the PFD and the 3.125 MHz LO. The PFD filter is designed for LO frequencies >9MHz. At 3.125 MHz there is only 12dB of attenuation which results in a 1V LO signal leaking into the servo board. We can add a notch filter by stuffing C14 and C19 with ~260pF on LIGO-D1002471.