Phased Array Processors for Submm VLBI

Transcription

1 Phased Array Processors for Submm VLBI SMITHSONIAN ASTROPHYSICAL OBSERVATORY SMA ACADEMIA SINICA IAA Event Horizon Telescope Workshop 26 January 2010 Jonathan Weintroub

2 Phased Array Processors for Submm VLBI SMITHSONIAN ASTROPHYSICAL OBSERVATORY SMA ACADEMIA SINICA IAA Event Horizon Telescope Workshop 26 January 2010 Jonathan Weintroub Collaborators: Rurik Primiani, Ken Young, Jim Moran, Ray Blundell, Bob Wilson John Test, Chris Schaab, Paul Yamaguchi (CfA/SMA), Dan Werthimer, and CASPER group at UC Berkeley VLBI Project PI Sheperd Doeleman, MIT-Haystack, and his colleagues notably Alan Rogers, Dan Smythe, Alan Hinton, Mike Titus, Ken Wilson & others SAO IR&D and REP funding 2006 to 2008, and SMA Support

3 Sensitivity is crucial: Submillimeter Valley, Mauna Kea, Hawaii CSO 10 m single dish (79 m 2 ) (aggregate area 482 m 2 equivalent of 25 m aperture) JCMT 15 m single dish (177 m 2 ) SMA eight 6 m dishes (compact configuration) (226 m 2 )

4 VLBI Phased Array Processor for SMA fielded April 2009 for the first time

5 ...using existing systems... SMA computer controls a DDS providing a 109 MHz phase-reference for locking the LO Used to implement fringe stopping, Walsh switching (not normal mode), and phase corrections 1024 MHz of IF tappedoff from each antenna 5

6 ...using existing systems MHz maser reference used to clock all digital hardware 1 PPS provided via GPS for VLBI clock syncing 100 Hz heartbeat used for data alignment 6

7 ...some new analog hardware... Two 512 MHz sub-bands, MHz (low) and MHz (high) Amplifiers and adjustable attenuators provide adjustable analog power levels 7

8 ...and CASPER electronics 5 iadcs for sampling at 1024 MHz and clocking fabric, 3 ibobs, for processing time-domain samples and shipping to storage terminal, and a single BEE2 corner chip for a 7-baseline calibration correlator 8

9 ibob delay compensation System deployment and test Jan 2009 BEE2 calibration correlator Mark5B+ # 1 Mark5B+ # 2 Linux control computer Analog phase compensation Analog Pre- Processors UPS power conditioning

10

11

12 Local VLBI Tests of Phaseup A lesson learned is that getting an unequivocal proof of concept result is comparable in effort to designing the actual system & bench testing Inconvenient to arrange true long baseline VLBI partner for each test, so developed local VLBI special mode Break out single antenna from phased array to be comparison antenna. Store this IF to one Mark5B+ via standard DBE Store phase-up via 10 Gbps custom DBE to other Mark5B+ Suppress all phase modulation on comparison antenna, and offset tuning by 117 Hz Scripting of scans in a comprehensive matrix of modes to tease out what was going on

13 Local VLBI phaseup results Source: 3C454.3

14 Fringe amplitude vs N

15 Postscript: Sunset and pointing Source: 3C454.3

16 Remaining issues in SMA PA XAUI latency alignment for eight antenna phase up Walshing for sideband separation and contemporaneous SMA correlator operation Correlator has DC offsets in lags (not affecting performance) Minor issues of convenience, such as AGC on IF levels, JTAG programming quirks Expand to 8 and then 16 Gbps, dual polarization, probably increase sample rate

17 Application to other arrays CARMA has similar systems to SMA, can adapt SMA array. 1 GHz sample rate and possible digital interface are complications. May feedback dewalsh developed at CARMA to SMA PdBI is, I think, fairly similar. However don t know implications of new back end: defer to discussion period. ALMA technically quite different, due to antenna digitization, DTS, hooks in correlator. Phase up principles are applicable, though. Polarization conversion for ALMA probably needed in digital domain, since not possible to fit 1/4 wave plates (at 230 GHz, 345 GHz may allow 1/4 wave plates)

18 ALMA Correlator hooks (ALMA memo 584 Escoffier Webber Lacasse Saez & Greenberg) Summation Logic FDM or TDM Modes

19 Straw-man ALMA Phased Array Figure 2: Functional Block Diagram of Phasing System for ALMA VLBI Operations.

20 Conclusions Sensitivity of the submillimeter VLBI array is crucial, to provide reliable detections on long baselines on which the source is partially resolved CASPER DSP technology enabled development of SMA phased array providing threefold improvement in sensitivity of the Mauna Kea station Lessons learned: do not underestimate the effort needed to develop test systems for unequivocal demonstration of performance Similar technology can enable phase-up of CARMA and PdBI. Similar principles can be applied at ALMA (though technology differs) Phase up of CARMA, PdBI, and ALMA is discussion period topic

21 Station Location A (m 2 ) λ(mm) Bandwidth (GHz) dual pol? maser? SMA Mauna Kea ; only Y JCMT Mauna Kea ; only (Y) CSO Mauna Kea ; ? (Y) Hawaii Mauna Kea ; only Y CARMA Cedar Flat, CA N Y CARMA-8 Cedar Flat, CA? N Y Mt. Graham SMTO Arizona ; & 0.8 Y LMT Volcán Sierra Negra Puebla, Mexico only N ASTE Atacama, Chile N N APEX Atacama, Chile ; & 0.8 N ALMA-10 Atacama, Chile ; 0.8 4/ & 0.8 N IRAM PdBI IRAM 30m Plateau de Bure, France ; only? Spain & 0.8? Pico Veleta

22 Supplementary Material

23 Hot spot models, T = 27 minutes, 230 GHz, ISM scattered 50 μas 50 μas no Spin 0.9 Spin Animations by Broderick and Loeb (2008)

24 Station Location A (m 2 ) λ(mm) Bandwidth (GHz) dual pol? maser? SMA Mauna Kea ; only Y JCMT Mauna Kea ; only (Y) CSO Mauna Kea ; ? (Y) Hawaii Mauna Kea ; only Y CARMA Cedar Flat, CA N Y CARMA-8 Cedar Flat, CA? N Y Mt. Graham SMTO Arizona ; & 0.8 Y LMT Volcán Sierra Negra Puebla, Mexico only N ASTE Atacama, Chile N N APEX Atacama, Chile ; & 0.8 N ALMA-10 Atacama, Chile ; 0.8 4/ & 0.8 N IRAM PdBI IRAM 30m Plateau de Bure, France ; only? Spain & 0.8? Pico Veleta

25 Hot spot models, T = 27 minutes, 230 GHz, ISM scattered 50 μas 50 μas no Spin 0.9 Spin Animations by Broderick and Loeb (2008)

26 !"#$%&$'&()'*+,-'.$/+0$'0$%*(-1')2$'-$3)'#(-$'(-')2(&'$4%(#5' A;(B$'C()D&E';FCGH%>&)%8BE'I'6=0(#'JKKLM'

27 Closure phase is a sensitive robust observable (alternative to true aperture synthesis imaging) Doeleman, Fish, Broderick, Loeb & Rogers, ApJ, 695, 59, 2009

28 Closure phase is a sensitive robust observable (alternative to true aperture synthesis imaging) Doeleman, Fish, Broderick, Loeb & Rogers, ApJ, 695, 59, 2009

29 Mark4 Fringe Plots Mk4 Fringe Plot 3C454_3.unlkxj, , SZ SMAP - SMAZ, fgroup U, pol LL Fringe quality 9 Amp. and Phase vs. time for each freq., 14 segs, 9 APs / seg (9.00 sec / seg.), time ticks 10 sec SNR PFD e Intg.time Amp Phase Sbdelay (us) Mbdelay (us) Fr. rate (Hz) Ref freq (MHz) AP (sec) Exp. Exper # Yr:day Start Stop FRT mm : Corr. date: 2009:328: Fourfit date: 2009:328: Position (J2000) 22h53m s " 29

30 Mark4 Fringe Plots Mk4 Fringe Plot 3C454_3.unlliv, , SZ SMAP - SMAZ, fgroup U, pol LL Fringe quality 9 Amp. and Phase vs. time for each freq., 14 segs, 9 APs / seg (9.00 sec / seg.), time ticks 10 sec SNR PFD e Intg.time Amp Phase Sbdelay (us) Mbdelay (us) Fr. rate (Hz) Ref freq (MHz) AP (sec) Exp. Exper # Yr:day Start Stop FRT mm : Corr. date: 2009:328: Fourfit date: 2009:328: Position (J2000) 22h53m s " 30

31 Mark4 Fringe Plots Mk4 Fringe Plot 3C454_3.unllll, , SZ SMAP - SMAZ, fgroup U, pol LL Fringe quality 9 Amp. and Phase vs. time for each freq., 14 segs, 9 APs / seg (9.00 sec / seg.), time ticks 10 sec SNR PFD e Intg.time Amp Phase Sbdelay (us) Mbdelay (us) Fr. rate (Hz) Ref freq (MHz) AP (sec) Exp. Exper # Yr:day Start Stop FRT mm : Corr. date: 2009:328: Fourfit date: 2009:328: Position (J2000) 22h53m s " 31

32 Mark4 Fringe Plots Mk4 Fringe Plot 3C454_3.unllob, , SZ SMAP - SMAZ, fgroup U, pol LL Fringe quality 9 Amp. and Phase vs. time for each freq., 14 segs, 9 APs / seg (9.00 sec / seg.), time ticks 10 sec SNR PFD e Intg.time Amp Phase Sbdelay (us) Mbdelay (us) Fr. rate (Hz) Ref freq (MHz) AP (sec) Exp. Exper # Yr:day Start Stop FRT mm : Corr. date: 2009:328: Fourfit date: 2009:328: Position (J2000) 22h53m s " 32