Alan E E Rogers coI MIT Haystack Raul A Monsalve Postdoc Arizona State University Judd D Bowman PI Arizona State University H1 low1 10x10 ground plane H2 low1 30x30 ground plane ID: 813126
Download The PPT/PDF document "aeer 31 July 2018 The design and cali..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
aeer 31 July 2018
The design and calibration of the 50 – 200 MHz receiver and spectrometer for 21-cm observations with a single antenna
Alan E. E. Rogers - co-I MIT Haystack
Raul A.
Monsalve
- Postdoc Arizona State University
Judd D. Bowman - PI Arizona State University
Slide2H1 – low-1 10x10 ground plane
H2 – low-1 30x30 ground plane
H3 – low-1 30x30 ground plane and recalibrated receiver
H4 – low-2 NS
H5 - low-2 EW
H6 – low-2 EW no balun shieldP8 - physical foreground (51 – 99 MHz) – others 5T poly (60-99) MHz
from: “An absorption profile centered at 78 megahertz in the sky-averaged spectrum” Nature 555, 67-70 (01 March 2018).
Slide3EDGES - 1
Antenna “Fourpoint” dipole
(Suh at al. 2003)
Antenna beamwidth ~ 80
º
Band (2:1) ~ 90 to 200 MHz
ADC 14-bit at 400 Ms/s
FFT samples in PC 32,768
Rogers, A.E.E., Bowman, J.D. 2008, "Spectral Index of the Diffuse Radio Background Measured from 100 to 200 MHz."
AJ
, 136, 2, pp. 641-648
Measured sky spectral index 2.5 +/- 0.1 with partial calibration
ed
g
e of wire mesh ground plane
Slide4EDGES – “2”
Broadband compact dipoles refl. < -15 dB 50 – 100 and 100 – 200 MHz with separate antennas
3-position switch to take out bandpass
Lab calibration of LNA
noise
waves and internal noise diode Noise diode calibration with hot and ambient loadLab performance verification using “antenna simulator”Temperature control of the
electronicsAutomated measurement of antenna S11Fully calibrated antenna and spectrometer
Slide5Slide6EDGES-2 installation at the MRO
Slide7Google view of EDGES-2 at the MRO in Western Australia
-26 42 53.5 116 36 17.9
Slide8Slide9Technical I
nnovations Uses VNA plus asymetric
2-port network to make measurements of
1] Short, Open, Load
2] SOL on asymmetric passive 2-port
3] SOL on asymmetric passive 2-port reversed4] DC resistance of Load 9-complex measurements + 1 real measurement to solve for 3 complex unknowns of SOL 3 complex unknowns of 2-port S11,S22,S12=S21
3 complex unknowns of VNA calibration“One-Port Direct/Reverse method for Characterizing VNA Calibration Standards” Monsalve et al. (2016) IEEE Transactions on Microwave Theory and Techniques 64(8): 2631-2639
Improved VNA calibration
Automated S11 measurement of antenna
Slide10LNA
ANTENNA
Γ
a
Γ
l
ref. plane
Antenna to Low Noise Amplifier mismatch
Slide11LNA noise waves reflected back from antenna
(T
sky
)
1/2
(1-
Γ
a
|
2
)
1/2
|F|
correlated noise
uncorrelated noise
T
0
2
nd
stage noise
antenna
LNA
Slide123 – position input switching – antenna, load, cal to take out “bandpass” and set temperature scale
Slide13Correction for losses
T =
T
sky
L
+ Tamb(1-L)
L = (1 - |
a|
2
)-1|S21|2(1-||2)/|1-S22|2 where: a = reflection coefficient on antenna measured from reference plane at LNA input = antenna reflection = (a-S11)/(S12S21-S11S22+S22a) S11,S22,S12,S21 = antenna balun
scattering coefficients
Slide14Calibration and processing procedure
In the Lab:
Measure S11 of LNA, hot and ambient loads, and open/shorted cable
Measure 3-position switched spectra of hot and ambient loads and open/shorted cable
Use the data above to calibrate internal noise diode and measure LNA noise
wavesMeasure the antenna balun lossIn the field:
Measure the antenna S11 and 3-position switched spectrumUsing EM simulationsEstimate the antenna and ground plane lossObtain absolute calibration of sky spectrum processing
Using lab calibration data, antenna S11 and
losses
obtain sky
spectrum Use weighted least squares with up to 6 physics based “basis” functions to remove foreground, ionosphere and solve for hydrogen line signatureError estimates from covariance matrix
Slide15Lab testing
Use an “antenna simulator” to test the accuracyA hot load at the end of a 3m cable
Assumes:
A mismatched load at uniform temperature is precisely equivalent to a lossless antenna observing a uniform sky at the same
temperature
If the temperature is not uniform the S-parameters between different regions needs to be knownFor the “hot load” we have the hot region and the ambient region in the cable to the reference plane. We need to S-parameters of the cable and the ambient temperature
Slide16Heated 50 ohm load with temperature probe
Corrections required for high accuracy:
1] S11 measurement vs temperature – as load changes with temperature
2] Input line loss – plus assumption of temperature gradient
Calibration HOT load of known temperature
Slide17Open/shorted cable used to calibrate LNA noise waves
open
short
Slide18Antenna S11 LNA S11 LNA noise waves
Slide19Slide20Slide21Slide22Slide23Calibrated receiver installed under the antenna 2015
Slide24function
Purpose
0
f
-2.5
Scale
1
log(f) f
-2.5
Spectral index
2
(log(f))
2
f
-2.5
foreground gamma
3
f
-4.5
Ion absorption
4
f
-2.0
Ion emission
Basis functions needed to remove
foreground and ionosphere
Slide25Difference spectra show the changes in ionosphere and track the change in opacity with local time
Rogers et al. Radio Sci. 50, 130-137 2015
Slide26Average magnitude and electron temperature of perturbations in ionospheric opacity
Slide2727
Blade Beam Chromaticity Correction
Frequency dependence of antenna beam at MRO
lat
= -26.7
lon
= 116.5
Mozdzen
et al. 2016
Slide289.8x9.8m
infinite
e
xtended antenna NS
e
xtended antenna EW
Slide29Slide30Relative sparseness of RFI from FM radio
Effects of FM radio on absorption signature
RFI from FM radio signals reflected from meteors which burn up at about 100 km. Sites need to be more than 2000 km from FM radio to completely avoid this source of RFI. Effects of FM radio reflected from the moon
about
0.05
K max when moon is close to the zenith.
RADIO FREQUENCY INTERFERENCE (RFI)
Slide31TESTS and CHECKS performed with EDGES-2
Sensitivity to receiver calibration S11 errorPerformed receiver recalibration – test suggested by Irwin ShapiroChanged antenna orientationData from 2 separate lowband
antennas on different ground planes
Changed ground plane size
Removed
balun shield and other checks for possible resonancesMeasured absorption over full range of LSTChecked for effects of ionosphere absorption and emissionChecked for RFI effects on absorption including reflection from moonChecked sensitivity to foreground fittingHave 2 independently developed processing pipelines Haystack’s and ASU’sChecked effects of processing data with calibration at made at different temperatures
Checked effects of beam correction with FEKO & HFSS and effect of no beam correctionChecked effect of making no balun and other loss corrections
Slide32Tungsten filament source
Calibrated spectrum and residuals to 1-term fit
Slide33Spectrum from Antenna simulator
Top plot are residuals to 5-term fit
Signature search
Slide34EDGES-3 built-in calibration
Slide35SARAS 2 110 – 200 MHz
50 – 110 MHz starting next month
Slide36Summary
1] Midband system 60 – 180 MHz – preliminary results confirm absorption2]
W
orkshop at Haystack 23 and 24 August to facilitate collaboration and to understand the details of 21-cm different systems
3] Continuation of tests at the MRO
4] Longer term midband system at the MRO. EDGES-3 built-in calibration and further improvements in accuracy
EDGES memos are at www.haystack.mit.edu/ast/arrays/Edges loco.lab.asu.edu/memos
Slide37Questions?
thank you
EDGES memos are at www.haystack.mit.edu/ast/arrays/Edges loco.lab.asu.edu/memos
Comments?