Reports 1-1 of 1 Clear search Modify search
AdV-DET (Local controls for all benches)
letendre, masserot - 23:48 Thursday 28 November 2024 (65654) Print this report
Excess of noise at SWEB.

We made some investigations on the additional noise observed  the SWEB local control error signals, To be noted that the same kind of problems occurred roughly one year ago: 2023-11-17

The attached plots shows

  • the evolution of this noise on all the WEB LVDT signals from the 2024-11-20 to the 2024-11-26 
  • the trial to increase the WEB_PCal minimal elapsed time of 14 us to 18 us, but without any improvements this time
    • 2024-11-28-19h56m52-UTC    WEB PCal external saved - masserot: Set the Min elapsed time to 16us instead  of 14us (rev. 257218)
    • 2024-11-28-19h57m01-UTC    'Reload External [file=/virgoData/VirgoOnline/WEB_PCal_External.cfg]' sent to WEB_PCal_fast
    • 2024-11-28-20h01m50-UTC    WEB PCal external saved - masserot: Set the Min elapsed time to 18us instead  of 14us (rev. 257219)
    • 2024-11-28-20h01m51-UTC    'Reload External [file=/virgoData/VirgoOnline/WEB_PCal_External.cfg]' sent to WEB_PCal_fast
    • 2024-11-28-20h11m14-UTC     WEB PCal external saved - masserot: Set the Min elapsed time to 16us instead  of 14us (rev. 257220)
    • 2024-11-28-20h11m18-UTC     'Reload External [file=/virgoData/VirgoOnline/WEB_PCal_External.cfg]' sent to WEB_PCal_fast
    • The min elapsed time remains to be restored to  14us
  • the last plot compare the ASD of the raw ADC channels and the ones read by the SWEB_LC Acl RT task
    • below the actions made on the SWEB_Fb server to acquire the  raw ADC channels
      • 2024-11-28-22h18m30-UTC   SWEB_Fb saved - masserot: Put on online frame the  "SWEB_ADC_LC_LVDT_out_0" channels (rev. 257222)
      • 2024-11-28-22h18m33-UTC   'Reload Configuration' sent to SWEB_Fb
      • 2024-11-28-22h22m17-UTC   SWEB_Fb saved - masserot: remove from online frame the  "SWEB_ADC_LC_LVDT_out_0" channels (rev. 257223)
      • 2024-11-28-22h22m20-UTC   'Reload Configuration' sent to SWEB_F
    • blue : raw ADC data with all the SWEB LC lines
    • red: the one read by the SWEB_LC server, applying a rejection filter at the frequency of the unwanted lines
    • The 2 channels see the same noise, confirming the result of the test related to the SWEB_LC elapsed time increase .

The issue is still there, and more time is required to investigate it

Images attached to this report
Comments to this report:
letendre, masserot - 12:05 Friday 29 November 2024 (65659) Print this report

One more plot showing  that the moving bump is visible in the B8_PD1_Audio channel acquired at 1MHz

Images attached to this comment
bonnand, masserot, sposito - 17:20 Friday 29 November 2024 (65663) Print this report

We continue the investigations related to moving bump on the LVDT channels with the support of Gianmatteo in the West End Building.

Below the detail of the operations performed

  • NIKHEF LVDT box
    • 2024-11-29-12h26m-UTC: switched OFF
    • 2024-11-29-12h29m-UTC: switched ON
    • None improvements observed : the line at 8500Hz remains at the same frequency and at the same level
  • SWEB_DBOX_LC (Dbox_SN43)
    • 2024-11-29-12h30m-UTC: switched OFF
    • 2024-11-29-12h32m-UTC: switched ON
    • 2024-11-29-12h32m56-UTC   Reconfigure SWEB_DBOX_LC in SWEB_dbox_rack
    • None improvements observed : the line at 8500Hz remains at the same frequency and at the same level
  • SWEB tank power supplies
    • 2024-11-29-12h38m14-UTC  Disable output of PowerUnit13 (Dboxes power supplies)
      • the line at 8500Hz  move at  9000Hz
    • 2024-11-29-12h40m24-UTC  Disable output of PowerUnit12 (Quadrants and Timing power supplies) 
    • 2024-11-29-12h41m21-UTC  Enable output of PowerUnit12 (Quadrants and Timing power supplies)
      • the noise is reduced when this power supply is OFF but return at the previous level when it is ON
      • second attempt of OFF/ON cycle
        • 2024-11-29-12h43m37-UTC  Disable output of PowerUnit12
        • 2024-11-29-12h44m12-UTC  Enable output of PowerUnit12
      • After this second attempt the noise floor is reduced and remains at the new level when this power supply is ON
    • 2024-11-29-12h46m25-UTC   Enable output of PowerUnit13 (Dboxes power supplies)
    • 2024-11-29-12h53m52-UTC   'Reload config' sent to SWEB_dbox_bench : reconfigure the SWEB tank Daq boxes
      • only the expect line at 9600Hz present with the same low noise floor
  • SWEB_LC stopped and restarted
    • 2024-11-29-12h57m11-UTC  process:SWEB_LC stopped
    • 2024-11-29-12h57m22-UTC  process:SWEB_LC started
    • SWEB_LC and SBE loops successfully closed at 2024-11-29-12h58m10-UTC

According to this time sequence, the noise seems related to the  PowerUnit12 or the devices (Quadrants and timing) powered by it .

The last plot show the evolution of some small lines around 10KHz after the restart of the bench

While we were performing these operations, we forgot to disable the WEB ALS BPC. As consequence, the ALS on-call had to realign the SWEB ALS  BPC to allow the ITF lock

 

Images attached to this comment
gouaty, masserot - 17:31 Tuesday 03 December 2024 (65699) Print this report

After  the power OFF/ON of the B8 quadrants , we forgot to restore its VBIas and to re-open its shutters . As consequence the SWEB_LC  bean drift control was not working  anymore .

The B8 Quadrant VBias have been restored at 2024-12-03-14h23m39-UTC

  • 2024-12-03-14h23m35-UTC    Command 'B8_QD1:enable vbias' executed
  • 2024-12-03-14h23m36-UTC    Command 'B8_QD2:enable vbias' executed

 

The 2 last plots show the 100s trend data over 1 year,   2024-01-01_to_2024-06-30 and 2024-07-01_2024-12-03 ,  for the SNEB and SWEB  TY_{min,max} and TY_err_{min,max} channels, where one can see

  • the SNEB_ TY_err_{min,max} is well below  0.1urad all the time
  • while for SWEB_TY_eer{min,max} there is very noisy periods and sometime long quiet period .
  • A flag has been added on the DMS to monitor the RMS of  the S{N,W}EB_LC_TY_err channels
Images attached to this comment
gouaty, masserot - 17:53 Tuesday 03 December 2024 (65702) Print this report

We also added the flags QPD_B7 and QPD_B8 in the DMS to monitor the status of the Vbias applied to these quadrants.

The attached figure shows the evolution of the angular position in TX and TY of the SWEB bench, as well as the values given by the B7/B8 quadrants since last friday (when the quadrants Vbias were switched off).

One can see that when the quadrants were off and the SWEB_LC process was restarted, the bench went to the setpoints written under the configuration file, which are about 2 urad away from the angular position obtained when the bench was under drift control.

Since the QPD Vbias was reactivated this afternoon (and the drift control consequently re-enabled), the bench almost returned to the same angular position as where it was last friday.

An email has been sent to the ALS on-call to notify him about this change that could potentially impact the ALS alignment.

Images attached to this comment
letendre - 11:58 Monday 09 December 2024 (65751) Print this report

The noise is correlated with variation of  the current of quadrant +12V power supply. Moreover, the +12V voltage of quadrant is very noisy. This voltage moves from a hundred mV around 12V, whereas it should vary around a few mV only.

The noise appeared on December 8, 2021, after an intervention in SWEB to bypass a short on the +12V used for the timing inside the bench. See: https://logbook.virgo-gw.eu/virgo/?r=54178

Images attached to this comment
Search Help
×

Warning

×