Transcript Virgo Commissioning update and VSR1 first month summary

Virgo commissioning update
VSR1 first month summary
E. Tournefier
WG1 meeting, Cascina
June 20th ,2007
Activities since last WG1 meeting (March 28th)
•
Virgo complete electrical shutdown (April 6th -> 18th)
Infrastructure works:
– Electrical works: UPS replacement, reorganisation,…
– Computing: civil work, re-organisation, new file server,…
Detector improvements
- Acoustic enclosure around external detection bench
- Replacement of Brewster window with a larger one
Restart went ok: ITF relocked within ~2days
•
1 month of commissioning before the run
– Acoustic enclosures at end benches
– Diffused light investigations + some noise reduction
– Improvements of controls (longitudinal, angular, suspensions)
– Change of naming convention (V1:channel_name)for easier data exchange with LSC
•
VSR1 started on May 18th as planned
– Short commissioning breaks for improvements / problem fixing
Acoustic enclosures
•
External detection bench:
– Displace bench from the tower (~50cm)
– Install acoustic panels (a room into the room)
•
End benches:
– Simpler: build a room around the benches
End buildings
Detection
Acoustic noise reduction: end benches
- Previous actions:
Improvement of the optical setup
But: projection of acoustic noise of
end benches was still close to the
measured sensitivity
Tentative acoustic noise projection before
 Acoustic enclosure needed
Result: impact reduced above 100 Hz
Tentative acoustic noise projection with acoustic enclosu
Seismic noise on the optical bench
- No acoustic enclosure
- With acoustic enclosure
New Brewster window
Why change the detection Brewster window:
- Brewster-detection tower was shown to be an area very sensitive to acoustic noise
- tentative noise projections: Brewster could be a limiting factor
- old Brewster didn’t respect the rule: diameter > 5 x beam waist
Tentative projection of ‘Brewster noise’
Detection tower
 Replace Brewster with a larger one (April
shutdown)
Followed by diffused light investigations
No obvious improvement in sensitivity
See Raffaele’s talk
Brewster link
Other actions for noise reductions
(last WG1 meeting)
BS actuator noise reduced
Replaced pump of SR tower with a smaller one
(noise reduced but path is unclear)
Mirror actuator noise reduction
•
Electronic noise of the coil drivers + DAC
 Shaping filters, more gain to reduce its impact
zCorr
I (A) to mirror coil
BS actuator noise reduced in May
The arm mirrors actuators need to be improved
can be done during VSR1
Actuators calibration
Simple cavity filtering
(500 Hz pole)
Actuation TFs Simple pendulum
More precise measurement of the actuators response
 frequency dependence could be due to Eddy current effects
 Change of 20% at high frequencies
10%
See Bastian’s talk
Longitudinal controls
Usual noise subtraction (online):
-MICH and PRCL corrections sent to the end mirrors to
subtract the noise reintroduced by these loops into the
dark fringe
- coefficient is frequency dependent
MICH noise reduction ~ 50
B2_ACpmix
PRCL
MICH
- with  =0
- with measured (f)
Longitudinal controls: noise budget
- B1_ACp
- BS control noise
- PR control noise
BS control noise limits the sensitivity below 40 Hz
=> What is the origin of this noise?
Origin of BS control noise ?
• Related with acoustic noise inside the laser lab
 Beam jitter seen by B2 photodiode (used for PR & BS)?
• Sometimes disappears … but acoustic noise does not change
Input beam jitter
Beam jitter vs dark fringe
dark fringe
Acoustic noise in laser lab
Beam jitter vs acoustic noise
Angular controls
•
Second modulation frequency
(8MHz) now used for Q22
=> Get rid of DC signal
• Control loops improvements
• Mirror centering ~ 1mm
Angular noise budget
-Angular controls do not limit the
sensitivity
- Reduction of electronic/shot noise
in error signal still possible
Suspension control: z ->  couplings
Compensation of non-linear coupling:
Large seismic activity
 large correction sent to marionetta (zM)
 y motion of F7 and payload
 Side effects on alignment
Solution: compensate for z -> y coupling
Need a quadratic compensation:
y=c*zM +d*zM2
Dark fringe
-No quadratic compensation
-With quadratic compensation
Suspension control: top stage control
Old problem: environmental noise (seism) reintroduced by Lvdt used for top stage
control
=> Use the ITF longitudinal error signal instead (Global Inverted Pendulum Control)
=> z correction to the mirror well reduced
GC (reconstructed z)
VSR1: enabled only for NE-WE
No GIPC
With GIPC
LVDT
LVDT
LVDT
LVDT
Sensitivity vs wind and sea activity
Matteo’s empirical formula: Horizon = H0(1 - a x seism – b x wind)
WSR1 (Sept 2006)
VSR1 (May 2007)
a= 0.37, b= 0.37
=> Sensitivity to bad weather conditions is well reduced
a= 0.07, b= 0.04
Lock acquisition: thermal transient troubles
•
Variations in the thermal transient : seems related to activities in laser lab
Sometimes need to wait few hours before being able to lock
Thermal effects: sidebands image
• Image of the sidebands (first step of phase camera):
scanning system on B5 beam:
- the beam is scanned on a pin-hole + photodiode
- demodulated signal is used to reconstruct the sidebands shape
Waiting for thermal compensation system + complete phase camera
VSR1 noise budget
Still magnetic noise?
Mystery noise
Mystery noise
•
Some observations:
– Coupling to DF increases with misalignments
– Sudden decreases/increases (also seen in frequency noise)
– Structures not excited with noise injections/tapping tests around benches/towers
Keep searching…
VSR1: first month summary
Thanks to Roberto for the material
Run organisation
•
Goal:
– Take science data for a minimum period of 4 months starting from the 18th
May 2007 at 23:00 (LT)
•
Shifts have been planned until 20th September 2007:
• 18 weeks, 378shifts.
– Shift team and time schedule:
• Team: 1 operator + 1 physicist,
• Shift period: 8 hours
• Shift start times: 07:00 – 15:00 – 23:00
Permanent ON Call:
– 11 Virgo ITF Sub-systems:
•
(Alignment, DAQ, Detection, Electronics, Locking, Global Control, Laser &Injection,
MSC, On line Computing &Network, Software, Suspension Electronics & Software)
– 4 EGO Site Services:
(Network & Global Security System, Site Infrastructure-Conditioning, Site
Infrastructure-Electricity, Vacuum)
Duty cycle
– Duty cycle in science mode ~ 85%
– 42 unlocks from “Science Mode” registered up to 17/06/07
– Longest Lock: about 58.5 hours , Longest Unlock: about 16.1 hours
Science mod recovery time
– Longest unlock: 16 hours
– Smallest ~ 30mn (time needed for lock acquisition)
Science Mode Unlock Time Periods
(18th May - 16th June: 42 Unlocks)
18
16
14
12
Hours
10
8
6
4
2
Days
/0
7
/0
6
/0
7
19
/0
6
/0
7
17
/0
6
/0
7
15
/0
6
/0
7
13
/0
6
/0
7
11
/0
6
/0
7
09
/0
6
/0
7
07
/0
6
/0
7
05
/0
6
/0
7
03
/0
6
/0
7
01
/0
5
/0
7
30
/0
5
/0
7
28
/0
5
/0
7
26
/0
5
/0
7
24
/0
5
/0
7
22
/0
5
/0
7
20
/0
5
18
16
/0
5
/0
7
0
16
/0
...
18
/0
...
20
/0
...
22
/0
...
24
/0
...
26
/0
...
28
/0
...
30
/0
...
01
/0
...
03
/0
...
05
/0
...
07
/0
...
09
/0
...
11
/0
...
13
/0
...
15
/0
...
17
/0
...
19
/0
...
Hours of Lockinng
Lock duration
– Longest lock: 58 hours
VSR1 Locking Step Time Periods
(18th May - 16th June: 42 Locking periods)
60
50
40
30
20
10
0
Date
Unlocks
•
Analysis of unlocks ongoing (preliminary statistics)
Main unlock reasons:
- IMC fast unlocks
Partially cured (loop gain reduced)
- Global control crashes
Pb = longer recovery (susp excited)
- Earthquakes
Manual unlocks for:
- Weekly maintenance
- Commissioning breaks
Only 3 unlocks due to bad weather
VSR1 Unlocks Type Distribution
(18th May - 16th June: 42 Unlocks)
Manual-Adjusting
2,4%
Maintenance
9,5%
SSFS
2,4%
Thunderstorms
2,4%
WI_LC Glitches
4,8%
12Hz Long. Oscill.
4,8%
Unknown
14,3%
ISYS
2,4%
Commissioning
4,8%
Fast Unlocks
16,7%
Bad Weather
4,8%
GxOB crash
2,4%
GC Crash
11,9%
Earthquakes
16,7%
Commissioning breaks
Alignment drifts observed during long locks:
related to Q81 quadrant mis-centering (used for BS control) => spoils error signal (AC)
 Temporary solution: realign quadrant from time to time
 Improvement: include Q81 asymmetry in the error signal for BS control
Small drifts still there
 Try to use another error signal (B1p?)
or
 Install galvanometer to keep centering
in science mode
IMC Fast unlocks
- Reduction of IMC loop gain
1 unlock per couple of day => ~ 1 per week
- more investigations ongoing (electronics)
Suspension controls: some improvements /tests
Dark fringe
Q81 asymmetry
realignment
Long term observations: Etalon effect!
•
Etalon effect:
small FP cavity inside input mirrors (due to AR face)
=> effective reflectivity modulated with mirror thickness (temperature)
AR
HR
HR
B7/B8 phd
LB
LA
r2
-FP transmitted power
-Input mirror temperature
r3
FP Transmitted power
r1
Temperature variation
Some short term plans
•
Continue noise investigations
– ‘Mystery noise’ (  200Hz)
– Magnetic noise (50 - 100 Hz)
– Low frequency noise in laser lab (=> BS error signal)
– Reduce mirror actuator noise (cavity mirrors)
•
Controls:
– Alignment drifts: get rid of Q81 or use galvanometer
– Suspension control improvements in preparation