Transcript RadioNet FP7

WP8: JRA “APRICOT”
All Purpose Radio Imaging
Cameras on Telescopes
Peter Wilkinson: U. Manchester
Objectives
• To develop the design and the sub-system technology for
multi-pixel focal-plane “radio cameras” in the range 33-50 GHz
(Q-band)
• To secure state-of-the-art High Electron Mobility Transistors
(HEMTs) and Monolithic Microwave Integrated Circuits (MMIC)
devices from within Europe.
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
2
Task 1: Architecture
Progress
 More thermal modelling & adding experience of building two
relevant receivers for the 100-m at Effelsberg.
o 2-beam K-band (mainly for spectroscopy)
o 2-beam Q-band (mainly for continuum)
 Deliverable 8.16 “Comparison of passive chain performance against
classical designs ” – main responsibility transferred to INAF-IRA
 Milestone 8.04 “Definition paper of functions and interfaces for
building blocks”– now drafted
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
3
Predicted thermal deformation of
the horn platelet array
When cooled from 300K to 15K At the
corners of the hexagon a deviation of
about 2mm is expected which has to be
taken into account when fixing the
structure to the cryostat vessel.
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
4
Radiation from the vacuum window
into the horn block structure
The demands to the vacuum window are
challenging: A test setup with these materials
has been measured under real conditions.
The window was placed between the 300K
environment and a 70K radiation shield under
vacuum conditions. The temperature at the
inside of the foil/foam window was measured to
236K. This value was basis for the calculations
in the next table .
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
5
Thermal power contributions
(assuming 25 horns)
Power item
power dissipation LNAs
waveguide lead
wire lead
radiation from surfaces
radiation from vacuum window
mechanical fixtures
Total:
300K to 70K
300K to 15K
70K to 15K
1,3
0,3
1,5
0,1
1,5
2,2
24
34
10
70,2
3
6,2
1,5
The cooling power needed to keep the system at reasonable temperatures
requires the largest cooling machines on the market.
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
6
Some other issues
DC supply in the dewar: lots of wiring to power & control the LNAs (5 wires
per amplifier)  ~250 wires are an infrastructure problem and power
dissipation is not good. MPIfR are working on an I2C approach to minimise
wires.
Mass: The table below summarizes the weight of the different parts of a
complete 25 pixel receiver camera.
Component
Cryostat incl. mechanical supports (tbc)
Horn
OMT
Polariser
Q-Band Front End Module
Down converter unit
Base band unit
DC Supply
Mechanical frame
Total
11 November 2011
kg/part
Nr. of parts kg total
150,00
1
150,0
0,39
25
9,6
0,23
25
5,7
0,05
25
1,3
0,30
25
7,5
0,50
25
12,5
0,20
100
20,0
1,00
25
25,0
25,00
1
25,0
0,0
0,0
0,0
257
FP7 RadioNet Board APRICOT
JRA, South Africa
7
Task 2: Passive Components
Progress
– Platelet array built and tested
– Innovative high performance polariser built and tested
– Platelet OMT designed and prototype being constructed
Deliverable 8.15 – changed
– “Designs for very low loss passive components, on low loss substrates” to
be replaced “The integration of low-loss passive and active components”
(joining together of programmes in Task 2 and Task 3 a better practical
step towards a working prototype receiver)
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
8
7-HORN PLATELET
SUMMARY
Concept: 45 no glued rings packed together
Material: Aluminum
Horn mouth diameter: 44mm
7-Horn length: 119mm
7-horn weight: 2.4kg
Construction: CNC machine
Prototype cost: 5380 €
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
9
7-HORN PLATELET
Summary of the Performance
MULTIHORN; B = 33-50 GHz; BW = 41%
Specification
Frequency
33 GHz
41.5 GHz
Achieved
50 GHz
Nr. of horn
7
Edge taper
S11
-10 dB@[email protected];
G/Tsys optimization for SRT optics
≤-25 dB
≤-25 dB
≤-25 dB
Crosspolarisation
≤-30 dB
≤-30 dB
SLL
≤-20 dB
Horn Isolation
Insertion loss
HPBW
Beam separation
11 November 2011
33 GHz
41.5 GHz
50 GHz
7
-6.2 dB
-10.6 dB
-18.1 dB
≤-30 dB
≤-30 dB
≤-30 dB
≤-30 dB
≤-28.9 dB
≤-32.8 dB
≤-30.9 dB
≤-20 dB
≤-20 dB
≤-22 dB
≤-22 dB
≤-20 dB
Best effort
Best effort
Best effort
≤-60 dB
≤-60 dB
≤-60 dB
≤0.3dB
≤0.3 dB
≤0.3 dB
≤0.1dB
≤0.1dB
≤0.1 dB
33 “
27.1 “
24.8 “
33 “
27.1 “
24.8 “
68.9 “
68.9 “
68.9 “
68.9 “
68.9 “
68.9 “
FP7 RadioNet Board APRICOT
JRA, South Africa
10
POLARISER
SUMMARY
Concept: single device iris polariser
Material: copper (golden inside)
Diameter: 10 mm
Length: 107 mm
Weight: 50gr
Construction: Electroforming
Prototype cost: 4100 €
POLARISER; B = 33-50 GHz; BW = 41%
Summary of the performance - the
best achieved anywhere across such a
wide (>40%) band
11 November 2011
Specification
Achieved
S11
≤-35 dB
≤-30 dB
S21
≤ 0.1dB
≤ 0.15dB
Phase Unbalance
  2o
 2.5o
Crosspolarisation
≤-35 dB
≤-33 dB
Ampl. Unbalance
≤ 0.05dB
≤ 0.05dB
FP7 RadioNet Board APRICOT
JRA, South Africa
11
OMT PLATELET
SUMMARY
Concept: 1 platelet repeatable turnstile OMT
Material: Aluminum
7-OMT diameter: 150 mm
7-OMT heigth: 47 mm
7-OMT weight: 1.6kg
Construction: CNC machine (flanges);
chemical erosion (sheets)
Prototype cost: TBD
Expandable OMTs array
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
12
OMT PLATELET - Simulation
Design Specification:
-Return loss:  -20dB
-Insertion loss:  0.3dB
-Isolation:  -50dB
-deliver the two outputs parallel each other,
parallel to the omt axis and possibly,
symmetrical with respect to the common port axis
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
13
Extension of Passive Concepts –
deliverable 8.15
Q band Front End Module (Q-FEM)
1) Each OMT has to be connected to LNA input:
Waveguide to microstrip transition
2) A Noise Cal signal has to be injected to each receiver chain:
Noise diode integration: No Cal distribution.
INTEGRATION OF ACTIVE DEVICES TO PASSIVES
Q-FEM
7-OMT
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
14
Q-FEM module – aspect of
deliverable 8.15
Output
Connectors
Bias
Connectors
Directional
Coupler
Noise Diode
room
Microstrip
LNAs
W to M Inputs
Q-FEM block diagram
W-M
Diode
W-M
11 November 2011
Coupler
SUMMARY
Concept: integration of many functions
Noise Cal: one diode per horn, no distribution
Injection method: twin directional coupler
Connection to OMTs: ridged transition
Connection among devices: microstrip
Mechanics: twisted OMT outputs for coplanar WG
inputs
Prototype cost: TBD
FP7 RadioNet Board APRICOT
JRA, South Africa
15
Task 3: Active Devices
Important Matching Efforts
• Contract between IRAM, MPIfR and IAF before the start of FP7
– assessment of the potential performance of IAF’s mHEMT process at
cryogenic temperatures.
– extending IAF’s mHEMT device model from room- to cryogenic
temperatures using MPIfR’s cryogenic s-parameters probe station
• Agreements in June 2008 & Sept 2009 between FG-IGN and the
University of Cantabria (UC) and IAF
– IGN is funding B. Aja (UC) to work at IAF on MMIC designs.
– FG-IGN gets access to space in IAF multi-project wafers to implement
MMIC designs of interest for FG-IGN.
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
16
Comparisons now available
FP5 FARADAY: NGST MMIC packaged into LNA
NGST 100nm InP from FARADAY project measured at 15K physical:
• TLNA 40K at 33 GHz reducing to 25K at 50 GHz. Average value is ~30K.
• TLNA of the packaged InP MMIC reduces by a factor 7 from RT (TLNA ~200K)
to 15K (TLNA ~30K
Cryogenic optimisation with IAF: design via IGN ; measurements
MPIfR
Very recent mHEMT MMIC using IAF cryo-optimised technology with 50nm
technology.
• TLNA 25K-60K going the other way to the NGST MMIC
• Gain is ~25 dB – perhaps could be higher?
Initial conclusion: average TLNA is close to the NGST InP technology
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
17
Q band LNA:
IAF MMICs (Beatriz Aja IGN)
•50 and 100 nm 31-50 GHz MMICs in Yebes;more finished designs on the way
•2× 31-50 GHz LNAs mounted and tested at room temperature
Design
Run /
Technology
Status
Measurements
Comments
V1 - coplanar
733c /
50 nm
9 units in Yebes
2 LNAs finished
On wafer
LNA RT noise
CT measurements
pending system
calibration
V1 - coplanar
734a /
100 nm
5 units in Yebes
On wafer
Error in selection of
chips sent, chosen
from wrong wafer *
V2 - coplanar
735Cryo /
50 nm
IAF
On wafer
V3 - µstrip
735Cryo /
50 nm
IAF
Resonance in
ground line
V4 - µstrip
740 /
50 nm
processing
V3 redesign without
resonance
11 November 2011
FP7 RadioNet Board APRICOT JRA, South Africa
18
Task 3: Low-noise transistors
• Industry standard Molecular Beam Epitaxy facility in UMAN
•Aim is to maximise performance at cryogenic temperatures
– by changing the spacer layer to increase the mobility and keep the
conductivity from saturating & hence to stop the Tlna flattening
off as the Tphys reduces (only a factor 7-10 reduction in noise
temperature compared with a factor 15 in physical temperature).
• Novel high breakdown InP pHEMT topology already established and
wafers all made
• Main hurdle has been establishing close cooperation with e-beam
partners
e-beam collaborations
IEMN collaboration:
• First pHEMT transistors fabricated with IEMN Lille’s standard lithography with gate
length 130nm (UMAN supplied material and IEMN did lithography).
• First wafer sent known to have inferior InP substrate.
• Nevertheless gave results which were reasonable but not competitive with NGST
Cryo-4 in the CAY test amplifier at 4-8 GHz.
• “Good” wafers sent to IEMN – but they have not yet delivered due to
unforseen problems with their furnace – now expect delivery in Q4 2011
Glasgow collaboration :
• Manchester does most of the lithography Glasgow writes the gates
• Trying 2 and 4 finger architecture and a range of gate widths.
• At the first iteration in summer 2011 there was a problem with incorrect metal
lift-off in Manchester – now redesigning the architecture slightly and learn lesson
of metal-lift-off – expect delivery in Q4 2011
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
20
MMIC down-converter
programme
Challenges
• Integrated down conversion module - quickly replaceable element
• Broad-band mixing stage with low power dissipation
• Multiplier function which allows distribution of a low frequency LO
• Commercial-off-the-shelf technology, stable and repeatable :
OMMIC 70nm metamorphic process
Progress:
- Chip now delivered and much testing/characterisation completed
Task 3 Down converter programme at OMMIC
Mixer & x8 Multiplier integrated onto the same 3x2mm2 chip:
Summary of results
LO section: original LO is injected at 4 GHz – for multiplication up to
32 GHz – chip also has option for injecting LO directly at 32 GHz with
external filter.
o Performance of sub-sections of circuit have been compared in
detail with simulations of a wide range of parameters – looks
pretty close in general.
•Amplifying-mixing section: comparisons of measured performance
and simulated performance are again pretty good.
o Some difference between isolations LO-RF and LO-IF –
explicable in terms of OMMIC wafer run characteristics not
being completely as expected (higher gain).
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
23
Task 4: Establishing accurate
performance of LNAs
FG-IGN (& INAF)
Progress:
• C and K band test amplifiers have circulated
through 4 labs
• Slower than hoped progress on Q band
cryogenic measurement system
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
24
Q band LNA:
First prototypes (UCan / Yebes)
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
25
First results of transfer amplifiers
11 November 2011
YCA (4-8 GHz)
YK (20-25 GHz)
• YK (20-25 GHz) and YCA (4-8 GHz) LNAs sent to 4 laboratories (MPG,
UCan next destinations):
– UTV
YK
Noise & S par. @ 300 K
– UMan
YK
Noise & S par. @ 300 K
– INAF-IRA YK & YCA
Noise
@ 300 K &15 K
– UCan
YCA
Noise
@ 44 K
FP7 RadioNet Board APRICOT
JRA, South Africa
26
First results of transfer amplifiers:
K band room temperature noise
YK22 005
T=300K
300
25
250
20
200
15
150
10
100
Yebes (NFM)
UTV (NFM)
UMan (NFM)
INAF (NFM)
5
Yebes (VNA)
UTV (VNA)
UMan (VNA)
Yebes
UTV
UMan
INAF
Noise Temperature (K)
Gain (dB)
30
50
0
0
16
17
18
19
20
21
22
23
24
25
26
Freq. (GHz)
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
27
First results of transfer amplifiers:
K band cryogenic noise
YK22 004
T=15K
30
30
25
24
20
18
15
12
10
6
Yebes (LED OFF)
Yebes (LED ON)
Noise Temperature (K)
Gain (dB)
36
5
Yebes (LED OFF)
Yebes (LED ON)
INAF (LED OFF)
0
0
16
17
18
19
20
21
22
23
24
25
26
Freq. (GHz)
11 November 2011
FP7 RadioNet Board
APRICOT JRA, South
Africa
28
Q band LNA:
Noise measurements system
• Measurement system still not ready
– Delays due to defective noise source
(already repaired and recalibrated by Agilent)
• Pending calibration of lines and components inside the Dewar
• New cryogenic attenuator
– Based on Chalmers designed ATC quartz chip
– High fidelity of temperature readings (due to material and design)
could simplify measurement system
– Commercial alternative not tested at cryogenic temperature
11 November 2011
29
FP7 RadioNet Board
APRICOT JRA, South
Task 5: Simulations etc
• Challenges
– Develop and test algorithms for the subtraction of atmospheric water
vapor without spatial switching
– Develop and test figures-of-merit to support queue-scheduling of the
receiver in both continuum and spectroscopic modes,
– To develop the hardware and software architecture for the back-end
of an APRICOT cameras.
11 November 2011
FP7 RadioNet Board APRICOT
JRA, South Africa
30
Deliverable:Atmospheric model
Assumed approach:
1) Hydrodynamical simulations of clouds formation. Numerical
determination of time dependent optical depth, and emissivity
variations (vs T(z) and wind speed).
Current status:
(Not started)
2) Reference analytical Kolmogorov-Taylor (K-T) model of
(In progress)
atmospheric surface brightness variations due to WV instabilities
3) OCRA-f receiver noise model
(Almost done)
4) Scanning strategy model
(Done)
5) Atmosphere sky field realizations (Beams model and uniform
elevation gain model)
(In progress)
6) Atmospheric noise observations with 8-beam OCRA-f system @
TCfA
(In progress)
7) TCfA atmosphere characterization by data fitting via MCMC
parameter estimation within combined K-T and receiver models.
Calibration of (1) and (2).
(In progress)
Deliverable: Atmospheric WV subtraction
without spatial beam switching
Assumed approach:
1) Implementation of various scanning strategies
Current status:
(Done)
2) Development of: RT-32 control system / Simulated RT-32 control (Almost done)
system to facilitate scanning procedures
3) Scanning strategies tests on RT-32/RT-32 simulator
(In progress)
4) OCRA-f receiver (total power) noise model (noise simulations)
(Almost done)
5) Atmospheric model (simulated atmosphere)
(In progress)
6) Optimization of scanning strategies to mitigate atmospheric
brightness fluctuations
(Not started)
Deliverable: Dynamic Queue
scheduling strategies
Assumed approach:
1) OCRA-f receiver noise model
Current status:
(Almost done)
2) Real-time OCRA-f noise parameter estimation (MCMC maximum (Almost done)
likelihood method)
3) Noise estimation processing server
(Not started)
4) Observation alert GUI monitor
(In progress)
Scanning software
Scanning strategies
WP8: Deliverables
Expect completion of all deliverables in standard period - with exception of
three involving realisation of hardware which will use the no cost extension to Q2
2012.
Additional work will also continue on Task 5 – matching -funded by TCfA.
Numbe
r
8.08
8.14
8.16
Name
Proposed Date
(months)
Forecast
date
Euro MICs: advanced technology
devices aimed at improved noise
performance (Task 3)
24
Q2-2012
Euro MMICs with improved noise
performance (Task 3)
33
Comparison of passive chain
performance against classical
designs (Task 1)
35
11 November 2011
Comments
Q2-2012
Q2-2012
FP7 RadioNet Board APRICOT
JRA, South Africa
36