STAR Light PDR

1

• STAR Light PDR 3 October 2001

SYSTEM REQUIREMENTS Roger De Roo 734-647-8779,
deroo_at_umich.edu
2

• Outline

• Outline
• Science requirements instrument concept
• STAR and DSDR technologies, instrument
  configuration
• Platform requirements (power/weight/balance)
• Flowdown requirements
• Noise Budget, sampling, interference rejection
• Calibration

3
Science Requirements
STAR-Light Design Goals Measurement Objectives

• Soil Moisture Monitoring (L-band radiometer w/ 4K
  accuracy)
• Land Surface Process Model Development (long term
  operation,
• plot scale )
• Polar Operations (airborne access only)

4
Platform Requirements
STAR-Light Design Goals Aircraft Sensor Concept
STAR-Light Control Module
STAR-Light Sensor Module
For weight stability, plane must be a
tail-dragger rather than equipt with tricycle
gear
5
Science Requirements
STAR-Light Design Goals Derivative Measurement
Objectives

• Soil Moisture Monitoring
• Radio astronomy band 1400 1427 MHz
• Noise Equivalent Brightness Uncertainty (NEDT) lt
  0.5 K
• Land Surface Process Model Development in Polar
  Regions
• Swath out to /- 35 deg from sensor normal
• Daily operations for 3 hours near dawn
• Synthetic beamwidth from 15 deg to 22 deg
• Ambient thermal environment 30C to 40C (243
  K to 313 K)

6
Platform Requirements
STAR-Light Design Goals Aerial Environment

• Max altitude about 3000m (higher requires
  oxygen)
• Min altitude about 300m (lower sacrifices
  safety)
• Surface to altitude temperature difference -30C
  typical
• Surface to altitude pressure change 1000mb to
  700mb typical

7
Sensor Concept Configuration
Mechanical arrangement on aircraft belly
Cold Plate
Receiver
Cross section
Receiver assembly is a field-replaceable unit
8
Sensor Concept
STAR-Light Aircraft Sensor Concept
Use Synthetic Thinned Array Radiometry
to -provide imaging capability -achieve multiple
angle of incidence electronically -keep the
sensor robust to partial failures
Use Direct Sampling Digital Radiometry to -move
complexity of STAR from analog to digital
domain -keep the sensor head compact -reduce
component count requiring thermal control
9
Sensor Concept STAR
STAR-Light Concept STAR Technology
Different antenna baselines sample different
spatial Fourier components of the scene
f
Baseline d
Vi j Vq Tb(f) F1(f) F2(f) exp(j 2 p sin f
d / l ) df

10
Sensor Concept DSDR
Direct Sampling Digital Receiver Technology

• Transfer
• Noise bandwidth definition
• I/Q detection (Hilbert transform)
• Complex correlation
• from analog to digital domain

11
Sensor Concept
STAR concept
Use a standard antenna array with missing
elements
7l/2
To simulate an array of larger dimensions, by
using each Element in turn as the phase center of
the array


14l/2
12
Sensor Concept
STAR-Light Antenna Configuration 1-D vs 2-D STAR
1-D requires long antenna elements to achieve
narrow beam -single angle of incidence (pushbroom
operation) -alias free spacing is 0.500
l -demonstrated (ESTAR)
2-D requires electrically small antenna elements
-multiple angles of incidence (snapshot
imaging) -many configurations 3-arm appears
optimal -alias free spacing is 0.577 l -proposed
(SMOS), but not yet demonstrated
13
Sensor Concept
STAR Issues
Huge sidelobes STAR requires an aperture taper
which increases synthesized beamwidth by a
factor of 2 (canceling the aperture
doubling) but the advantages of thinning
remain Optimal taper is Blackmann Camps etal
98 Increased noise Noise in STAR image
Real Aperture Area .
Noise in real aperture pixel
Actual Aperture in STAR but longer dwell
time for STAR to reduce noise equals time
required to scan the real array or real aperture
LeVine 90, Ruf 88
14
Flowdown Rqmt Antenna Spacing
15
Sensor Concept
STAR Image Generation Gain Correction
V(GT)F(GT)
TF-1(V(GT))/G
TF-1(V(GT))/F-1(V(G))
G is gain pattern of commercial patch
antenna Correction is not as pronounced for
Gcosnq
FOV35o
16
Flowdown Rqmt Antenna Elements
Pattern knowledge requirement
Errors induced by imperfect knowledge of
antenna gain patterns Image DC offset
30mK/K/deg2 12mK/K/2 Image rms error
/- 0.4 mK/K/deg /- 0.35 mK/K/
Constant brightness temperature scene inverted
by system with gain pattern uncertainty of 1dB
and 10o Goal is 0.5dB and 5o
17
Sensor Concept
STAR Image Generation Impulse Response
cos2 q
0o
30o
60o
89o
patch antenna
Array spacing driven by horizon alias generation
d0.68l 14.4 cm
18
Sensor Concept Thermal
Heat Dissipation and Thermal Control
150 mW steady 2.9 W intermittent
54 W typical 70 W maximum
27 W steady
19
Sensor Concept Geometry
Mechanical arrangement
Preferred Orientation for Cold Plate easy
side access for cooling fluid conduits
Required Orientation for Linear Pol
Antennas Parallel or anti-parallel
A D
Analog side needs high precision control,
moderate heat removal Digital side needs low
precision control, large heat removal
A D
Problem orientation of cold plate to antenna
20
Sensor Concept Receiver Module
Solutions to Cold Plate / Antenna Orientation
Conflict
Solution 0 disconnect Antenna from Receiver to
allow Receiver orientation to Cold Plate Very
difficult field cal
Solution 2 multiple fixed Antenna Receiver
modules Expensive
Cold Plate
D
A
Antenna
Solution 1 flexible connection between
Antenna Receiver to allow Receiver
orientation to Cold Plate Questionable quality
Solution 3 Circular Polarized Antennas Tricky
Cold Plate
D
A
Antenna
21
Flowdown Rqmt Antenna Elements
Single Feed Circular Polarization Patch
Notches create two modes w/ different
resonances Proper feed allows these two modes to
be fed w/ equal amplitude and 90o phase 1.4
circular polarization bandwidth at AR1dB while
11 VSWR bandwidth (VSWR2) Q8.6
Eff90 Cupped design to reduce mutual
coupling Parameters shown from design paper
must be modeled w/ EM analysis SW
14cm
7.75cm
e2.2, t4.6mm
22
Platform Capabilities
STAR-Light Design Goals Aircraft Capabilities
Parameter Aviat Husky A-1B Piper Super Cub PA-18 150
Power available 420 W ?
Carrying capacity (pilot instrument) 810 lbs 767 lbs
Min Safe Speed (Stall Speed X2) 110 mph 50 m/s 74 kts 38 m/s (40 deg flaps)
Availability New or used Used only
Aircraft acquisition costs and aircraft
integration are not part of STAR-Light project
23
Platform Requirements Weight
Aviat Husky Weight Limitations
Design Empty Weight 1190 lbs
Equipment Changes 80 lbs
Std Zero Fuel Empty Weight 1270 lbs
Oil and Unusable Fuel 27 lbs
Equipped Weight Empty 1297 lbs
Fuel (50 Gal max) 300 lbs
Useful Load (excl. Fuel) 397 lbs
Gross Loaded Weight 1994 lbs
Max. Gross Weight 2000 lbs (normal category)
24
Platform Requirements Weight
Useful Load Weight Breakdown
Pilot 200 lbs
Survival Package 20 lbs
Sensor Module 83 lbs
Control Module 70 lbs
Cabling 20 lbs
Pilot Interface 4 lbs
Useful Load (excl. Fuel) 397 lbs
Present estimate 10 lbs Weight Margin 4
lbs (from previous viewgraph)
25
Platform Requirements Balance
Weight and Balance
w/ full fuel tanks
w/ empty fuel tanks
26
Platform Requirements Power
Constant Power Requirements
Available Power (70 A _at_ 12 V) 840 W
Essential Flight Loads (33.1A) 400 W
Power Available for STAR-Light 440 W
S-L Sensor Module (RF Amps) 27 W
S-L Sensor Module (Digital) 76 W
STAR-Light Control Module 15 W
STAR-Light Thermal Control 85 W
STAR-Light Direct Power Rqmt 203 W
STAR-Light Power Supply Losses 31 W
STAR-Light Total 232 W
Power Margin 208 W
27
Platform Requirements
Intermittent Power Requirements
Aircraft systems Taxi/Landing Lights (14.2A _at_
12V) 170.4 W Radio Transmissions (6A _at_ 12 V)
72 W STAR-Light Components RF switches 2.9W
at 0.3 duty cycle 10mW Cooling System on climb
to altitude
28
Sensor Concept Thermal
Increase in altitude to 3000 m
Cooling Control Setpoints
Ground Ambient
Airborne Ambient
29
Flowdown Requirements
Integration Time for STAR-Light 2x Husky no-flap
stall speed
30
Flowdown Requirements
Integration Time for STAR-Light Slower speed
31
Flowdown Rqmt Noise Figure
For any taper Camps, 98 DT
dWconstant NEDT(uniform) dW(Blackman)/dW(unifor
m) NEDT(Blackman)
(15deg)2 / (10deg)2 0.5 K
1.12K For uniform taper LeVine,
90, NEDT Tsys Asyn Trec
300K 73 sqrt( B t )
n Ael sqrt( 20e6 . 1.5 ) 10 For NEDT
lt 1 K, Tsyslt 750 K or Treclt450 K (NF lt 4.1
dB)
32
Flowdown Rqmt Noise Figure
Antenna
Low Noise Amp Miteq
Cal injection Teledyne switch
Interference Reject Filter IMC
IL0.60dB
NF0.80dB
IL0.45dB
IL0.25dB
Interconnect losses lt 0.5dB
Downstream components add 0.1dB
System Noise Figure 2.7 dB (Trec250K)
33
Flowdown Rqmt Gain
Signal amplitude at ADC must be gt 4 levels (2
bits) for bias levels to not matter Fischman,
01 At Tsys250K, k Tsys B -101.6
dBm LSB15.63mV for typical ADC (SPT7610) gt
Padc-26.6 dBm Overall gain must be gt 75 dB For
amplifier w/ G26dB, 3 amplification stages
minimum (to allow for losses in receiver, use 4
stages)
34
Flowdown Rqmt Gain Fluctuations
Temperature fluctuations gt Gain fluctuations gt
system noise
dG/dT -0.02 dB/K per gain stage
dG/dT -0.08 dB/K for system requires 2mK rms
to keep gain fluctuation component lt fundamental
NEDT
Thermopad temperature compensating attenuator
Thermopads come in loss coefficient increments of
0.01 dB/K Goal Use Thermopads to get system to
/- 0.015 dB/K thermal control to 11mK rms
35
Flowdown Rqmt ADC levels
Need a minimum of 4 levels for darkest target
Fischman 01 Is a 3-bit Analog to Digital
Converter (ADC) enough? Tsys(max) / Tsys(min)
lt (8 levels)2 / (4 levels)2 4 where
TsysTbTrec If we wish to look at the sky,
Tb(min)0K On Earth, Tb(max)300K Then,
Trecgt100K or we need more bits Therefore, 3 bit
ADC is enough
36
Flowdown Rqmt Pre-Sampling Filter
IMC Ceramic Filter
The Fringe Wash Function measures the
differences between bandpass filters, and
reduction in measurable visibility due to
receiver differences

• A pre-sampling filter is used to
• define sampled bandwidth
• interference rejection
• out-of-band noise rejection

FWF0.996
37
Flowdown Rqmt Pre-Sampling Filter
IMC Ceramic Filter
The half-bit level for a 3-bit ADC is
24dB Variations over temperature define the
bandwidth extent for sampling
38
Flowdown Rqmt ADC sampling
Sampling Rate considerations Feixure etal
98 For a noise bandwidth (approx -3dB BW) of
1403 1423 MHz, the sampled bandwidth (approx
24dB BW) is 1390 1435 MHz For I/Q
demodulation, 2fH/m lt fs lt 2fL/(m-1), where
m1,2,mmax and mmaxfloorfH/(fH-fL) 92.58 MHz
lt fs lt 92.66 MHz or 95.67 MHz lt fs lt 95.86 MHz
or 98.97 MHz lt fs lt 99.29 MHz or 102.5 MHz lt fs
lt 102.96 MHz fs102.8 MHz
39
Flowdown Rqmt ADC sampling
Sampling skew If tskew lt 6.7 ns, reduction in
visibility envelope is less than 3
ENVsinc(Btskew) ViENVcos(2pf0tskew) VqENVsin(
2pf0tskew)
Fischman was unable to verify this form for the
envelope Verification is a primary objective of
the two channel system
40
Flowdown Rqmt ADC sampling
Sampling jitter sd produces a Coherence Loss
(CL) in a visibility value Fischman 01 CL
10 log( 1 ( 2 p f0 sd)2 ) for sd 20 ps,
CL 0.14 dB, or, in other words, 20 ps jitter
reduces a visibility value by 3 over a zero
jitter visibility
41
Flowdown Rqmt Noise Budget
Noise source Noise (K rms) RCVRAntenna Noise (K rms) Image (Blackman) conditions
Fundamental .101 .326 Tsys600 K B20MHz t1.5s
Gain Fluctuations .080 .260 dG/dT -0.015 dB/C DTo 11 mC rms
Passive Parts .004 .013 IL 1.75 dB DTo 11 mC rms
Antenna .022 .072 Efficiency 90 (IL0.45dB) DTo 200 mC rms
Radome .022 .072 Efficiency 95 (IL0.22dB) DTo 400 mC rms
ADC .048 .156 dSpan/dT 50 ppm/C DTo 200 mC rms
Total (RSS) .141 .460
x (73/10)x(0.45) x 3.25
42
Flowdown Rqmt Interference

• Keep cultural sources of RFI out of receiver
  chain to the extent that
• Amplifiers do not saturate
• Intermodulation products do not get generated in
  Radio Astronomy band
• RFI does not alias into ADC sampling window
• Some worst-case sources of interference
• 1. Air Traffic Control Radar Beacon System
  (ATCRBS) Transponder
• Responds to 1030 MHz radar pings, reporting
  aircraft altitude to ATC
• Transmits from the STAR-Light aircraft at 1090
  MHz w/ peak power
• between 70 and 500 W (48 dBm to
  57 dBm)
• Air Route Surveillance Radar (ARSR)
• Transmits from the ground from 1250 to 1350
  MHz w/ peak power up
• to 5 MW (97 dBm)
• Some similar military systems have high
  resolution modes which use
• up to 1375 MHz, 1380 MHz, or 1400 MHz

43
Flowdown Rqmt Interference
Keeping the ATCRBS Transponder from saturating
STAR-Light amplifiers
53 dBm
Moving the transponder antenna to the top of the
tail gives a distance of 4 m to
STAR-Light Coupling lt 42 dB at 4 m Typical
model (Garmin GTX 320A) transmits 200W (53dBm)
at 1090 MHz
4 m

Cumulative Rejection Needed
P1dB8dBm
P1dB14dBm
44
Flowdown Rqmt Interference
Keeping the ARSR 1250 - 1350 MHz intermodulation
products out of the 1400 1427 MHz Radio
Astronomy band
Rqmt Keep PIM lt -140dBm At 50 km, ARSR-3
power at antenna terminals is Pr 5dBm
(assuming gain is down by 8dB, and
polarization match 50) Miteq IIP3
-9dBm Requires filtering of F 37dB at 1350 MHz
P2Pr-F
P1Pr-2F
PIM2P2P1-2IIP3
f (MHz)
1345/-5 f2
1277/-5 f1
1413/-10 2f2-f1
We will get hit w/ intermodulation interference
from ARSRs. ARSRs sweep at 5 rpm, and our
recovery time is on the order of
microseconds. (Subsequent stages also need
protection from amplified f1 and f2 M/A Com amp
has IIP3-2dBm)
45
Flowdown Rqmt Interference
Quadrant Engineering, Inc. Experience Scanning
Low Frequency Microwave Radiometer (SLFMR)
Goodberlet 00

• SLFMR system
• f 1413 MHz B 100 MHz
• Phased Array antenna, not STAR
• Designed NEDT0.3K verified in lab
• Observed NEDT5K over water (Tb100K)
• in field tests 20 miles from interference source
• (Norfolk, VA)

STAR-Light Implication With just 15dB of
Interference Rejection Filtering, we can drive
that interference NEDT down to 0.15K
46
Calibration Hardware List
STAR-Light Calibration Design Pre-flight /
In-flight Calibration
To calibrate each antenna-receiver channel, we
need a hot load a cold load to estimate
the receiver temperature and overall receiver
gain To calibrate each pair of channels, we
need correlated noise uncorrelated noise
to estimate the receiver correlation in magnitude
and phase
47
Calibration Warm Cold
Slope a Gain
V(d0)
x2
Tb300K
Tb 77K
V(d0)
Tb
-Trec
0K
77K
300K
To calibrate each antenna-receiver channel, we
need a warm load a cold load to estimate
the receiver temperature and overall receiver gain
48
Calibration Warm Hot
Delay t
Vi
DSP
Tb300K
Vq
Dt
Delay t Dt
To calibrate each pair of channels, we need
correlated noise uncorrelated noise (to
determine Vi, Vq offsets) to estimate the
receiver correlation in magnitude and phase
49
Calibration Receiver Two-Point Cal
STAR-Light Calibration Design Two-Point
Calibration of a single channel
Trec300K B20MHz t1.5 s
50
Calibration Cold Noise Source
STAR-Light Calibration Design Quality of Cold
Load
Phase Uncertainties Reflection /- 6
deg Transmission /- 12 deg
L0.3 dB VSWR1.1
L0.4dB VSWR1.1
RCVR
VSWR1.1
50W at 77K VSWR1.05
51
Calibration Correlated Noise Source
STAR-Light Calibration Design Correlated Noise
Distribution Network
3-diode design allows any one diode failure
while maintaining calibration
r0.743
r0.754
52
Two Channel Prototype

• Two Channel prototype tests
• NEDT verification (Dec 01)
• End-to-end fringe wash function measurement (Jan
  02)
• Receiver calibration validation (Feb 02)
• Antenna radiation efficiency measurement (Spring
  02)
• Tasks to be done prior to CDR
• Antenna specification and design (Oct 01)
• Antenna manufacture and integration (Nov 01 Mar
  02)
• STAR model evolution (continuous)
• Cold load final design (Oct-Dec 01)
• Post CDR
• Antenna characterization
• System validation