Reports 1-1 of 1 Clear search Modify search
AdV-SGD (FDS commissioning)
polini, sorrentino, vardaro - 20:48 Monday 10 January 2022 (54442) Print this report
Stray light active loop: higher bump

Last Friday we continued the work on the active stray light loop.

0. Alignment and visibility optimization

We started by cheking the HD visibility between LO and BAB.

We have the following powers:

  • BAB:
    • 0,39 V on HD PD1
    • 1.75 V on PD_MONI
  • LO:
    • 0.78 V on PD1
    • 3.69 V on PD_MONI

With this difference in power we can get a maximum visibility of 94%.

We started with 57% of visibility and then by moving SQB1_M13 and EQB1_M6, we reached ~80% (normalized by the powers).

 

1. Higher stray light bump on HD AUDIO DIFF channel

We noticed that the bump in the AUDIO DIFF channel of the homodyne, is larger than the previous time.

We compare the bump of the 7th January 2022 with what we had on 14th December 2021: see Fig. 1.

We can clearly see that the bump is larger, meaning that:

  • our system is moving more than last time, see Fig. 2: the seismic activity is larger the 7th January (blue)
  • we have a clip on the IR as shown from the fact that we are not able to restore the 1.5V in transmission from the cavity, but oly 0.9V.

We will further investigate the clipping that can be due to a different configuration of our system after Christmass holidays.
 

2. Dither line amplitude optimization

We optimized the amplitude of the dither line in order to have a good SNR on the error signal.

We acquired the following data on the 7th January 2022:

  • 10h56m12: sensing noise, no dither line;
  • 10h57m19: line amplitude = 5mV;
  • 10h59m19: line amplitude = 10mV;
  • 11h01m20: line amplitude = 15mV;
  • 11h03m08: line amplitude = 20mV;
  • 11h06m08: line amplitude = 25mV.

In Fig. 3 we have the signals increasing by increasing the line amplitude. If we increase it more than 15 mV we don't improve the SNR. We set the line amplitude to 15mV that corresponds to ~10 of SNR.

 

3. Correction filter modification attempt

We needed to tune the demodulation phase after the reconfiguration of the DAQ system: phi0 = 2.9 rad.

We can close the loop but this time we are not able to completely correct the bump, as shown in Fig. 4. In fact we are correcting only ~15 Hz but it is like if the bump is too large to be entirely corrected.

We tried to modify the correction filter putting a boost or a resonant filter and tuning the gain. We didn't manage to improve the loop in this way.

 

4. Attempt to close the stray light active loop on the input mirror of the Filter Cavity

Since the mirror M4_Z is used also on the CC coarse loop, we should move one of the actuations of these loops on the input mirror of the filter cavity.

We tried to move the stray light loop actuation on the input mirror and we checked the behavior of the stray light loop in this configuration.

We switched off the weight of the correction on M4_Z and switch on the weight of the correction in FCIM_LC with python:

  • EQB1_HD_AA.HD_M4_Z_cmd.HD_M4_Z_CORR = 0

  • FCIM_LC.MIR_Z_ACT.EQB1_HD_M4_Z_CORR = 1

And we verified that closing the stray light loop on the input mirror, we were sending the actuation on FCIM_LC_MIR_Z_ACT instead of EQB1_HD_M4_Z_cmd.

We were not able to correct the bump in this configuration. We switched back the correction to M4_Z.

 

5. Correlation between EQB1_HD_M4_Z_cmd and FCIM_LC_MIR_Z

We noticed a correlation between the correction of the stray light loop EQB1_HD_M4_Z_cmd and the movement of the input mirror of the filter cavity, see Fig. 5.

This could suggest that the residual motion of the cavity is generating most of the stray light. We will discuss with SBE expert in order to better investigate this possible issue.

 

6. Spectrogram of AUDIO DIFF channel during the weekend

In order to better understand the bump behavior, we left the system with the LO during the weekend in order to acquire data of the bump.

In Fig. 6, the seismic activity over the weekend: on Sunday 9th there was more seismic activity wrt to Saturday 8th.

In Fig. 7, the comparison of the AUDIO DIFF channel spectra between Sunday (blue) and Saturday (magenta): the blue bump is bigger.

We also analized the spectrograms for different times:

  • Jan 9th 2022 @ 5h30 UTC (Fig. 8)
  • Jan 8th 2022 @ 16h00 UTC (Fig. 9)
  • Dec 14th 2021 @ 14h30 UTC (Fig. 10)

We can see that in December the stray light arches are smaller (and less spead in frequency) wrt to January spectrograms.

We also can see that in Jan 9th the arches are more spead in frequency than in Jan 8th.

As we already expected the bump is strongly influenced by ENV conditions.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------

We will go on with this activity this week, starting from recovering the optimal alignment of the system.

Images attached to this report
Comments to this report:
polini, sorrentino - 12:07 Tuesday 11 January 2022 (54448) Print this report

The attached plot shows the impact of the low frequency boost. Magenta: open loop; yellow: closed loop, no boost; blue: closed with boost; green: sensing noise

Images attached to this comment
bertolini, sorrentino - 11:01 Thursday 13 January 2022 (54469) Print this report

About the plot in fig. 5, where the oscillation of the optical path length seems correlated with FCIM_LC_MIR_Z, indeed the ~6 V p-p oscillation of the EQB1_HD_MZ_Z correction corresponds to ~40 µm change of the optical path, which is much larger than the ~10 µm oscillation of FCIM_LC_MIR_Z. See also the plot in attached Fringe_scan.png where we do not compensate the optical path change: the HD interference  scans about 45 fringes at 1064 nm for a semioscillation of FCIM_LC_MIR_Z.

Indeed the microsesimic noise was extrimely large at the time. However the residual motion of FC mirrors is much smaller than the LC_MIR_Z signals.  The plots in FC_IP_controls.jpg show a comparison between the IP LVDT, LC_MIR_Z and the IP error signal (blending between LVDT and on-board seismometer) for the two microtowers. What counts for the real mirror motion is the blue line, since the other two signals just measure the relative motion between the ground and IP (LVDT) or mirror (Oplev). The FCIM control is a bit worse with less suppression of the 120mHz peak but this is expected based on the settings of the blending filters; however both suspensions seem to work quite well.

In FCIM the cross calibration between LVDT and Mirror oplev is pretty good; it's worse in FCEM but not dramatically. MIR_Z is simply measuring the motion of the ground since the mirror is much quieter. This is confirmed by the signals of the environmental accelerometer placed underneath each microtower, see FC_env_seism.jpg (only take into account the bold lines - double integration of the accelerometer signal done by DDisplay); one can clearly see that the MIR_Z oplev matches pretty well the signal of the ground seismometer.

In conclusion the fringes that we observe are not due to the common mode motion of the cavity.

Images attached to this comment
Search Help
×

Warning

×