Transcript Menardi

Fringe Sensor Unit
ESO/PAOS progress meeting
Leiden, 29 September, 2004
S. Menardi, ESO
Overview (1)
Scope of the contract with
Alenia: development of FSU
A and B operating in K band
with provisions for H band.
The FSU combines the light
of each object (Primary Star
and Secondary Star),
collected by two VLTI
telescopes and delivers and
records measurements of:
• Optical Path Difference
[modulo 0]
PS
s1
SES
PS
s1
s2
Telescope T1
(UT or AT)
Metrology end point E1
s2
Telescope T2
(UT or AT)
Baseline, B
Metrology end point E2
Star Separator 1
(STS1)
Beam A1
(SES)
SES
Star Separator 2
(STS2)
Beam B2
(PS)
Beam B1
(PS)
Delay Line 1
(DL1)
OPD
Controller
OPDPS mod 0, GDPS, APS
Beam A2
(SES)
Delay Line 2
(DL2)
OPDPS mod 0, GDPS
data storage
Differential Delay
Line 1B (DDL1B)
• Group Delay
Fringe Sensor
Unit B (FSUB)
Metrology
System
Differential Delay
Line 2B (DDL2B)
GD
L
data storage
OPDSES mod 0, GDSES
• Fringe Amplitude
data storage
Differential Delay
Line 1A (DDL1A)
Fringe Sensor
Unit A (FSUA)
(or MIDI,
or AMBER)
Differential Delay
Line 2A (DDL2A)
OPDSES mod 0, GDSES, ASES
OPD
Controller
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 2
Overview (2)
FSU operating principle
B
achromatic

light
from T2
p2 & s2
 
s1 + s2
p2 - s2| = 90°
p1 + p2
s1 + s2
p1 + p2
A
BC
p1 + p2
s1 + s2
PBS
PBS

p1 + p2
compensator
light
from T1
C
 
p1 & s1
s1 + s2
D
Ck

 

ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 3
Overview (3)
FSU main components
OPD Controller
PRIMA WS
FSU LCU 1
FSU LCU 2
IRACE
Shutter
System
Alignment
System
PRIMA
Metrology
(1319 nm laser)
Re-imaging
Optics
B
A
Beam
Combiner
Polarising
Beamsplitters
C
D
PRIMA
Metrology
(1319 nm laser)
Single-mode
fibers
(Spatial filters)
IR Array
Detector
Dispersive
element
Cryostat
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 4
Project organisation
Alenia Spazio, prime contractor

opto-mechanics design and procurement

system engineering

system assembly, integration and verification
Osservatorio Astronomico Torino (OATo), main sub-contractor

Cryostat design and procurement

Measurement algorithms and performance analysis

Software development (LCU level)
ESO furnished equipment

2 x PICNIC detectors & IRACE systems

Control Electronics Hardware
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 5
Overall Configuration (1)
FSU-A and FSU-B Overview (K band only)
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 6
Overall Configuration (2)
Shutter System and Alignment System
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 7
Overall Configuration (3)
K-PRISM and Compensator Assembly
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 8
Overall Configuration (4)
Beam Combiner Assembly
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 9
Overall Configuration (5)
Polarising Beamsplitters, Doublets and Fibers
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 10
Overall Configuration (6)
K+H
band
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 11
FSU Optics (1)
Glass compensator
Task: Compensation of LAD introduced by differential
air path (+/- 120 meters, 5 x 48 m regions)
Dimensions in mm
Description: Infrasil® plano parallel plates
with suitable thicknesses
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 12
FSU Optics (2)
Alignment Unit Mirrors
Task: Alignment of input beams (pupil, image, OPD)
w.r.t. VLTI artificial source Leonardo
Dimensions in mm
Description: 2 actuated flat mirrors on each beam
(2 x 5 degrees of freedom)
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 13
FSU Optics (3)
Achromatic Retarder and Compensator
Task: Create a /2 phase delay between p and s (Retarder)
and Compensate for OPL inside the K-prism
(Compensator)
Dimensions in mm
Description: Retarder is a K-Prism (3 internal reflections)
Compensator is a parallelepiped.
Both in Infrasil®
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 14
FSU Optics (4)
Beam Combiner
Dimensions in mm
Tasks: combines both telescope beams, introduces a 
phase delay between transmitted and reflected beams,
combines both metrology beams, reject unwanted
polarisation component of metrology beams,
reject metrology laser stray light (angular deviation).
Description: Beamsplitter cube 50/50,
linear polarisers @ 1319 nm in the 2.5 mm central area,
wedges to reflect metrology laser in different direction.
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 15
Beam combiner
s_MET _in
p_MET _in
Re1
T r2
Wedged
Polarizer
h
b
Beam-Splitt er
Cube
beam-split ter
cube
Wedged
Polarizer
c
T r1 Re2

s_MET _out
a
clear
apert ure
In1

Y
Z


d
Y
X
Wedged
Polarizer
p_MET _out
In2
Side view
Top view

Wedged
Polarizer
Y
X
Wedged
Polarizer

ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 16
FSU Optics (5)
Metrology interface
Task: Inject/extract metrology beams from the
stellar beams path (in central obscuration)
Dimensions in mm
Description: two holed mirrors, reflect stellar beams
and transmit the metrology beams
Metrology and stellar beams are common-mode
up to beam-combiner
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 17
FSU Optics (6)
H/K Dichroics
Task: splits H band (reflected) and K band (transmitted)
and reject metrology laser light
H + K + 1319 nmK1319 nmH
Dimensions in mm
1319 nm
K
Description: Dichroic coatings on both sides,
Substrate with small wedge
H
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 18
FSU Optics (7)
Polarising beam splitters
Task: splits p and s polarisation components
Dimensions in mm
Description: 2 PBS designed in K band, with 2 corner cube
retroreflectors to minimize thermal background
(the fiber “sees” its own core)
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 19
FSU Optics (8)
Injection doublet
Task: Injection of K band in the optical fiber
Dimensions in mm
Description: Achromatic doublet for fiber injection
Manufacturing: Fused Silica and Zinc Selenide
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 20
FSU Optics (9)
Optical fibers
Task: Spatially filter the combined beams,
transport flux inside the cryostat
Dimensions in mm
Description: Optical fibers for K band (LeVerre Fluore).
Fiber positioners (New-Focus)
Manufacturing: Single-mode fiber in ZrF4.
On the cold side, the 4 fibers are glued in a metallic block
to from a square array (3 um accuracy). NA = 0.17
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 21
Cryostat
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 22
Cryostat (2)
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 23
FSU Optics (10)
Cold collimator
Task: Collimation of fiber output beams
Description: Achromatic doublet
Fused silica and Zinc Selenide
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 24
FSU Optics (11)
Cold prism
Task: disperse A, B, C, D beams
reject >2.5 um (cold K filter)
reject MET laser straylight (1/500)
Description: Fused silica prism, Wedge ~ 12º,
dichroic coatings
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 25
FSU Optics (12)
Cold camera
Task: Projects dispersed spots on the array detector
Description: Single aspheric lens
Manufacturing: Zinc Selenide
Front surface is aspherical, rear surface is spherical
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 26
Cryostat (3)
Cold plate
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 27
Detection Algorithm Software Architecture
TAC Standard blocks
tacTIMBlock: IRACE Timing and
Algorithm Scheduling
Probe: Algorithm results storage by
callback function , max rate 8 KHz
Monitor: Real Time Display of last
computed quantities up to 100Hz
Custom blocks
tacIRVMEBlock: manage the
detector raw data in CDS and NDRO
readout mode
tacOPDAmpBlock: produce OPD
and squared Amps at up to 8 kHz
tacGDBlock: implement the
algorithm for GD estimate at user
selectable rate up to 200Hz
tacFluxBlock: provide the flux
estimates in the 3 spectral channels
tacRTNBlock: deliver OPD, GD and
squared Amps to the OPD Controller
at the rates applicable to each
quantity.
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 28
Detector readout modes
DWT
DWT
Sequencenr loop Run-Time
Sequencenr loop Run-Time
Pixel Charge
Pixel Charge
RD1
RD2
RD2
RD1
RD1
RD2
RDn-1
RDn
RDn
RD2
RD2
DIT
Trigger by TIM
time
DIT
RST
Trigger by TIM
RST
RST
RD1
RDn-1
RDn
RD1
DIT
time
RST
DIT
min Tr=0.125 ms
min Tr=0.125 ms
Tr= Readout Interval
Tr= Readout Interval
IRQ
IRQ
IRQ
IRQ
IRQ
IRQ
time
time
Algorithm results
available
Algorithm Results
Available
TAC Algorithm Run-Time < Tr
TAC Algorithm Run-Time < Tr
Read-Reset-Read
Multiple Non Destructive Readouts
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 29
Measurement algorithms: OPD
OPD estimate:
iterative procedure of least square fit of measured data sk to nominal data fk
First step: linear range identification
Selection of minimum error position
among three initial points (x1, x2, x3)
in the fringe period
Subsequent steps: iterations of zero
crossing estimate formula using
tabulated functions f, g, l, h’
e( x, z1 )  
( sk  f k ( x )) 2
 k2
k
zn 1  zn 
h ( zn )
h ' ( zn )
h( x )   sk  gk ( x )  l ( x ) signature function
k
g n ( x)  2 
f n' ( x) /  n2
l ( x )   sn ( x )  g n ( x )
weight function
Linear iterations required: 3
Template resolution:
1 - 5 nm
bias function
n
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 30
OPD performance – Noise
Simultation results Source = 3500 K
Evaluated vs. signal + background photon noise, read-out noise
Integration time
[ms]
0.25
0.25
0.25
0.25
2
2
Assumed Readout noise [e rms]
11
11
11
11
4
4
Magnitude (UTs)
[mag]
7
8
10
11
13
14
Req. on OPD noise
[nm]
11.1
19.0
71.1
159
72
140
OPD noise, p = 0 m
[nm]
6.2
11.2
44.5
109
46
98
OPD noise, p = 30 m
[nm]
6.2
11.3
47.0
114
48
105
OPD noise, p = 60 m
[nm]
6.8
12.3
50.8
124
49
110
OPD noise, p = 90 m
[nm]
7.4
13.3
54.6
133
55
121
OPD noise, p = 120 m [nm]
8.5
15.3
61.6
163
64
147
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 31
OPD performance – Linearity
OPD mod  measurement linearity - direct I/O comparison
Requirement:
d m
 1  10%
d
Fulfilled from zenith to
120 m air path in delay line
Simulation results:
OPD derivative – 1 Mean Peak RMS
[%]
[%]
[%]
p=0m
0.000 0.063 0.028
p = 30 m
0.000 0.212 0.052
p = 60 m
0.000 0.158 0.044
p = 90 m
0.000 0.133 0.035
p = 120 m
0.000 0.114 0.026
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 32
Measurement algorithms: Group Delay
GD estimate: by least square fit of measured data sk to nominal data fk
Montecarlo simulation over GD range [-6.1 mm, 6.1 mm]
4067 cases uniformly distributed - data resolution 3 nm
Source: point-like, T = 3’500 K and T = 25’000 K
Nominal FSU configuration - Spatial template resolution: 5 nm
Fringe jumps included (not removed)
Implementation approach:

find local minimum of error in central fringe using OPD algorithm

find global minimum by error comparison over 6 fringes:
z1±, z1±2, z1±3

 z1
 z2
adjust local minimum around z2
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 33
Group Delay – Noise
Simultation results Source = 3500 K
Evaluated vs. signal + background photon noise, read-out noise
Integration time
Readout noise
[ms]
[e rms]
Magnitude UTs
5
5
200
200
2000
2000
5.1
5.1
0.8
0.8
0.3
0.3
10
12
14
16
16
19
Req. on GD noise
[nm]
900
3300
800
1900
600
2300
GD noise, p = 0 m
[nm]
5.8
346
5.8
398
4.7
641
GD noise, p = 30 m
[nm]
5.9
370
6.0
383
4.9
551
GD noise, p = 60 m
[nm]
68
504
237
592
132
698
GD noise, p = 90 m
[nm]
500
799
771
811
796
917
GD noise, p = 120 m
[nm]
1171
1208
1012
1104
964
1186
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 34
Group Delay – Linearity
GD measurement linearity - direct I/O comparison
Noiseless FSU GD output computed for a set of ~2500 input GD values,
uniformly distributed in the range [-6.1 mm, 6.1 mm]
Point-like source @ T = 3’500 K
Template resolution: 5 nm
GD derivative – 1
p=0m
Mean
[%]
0.000
Peak
[%]
0.036
RMS
[%]
0.009
p = 30 m
0.000
0.033
0.009
p = 60 m
0.000
0.026
0.009
Requirement:
d m
 1  20%
d
Fulfilled from zenith to
60 m air path in delay line
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 35
Group Delay – Bias
GD measurement bias - direct I/O comparison
Noiseless FSU GD output computed for a set of ~1000 input GD values,
uniformly distributed in the range [-0/2, 0/2]
Point-like source @ T = 3’500 K
Template resolution: 5 nm
GD bias
p=0m
Mean
[nm]
0.000
Peak
[nm]
0.014
RMS
[nm]
0.006
p = 30 m
0.000
0.013
0.005
p = 60 m
0.000
0.010
0.004
Requirement:
d m  d  2.5 nm,
d  d 0  0 / 2
Restricted GD range derived
from recent definition of GD bias
specification on central fringe
Fulfilled from zenith to 60 m air path in delay line
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 36
Group Delay – Bias
p=0m
p = 60 m
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 37
Sensitivity analysis
Method:
modify FSU parameters (A beam only) & evaluate FSU output variation
Required knowledge of the transmission spectral distribution:
0.5% on transmission over full K band
2% on single 100 nm spectral region
Required knowledge of the phase spectral distribution:
1º over full K band
5º on single 100 nm spectral region
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 38
Sensitivity analysis – warm fiber end alignment
Questions:
How well shall A, B, C, D fibers be mutually aligned?
What is the differential instant coupling efficiency (for a
given misalignment)?
Results:
Differential coupling efficiency for 1 mm misalignment:
~0.5% average/PTV, 0.1% RMS
Conclusion:
Assuming a uniform distribution of transmission
perturbation of ±0.26% (independent for each fiber), the
fraction of configurations exceeding the specified 2.5 nm
peak GD error is below 5% (acceptable).
 0.5 mm fibre misalignment is acceptable
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 39
Sensitivity analysis – Cold Camera Alignment
Fiber alignment along dispersion is critical, as it affects the spectral
response of the FSU A, B, C, D channels.
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 40
Sensitivity analysis – Cold Camera Alignment (2)
Simulation: 1117 GD values in [-1.11 mm, 1.11 mm] - resolution 2 nm
Peak I/O discrepancy: ±2.42 nm
RMS: 1.26 nm
Linear discrepancy: variation in apparent effective wavelength
Peak value compatible with GD bias requirement
Conclusion: ~1 mm alignment stability of cold camera is acceptable
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 41
FSU calibration (1)
FSU Calibration procedure
Rationale:
detailed characterisation of instrument parameters
Global approach:
includes VLTI optical train and average atmosphere
Method:
FSU A + B in calibration mode, OPD scan
(Fourier Transform Spectroscopy)
FSU A on Fringe Tracking loop, FSU B measuring for self-calibration
A  B roles (tracking / calibration) exchanged for calibration of FSU A
Requirements:
FSU A + B; AT (UT); DL; STS;
MET; known bright star
Purpose:
FSU spectral response
Target effective 
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 42
FSU calibration (2)
Source requirements: knowledge / stability temperature or K magnitude
Requirements more relaxed for higher temperature sources
Not unreasonable for coldest stars:
Requirements on lab source:
few 10 K, 0.02 mag
1 K @ T = 800 K
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 43
FSU calibration (3)
On-sky calibration sequence for FSU B (FSU A tracking)
1.
Configure both STSs in calibration mode (+ telescope pointing etc.)
2.
Acquire and centre stellar fringes on FSU A and FSU B independently
3.
Close fringe tracking loop on FSU A: DL A driven by FSU A
4.
Reset PRIMA metrology
5.
DL B driven by FSU A + MET to cancel internal dOPD + OPD scan offsets
6.
FSU B outputs recorded during OPD scan
7.
OPD noise on FSU A and on MET supposed to average down to nm level
8.
Fourier transform of FSU B output
9.
Removal of source spectrum
10. Computation of transmission (modulus) and phase (argument) distributions
Procedure verified vs. photon + readout noise on FSU B
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 44
FSU calibration (4)
Plot of Fsu output A, 100 mm OPD scan, T = 6000 K source
Requirements for good spectral sampling (31 points) : 300 mm scan
Exposure requirements by Montecarlo evaluation of noise on
measured transmission (req. < 0.5%) and effective  (req. < 0.5e-4)
K = 10 mag, TI = 400 ms  240 s total (100 ms OPD step actuation)
Nearly independent from source spectral type
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 45
On-line diagnostics
Colour index
Spectral resolution of FSU detection system  photometric diagnostics
Spectral changes in measurement conditions: channel balance variation
CI  100
I 3  I1
,
I2
I 1  A1  B1  C1  D1, ...
Colour index variation in 7 transmission perturbation cases (sensitivity
analysis) and cold camera alignment, 1 s equivalent integration:
Colour index
T1
Mean
nominal
0.92
T2
RMS
0.015
Mean
perturbed
0.97
Variation
[]
3.15
0.92
0.013
0.96
3.03
T3
0.92
0.015
0.93
0.66
+1 mm
0.92
0.015
-2.65
-241.14
-1 mm
0.92
0.015
4.44
237.64
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 46
System Verification Facility (Alenia)
SVF Configuration:
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 47
FSU Calibration Facility (Paranal)
Collimator
Fiber Head
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 48
Achieved milestones and next steps
Achieved

Contract Kick-off: July 2002

Final Design Review: September 2003

Order of fiber bundle (critical long lead item): Feb. 2004

First release of Ali LCU software and FSU WS software: May 2004

Cryostat acceptance tests and PICNIC detector integration: Sept. 2004
Next steps

Finalise procurement of Beam Combiners (prototypes are available)

Complete opto-mechanical mechanical integration (by end 2004)

SVF and Calibration Faciliy manufacturing

Acceptance testing, scheduled in March 2005

Delivery of FSU A and B, in June 2005
ESO/PAOS progress meeting
Leiden, 29 September 2004 Page 49