Upload
hoangdung
View
216
Download
4
Embed Size (px)
Citation preview
IVTW -
Haystack 12012 October 22 1
The VLBI2010 Broadband System: First Geodetic Results
Reported by Arthur NiellMIT Haystack Observatory
IVTW -
Haystack 22012 October 22
GGAO12M Development TeamChris Beaudoin
1
, Bruce Whittier1, Mike Titus1, Jason SooHoo1, Dan Smythe1, Chet Ruszczyk
1, Alan Rogers1, Mike Poirier1, Arthur Niell1, Russ McWhirter
1, Alan Hinton1, Brian Corey1, Jon Byford, Alan Whitney
1
Chopo Ma
2
Ed Himwich3, Tom Clark3
Jay Redmond4, Skip Gordon4, Mark Evangelista4, Irv Diegel4, Paul Christopolous
3
Wendy Avelar
5, Chuck Kodak5, Roger Allshouse
5, Katie Pazamickas
5, Ricky Figueroa
5
1
MIT Haystack Observatory, 2
NASA GSFC,3
GSFC/NVI ,4
HTSI, 5
ITT
Science driver
Accurate terrestrial and celestial reference frames
1 mm position and 0.1 mm/yr velocity
Requirement for accurate sea level determination
Achieved by combination of VLBI, SLR, and GNSS
Next generation VLBI development
Fast slewing antennas to rapidly sample the atmosphere
High precision delay observable
Wide spanned bandwidth: 2 –
14 GHz
Four bands and two polarizations
High data rate: 8 Gbps
IVTW -
Haystack 32012 October 22
Background
IVTW -
Haystack 42012 October 22
12m antenna at Goddard Geophysical and Astronomical Observatory, Greenbelt, Maryland
IVTW -
Haystack 52012 October 22 5
~ 2.2 – 15 GHz Spanned RF Bandwidth
OLDOLD
Phase
Frequency (GHz) 142
Observing Frequency Bands
IVTW -
Haystack 62012 October 22
Odd channels from each pol’n for one band output to each Mk5C.
2 Gigabits/sec recorded on each Mk5C.
Total data rate: 8 Gbps
Feed and LNAs cooled to ~20K
Both senses of linear polarization usedAntenna
Control room
RF filterphase/noise cal
IVTW -
Haystack 72012 October 22
Caltech supplied
QRFH feed
2-12 GHz LNAs
IVTW -
Haystack 82012 October 22
Westford QRFH feed
All backend equipment in one rack
Picture (oops! DBE compatibility test in progress)
IVTW -
Haystack 92012 October 22
Broadband 8 Gbps Rack
IVTW -
Haystack 102012 October 22
GGAO12M SEFD -
2012 May
Dewar (~20K)
QRFH feed
Two broadband LNAs
Couplers for phase and noise cal injection
Phase cal
5 MHz spacing
Use all tones for each channel to obtain delay correction
Noise source for Tsys
80 Hz switching controlled by RDBE
Not implemented for these observations
RF to control room over optical fiber
Caused implementation delay because manufacturer did not meet advertised specs
IVTW -
Haystack 112012 October 22
Signal Chain Frontend
UpDown
Converter
Select frequencies and Nyquist
zone (NZ2 used)
Band frequencies (lower edge)
3.2 GHz, 5.3 GHz, 6.3 GHz, 9.3 GHz
Band width 512 MHz
RDBE
Joint NRAO-Haystack development (also in use at VLBA)
PFB: 32 MHz channels (sub-bands)
Record eight horizontal and eight vertical polarization channels
2 Gbps
data rate
Mark5C recorder
2 Gbps
1 module
IVTW -
Haystack 122012 October 22
Signal Chain Backend
Schedule using sked (simplified)
Select sources, band frequencies, SNR target per band
Specify scan parameters and antenna characteristics
Push GO to generate 24 hour schedule
Observing
Convert sked output to scripts for controlling RDBEs
and Mark5Cs
Make usual checks of frequencies and time (but four sets!)
Antenna is run by Field System
RDBEs
and Mark5Cs are run from scripts until Field System ready
Quantization threshold set for each scan by script
IVTW -
Haystack 132012 October 22
Geodetic Session Components
Correlation and observable estimation
DiFX
Difx2mark4
fourfit
coherent fit to 4 bands*2 polarizations + ionosphere
Geodetic estimation
nuSolve
with stochastic processes for atmosphere and clocks
IVTW -
Haystack 142012 October 22
Geodetic Session Components
2012 May six hour session
30 scans/hour
30 second scans
50 source s ≥
0.6 Jy
Median formal delay error ~0.5 psec
Scaled delay error ~3 psec
(1 mm)
Position estimation for GGAO12M
Delay residual RMS using νSolve 10 psec
Position uncertainties ~ 2/3/9 mm in E/N/U
Probably atmosphere dominated
Phase cal delay variation in azimuth of 4 psec
Possible 0.5 mm horizontal position error
IVTW -
Haystack 152012 October 22
Results
IVTW -
Haystack 162012 October 22
IVTW -
Haystack 172012 October 22
External sources of RFI
Usual S-band sources
Unknown 4 GHz at GGAO (NSA or CIA?)
Local comm
link at 6 GHz at Westford
Intra-technique RFI for geodetic Core Sites
Satellite Laser Ranging aircraft avoidance radar at 9.4 Ghz
Potential to damage LNAs
Must keep pointing direction 90º
apart
Coordinated observing appears difficult
Attempting to mitigate by physical blockage near radar
DORIS transmission near 2 GHz
IVTW -
Haystack 182012 October 22
RFI
IVTW -
Haystack 192012 October 22
Sky coverage no mask
0
30
60
90
60
30
0
030609060300
Observations from schedule file ./12279o.skd for experiment 121005 ( 774 scans)
WESTFORD (Wf) 774 scans0
30
60
90
60
30
0
030609060300
Observations from schedule file ./12279o.skd for experiment 121005 ( 774 scans)
GGAO7108 (Gg) 774 scans
IVTW -
Haystack 202012 October 22
Sky coverage mask on
0
30
60
90
60
30
0
030609060300
Observations from schedule file ./12278d.skd for experiment 121005 ( 785 scans)
WESTFORD (Wf) 785 scans0
30
60
90
60
30
0
030609060300
Observations from schedule file ./12278d.skd for experiment 121005 ( 785 scans)
GGAO7108 (Gg) 785 scans
2012 October 4-5
Two six hour sessions
34 scans/hour
30 second scans
SLR radar active on Oct 4; mask on
SLR radar off on Oct 5; mask off
Fringes both days
Both days correlated on Mark4; only three phase cal tones/channel
No post-correlation processing will be done till correlated on DiFX
Anxiously awaiting results to look at baseline repeatability
IVTW -
Haystack 212012 October 22
Results
12m antenna
Install motorized positioner
for Dewar/feed/LNA package
Implement noise cal for Tsys
Signal chain
Replace 4 RDBEs
by 2 quad-RDBEs
(4 IFs
per ROACH)
Replace 4 Mark5Cs by one Mark 6
IVTW -
Haystack 222012 October 22
Next steps
Broadband system
Successful implementation on the 12m antenna
Broadband RF from 2 –
12 GHz achieved with QRFH and two LNAs
Obtained anticipated sensitivity of 2500 Jy
for SEFD
Phase calibration allows phase connection over full range of frequency
Obtained anticipated delay precision of less than 4 picoseconds
Operations
UDC/RDBE/Mark5Cs operated remotely
Geodetic sessions run unattended after setup
Correlation/post-processing
DiFX
correlation
Coherent fitting of four bands dual polarization with ionosphere
estimation
Full end-to-end geodetic observations IVTW -
Haystack 232012 October 22
Summary
IVTW -
Haystack2012 October 22
Courtesy Wendy Avelar
24
IVTW -
Haystack 252012 October 22
IVTW -
Haystack 262012 October 22
RDBE-Q
Input four 512 MHz IFs
Output 2 x 2 Gbps
(current geodesy mode) or 2 x 4 Gbps
Output in VDIF with complex samples
Tested with test vector generator and test tone inputs
Mark6
8 Gbps
to one module (now)
Anticipate first observations in December
RDBE-X
Design begun (Haystack and Petrachenko)
ROACH2 based
Four 1 GHz IFs
16 Gbps
IVTW -
Haystack 272012 October 22
Signal chain development
GGAO12M
Elevation motion –
problem has not returned
Phase cal peculiarities –
most of system to be replaced
Positioner
Moves under motor control
Dewar holds vacuum
Testing for cooling right now
Install on 12m mid-November
Westford
QRFH feed and new hardware in Dewar performed well
Tsys
~ 30K
Efficiency very low
IVTW -
Haystack 282012 October 22
Haystack-NASA VGOS Systems
Kokee
20m
Haystack will install signal chain by next summer if observing permits
Kazan
Proposal for 12m with Broadband Signal Chain submitted by HTSI and Haystack
Too expensive, so fall back to Intertronics
S/X feed and LNAs
Receiver and formatter/recorder not discussed yet
USNO
Expect RFP within a month for 12m at Kokee
NASA Request for Information (RfI) for Core Sites (up to 10)
Response by HTSI/Haystack for combined VLBI/SLR/GPS/DORIS
Separate response by Haystack to provide VLBI systemIVTW -
Haystack 292012 October 22
Other VGOS development
IVTW -
Haystack 302012 October 22
Geodetic VLBI Highlights
VLBI2010 broadband system at GSFC
12m antenna with all major components
First geodetic session scheduled, observed, and correlated
DiFX
correlator
and fourfit
Developing new capabilities for broadband operation
Coherent fitting across polarizations and bands
Correlation of mixed channel widths , e.g. 8 MHz/32MHz
Mark6 recorder
16 Gbps
will reduce four recorders to one
Kazan University VLBI2010 signal chain (proposed)
May be second operational system
Cooperative development with HTSI (Honeywell)
IVTW -
Haystack 312012 October 22
Program Overview
NASA (2010-2014)
Engineering
Operations
Research and development
Technique improvement/science
USNO correlator
(2009-2012 (2013 likely))
Engineering and development
Operations backup
USNO VLBI2010 implementation (-2012)
VLBI2010 signal chain for Kokee
20m antenna
Kazan University VLBI2010 signal chain (proposed)
Sub-contract to HTSI for 12m antenna
IVTW -
Haystack 322012 October 22
NASA Contract
Engineering support
Field station diagnostics and repair
Analyse
station performance from correlator
results
Respond to problems from stations
Maintain spares depot; repair modules; ship replacements
Develop replacement equipment
Correlator
development (Mark4 and DiFX)
Operations
Mark4 and DiFX
correlators
R&D and Reference Frame sessions
New broadband data
Westford antenna
GGAO12M
Other sites as needed
IVTW -
Haystack 332012 October 22
NASA Contract (cont’d)
Research and development
Frontends (Cryogenic feeds and amplifiers)
Receivers (RF to video)
Digital backends
Recorders
VLBI2010 broadband system
12m antenna at Goddard Space Flight Center
Broadband system on Westford
Technique improvement and science
RFI mitigation
Observation planning
Atmosphere
Radio astronomy ‘effects’
IVTW -
Haystack 342012 October 22
Other Active Contracts
USNO: Washington Correlator
engineering support
through 2013 probably
Maintenance of Mark4 hardware correlator
Haystack correlator
as backup
USNO: VLBI2010 signal chain implementation
UpDown
converters, RDBEs, and Mark5Cs –
1st
contract
Dewar/feed/LNAs/calibration box, optical fiber links –
current
Installation on Kokee
20m antenna
IVTW -
Haystack 352012 October 22
Potential Contracts (cont’d)
Kazan University (Russia) VLBI2010 system (proposed)
Will be under a contract from HTSI
New 12m to be purchased from Intertronics
Antennas
HTSI (Honeywell) responsible for everything else
Haystack to provide the broadband signal chain and checkout of geodetic VLBI functionality, first in US then in Russia
Moscow State University (Russia) VLBI2010 systems
HTSI: same arrangement
Proposal under discussion
GSI (Japan)
Interest in obtaining front end (Dewar/LNAs/feed)
Others interested in VLBI2010 Signal Chain
Norway, Sweden, Spain/Azores/Canary Islands
IVTW -
Haystack 362012 October 22
NASA Geodesy and Haystack
Motivation for new system development (IVS WG3)
Replace aging antennas
Improve accuracy to 1 mm
Provide for continuous, unattended operation
Resulted in VLBI2010 specifications
Background for Haystack involvement
Provided majority of data acquisition systems for global program
Significant contributions to atmosphere and astronomy modeling
2012 October 22 IVTW -
Haystack37
http://ivscc.gsfc.nasa.gov/about/wg/wg3/IVS_WG3_report_050916.pdf
ftp://ivscc.gsfc.nasa.gov/pub/m isc/V2C/TM-2009-214180.pdf
Limiting error sources
Varying atmosphere delay
Sensitivity
Strategy
Use fast-slewing antennas (5º/sec-12º/sec slew rate)
Obtain delay sensitivity through high data rate and wide spanned
bandwidth (Broadband Delay)
Design goals
Antennas of ≥
12m diameter
Data rates ≥
8 Gbps
using four bands of 0.5 GHz to 1 GHz each
Spanned bandwidth 2.2 GHz to ~14 GHz: delay uncertainty ~4 psec
BUT maintain observing compatibility with current S/X systems
IVTW -
Haystack 382012 October 22
VLBI2010 development
IVTW -
Haystack 392012 October 22
VLBI2010 signal chain
Cooled broadband QRFH feed and LNAs
Purchased from Caltech
Tested and packaged by Haystack
Phase and noise calibration
Phase cal generator developed by Haystack
Packaged by HTSI
UpDown
Converters (4) (Haystack)
In use for 2-12 GHz
Possible extension to 2-14 GHz (or higher)
IVTW -
Haystack 402012 October 22
VLBI2010 signal chain
RDBE digital back end (4)
Developed by NRAO and Haystack
Produced by Digicom
(California)
In use by VLBA and NASA
Mark5C recorder (4) (Conduant)
In use by VLBA and NASA
In use by EVN
Mark6 recorder (1) (Conduant)
Up to 16 Gbps
from four 4 Gbps
inputs
Will replace 4 Mark5C recorders
IVTW -
Haystack 412012 October 22
VLBI2010 correlation/processing
Correlation on DiFX
software correlator
Four separate correlations (four bands)
Working on 8MHz within 32MHz correlation for compatibility with legacy S/X systems
Post-correlation processing
difx2mark4 –
convert native output to Mark4 input
fourfit
Coherent fit across bands and polarizations while estimating ionosphere
IVTW -
Haystack 422012 October 22
VLBI2010 Status
12m system
Operational at Goddard Space Flight Center
Using deliverable equipment
Some components not tested or not implemented
First geodetic observing session just completed
Westford
Using deliverable equipment
Low efficiency (≤
20%) above 5 GHz not understood
Kokee
20m
Most of signal chain completed
Funding/mechanism for mounting Dewar unclear
IVTW -
Haystack 432012 October 22
NASA Space Geodesy Project (SGP)
Project level at NASA Headquarters
Viewed within NASA HQ as best way to get support and funding
Planning for ten sites
Not all in North America
Haystack involvement
Short term requirements
Complete and document 12m broadband system
Potential role of Haystack in SGP future
Not known at this time
Role of Haystack outside of SGP
Expected to continue as currently defined
IVTW -
Haystack 442012 October 22
Current Technical Development
Mark6
RDBE
RDBE-Q (four 512-MHz bands per ROACH)
RDBE-V (two 1-GHz bands per ROACH)
Future: four 1-GHz bands per ROACH
MCI (complete monitor and control)
RFI mitigation
Vector Tie System
High speed networking (real-time and e-xfr)
IVTW -
Haystack 452012 October 22
Technique Improvement
Implement end-to-end scheduling/analysis capability
Develop improved atmosphere modeling
Study effect of radio source structure
Analysis, modeling, and delay correction
IVTW -
Haystack 462012 October 22
IVTW -
Haystack 472012 October 22
Astro/Geo Common Dev’ment
Shared goals (simplified)
Improve sensitivity
Increase data storage capacity
Differences
Astronomy –higher frequency for better angular resolution
Geodesy –
wider spanned bandwidth for better delay sensitivity
Astronomy –
campaign observing few times per year, several antennas
Geodesy –
sustained observing several times per week, many antennas
Astronomy –
mobilize personnel & equipment for limited time with close to best possible capability
Geodesy –
operational reliability through less than best achievable performance
Unique measurements by VLBI
Earth rotation angle (time): UT1-UTC (keep GPS up)
Scale of terrestrial reference frame: sea level change
Nutation: (e.g.) dynamics of the Earth’s core
Mandate to NASA for accuracy improvement
National Research Council decadal survey
1 mm position and 0.1 mm/year velocities
Implementation proposed through Space Geodesy Project
VLBI, SLR, GPS, and DORIS
Unified station
Vector ties among techniques to 0.1 mmIVTW -
Haystack 482012 October 22
VLBI geodesy
Observations
One or two scans per minute scattered around the sky
Scan lengths only long enough to get necessary SNR
Duty cycle only about 1/3 so can use burst mode to maximize instantaneous data rate.
IVTW -
Haystack 492012 October 22
VLBI2010
Antenna and data acquisition
Cooled broadband frontend 2 –
14 GHz
Flexible RF to IF frequency conversion
Digital backends
High data rate recorder(s)
DiFX
software correlator
Extract all phase-cal tones
Correlate 8 GHz S/X legacy data and 32 MHz broadband channels
Post-correlation
Coherent fitting of all polarizations
Estimate differential ionosphere
IVTW -
Haystack 502012 October 22
VLBI2010 System
Haystack as supplier
Worked with Digicom
to make RDBE available
Working with Conduant
to make Mark6 available
Building Dewar assembly and UDCs
Complete for USNO 20m this year
Kazan University through HTSI (proposed)
Potential customers
Moscow U
GSI Japan
Auscope, other Patriot 12m
IVTW -
Haystack 512012 October 22
VLBI2010 Implementation
Instrumentation development
Digital back end
1 GHz bandwidth (2*512 MHz)
Higher data rate per chassis (2 ADCs/ROACH)
Recorder: 16 and 32 Gbps
reliable recording
Increase spanned bandwidth
Haystack’s role in VLBI2010 implementation ?
Science
Atmosphere limitation –
turbulence and asymmetry
Radio source structure effects
VLBI-unique geophysical applicationsIVTW -
Haystack 522012 October 22
Going forward
Coherently fit all four bands and all four cross-hands for phase delay, amplitude, rate and differential ionosphere.
Incorporate source structure phase.May have to be incorporated in the coherent fit.
IVTW -
Haystack 532012 October 22
Post-correlation
Four independent 512 MHz RF bands
Each recorded on separate Mark5C at 2 Gbps
(Mk5B fmt)
Two linear polarizations for each 512 MHz band
Each pol’n/IF becomes sixteen 32-MHz LSB channels.
The channel set to be recorded is selectable.
Half of the channels from each polarization are recorded.
Currently: H1 V1 H3 V3 …
H15 V15
Switch soon to non-redundant spacing
e.g. 1 2 4 7 9 13 14 15
IVTW -
Haystack 542012 October 22
VLBI2010 Broadband System
IVTW -
Haystack 552012 October 22
VLBI2010 signal chain (cont’d)
Mark6
16 Gbps
demonstrated on prototype
Replace four Mark5Cs with one Mark6
Hardware prototype to be delivered in next two months
Resilient file system development beginning
Will accommodate slow or failed disks
Control plane software development beginning
Hope to begin testing in two months
Quadridge Feed Horn (QRFH)
California Instit
ute of Technology Design
Sandy Weinreb Ahmed Akgiray
Bill Imbriale
2‐14 GHz Unbalanced Design One single‐ended port per polarization
Designed for Patriot
12m Antenna Shaped Optics
Adaptable to other antenna optics
Overcomes limitations of commercial quadridge designs
Have observed 60% efficiency on 12m antenna
2012 October 22 56IVTW -
Haystack
2012 October 22 57IVTW -
Haystack
12m performance
fiber link noise
Current
Mark4 (RCP)
6 S-band channels
10 X-band channels: 8 upper sideband plus lower sideband on outer two.
DBE (when S/X stations have them)
Probably 3 DBEs
(2 for X and 1 for S)
Single pol’n: 15 channels per DBE
Dual pol’n: 8 channels per pol’n
per DBE
IVTW -
Haystack 582012 October 22
S/X System
Correlate all cross-hands
IVTW -
Haystack 592012 October 22
Correlation –
DBE to DBE
Broadband to Mark4 cross-correlation
8 MHz Mark4 channels 14 USB + 2 LSB
32 MHz PFB channels LSB
May have multiple 8 MHz channels within 32 MHz channel
Each type to be correlated against both types in same scan
Polarization
Mark4 channels RCP
Broadband channels linear polarization
IVTW -
Haystack 602012 October 22
Correlation –
Mk4 to DBE
IVTW -
Haystack 612012 October 22
Mark4 vs
DBE8-MHz analog channels (Mark5A or Mark5B – one module) S1U
S2U
S3U
S4U
S5U
S6U
X1L
X1U
X2U
X3U
X4U
X5U
X6U
X7U
X8L
X8U
PFB 32 MHz channels (Mark5C - 3 modules Mk5B format)|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---| S-band (2 pol’n)
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---| Xlow-band (512 MHz) (2 pol’n)
Xhigh-band (512 MHz) (2 pol’n) |---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Correlate Mark4 against both linear polarizations of the broadband system
X1L X1U X2U Mark4 (8 MHz)--------
--------
--------
RCP
PFB (32 MHz LSB)--------
--------
--------
--------
Vertical pol’n
--------
--------
--------
-------- Horizontal pol’n
0 frequency(MHz) 32
IVTW -
Haystack 622012 October 22
Example of 8 MHz vs
32 MHz
Mark4 Mark4
PFB PFB
IVTW -
Haystack 632012 October 22
Correlation