evla front-end cdr · lna lna lcp rcp ghz ln2 liquid nitrogen ... • not always as accurate as...
TRANSCRIPT
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
1
EVLA Front-End CDR
LaboratoryReceiverTesting
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
2
EVLA Ka-Band ReceiverOverview
1) Rx Characterization - SOIDA Rack- EVERETT Rack
2) PNA Measurements - OMT Tweaking- Phase Matching
3) SNA Measurements - General Purpose Measurements- Sensitivity (Y-Factor)- Axial Ratio- Total Power Stability
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
3
The SOIDA Rack
• The SOIDA Rack consists of a data acquisition PC, two commercial frequency synthesizers, a power meter plus a custom designed Baseband Converter (BBC) unit.
• The system is designed to:• Measure the receiver sensitivity
(i.e., TRx) vs. frequency• Measure the noise diode calibration
(i.e., TCal) vs. frequency• We also have a 2nd test rack which is
affectionately known as the SOS Rack, for “Son of SOIDA”
SOIDA Testing Q-Band Rx
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
4
Cold Load
77°K
Hot Load
295°K
“SOIDA” Hot/Cold LoadEVLA Q-Band - Swept LO1
CharacterizationCV-3
MicrowaveAbsorber
Dewar
RF/IF Box
SOIDA
Key:
ChannelNot Shown
30 dB
30 dB
40-50
NoiseDiode
Pol
LNALNA
RCPLCP
GHz
LN2Liquid
Nitrogen
NoiseDiodeNoiseDiode
GigatronicsSynthesizer1-26 GHz
20 dB 10 dB
HP 8671ASynthesizer2-18 GHz
HP436APowerMeter
DataAcqPC
GPIB Interface
DC-20MHz
LO1 =16.133 to 19.467
GHz
IF = 8.4GHz
LO2 = 8.4 GHz
Baseband Converter
35 dB
39-51 GHz
Warm RF/IF Sub-system
x3
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
5
What SOIDA Sees
)1()(
/
/
−⋅−
=CH
ColdCHHotRx
YTYTT
Hot LoadPower(dBm)
0 10 20 30 40 50
Freq(MHz)
Cold Load
Cold Load + Noise Diode
YH/C
YND/C
)1()( / −⋅+= CNDColdRxCal YTTT
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
6
A Typical SOIDA PlotEVLA Q-Band Receiver
The SOIDA plot is designed to
display thefollowing
information:PHotPColdGainTRxTCal
(SCal)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
7
The Next Generation Test
Rack• EVLA requires 240 upgraded or new receivers.• Both SOIDA & SOS use obsolete test equipment
- synthesizers & power meters.• The PC’s are ancient 33 MHz machines.• Only able to do one channel at a time.• Takes ~1½ hour to do a single W-Band channel.• Developing a new dual-channel replacement rack
which uses a modern laptop, E4119B power meter, 83630B/24L synthesizers and a new broadband dual-channel BBC.
• Cost $90K each• EVERETT++ = “Expanded Vla Enhanced
Receiver Evaluation & Test Terminal”.• Plan on 2 EVERETT Racks as well as 1 (or 2)
SOIDA Racks (depending on longevity).
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
8
E8364BPrecision Network Analyzer
(PNA)• Agilent’s newest generation of Vector Network Analyzer
• DC – 50 GHz• Electronic Calibration Unit (to 50 GHz)• Mechanical Cal kits
• Coaxial - 3.5 mm• Waveguide - WR-42, WR-28 & WR-22
• Frequency Translation Option• Higher Dynamic Range and faster than the 8510 VNA• Windows Platform• Bought to facilitate Ka & Q-Band MMIC-based modules• Cost -$150K
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
9
Precision Network AnalyzerEVLA Related Measurements
• Return Loss measurements of OMT during tweaking
• Phase Balancing of Cables between OMT & Hybrid
• Measurement of MMIC device and multi-function module S-Parameters using Wafer Probe Station
• Numerous General Purpose Measurements of RF & microwave components and sub-systems
PNA measuring Return Loss on 2nd L-Band OMT
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
10
Scalar Network Analyzer
• Coaxial SNA = 8757C + 83630B Synthesizer10 MHz to 26.5 GHz
• W/G SNA = 8757E + 83623A Synthesizer2 to 20 GHz
Plus Source Modules atKa, Q & W-Band
Acquired enough surplus 8757C’s to build one intothe new EVERETT Racks as well.
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
11
SNA GP MeasurementsFrequency Response
Ka-Band Down-converter
CoaxialSNA
WaveguideSNA
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
12
SNA GP MeasurementsKaDCM Frequency Response
Close up
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
13
Y-Factor Measurement with SNA
• SNA’s can be configured to measurement Y-Factor sensitivity essentially in real-time – 1 minute vs. 15 to 90 minutes on SOIDA– Dual channels simultaneously
• For Lower Frequency Rx’s with no internal frequency conversion or for the EVLA Block Converter Mode receivers, use with the BBC on the SOIDA Rack.
• For Higher Frequency Rx’s with internal mixers, use a narrow filter and post-amp n the IF output.– 8400, 11000, 14200 MHz filters with 100 MHz bandwidth
• Not always as accurate as SOIDA but allows quick comparison measurements (TRx vs. LO Power, bias settings, cpt changes, etc.)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
14
SOIDA & SNA Measurement Comparison of Receiver Temperature Block Converter Mode on Receiver Q#17
LCP - SOIDA 100 MHz Step Size, 40 MHz Filter Bandwidth & 32 Sample AverageLCP - SNA 25 MHz Step Size. 40 MHz Filter Bandwidth & 8 Sweep AverageRCP - SOIDA 100 MHz Step Size, 40 MHz Filter Bandwidth & 32 Sample AverageLCP - SNA 25 MHz Step Size, 40 MHz Filter Bandwidth & 8 Sweep Average
40 41 42 43 44 45 46 47 48 49 50Frequency (MHz)
20
30
40
50
Rec
eive
r T
empe
ratu
re (K
)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
15
Lab Measurements of Circular Polarization
• Attempt to measure the response of the receiver to a linearly polarized signal rotated through 360E by injecting it into the receiver and detecting the LCP & RCP output power.
• If the polarizer is perfect, we should see no variation.• If the polarizer has a 1 dB axial ratio, we should see the output power
change by 1 dB.• Best done on an antenna test range
• Would be a pain at L-Band with its 16 ft feed• So lab measurements done using a circular waveguide test fixture
that gives us both a linear polarized signal and a rotary joint.• Initially measurements were taken with a power meter at each
Position Angle (typically every 10 to 20E) for a fixed frequency.• Polarization plotted up in Polar Plots (i.e., “Peanuts”)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
16
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
1200 MHzRCP = 18 dB ?LCP = 16 dB ?
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
1300 MHzRCP = 18 dB ?LCP = 20 dB ?
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2Axial Ratio (dB)
L-Band S/N 02Axial Ratio (dB) and Position Angle of Major Axis versus Frequency
Cold Receiver with Bradley Prototype Balanced AmplifiersDrain Voltages & Currents Tweaked to Eliminate LNA Oscillations
(28 Jan 2004)
1400 MHzRCP = 15 dB ?LCP = 17 dB ?
1500 MHzRCP = 17 dB ?LCP = 22 dB ?
1600 MHzRCP = 21 dB ?LCP = 18 dB ?
1700 MHzRCP = 12 dB ?LCP = 13 dB ?
1800 MHzRCP = 10 dB ?LCP = 11 dB ?
FrequencyRCP LNA IRLLCP LNA IRL
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
17
SNA EllipticityMeasurements
• Power Meter measurements truly painful and we were under sampling the frequency response
• Developed new test procedure using the SNA– Full frequency sweep taken at each Position Angle– SNA Reference channel eliminates power variations from
RF cable being twisted.– SNA data “grabbed” by laptop & written to Excel file– Super spreadsheet imports the -100,000 data points and
then calculates the axial ratio for each frequency point • 2.5 MHz @ L-Band or 20 MHz at K-Band
– Scheme also allows smoothing of data before AR calculated so effect of freq ripple in the test setup can be reduced.
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
18
Scalar Network AnalyzerAxial Ratio vs. Freq Measurements
K-Band Ellipticity Test Setup
35 dB
35 dB
18-26 GHz
NoiseDiode
OMT
LNA
LNA
Mod
eC
onve
rter
90 DegreePhase-Shifter
Cir
cula
rT
oSq
uare
Vac
uum
Win
dow
Squa
reto
Cir
cula
r
OMT
Ter
min
atio
n
SynthesizerHP 83630B
Scalar NetworkAnalyzer
HP 8757D
DetectorHP 85025E
Det
ecto
rH
P 85
025EDewar
Det
ecto
rH
P 85
025E
Chan A
Chan B
LCP
RCP
Rotary Joint
Ref
GPIBRF Out-15 dBm
20 dB
Power Divider
HP 11667B
K-BandEllipticity
TestFixture
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
19
K-Band S/N 28Under Test
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
20
K-Band Ellipticity Test Fixtures (KETF’s)
OMTKETFusesspare
WollackOMT
IsolatorKETFuses
FaradayIsolator
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
21
High Resolution SNA Axial Ratio
Scan
18000 20000 22000 24000 26000Frequency (MHz)
0
0.5
1
1.5A
xial
Rat
io (d
B)
0
0.5
1
1.5
Axi
al R
atio
(dB
)
0
0.5
1
1.5
Axi
al R
atio
(dB
)
0
0.5
1
1.5
Axi
al R
atio
(dB
) Iso-KETF
Iso-KETF with Spacer
OMT-KETF
OMT-KETF with Spacer
K#22 Axial Ratio for Various K-Band Ellipticity Test Fixtures
LCP & RCP(15 April 2005)
LCPRCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
22
Typical K-Band Axial Ratio Measurement
Measured Axial Ratio on K#22 LCP & RCP Channels using the Iso-KETFversus
Calculated Axial Ratio for Phase-Shifter #23 based on Differential Phase Shift Test Data(Assumes 0 dB Amplitude Inbalance)
15 April 20051 dB Axial Ratio SpecLCP (Unsmoothed)LCP SmoothedRCP (Unsmoothed)RCP (Smoothed)Phase-Shifter #23
18000 19000 20000 21000 22000 23000 24000 25000 26000
Frequency (GHz)
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
3
Axi
al R
atio
(dB
)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
23
AR Test on L#21using VLA-Style OMT
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
24
Simulated LNA Return Loss Measurements
Hybrid Coupler Polarizer in anEVLA Interim L-Band Rx (L#32)
Simulated LNA’s using Terminations and/or Unterminated Pads (L#21)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
25
Simulated LNA Return Loss
Measurements
1300 1400 1500 1600 1700 1800Frequency (MHz)
0123
Term
inat
ions
0123
IRL=
20 d
B
0123
IRL=
18 d
B
0123
IRL=
16 d
B
0123
IRL=
14 d
B
0123
IRL=
12 d
B
0123
IRL=
10 d
B
0123
IRL=
8 dB
0123
IRL=
6 dB
L#21 Circular Polarizer - Quadridge OMT & Hybrid Coupler Axial Ratio vs. Simulated LNA Input Return Loss
Axial Ratio Key = LCP , RCP & 1 dB Spec (17 Feb 2006)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
26
SNA Measurement of Total Power Stability
Q-Band S/N 12
• Measure noise power (in 100 MHz BW) across Rx frequency response
• Save in Memory• Look at “Meas-Mem”• Microphonics will
appear as deviations from a flat line
• Whack on dewar
LCP
RCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
27
Total Power StabilityLong & Short Term at 48 GHz
Q-Band S/N 12
10 sec sweep 200 ms sweep
Microphonics - 17 ms period
LCP
RCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
28
Total Power StabilityAdd Restraining Brackets across
Output WR-22 FlexguideQ-Band S/N 12
RestrainingBracket
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
29
Total Power StabilityShort Term after Fix
Q-Band S/N 12
LCP
RCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
30
Summary
• We believe we have the appropriate test equipment and have developed the requisite test techniques to characterize all of the EVLA receivers that will be built over the next 6 years.
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
31
Questions ?
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
32
Backup Slides
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
33
Laboratory ReceiverCharacterization
• The first automated amplifier & receiver test measurement system built at the NRAO was designed by Sandy Weinreb at CDL in the late 70’s & early ’80’s.
• This system used the foremost microcomputer of the day - the Apple II.
• The software for the system was written in BASIC by Sandy’s son, Glenn.
• They gave it the name ADIOS, which stands for “Analog Digital Input Output System”.
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
34
ADIOS, Amigos
• In the 1990’s, as the AOC took on an increasing role in VLA/VLBA receiver development & maintenance, Paul Lilie(former Head of the FE Group) designed a new PC-based test system.
• Ever one for a pun, he endowed it with a name that gave tribute to the original ADIOS system…
• The word SOIDA has become ubiquitous.• Adjective : “Did you do a SOIDA test on the receiver?”• Noun : “I did a SOIDA on it yesterday?”• Verb : “When will you SOIDA this receiver?”• Adverb : “I was SOIDA testing the receiver when it died?”
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
35
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
36
Hot/Cold LoadsHigh Frequency Rx’s
Ku, K, Ka & Q-Band• Feeds are considered to
be part of the receiver.• Hot/Cold Loads use
CV-3 Absorber.• Boxes made from
Styrofoam or Zotefoam.• K-Band has a metal plate
to make Cold Load look “colder” (reduce leakage from 300K ambient).
• K-Band Cold Load elevated with fan to eliminate moisture build up on bottom surface.
K-Band Cold Load
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
37
Hot/Cold LoadsLower Frequency Rx’s
L, S, C & X-Band• The feeds for these
receivers are built into the Vertex Cabin.
• So they need circular waveguide Hot/Cold Loads.
• Built from OMT’s with coaxial terminations at ambient or immersed in LN2
• Old VLBA Hot/Cold Loads not broadband enoughSOS Testing a VLBA Ku-Band Rx L-Band Proto with 40” Section
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
38
Hot/Cold LoadsLower Frequency Rx’s
Examples
L-Band Hot/Cold Load Lilie Noise Standard C-Band Hot/Cold Load
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
39
Typical L-Band Axial Ratio Measurement
L#32 : AR & PA vs. Frequency (Unsmoothed) - With Reference Subtraction23 Nov 2004
0
0.5
1
1.5
2
2.5
3
1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000
Frequency (MHz)
Axi
al R
atio
(dB
)
-180
-160
-140
-120
-100
-80
-60
-40
-20
0
20
40
60
80
100
120
140
160
180
Posi
tion
Ang
le (N
orm
aliz
ed)
AR - LCPAR - RCPPAN - RCPPAN - RCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
40
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
1000 MHz
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
1100 MHz
045
90
135180
225
270
315
-1 0 1 2
045
90
135180
225
270
315
-1 0 1 2
1200 MHz
1300 MHz 1400 MHz 1500 MHz
1600 MHz0
45
90
135180
225
270
315
-1 0 1 2
1700 MHz 1800 MHz0
45
90
135180
225
270
315
-1 0 1 2
RCP
LCP
L-Band S/N 32 (23 Nov 2004) - With Ref SubtractionAxial Ratio (dB) and Position Angle versus Frequency
Cold Receiver with Balanced LN+HP Gain BlocksSNA Captured Data (Unsmoothed)
045
90
135180
225
270
315
-1 0 1 2
1900 MHz0
45
90
135180
225
270
315
-1 0 1 2
2000 MHz0
45
90
135180
225
270
315
Position Angle
-1 0 1 2Axial Ratio (dB)
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
41
L#21 Circular Polarizer - Quadridge OMT & Hybrid Coupler Axial Ratio vs. Simulated LNA Input Return Loss
(19 Feb 2006)IRL = TerminationsIRL = 20 dBIRL = 18 dBIRL = 16 dBIRL = 14 dBIRL = 12 dBIRL = 10 dBIRL = 8 dBIRL = 6 dBAxial Ratio 1 dB Spec
-25
-20
-15
-10
-5
0
nput
Retur
n Los
s (d
B)
1300 1400 1500 1600 1700 1800Frequency (MHz)
-40
-35
-30
I
1300 1400 1500 1600 1700 1800Frequency (MHz)
-40
-35
-30
B)
-25
-20
-15
-10
-5
0
Input Return L
oss (d
LCP RCP
1300 1400 1500 1600 1700 1800Frequency (MHz)
0
1
2
3
Axi
al R
atio
(dB
)
1300 1400 1500 1600 1700 1800Frequency (MHz)
0
1
2
3
Axial R
atio (dB)
LCP RCP
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
42
L#21 Circular Polarizer - Quadridge OMT & Hybrid Coupler Axial Ratio vs. Simulated LNA Input Return LossLCP & RCP IRL = Termination and/or IRL10 dB
Axial Ratio Key = LCP , RCP & 1 dB Spec (17 Feb 2006)
1300 1400 1500 1600 1700 1800Frequency (MHz)
0
1
2
3
LCP=
Term
& R
CP=
Term
0
1
2
3
LCP=
Term
& R
CP=
10dB
0
1
2
3
LCP=
10db
& R
CP=
Term
0
1
2
3
LCP=
10db
& R
CP=
10dB
R. Hayward EVLA Front-End CDR – Laboratory Receiver Testing 24 April 2006
43
General Purpose Test Equipment
• PNA DC-50 GHz (Agilent E8463B)• Coaxial SNA DC-26 GHz (HP8757C + HP83630B)• Waveguide SNA (HP8758E + HP83624A)
• Ka, Q & W-Band Source Modules
• Spectrum Analyzer (Agilent 8563EC)• DC-26, 33-50, 75-110 GHz
• Power Meter (Agilent E4419B)• DC to 110 GHz
• SiteMasters (Anritsu S332B & S820A)