1. Science at Q-band Manchester 14-15 September 2009
APRICOT
All Purpose Radio Imaging Cameras On Telescopes
Peter Wilkinson
U. Manchester
Design studies & sub-system prototyping for future large-format Q-band FPA “cameras” on large telescopes
2) MIC + MMIC devices from within Europe.
Science at Q-band Manchester September 2009

2. EC Framework7 “RadioNet” Instrumention R&D
APRICOT : Q-band camera subsystems + European
MIC/MMICs
(Caltech/JPL MMIC array spectrographs &
NRAO FPA goals overlap with APRICOT)
AMSTAR+ : mm/sub-mm cameras subsytems (SIS, MMIC)
(W-band FPAs overlaps with APRICOT)
(UNIBOARD: multi-purpose digital backends)
Science at Q-band Manchester September 2009

3. Some European Q-band telescopes
Yebes-40m
SRT 64-m
Effelsberg-100m
(new active 2ry
mirror)
Science at Q-band Manchester September 2009

4. Science at Q-band Manchester 14-15 September 2009
Science Strategy
“Cameras” enable these telescopes to make new surveys in scientifically rich range GHz –
In “intermediate” gap between SKA and ALMA
Follow-up with EVLA, VLBI, VSOP-2, ALMA Band 1
Continuum and spectroscopy: observations in
different weather conditions
Complement other mmsub-mm observations
Complement CMBR observations e.g. Planck
Science at Q-band Manchester September 2009

5. Science at Q-band Manchester 14-15 September 2009
APRICOT: Basics
Operating range: GHz
spectrally rich + many continuum applications
All Stokes parameters + spectroscopy at same time
Continuum band split into (n x few GHz) sub-bands for
atmospheric & spectral discrimination.
Broad-band IF output selected from anywhere within
overall band, sent to high-speed digital FTS (UNIBOARD)
Science at Q-band Manchester September 2009

6. Molecular Line Spectroscopy
Star-forming regions & circumstellar envelopes
Imaging plus modelling  temperature & density
Many carbon-chain species in the GHz band (HCnN n=3,5,7; CnH n=5,6; CnS n=1,3,5) diagnostic of cold dense quiescent gas
Other species:
SiO (shock tracer); OCS (sulphur sink)
CH3CN (hot core species); SO (Zeeman sensitive)
SiO masers in CSEs close to the star
With large format cameras could survey
complete clouds in one day
Blind surveys in redshifted CO (1-0)
Distances & mass estimates of dusty galaxies
Science at Q-band Manchester September 2009

7. Science at Q-band Manchester 14-15 September 2009
Continuum Studies
Surveys for discrete sources (in polarisation)
Find new types of AGN e.g. youngest CSOs
follow-up with mm-VLBI & VSOP-2 imaging
follow-up of GLAST transients
Provide net of calibration sources for EVLA and mm-VLBI
Support of Planck and all high-sensitivity CMBR experiments
Surveys for/of clusters of galaxies via the S-Z effect
Surveys of diffuse Galactic emission (in polarisation)
Synchrotron; free-free; anomalous dust; thermal dust
Need to dissect out the contributions: for ISM astrophysics
and CMBR polarised foregrounds
In compact regions e.g. YSOs - diagnostics of dust
agglomeration in protoplanetary disks
Science at Q-band Manchester September 2009

8. Scientific Synergy Surveys with radio cameras link these fields
a wealth of astrophysics
becomes accessible if we study
both galactic and extragalactic
foregrounds with precision
for progress in cosmology must
understand the unpolarized
and polarized foregrounds
with exquisite precision
Surveys with radio cameras
link these fields
MAS Meeting Pasadena, 27 April 2005
Science at Q-band Manchester September 2009

9. Relevant FPA experience
Science at Q-band Manchester September 2009

10. Science at Q-band Manchester 14-15 September 2009
EMBA: 32 GHz (MPIfR)
Dewar interior
7-beam horn array with different properties
Science at Q-band Manchester September 2009

11. Science at Q-band Manchester 14-15 September 2009
FARADAY GHz (IRA)
7 feed, hexagonal configuration with central feed
14 x 2 GHz IF outputs right and left polarization;
Feeds and LNAs cooled at 20 K;
Mechanical de-rotator to track the parallactic angle
Science at Q-band Manchester September 2009

12. Science at Q-band Manchester 14-15 September 2009
OCRA GHz (U. MAN)
Single polarisation
continuum GHz
Direct detection
All-MMIC receiver
based on WMAP/Planck-
LFI design
NGST InP MMICs
Horns all cold – behind
self-supporting 480mm
vacuum window
26-36 GHz 16-beams
Continuum (UMAN)
Science at Q-band Manchester September 2009

13. Science at Q-band Manchester 14-15 September 2009
EXPERIENCE......
NGST MMIC performance: very good - ITAR produced great delays
MMIC chips to “mass production” of LNAs  extraordinarily long
Dewar: many devices, cables, wires, connectors  big, complicated
Noise marker and LO distribution: simple in principle but complicated
LNA power supply: complicated, huge amount of thin wires
Weight: great
Maintenance: maybe lots
Science at Q-band Manchester September 2009

14. Hence the need for major changes in approach…
Science at Q-band Manchester September 2009

15. WP1 Receiver Architecture
MPIfR; IRA, UMAN, CAY, TCfA
Architectures for highly integrated multi-pixel receivers
Modular design with well-defined interfaces
Design for mechanical and cryogenic stability
Optimise layout for maintenance/fault-fixing
Design of monitor, control and calibration systems
Integration of direct detection and heterodyne systems
LO generation and distribution
Design, packaging and integration of RF, IF, LO systems
Establish capability to batch-produce RF, IF modules
Deliverables are mainly design study reports
Science at Q-band Manchester September 2009

16. WP2: passive components
IRA; MPIfR, UMAN
Highly integrated chain with OMT, hybrids,
transmission sections etc
Low-loss, low size/weight, low cost, ease of
manufacture
Standard waveguide technology too expensive
Needs technology shift
 Planar technology, microstrip transmission lines
and filters etc
Deliverables - design study reports
- few pixels hardware comparing performance
of conventional and “innovative” approaches
(with WP1 and WP3)
Science at Q-band Manchester September 2009

17. WP3: MIC/MMIC development
UMAN; IRA; MPIfR; CAY; URomeTV
To develop and secure European supply of world-standard MMIC devices for astronomy
To seek improved noise performance:
TLNA stuck at 5-6 x Quantum Limit for >10 years: why?
To explore/achieve increased levels of integration & multi-function capability within a MMIC circuit.
Science at Q-band Manchester September 2009

18. Science at Q-band Manchester 14-15 September 2009
MIC/MMIC producers
Fraunhofer Institute (IAF) Freiburg
100 & 50 nm GaAs mHEMT technology
Multi-function MMICs
Experimental 35nm processes
Now interested in low-noise at cryo temps
Link
with AMSTAR+
U. Manchester
100nm InP HEMT technology
innovation in materials and architecture for low-noise
rapid response to new design inputs
OMMIC company
70nm GaAs mHEMT technology
Science at Q-band Manchester September 2009

19. Science at Q-band Manchester 14-15 September 2009
WP4: Device Testing
CAY; UMAN; MPIfR, IRA; URomeTV
Accurate measurement of noise temperature and gain
fluctuation of devices at cryo temperatures not easy
Results from well-respected labs often differ !
CAY have lots of experience in this arena from LNA
work for Herschel (HIFI), ALMA, IRAM, ESOC etc
Science at Q-band Manchester September 2009

20. Synergy with “MMIC Array Spectrographs”
Caltech/JPL: aims similar to APRICOT/AMSTAR+
Push for lowering InP pHEMT noise closer to
quantum limit in collaboration with NGST
Why don’t MICs and MMICs perform the same
from wafer to wafer?
Why don’t MMICs in modules perform as well as
expected ? (QUIET experience)
Science at Q-band Manchester September 2009

21. Science at Q-band Manchester 14-15 September 2009
WP5: Data handling
TCfA: UMAN, IRA, MPIfR
Develop and test algorithms using the full range of
multi-pixel and multi-spectral data to reduce
effects of 1/f noise and atmosphere without spatial
switching,
(“on the fly-mapping” – strongly affects receiver concept)
Develop and test figures-of-merit to support queue-
scheduling of the receiver in both continuum and
spectroscopic modes
Combine knowledge with other mm-wave users
Science at Q-band Manchester September 2009

22. Science at Q-band Manchester 14-15 September 2009
Emerging Specification
Npixels and configuration
25100
Limits from optics: SRT ~30 beams for 0.5db loss
Configuration affects survey efficiency: hexagonal may not be best for beam-beam switching
Tsys
<50K
RX temperature <25K + sky ~ 25K .
Higher Tsys (>5 K) could be tolerated for larger Npix
Total RF Bandwidth
33-50 GHz
Waveguide band: Goal is 17 GHz BW to backend
Which few GHz is least important?
Polarisation type
LCP/RCP or X/Y
Needs detailed discussion
Polarisation purity
30dB
Set by science demands
Hard to maintain 30 db over 46% BW with circular?
Continuum channels per pixel 8
For atmospheric diagnosis and instant spectra.
Critical requirement for the architecture
Spectroscopic coverage
17 GHz instant
>2 GHz RF bands
Architecture possible to deliver all RF band, split in
channels per pixel, to spectrometer with no tuning
Gain stability /knee frequency
???
Acceptability? Depends on observation technique
Rapid scan rates allow on-the-fly mapping
major issue for WP5 simulations
Spectroscopic channels per pixel
64k for each 2 GHz band
Demand on backend to UNIBOARD
Resolution ~30 kHz – science “would like” 15 kHz
Science at Q-band Manchester September 2009

23. Science at Q-band Manchester 14-15 September 2009
Advice requested
Bandwidth goal is 17 GHz (46%) split into ~8 continuum bands
- which end is scientifically more important if compromises needed?
Receiver is simpler if the internal switching schemes are minimised/nil
– challenge is observing with 1/f gain variations and weather –
experience on antenna driving requirements ?
Pixel calibration:
- acceptable variation between pixels?
Spectroscopy
- is ~2 GHz instantaneous enough?
- is spectral resolution of 15 kHz = 0.1km/sec within a 2 GHz band OK?
- what is the specification on bandpass ripple within channels?
Polarisation specifications:
- linear or circular – any comments?
- separation purity: 30 dB ?- maybe hard to maintain withm circular
polarisation across a 46% BW)
Science at Q-band Manchester September 2009