Transcript Interferometer Control

Interferometer Control
Matt Evans
…talk mostly taken from…
Università degli Studi di Pisa
Scuola di Dottorato Galileo Galilei
Ph.D. in Applied Physics
The control of the VIRGO
interferometer for gravitational
wave detection
Lisa Barsotti
Pisa, 20th April 2006
European Gravitational Observatory (EGO)
(Cascina-Pisa)
3 km
Things to Know
-> If you have a question,
raise your hand and
wave, or take some other
action to draw my
attention. This may be
difficult because I haven’t
slept much this week, so
don’t be shy.
->The topic of this talk is
“Interferometer Control”,
but after years of work in
this field I still don’t really
know what I’m doing, so I
tried to make this talk
easy to understand. I
probably failed… ask
questions.
-> I am American, and I
suffer from the common
misconception that
Americans speak English.
This means that I talk
fast (because I think I
speak English well), but I
am difficult to understand
(because I’m not really
speaking English).
->This talk will require
about 15 minutes if I talk
fast and nobody stops
me.
->I like talking to myself,
but not in public…
The Virgo Interferometer
-> High quality optics with low
absorption, suspended in vacuum
-> Input Mode Cleaner
-> Laser Beam 20 W
EOM
Injection System
-> It provides the beam entering the
ITF with the required power and
frequency stability
-> Beam RF modulation
Detection System
-> Output Mode-Cleaner to improve
the contrast
-> Detection, amplification and
demodulation
Virgo Design Sensitivity
Seismic wall @ 4 Hz
Operating Point
The ITF has its nominal sensitivity only at
its operating point
Constraints on the tolerable fluctuations
of the relative position of the mirrors
 resonant light inside the cavities to increase the phase
sensitivity
 dL < 5x10-9 m RMS (integrated DC-10 kHz)
 anti-symmetric port on the dark fringe in order to
prevent intensity noise from dominating over shot noise
 dL < 10-12 m RMS
Suspension System
 The Superattenuator is a multi-stage pendulum, with passive attenuation:
10 14 @ 10 Hz
At lower frequencies the noise
is instead totally transferred to
the mirror, even amplified by
the pendulum resonances
Local active control
10
14
of the Superattenuator
reduces mirror motion
below a few Hz
Residual longitudinal
motion of the mirror
dL ~ 10-6 m RMS
Length Control: Why
Residual longitudinal
motion of the mirror
dL ~10-6 m RMS
Intensity noise based
requirement
dL < 10-12 m RMS
A global control system is needed to hold
the ITF on its operating point by
controlling relative mirror positions
Length Control: Why
Transmitted Power
Length Control: What
Actuation
CA
SB SB
 Correction signals are sent
to the optics by means of
coil-magnet actuators
Length Sensing
Filtering
Gain
Hz
 Pound-Drever-Hall error signals giving
the deviation from the operating point are
extracted at the output ports of the ITF
 Error signals are filtered to
compute correction signals
Control Example: Filtering
Actuation
 Correction signals are sent
to the optics by means of a
bio-actuator (hand)
Length Sensing
 Error signal giving the
deviation from the
operating point are
extracted from our
volunteer (eyes)
Filtering
Gain
Hz
 Error signals are filtered to compute
correction signals (brain)
 Different mechanical systems require
different filters
The Length Control
Chain

Signals are acquired with 16-bit
ADCs @ 20 kHz

Data are transferred via optical
links to the Global Control which
computes correction signals

Corrections signals are sent to the
DSPs of the involved suspension,
passed through DACs and applied
to the mirror
Global Control
Control Example: Delay
Actuation
 Correction signals are sent
to the optics by means of a
bio-actuator (hand)
Length Sensing
Filtering
Gain
Hz
 Error signal giving the deviation from the
operating point are extracted from volunteer
1 (eyes)
 Error signals are filtered to
compute correction signals
(brain 1? brain 2?)
Length Control: Data
The Lock Acquisition Problem
Error signals are available only when the ITF is around resonance
 no signals available far from resonance
Transmitted power
 Only 1 degree of freedom
 Correction signal sent to
the mirror at a resonance
crossing
Correction signal
Recombined ITF: lock acquisition


More complex optical scheme
(3 degrees of freedom: the two cavity lengths and the Michelson length)

Lock acquisition can be made similar to the single cavity by using the end
photodiodes
Lock of the two arms indipendently
 Lock of the michelson
Recombined ITF: linear lock
 Once the ITF is locked on its
operating point, the
longitudinal control scheme is
optimized in order to reduce
the control noise:
* use of less noisy error signals
* use of more aggressive filters
linear lock control scheme
Recycled ITF: after locking
 Frequency Servo
used for common
arms
 GW signal used to
control differential
arms
 BS controlled to
keep anti-symmetric
port dark
 PR controlled to keep
power level high
Angular Problem
 The end mirrors are 3km away
 The beam travels this distance
many times
 Small angles (1 micro-radian)
cause big problems
Angular Sensors
P
P
Y
P
Y
Y
Recycled ITF: angles
 There are 6 mirrors to control,
each with 2 degrees of
freedom
 The input beam has 4 DOFs
 16 total DOFs
Angular Matrix
P
Y
Other Loops…
 Laser System
Beam position
Intensity
Modulation Frequency
 Infrastructure
Building temperature
Vacuum pressure
 Suspension Systems
Inertial damping
Local Control
Conclusions
 Control loops should be
avoided
 Coupled systems should
be placed firmly in the
rubbish bin without
hesitation or remorse
 Interferometers are evil
 Sleep is good