# On 13/08/19 12H UTC we have installed the Master laser frequency monitor (MLFM).

This system is meant to independently monitor the freq of the ML and make some correlation with EOM_CORR glitches in order to state whether the ML is part of the issue in the fast unlocks saga.

- It is a Mach-Zehnder with an arm made with a 105 m long silica fiber (refractive index = 1,44).

- The releavant signals of the MLFM are the Error signal: ML_FREQ_Q_MONIT_FS and the correction signal: ML_FREQ_I_MONIT_FS

ML_FREQ_Q_MONIT_FS = 0.1 V * cos(phi) - 0.025 V  (1)

with phi = 2pi.d_nu/c*105*1.44 = 3.16e-6rd/Hz  or d_nu = 1.98 MHz per fringe (2.pi phase)  (2) 

d_nu is the frequency drift of the laser undr study (ML)

The MLFM can be locked using a fiber stretcher integrated in the fibered long arm. this associated correction signal is ML_FREQ_I_MONIT_FS. Its sensitivity is 50*0.305 rd/V.

The MLFM can be let unlocked with the error signal freely swinging. This config avoids the possible saturation of the fiber stretcher actuation when long, slow drift occurs (Temp, pressure...).

at ~9H LT we let the MLFM locked. The UGF is ~150 Hz. In this config the frequency drift is visible in the corrrection signal. This drift comes from the frequency of the laser but also from the long fiber susceptibility wrt the environenment (Temp, Hu, Pressure). Those drifts are quite slow wrt to the EOM glitches we want to monitor and shall not be considered as showstoppers.

It ends up into saturation of the fiber stretcher actuation after a couple of hours. The error signal is anyway still usable. One just needs to consider the formula (1, 2) to get "phi" or "d_nu".

We wait for a hardware logic which will reset the servo once clamped into saturation, and will allow to get the error signal always aroud zero. So far we let the system as it is (correction saturated, Error signal is swinging).

At the next ITF unlock It is worth opening the loop in order to avoid possible unexpected feedback of the actuation on the MLFM.

# the plot shows the "lg term" drift of the MLFM (due to both fiber and laser). We see it is well correlated with the lab pressure, as expected. The senitivity is quite large : 10 V on the stretcher (that is 500V*0.305 rd/V = 150 rd) for a pressure variaiton of 0,2 hPa. It demonstrates there is no way that the actuation dynamic could be sufficient to keep the sytem locked for very long period.

Again, this is no showstopper since the goal being to get an error signal during fast EOM glitches occurence.