LWA Station Design

S. Ellingson, Virginia Tech N. Kassim, U.S. Naval Research Laboratory

URSI General Assembly – Chicago – Aug 11, 2008

JPL

Long Wavelength Array (LWA)

An LWA Station

State of New Mexico, USA

Technical Goals:

10-88 MHz tuning range

Baselines up to 400 km for resolution [8,2]’’ @ [20,80] MHz

52 “stations”; (mJy-class sensitivity

Each station is an array of dipole-like elements in ~100 m diameter aperture for FOV = [8,2]°

Access to Galactic Center (low gain antennas)

LWA ScienceAstrophysics

CosmologyHigh redshift radio galaxies, containing the earliest black holesEvolution of dark matter & dark energy by differentiating relaxed & merging clusters

Acceleration, Propagation & Turbulence in the Interstellar MediumOrigin, spectrum & distribution of Galactic cosmic raysSupernova remnants & Galactic evolutionPulsars

Solar Science & Space WeatherRadio heliography of solar bursts & coronal mass ejectionsSolar radar

Exploration of the Transient UniverseNew coherent sources (More GCRT J1745-3009s?)GRB Prompt EmissionMagnetar FlaresExtra-Solar Jupiters: Detect magnetic field; conditions for life?Poorly explored parameter space…new sources

Ionospheric Physics

Unprecedented continuous spatial & temporal imaging of the ionosphere

Test and improve global ionosphericmodels

Dual Polarized Antennas (“Stands”) Per Station

Each station needs to contribute sufficient collecting area to ensure calibratibility.

Estimates of # of dual-pol antenna elements (stands) required per station, extrapolating from VLA 74 MHz experience

Na = 256 selected

LWA Memo 94

“Conservative”

“Reasonable”

Galactic center at maximum elevation; @ 74 MHz

Station Antenna Array Geometry• Every element is digitized to allow

unconstrained pointing of beams (among other things)

• Cost ∝ Na, so prefer to minimize Na

• Using 256 stands results in spacings 3 x Nyquist at 80 MHz

• Therefore, array has to have irregular spacings to mitigate against aliasing

• Possible elliptical (extended N-S) geometry to reduce variation in beam shape for low-elevation transit beam pointing

LWA Memos 73, 139

110 m (N-S) x 92 m (E-W)4 m min. stand separation

Active Antenna“Big Blade”

“Tied Fork”

Front end noise temp required to achieve indicated level of G.N.D.

Current front end noise temp (~250K)

Goal front end noise temp (120K)

Confirmed (approximately)In field measurements

Mutual Coupling – Collecting AreaCircular 100 m dia station,Irregular geometry, Min. 4 m between standsSimple dipoles, 38 MHz

SingleDipole,Simple model

SingleDipole,Rx-modeNEC2

SingleDipole,Tx-modeNEC2

Array,Rx-modeNEC2(stand average)

ApertureEfficiency

LWA Memo 73

Effect of mutual coupling

[m2] [m2]

Mutual Coupling – Beam Shape

LWA Memo 67

Collecting Area, Pointing Zenith Phase, Pointing Zenith

Collecting Area, Pointing 45o Phase, Pointing 45o

• Shown here: Magnitude and phase of current induced at each feedpoint(moment method)

• Only a “small” effect on beam shape and sidelobe level for any given pointing

• Bigger concern is “rumbling” of beam as a function of direction of arrival –Rumbling doesn’t stop even if you fix the beam!

• Actual impact on imaging not yet clear…

2x2MatrxMult.FIR

Station Electronics Architecture

196 MSPS 12 bits

FIRT&ZFIFO

3 more beams

T & Z

FIFO Sum

to

Form

Bea

m 1

Storage(continuous)

(TBN)

Storage (one-shot)

(TBW)57 ms

100 kHz from passband

ActiveBaluns

LongCoax

Gain &Filter

A/DARX

Stand 1

ABAB

T & Z

0.4-8.0 MHz from passband;4096 channels

T & Z

• Available to correlator:• 4 completely independent beams• 2 calibrated polarizations per beam• 2 tunings per beam, 0.4-8.0 MHz each, 4096 channels each

• Other products:• Full bandwidth beams• All dipoles, full bandwidth for 57 ms (“TBW”) • All dipoles, 100 kHz continuously (“TBN”)

84 samplesCoarse delay

I/Q &Dec by 2

Delay (for BF), Dispersion, Polarization

Sum

to

Form

Bea

m 2

Sum

to

Form

Bea

m 3

Sum

to

Form

Bea

m 4

98MSPS

“T&Z” = Tune within passband, filter, and reduce sample rate (“tune & zoom”)

78 MHz BW, 98 MSPS

Analog Receiver (ARX)

• Gain & selectivity only (Direct sampling architecture)

• No LOs

• 4 channels (2 stands) per board (128/station)

• Gain Control + Reconfigurable Bandpass:

(1) 10-88 MHz(2) 41 MHz highpass

“shelf” filter (x 2)(equalizes HF)

(3) 28-54 MHz(safe(r) mode)

LWA Memo 121

Direct Sampling A/D

• Confirmed performance of direct sampling is consistent with LWA specs

• 200 MSPS, 12 bit A/D prototype board (Analog Devices AD9230-250)

• Evaluated also in lab; found OK

• Quick & dirty front end using ETA active antenna + ETA ARX modified for 20-80 MHz; Site near Blacksburg, VA

ATSCCarrier NTSC

Carrier

FMBroad-

castHF

LWA Memos 127, 130

Frequency Plan

LWA Memo 101

Fs/4 Shift LeftMultirate LPF + Decimate by 2

88-108 MHz aliases

onto itself

0-10 MHz aliases

onto itself

Challenge here is to achieve:

• Best possible rejection of strong out of band signals, consistent with:

• Bandwidth (78 MHz) and

• Low complexity

A/D output(196 MSPS real)

Beamformer(98 MSPS complex)

Polarization & Dispersion Calibration• Beams should be not only “full bandwidth” (78 MHz) and

fully independent, but also well calibrated. Xpol!• “Perfect” calibration possible, but only for a single

frequency and beam pointing, or if FIR filters of infinite length are available

• Cable dispersion further complicates this:

• “Reasonable” performance seems possible with M=16 (98 MSPS) FIR filters

LWA Memo 138

Z=74°, φ=45°, M=16

Z=74°, φ=45°, M=16 (calibrated for Z=0)Z=74°, φ=45°, M=4

XPD 5-20 dB

XPD negl.XPD negl. ~ 10 dB

Effect of Fence• 120 m x 120 m security fence required

around array – effect?• Biggest impact is for H-plane pattern,

when collinear (as shown in these moment method simulations)

• < 1 dB gain variation, but oscillates• Effect depends on ground type

38 MHz 80 MHz

Acknowledgements

JPL

Site infrastructure, cable system, electronic shelter (J. Copeland, A. Kerkhoff, D. Munton, J. York)

Program management, systems engineering, analog receivers, civil engineering (J. Craig, W. Gerstle, Y. Pihlstrom, L. Rickard, G. Taylor)

Antennas, front end, array geometry (T. Clarke, A. Cohen, B. Hicks, N. Paravastu, P. Ray)

Digital electronics (L. D’Addario, R. Navarro)

Array, signal processing, calibration, monitoring & control, architecture (S. Ellingson, M. Harun, K. Lee)

+ Many others at these institutions also involved in the LWA project

+ Many others at other institutions helping out

Office of Naval Research

http://lwa.unm.edu