Transcript E.Tournefier - Virgo progress

Virgo commissioning since the end of VSR1
E. Tournefier for the commissioning team
ILIAS WG1, Cascina
March 5th ,2008
1
The end of VSR1: mystery death
•
Noise structures (200-300 Hz + 600-1000 Hz) highly non-stationary:
depend on the quality of the ITF alignment
realignments
2
Mystery noise investigations
Suspended bench
•
Noise also seen in the suspended injection bench error signals
 investigations in laser lab, around injection tower
(electronics, seismic noise, laser chiller, tapping tests,…)
 investigations on the injection suspended bench
Guilty = piezo-actuators on the suspended bench
Mechanism:
Piezo driver noise => actuator motion => IB shaken
=> IMC length noise + jitter noise
Tower window
External injection bench
Laser bench
3
Mystery noise = beam jitter
•
Solution: low pass filter the signal of the piezo drivers
 impressive decrease of frequency noise and beam jitter noise
 (no more) mystery structures disappear from DF
Dark fringe
Dihedron resonance
Frequency noise
4
Noise budget at the end of VSR1
Control noises
Environmental noise
Actuacor noise
Shot noise
Eddy currents
5
Noise budget now
Actuator noise
Environmental noise
Control noises
Shot noise
Eddy currents
6
Summary
•
Control noise reduction:
– Longitudinal controls: use a second modulation frequency, noise subtraction,…
– Angular controls: more efficient filters, mirrors centering,…
•
Actuator noise:
– DAC noise filtering
– Magnetic noise: replaced magnets on input mirrors (strength/5.5 + good polarity)
•
Environmental noise investigations
– Noise sources: air conditioning (<150 Hz), electronic racks
– Path to dark fringe:
injection beam jitter, diffused light on benches, detection/Brewster
•
Environmental noise reduction:
– Beam jitter reduced
– Diffused light mitigation on end benches, improved telescopes
– Air conditioning improvement started
– Brewster => cryotrap to be installed soon
7
Alignment noise reduction
•
•
Post VSR1 alignment noise reductions:
see Maddalena’s talk
– More efficient filters
– Mirror centering improved
– NI+WI local control noise reduced
 The alignment noise is below V+ design above 20 Hz
Ongoing improvements:
– Find cleaner signals for BS and input mirrors (under drift control)
 New end benches and new quadrant on B1p have been installed
-> Drift control (10mHz)
8MHz
-> Drift control
8
Upgrade of the end benches
•
New end benches telescopes + rearrangement
– Hoping for better quality alignment signals
– Galvos installed at that occasion (centring during science operation)
Galvos
– Better telescope helps to reduce diffused light
Remaining problem: quadrant signals polluted by air flux and bench motion
Now
Virgo+
Galvos
Telescope
Quadrant
setting
9
Longitudinal noise reduction
Post VSR1 activities:
1/ Characterization of the ITF response (2 weeks immediately after the run):
- response of actuators (mechanics + electronics) / also a cross check for calibration
- optical gains (W/m) : mirror displacement  photodiode signals
2/ Improvement of the controls for noise reduction:
- adapt filters and driving matrix to the actuator response
- adapt the sensing matrix to the optical response
- use cleaner signal from the second modulation frequency (8MHz) for MICH control
- new filters for marionetta to avoid exciting 16/46 Hz resonances
- and: better noise subtraction, more flexible Gc algorithm,…
8MHz
B2_8MHz
10
Longitudinal control noise reduction
Sensing for central cavity:
New B2 signal (8MHz) + new sensing matrix
•
•
see Gabriele’s talk
MHz
0  B 2 _ ACp
ACp__8mix
 PRCL   0A C


  

 
MICH
D
0
B
5
_
ACq
d
B
B
5
_
ACq

 


Longitudinal control noises are not anymore limiting the sensitivity !
Longitudinal control noises are now limited by environmental noises
– Beam jitter subtraction imlemented for B2_8MHz
– Long term solution: reduction of environmental noises
control signals
error signals
Environmental noise
-VSR1
- now
11
Environmental noises
• Test: switch OFF the air conditioning in ALL buildings (NE,WE, CE, MC)
 Identifed sources and path to dark fringe
- diffused light from end benches (now almost completely cured)
- beam jitter induced by AC in laser lab
- diffused light from Brewster/detection induced by Central hall + DAQ room AC
DAQ room (+CE hall)
AC ON
AC OFF
North End AC
CleanRooms AC
 Laser lab
Central hall AC
Water pump of CE
Other identified sources: electronic racks (peaks above 200 Hz)
12
Beam jitter at the injection
• Mechanical resonances of the Beam Monitoring System mount + PZT seen in dark fringe
 Mount + PZT replaced with better one:
Move resonances at higher frequencies + they are less easily excited
new mount 2"
580 Hz
A.U.
bench
Y Axis
0
200
630 Hz
400
new mount 1"
600
800
1000
frequency (Hz)
1200
1400
• Identified vertical/horizontal resonance (~45/18Hz) of the injection bench
 beam jitter
 Damping of these resonances under study
13
Diffused light at the detection
•
•
Many structures above ~100 Hz are coherent with the motion of the Brewster window
Tapping tests  Brewster is the most sensitive of all vacuum pieces
But difficult to make separate what comes from detection tower and from Brewster
 Diffused light / modulation by Brewster?
 Diffused light by detection bench on tower walls or Brewster link / back into the ITF?
Dark fringe
BS
Detection
benches
Coherence with Brewster Seismometer
14
Diffused light at detection port
•
Investigate the path for diffused light
– Diffused light on the tower walls?  baffles installed – no visible effect
– Diffused light inside the Brewster link?  better baffles have been installed (any
effect?)
– Light diffused by detection optics?
 bench rotation experiment-some unclear effect
– Coupling through the Brewster window?  replace it with cryotrap next week
BS
Detection
benches
15
Diffused light reduction (?) at detection
Indication of light reflected/diffused by the detection bench:
Dedicated experiment: rotate the suspended bench by 100 rad
OB rotation 100 rad
•
 30 Hz and 45 Hz bumps disappeared (related to air conditioning noise in central hall)
 No clear effect at higher frequencies
Other rotation experiments were not so clear (??)
16
Which is the mechanism?
‘Brewster noise’ variations with alignment
•
•
•
Noise varies with the alignment/power on dark fringe
True only above 100 Hz
What about the low frequency which seemed to change with bench
rotation?
17
Environmental noise reductions
• Diffused light at end benches well reduced (beam dump + new optics)
• Beam jitter reduced (new BMS)
• Air conditioning improved inside the DAQ room
• Some pumps used for AC are OFF (do not seem needed…)
=> Impact of AC well reduced
18
New DAQ room air conditioning machine
301
120
200
32
70
119
90
232.5
233.5
60
UTA
157
137.5
430
19
Actuator noise reduction
•
Reduction of the mirror actuator electronics noise during VSR1 (arm mirrors)
DSP
Emphasis Emphasis
G
DAC
DeEmphasis
CoilDriver
DeEmphasis
Impact of DAC noise reduced:
• BS:
Air conditioning OFF
– marionetta reallocation now working
=> CD gain decreased by 4 (DAC + CD noise reduction)
• Arms:
gain decreased by factor 4 (reduced dynamic)
=> new coil driver with several (switchable) low noise sections
next step: better emphasis filters
+ Improved measurement of low frequency DAC noise performed
No clear improvement observed…
20
Actuator noise now
•
•
Marionetta DAC noise below 20 Hz?
We’ll add some filtering this week, let’s see…
Eddy currents
21
Magnetic noise?
•
•
Direct coupling of magnetic noise through the magnets attached to the mirror
Magnetic lines injections, modelling => tentative projection (upper limit)
see Bas’s talk
22
Magnetic noise?
•
We replaced the magnets of the input towers with less strong ones (/5)
+ restored the good polarity
No real improvement observed…
AC OFF
Dec 21st
Mar 3rd
23
Thermal effects
•
Cleaning of the input mirrors was performed at the end of November
– Some visual improvement was observed for both
see Michele’s talk
WI mirror :
 Observation before cleaning : with an halogen lamp coupled to a fiber
some large particules on the coatings
center part : a little bit ‘opaque’ (~ Ø1 cm)
a lot of condensation points but not on the coating
 Observation after cleaning : no more large particles
center part : clearer(# Ø1 cm)
visual aspect better than before cleaning
– But no clear improvement in the thermal effects:
- thermal transient similar to before
- increase of the sidebands power ~ 25% only
24
Thermal effects
•
Before / after cleaning
Locking step
Dark fringe
Sidebands power
B5, carrier
25
Plans before Virgo+ shutdown
•
Replace Brewster with cryotrap next week
•
Thermal Compensation System installation this month
– TCS commissioning
•
Understand effect of the magnets replacement (magnetic noise, Eddy currents)
– Also replace those of the end mirrors?
26