AA-mid demonstrator
Dion Kant
AAVP 2010
8 – 10 December 2010, Cambridge, UK.
Outline
• Project target
• Description of Work
• Project partnerships
• Current activities
• Evaluation results
• Conclusion
AAVP 2010
AA-mid 2 - GWK, 2010/12/10
AA-mid (draft) key requirements
Physical collecting area
~2000 m^2
Number of stations
10 to 15
Polarisation
Dual polarisation
Frequency range
400 – 1450 MHz
Architecture
Based on EMBRACE
T_sys
50K Target
Dynamic range @ ADC
42 dB
Instantaneous bandwidth
~250 MHz
Number of independent FoVs
2
HPBW (FoV)
16 degrees @ 1 GHz
Station output bandwidth
500 MHz
Requirements may still be adapted
AAVP 2010
AA-mid 3 - GWK, 2010/12/10
AA-mid time line, phases and
• Full project (~2000 m^2 AA) not funded today
• AAVS1 = first phase
• Funding for design phase within AAVS1
– 2010 to 2013
– Delivers design + small number of front end proto types
– Use EMBRACE evaluation as input for design
– Use EMBRACE platform as test-bed for new tile developments
– New design based on EMBRACE
• Site selection will be decided on later
• First testing will be performed at current EMBRACE stations
• Second stage testing: Use Portugal site (SKA like environment)
• Second phase (AAVS2) (beyond 2013) full roll-out AA-mid
demonstrator
AAVP 2010
AA-mid 4 - GWK, 2010/12/10
Design approach
• Base new design on EMBRACE architecture and other
demonstrators
• Incremental approach
– Level 1 topics (should)
– Level 2 topics (may)
• Main enhancements are (level 1 topics):
– Introduce dual polarisation signal processing
– Reduce power consumption per signal path
– Improve tile control system
– Reduce system noise
– Second generation beam former chip (4-bit phase control)
– Improve mechanical design, this includes housing & radome
– Improve design for manufacturability
AAVP 2010
AA-mid 5 - GWK, 2010/12/10
Design approach level 2 topics
• Investigate LNA above ground plane (integrated on feed board).
Main goal is to be able to use FR4 for feed board (element cost
reduction)
• Investigate balanced LNA approach
• Improve station processing bandwidth
– Input bandwidth (250 MHz per channel)
– Output bandwidth >2x250 MHz
– As a fall back current EMBRACE
processing hardware will be used
• Investigate non-continues
Unit/substation approach;
• Obtain better EM model of tile to predict radiation pattern
including polarisation behavior.
• ….
AAVP 2010
AA-mid 6 - GWK, 2010/12/10
Non-continues substations
• XX “tiles” in a substation.
• Every substation slightly tilted.
• Substation closed by EPS wall en top.
• Passive cooling by ground connection
through poles.
• SKA station formed by multiple
substations.
• Walkable spaces between substations.
• 56m diameter
288.000 antenna elements
AAVP 2010
7
AA-mid 7 - GWK, 2010/12/10
Design approach level 2 topics
Topics to be solved :
• Need software development partnership
• Need correlator suitable for 15x2x1GS/s inputs
• …
Can survive with current EMBRACE system in first phase of the
project
AAVP 2010
AA-mid 8 - GWK, 2010/12/10
Project partnerships
• Established members:
– OPAR, France: Part of system design, beam former chip design,
EMBRACE evaluation;
– ASTRON, the Netherlands: Part of system design, focus on level 1
front-end topics, industrialisation and EMBRACE evaluation
– …
• Members (becoming established):
– Italy, INAF: Possible topics: Receiver, LO system, …
– UK Several Institutes: Antenna design alternatives including electronic
integration and testing, Digital processing, …
– Germany, MPIfR: Signal distribution and multiplexing, digital
processing;
– Portugal, Preparation for Portugal station roll out and site development
– France, AD converter including poly phase filter bank integration
– …
AAVP 2010
AA-mid 9 - GWK, 2010/12/10
Current Activities and Progress
• First AAmid proto type tile underway
Octa boards
Aluminum ground planeConnection boards
AAVP 2010
Center board
AA-mid 10 - GWK, 2010/12/10
Current Activities and Progress
• First AAmid proto type tile underway
Bottom view
Octa concept
AAVP 2010
AA-mid 11 - GWK, 2010/12/10
Current Activities and Progress
AAmid Octa Concept; 3D bottom view
AAVP 2010
AA-mid 12 - GWK, 2010/12/10
Improved element mechanics
Multiple vivaldi
antennas in one sheet
Plastic support to keep tops
in place
LNA close to pick-up
point
AAVP 2010
SKA Station Design
AA-mid 13 - GWK, 2010/12/10
Aluminum Large Plates to decrease assembly time
Also expandable
(connectable) in other
direction
Metal-Metal joining by
clinching
AAVP 2010
SKA Station Design
AA-mid 14 - GWK, 2010/12/10
EM analysis of non-continues substations
• XX “tiles” in a substation.
• Every substation slightly tilted.
• Substation closed by EPS wall en top.
• Passive cooling by ground connection
through poles.
• SKA station formed by multiple
substations.
• Walkable spaces between substations.
• 56m diameter
288.000 antenna elements
AAVP 2010
15
AA-mid 15 - GWK, 2010/12/10
EM analysis of non-continues substations
Vacuum formed
“tile” bottom for
six modules
(approx. 1.5 x
2m)
Placement in field
by forklift or
crane, like LOFAR
HBA.
AAVP 2010
AA-mid 16 - GWK, 2010/12/10
Small tile: Simulate and measure
AAVP 2010
AA-mid 17 - GWK, 2010/12/10
Good resemblance model and measurements
Active reflection coefficients
Element 1
Element 3
5
MoM Simulation
Measurement
0
[dB]
[dB]
5
-10
-15
-20
-25
0
-5
| act(=0,=0)|
| act(=0,=0)|
-5
MoM Simulation
Measurement
0.5
1
Frequency [GHz]
1 = Edge
AAVP 2010
1.5
-10
-15
-20
-25
0.5
1
Frequency [GHz]
1.5
3 = Central
AA-mid 18 - GWK, 2010/12/10
Good resemblance model and measurements
Measured and simulated phases
S11
S12
-100
1
Frequency (GHz)
S21
1.5
0.5
-100
1
Frequency (GHz)
S31
1.5
1
Frequency (GHz)
S32
-100
0.5
1
Frequency (GHz)
S41
1.5
1
Frequency (GHz)
S42
-100
0.5
1
Frequency (GHz)
1.5
1
Frequency (GHz)
S43
1
Frequency (GHz)
1.5
1.5
0.5
1
Frequency (GHz)
S44
1.5
0.5
1
Frequency (GHz)
1.5
0
1.5
100
0
-100
0.5
1
Frequency (GHz)
S34
-100
0.5
degrees
degrees
-100
0.5
100
100
0
1.5
0
1.5
0
1.5
100
0
1
Frequency (GHz)
S33
-100
0.5
100
degrees
0.5
degrees
degrees
-100
1
Frequency (GHz)
S24
-100
100
0
0.5
100
0
1.5
100
0
1.5
-100
0.5
100
1
Frequency (GHz)
S23
100
0
0
-100
0.5
-100
0.5
degrees
1.5
degrees
0
degrees
1
Frequency (GHz)
S22
100
degrees
degrees
100
AAVP 2010
-100
degrees
0.5
0
degrees
-100
0
100
degrees
0
S14
100
degrees
100
degrees
degrees
100
S13
0
-100
0.5
1
Frequency (GHz)
1.5
AA-mid 19 - GWK, 2010/12/10
Current Activities and Progress
• EMBRACE Evaluation and test-bed development
AAVP 2010
AA-mid 20 - GWK, 2010/12/10
Current Activities and Progress
Status Nançay Station
• ~ 62 m2 area 56 Tiles
• 4032 Vivaldi elements (single linear polarisation)
• Currently 48 tiles cabled up end-to-end
• Grouped into tilesets of 4 tiles
• 64 tiles to be installed by end 2010
• Ultimately, 80 tiles in early 2011
AAVP 2010
AA-mid 21 - GWK, 2010/12/10
Current Activities and Progress
Status Westerbork Station
• 64 Tiles, (~70 m^2)
• 25 Tiles online in a 3x3 and 4x4 array configuration
• Ultimately 144 Tiles (160 m^2) in 2011
AAVP 2010
AA-mid 22 - GWK, 2010/12/10
Digital beam former details
AAVP 2010
AA-mid 23 - GWK, 2010/12/10
Modified architecture to more generic one
•
Full data rate can be exploited:
•
AAVP 2010
2x62 Ebeamlets per CDO output;
e.g. two beams of 12 MHz each
•
Digital beam former software less
complex (can use Matrix formalism now)
•
Increased update rate digital beam
weights
AA-mid 24 - GWK, 2010/12/10
Current Activities and Progress
EMBRACE station experiment objectives and results
• Detection of GPS satellite signals
• Scanned Beam Pattern on Afristar
• Redundant base line experiments on Afristar
• Solar Fringes and system noise verification
• Fringes of Cas A and Cygnus
• RECENT RESULT: Detection of pulsar
Plan
• Simultaneous detection of two pulsars within one FoV
• Simultaneous detection of two widely separated pulsars
using independent FoVs
• Detection of HI
• Simultaneous detection of HI and pulsars
• Beam swapping with beams on Cas A and Cygnus
• Correlate with WSRT dish
AAVP 2010
AA-mid 25 - GWK, 2010/12/10
Drift scan of GPS satellite
• Strong carrier at 1227MHz
• Subband statistics
• Pointing offset between
tilesets
AAVP 2010
AA-mid 26 - GWK, 2010/12/10
Early highlights: solar fringes
Interferometric measurement with 10 tiles

two E-W ULAs

frequency: 1179 MHz

integration: 10 s

bandwidth: 195 kHz
Initial performance est.

A/T = 4.7·10-3 (to sun)

Tsys = 103 – 117 K (to zenith)
AAVP 2010
AA-mid 27 - GWK, 2010/12/10
Early highlights: instrument quality
• Redundant baselines produce redundant visibilities
AAVP 2010
AA-mid 28 - GWK, 2010/12/10
Scanned Tile Beam pattern
measured on Afristar satellite
Model
Measurement
Fringes of Cas A!
• Experimental settings
• 3x3 array
• 1254 MHz
• 30 s integration
• 195 kHz bandwidth
• Initial conclusions
• Confirms A/T
• Correlation offsets
AAVP 2010
AA-mid 30 - GWK, 2010/12/10
Pulsar detection Setup details
AAVP 2010
AA-mid 31 - GWK, 2010/12/10
EMBRACE first pulsar result
AAVP 2010
AA-mid 32 - GWK, 2010/12/10
Conclusion
• AA-mid is ramping up
• EMBRACE Pulsar Detection demonstrates first Lband AA detection ever! It demonstrates:
•
Complete data path and processing is under control
•
Hierarchical beam forming is working
• and much more to come
Thank you
AAVP 2010
AA-mid 33 - GWK, 2010/12/10