Presentazione di PowerPoint

COMMISSIONING PROGRESS
G.LOSURDO – INFN Firenze
for the Virgo Collaboration
Progress since last meeting
3 more Weekly Science Runs performed
4 months science run due to start on May 18th
WSR #
date
duty cycle
1
Sep 8-11
90
2
Sep 22-25
65
3
Oct 6-9
cancelled
4
Oct 13-16
cancelled
5
Nov 10-13
65
6
Dec 1-4
80
7
Jan 12-15
76
8
Feb 9-12
94
9
Feb 17-19
100
10
Mar 9-12
70
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
2/25
Virgo vs LSC
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
3/25
Detector plans
2006
Commissioning
2007
Run
4 months
FULL LSC-Virgo DATA EXCHANGE
will start on May 18th
LSC Council resolution - March 20, 2007
The LSC Council endorses the implementation of the data sharing provisions of our
agreement with Virgo, starting on 18 May 2007.
We look with favor on exchanges of data before then, for technical purposes.
The Council asks the joint Run Planning Committee to meet intensively in the
coming weeks, and to present at the May Collaboration meeting a set of options for
joint running of the Virgo, GEO, and LIGO interferometers over the next few years.
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
4/25
COMMISSIONING ISSUES:
longitudinal control
Longitudinal control noise - PRCL
 B2_3f has bad SNR:
control noise limiting (or close to
limit) the sensitivity below 200 Hz
WSR6
B2_ACp
Switch to B2 to reduce the noise
Problem: the switch to B2 is stable if the “good” operation point is found. This is
not easy and not reproducible.
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
6/25
B2_mix - PRCL

Combine stability at low frequency and good noise performance above
10 Hz

B2_3f replaced by a mix of B2_3f (low passed) and B2 (high passed)

Good stability achieved. Noise performance comparable with B2 only
B2_3f
B2
@Copyright
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
7/25
Alpha and Beta – MICH and PRCL
MICH and PRCL corrections sent to the end
mirrors to subtract the noise reintroduced by
these loops into the dark fringe
PRCL
MICH
Linear Coupling Approximation:
MICH  coupling coeff is called ALPHA
PRCL  coupling coeff is called BETA
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
8/25
Frequency dependent Alpha - MICH
A frequency dependent a(w) has been introduced, improving the
result: a factor ~20 of cancellation everywhere
MICH Noise Injection (alpha=0)
MICH Noise Injection (const alpha)
MICH Noise Injection (6th order filter)
ILIAS WG1 - CERN, Mar. 30th, 2007
Effect on sensitivity
G.Losurdo – INFN Firenze
9/25
Beta - PRCL
Problem with the Beta subtraction:
the coupling PRCL  dark fringe varies considerably with time
A frequency dependent b(w) does not help
Solution:
automatic Beta tuning
Alp macro done and tested
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
10/25
New MICH filter
Modified roll-off:
MICH control noise reduced
around 50 Hz
before
ILIAS WG1 - CERN, Mar. 30th, 2007
after
G.Losurdo – INFN Firenze
11/25
COMMISSIONING ISSUES:
angular control
WSR10 alignment layout
6.26 MHz
8.35 MHz
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
13/25
2nd modulation frequency
2nd mod.
(8.35 MHz)
IMC mod.
(22.4 MHz)
Virgo mod. (6.26 MHz)
2nd modulation frequency
needed for having absolute
(demodulated) reference for
common end mirror alignment
Previously aligned with ITF
reflected light (Q21) DC signal
2nd mod. on the same EOM =>line
forest on EOM => perturbed
locking signals
Solution: added 2nd EOM =>
Much less problems but weaker
modulation index to avoid
perturbations. To be investigated
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
14/25
End mirror beam centering
corr

Shake end mirror tx/ty

If mirror is miscentered: cavity length oscillates

Lines visible in the locking correction signal

Tune input mirror alignment  null the line in
Zcorr (repeatable to 0.5 urad)
 New input mirror alignment loops (Alp)

very slow (some mHz)

mirror center as reference

before: odd combinations of end quadrant DC’s
10 Hz
before
after
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
15/25
Automatic control noise budget
Automatic procedure to calculate the noise budget for longitudinal
and angular control (a few minutes of noise injections)
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
16/25
COMMISSIONING ISSUES:
Suspension controls
Stability evolution
Sensitivity to environmental conditions considerably decreased
– larger gain in automatic alignment loops
– better cut-off in damping of inverted pendulum respect to ground
– damping of inverted pendulum respect to ground switched OFF on 4
suspensions (GIPC)
– Reallocation of DARM correction split over both marionettes
see talk by Ettore
Improved stability of BNS horizon
(with similar environmental conditions)
Darker and more stable dark fringe
WSR8
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
18/25
Detection bench local contol
The big haircut:
28 Hz line with “hair” cancelled with better roll-off
on detection bench local control
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
19/25
COMMISSIONING ISSUES:
pollutions
Stray light
Environmental noise acting on in air components +
Scattered light from in air components
=
------------------------------------------------------------Phase noise in the interferometer
(with the typical environmental-like noise structures)
End benches
Injection bench
Laser
Brewsters
See talk by Julien
ILIAS WG1 - CERN, Mar. 30th, 2007
Detection bench
G.Losurdo – INFN Firenze
21/25
Acoustic noise reduction
Acoustic isolation box
– installed around laser and injection tables
– to be installed around detection table (April)
– ordered also for terminal benches (to be installed when possible)
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
22/25
External bench reshuffling
General re-organisation
– simplified setup
– 2” optics
– dump the photodiode
reflection
– dust cover
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
23/25
Electromagnetic noise
Electromagnetic noise hunting
– coherence with electromagnetic sensors observed
– sources of magnetic field near the mirror tracked
– noisy power supplies identified and removed
The original sin: the 4 magnets
have been glued with the same
polarity on NI/WI mirrors
dark fringe spectrum
Possible solution: null the dipole
gluing 4 magnets with opposite
polarity on the front face
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
24/25
Effects of turbo pumps

Noise reduction after switch off of the turbo pumps and some vacuum
racks in the central area

Noise reduction observed also on the accelerometer on the brewster
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
25/25
New SSFS board
old SSFS board

More professional electronics

More gain in the 1-5 kHz
range

Low freq gain to be increased
soon
sensitivity improvement
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
26/25
Sensitivity today
Hardware Limit
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
27/25
Short term plans

Noise hunting
–
–
–
–
replacement of Brewster window with larger one
installation of more acoustic isolation
actuator noise reduction
control noise investigations

Run preparation

Virgo Science Run 1 to start May the 18th
ILIAS WG1 - CERN, Mar. 30th, 2007
G.Losurdo – INFN Firenze
28/25