Transcript Luciano Di Fiore

Development of an Optical Read-Out system
for the LISA/NGO gravitational reference
sensor: a status report
Rosario De Rosa, Luciano Di Fiore, Fabio Garufi,
Aniello Grado, Leopoldo Milano and Giuliana Russano1)
This R&D activity is supported by INFN Commission II
1) Present address: University of Trento
Paris May23rd 2012
L. Di Fiore
1
Talk overview
• goal of the activity
• proposed ORO set-up
• tests on sensitivity
1. Bench top
2. Suspended
• layout for implementation in LISA
• next steps for space qualification
Paris May23rd 2012
L. Di Fiore
2
Goal of the R&D activity
development of an optical read-out (ORO) system for the LISA
(NGO) inertial sensor, to be integrated in the present design of
the GRS, together with the capacitive sensor.
the motivations are:
• risk reduction: a back-up sensor in case the capacitive one
fails after the launch  this becomes still more important for
NGO because in this case, with only two arms, the failure of a
single Inertial sensor would compromise the mission.
• improved sensitivity  relaxed specifications on cross
couplings - Present requirement on C.C. is 0.1 % that is a very
strong specification!
Paris May23rd 2012
L. Di Fiore
3
Goal of the R&D activity (II)
Target sensitivity
For a back-up sensor the sensitivity
should be at least comparable with the
one of the main sensor
Of course any improvement in
sensitivity is useful but there is not
a specific requirement
2 nm/Hz1/2
200 nrad/Hz1/2
Factor 2-3 already interesting
Factor 10 more would give big
advantage
Limited complexity
We wont to keep everything as simple as possible
(compatibly with requirements)
Paris May23rd 2012
L. Di Fiore
4
Optical lever sensor
With these requirements in mind we have
selected, as a simple solution, the usage of
optical levers:
A laser beam is sent trough a SM optical
fibre to the test mass and the position of the
reflected beam is measured with a position
sensor (Quadrant photodiode of PSD)
the sensitivity depends on input power and
measurement range (beam size for QPD or
detector size for PSD)
with a suitable combination of three beams
and sensors we can recover the six DOF of
the test mass
Paris May23rd 2012
L. Di Fiore
5
The activity was performed in parallel developments
• Study of the sensitivity to demonstrate that we can
reach the target sensitivity
• Test on torsion pendulum in Trento
• Design of a sensor layout compatible with the
present inertial sensor design
• Study of space qualified parts (just started)
Paris May23rd 2012
L. Di Fiore
6
Expected sensitvity
If with start with reasonable assumptions : Low light power (belor 1 mW)
DC detection (for simplicity)
We expect that limiting noise is the input noise current of the trans-impedance amplifier
used for photodiode readout 𝐼𝑛 :
𝐼𝑛
𝐼𝑛 ∙ 𝑅𝑥
𝑥𝑛 =
~
𝑑𝐼
𝑃∙𝛼
𝑑𝑥
With:
𝑥𝑛 = sensitivity to spot displacement on the sensor
P = light power
a = photodiode responsivity
𝑅𝑥 = measurement range:
 ~ spot size for QPD
 ~ detector size for PSD
The TM displacement noise is
with AG a geometrical factor
Paris May23rd 2012
𝑥𝑇𝑀 = 𝑥𝑛 ∙ 𝐴𝐺 = 𝑥𝑛 / 2sin 𝜃
L. Di Fiore
7
Adopted components
• Light source: S-LED coupled to SM optical fibers
Spot size ≈ 400 mm
l ≈ 830 nm (longer wavelength should be OK)
• Sensor: Quadrant PD of PSD (Hamamatsu)
• Trans-impedance amplifier OP27EP
Paris May23rd 2012
L. Di Fiore
8
Electronic noise
elect. noise X
elect. noise Y
noise model typical
Noise model maximum
-9
A/Hz1/2
10
-10
10
-11
10
-3
10
-2
-1
10
10
Hz
Electronic noise (with light-off) agrees with the model according to
component characteristics (1/f1/2 slope)
Paris May23rd 2012
L. Di Fiore
9
Sensitivity measurement with a rigid set-up
the bench is machined from
a single block of stainlesssteel and has some interfaces
for fiber couplers and
sensors.
the "test mass" mounts some
mirrors and can be moved
for calibration
the system is symmetric for
differential measurements
(if necessary)
the whole set-up is closed in
a box to reduce thermal
variations and prevent effect
of air flows etc.
Paris May23rd 2012
L. Di Fiore
10
By increasing the power the noise decreases less than expected
Normalized asimmetry [1/Hz1/2]
10
10
10
10
22 uW
mode 21 uW
330 uW
model 330 uW
580 uW
model 550 uW
-3
-4
-5
-6
10
-3
Paris May23rd 2012
-2
10
Hz
L. Di Fiore
10
11
-1
Dependence of sensitivity on power
-5
10
Normalizer asimmetry
1/2
1/Hz
QPD measured 1
QPD measured 2
PSD measured
noise model
-6
10
-7
10
-8
10 -5
10
-4
10
power [W]
-3
10
With the QPD we observed en excess noise above about 0.15 mW
With the PSD the noise follows the model up to 0.5 mW
Paris May23rd 2012
L. Di Fiore
12
Noise at 1 mHz
-7
10
QPD 1
QPD2
PSD
PSD model
QPD model
capacitive sensitivity
-8
10
m/Hz1/2
-9
10
-10
10
-5
10
Paris May23rd 2012
-4
10
L. Di Fiore
-3
10
13
Comparison of QPD versus PSD sensitivity
-8
10
QPD 560 uW
PSD 550 uW
Capacitive readout
-9
m/Hz
1/2
10
-10
10
-11
10
-3
10
-2
-1
10
10
Hz
Measurement range 
QPD ~ spot size ~ 400 mm
PSD ~ detector size ~ 4.7 mm
NB: with PSD we can still gain a factor 2 (or more) by using a smaller sensor
Paris May23rd 2012
L. Di Fiore
14
Test on torsion pendulum
(in collaboration with the Trento group)
We implement an optical read-out system on the four masses
torsion pendulum in Trento
The goal was
• Check of performances,
reliability and sensitivity
• Check of back action level
• Contribute to improve the
performance of the facility (if
possible)
Paris May23rd 2012
L. Di Fiore
15
the ORO during the assembling phase
Paris May23rd 2012
L. Di Fiore
16
Angular measurement: j (2008)
10
rad/sqrt(Hz)
10
10
10
10
10
phiEM
Pendulum thermal noise
phiORO
DAQ noise
ORO expected noise
Tot noise
-4
-5
-6
EM
j =a(Z1 - h Z2)
ORO
j =1/(4lcosq)·(Dx1 + k Dx2)
-7
-8
-9
10
-4
10
-3
10
-2
10
-1
Frequency(Hz)
ORO sensitivity (~ 2·10-8 rad/sqrt(Hz)
EM sensitivity ((~ 3·10-7 rad/sqrt(Hz)
Paris May23rd 2012
L. Di Fiore
17
Force measurement (2008)
The ORO signal can be used, as the capacitive one, for putting upper limits to
the force noise
10
10
N/sqrt(Hz)
10
10
10
10
10
-8
-9
Pendulum thermal noise
j EM
EM
ORO
STC-EM
ORO-EM
j ORO
-10
(Xem – Xstc)
-11
Xoro – Xstc
-12
-13
-14
10
-4
-3
10
Frequency(Hz)
10
-2
10
-1
We get the same (or slightly lower) limit with the ORO
Paris May23rd 2012
L. Di Fiore
18
ORO as a readout for torsion pendulum facility
After this test we designed, in collaboration with the Trento group, an
auxiliary readout system for the 4 mass torsion pendulum facility based on
the same simple technology and using multiple reflections.
This increases the force readout sensitivity of the facility by almost one order of magnitude
Paris May23rd 2012
L. Di Fiore
19
Integration in LISA (eLISA/NGO)
• As a starting point, we studied the possible integration of the
ORO in the present design of the LISA Pathfinder inertial
sensor.
• The main problem is the little space left between the
electrodes to let the light reach the surface of the proof mass.
• the final LISA design should not be different from LISA-PF
electrode configuration.
Paris May23rd 2012
L. Di Fiore
20
Placing an ORO inside the IS is not an easy task
Paris May23rd 2012
L. Di Fiore
21
Proposed solution
position
sensors
Paris May23rd 2012
fiber
couplers
L. Di Fiore
22
The idea is to use the electrodes as mirrors for directing
the beams to the test mass surface and to the sensors
Paris May23rd 2012
L. Di Fiore
23
Optical fiber couplers
Position detectors
Paris May23rd 2012
L. Di Fiore
24
Bench top prototype (with PZT actuation in translation)
Photodiodes
Optical Fibers
Test mass
Calibrated
X,Y,Z PZT
actuator
The prototype was assembled successfully, there maximum correcrion
necessary for getting the beams at the center of the detectors was of about
0.2 mm, compatible with machining and assembling tolerances
Paris May23rd 2012
L. Di Fiore
25
A particular with
the plate where the
output fiber
couplers are
attached
Paris May23rd 2012
L. Di Fiore
26
Results:
• check of the design (passage of the beams etc.).
• measurement of the 6X6 sensing matrix in agreement to the
analytical model (within few %)  validation of the analytical model
• the measurement was only performed for the longitudinal DOFs (18
out of 36 matrix elements) because only a 3 DOF PZT system was
available
 X h    1.94  0.013
  
 X   0.012  0.002
 v 
 Y    0.04
0.34
 h 
Y  
 0.01
 v   0.01
  
0.03
 Z h   0.06
  
 0.02
 Z v   0.03
Measured
Paris May23rd 2012
0.17
0.03
 0.04
 0.06
 0.11
 1.83
? ? ?  X 
  
.. .. ..   Y 
  
.. .. ..   Z 
 
.. .. ..   a 
  
.. .. ..    
  
.. .. ?   q 
0
0
0
0
0.090   X 
 X h    1.932
  
  
X   0
0
0
0
 0.078
0  Y
v
  
  
Y   0
0.347
0
0
0
 0.054   Z 
 h 
 
Y  
0
0
 0.010
0
0   a 
 v  0
  
  
0
0
0
 0.031
0   
 Zh   0
  
  
Z
0
0

1
.
922

0
.
102

0
.
012
0
v

 q
 
Computed with analytical model
L. Di Fiore
27
Other configurations non using x face
Front view
Top view
3 beams  Z, Y, a, , q
Paris May23rd 2012
2 beams  Z, a, q
L. Di Fiore
28
We tested this new configuration with a new prototype. In this case we
addes an angular PZT staged so we checked 5 DOFs out of 6 (X, Y, Z, a, q)
TM with X,Y,Z
PZT actuator
inside
Calibrated a,q
PZT actuator
Paris May23rd 2012
L. Di Fiore
29
Also in this case we measured the sensing matrix (5 DOF) that is in good
agreement with the analytical model
Measured
Analytical
Paris May23rd 2012
 Yh   0.0053

 
 Yv   - 0.0293
 Z1   0.0277
 h 
 Z1v   0.0778
Z2  
 h   0.0162
 Z 2   0.0283
 v 
 Yh  

 
Y
 v  
 Z1  
 h 
 Z1v  
Z2  
 h 
Z2  
 v 
- 0.0007
0.3433
- 0.0001
- 0.0007
0.0005
- 0.0134
0.0026
- 0.0222
- 0.0003
- 1.9028
- 0.0004
1.9106
0.0021
0.0512
- 0.0002
0.1123
0.0004
0.0964
?
?
?
?
?
?
0.0551   X 
  
- 0.0041  Y 
- 0.0004   Z 
 
0.0005   a 
0.0004    
0.0029   q 
0
0
0
0
0 0.0538   X 
  
0 0.3473 0
0.0546
0
0  Y 
0
0
0
0
0.0281 0   Z 
 
0
0 - 1.9225 0.1020 0.0115 0   a 
0
0
0
0
- 0.0281 0    
0
0 1.9225 0.1020 - 0.0115 0   q 
L. Di Fiore
30
Search for space qualified component
• We started only recently to work in this direction
• It should be clear that we don’t wont to build flight hardware: industries
do that
• Our goal is to check if they already exist SQ components to be used for
the ORO, and to identify possible criticalities
• The main points are:
Paris May23rd 2012
electronics
light sources
fiber components and collimators
light detectors
L. Di Fiore
31
ORO readout- electronics
The electronic used for processing QPD signals is based on OP27EP op-amp.
There is a SQ equivalent component OP27AJ/QMLR.
At the beginning of 2011 we procured some samples and tested them in a
photodiode readout card.
The noise performance are
exactly the same as the standard
components
Il looks that PD readout will not
be a problem.
Care must be putted on the rest
of electronics (signal processing)
In order to maintain low power
consumption
Paris May23rd 2012
L. Di Fiore
32
Next steps:
We had some very useful discussion with ESA people. and got so very
useful suggestions:
We didn’t yet identified other components, but is seems that:
SQ SLED have been already used in some NASA mission: hard to gat
details
QPD and PDS have been used in space (includiong LISA-PF) and should
not be a problem
Fiber collimators looks a delicate point, because we cannot use standard
component close to the TM so a dedicated development will probably be
necessary.
Assembling procedure inside the vacuum chamber needs to be
investigated and could result very complex.
Paris May23rd 2012
L. Di Fiore
33
Conclusions
• Both bench top and suspended tests confirm that the ORO sensitivity can
be better than the capacitive one, above 1 mHz
• The noise level is well characterized, even if not completely understood
and allows to make predictions and trade-off between sensitivity and
measurement range
• There are possible layouts for the integration in the present design of the
inertial sensor, verified with bench-top models.
• study of space compatible parts is just started: electronics is non a problem
and it looks that there are available component already tested on flight:
Further studies are required.
• The ORO is a good candidate as a back-up sensor for the eLISA/NGO
inertial sensors, with possible sensitivity improvement.
Paris May23rd 2012
L. Di Fiore
34
Thank you for your attention
Paris May23rd 2012
L. Di Fiore
35