Beyond S-parameters: Modern Network

Analyzer Architectures and Algorithms

Presented by:
Ernie Jackson

Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Objectives
• Examine architectures of modern vector network analyzers (VNAs)
• Provide insight into nonlinear characterization
of amplifiers, mixers, and converters using a vector network analyzer
• Understand associated calibrations for non-S-parameter tests

Page 2
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Agenda
• Overview of Component Testing Using a VNA
• Modern VNA Architectures
• Nonlinear Amplifier Tests
– Intermodulation distortion LO

Source 2
– Phase versus drive To receivers
Source 1
OUT 1 OUT 2
– Hot S22 OUT 1 OUT 2

– True-mode stimulus R1 R2
A B
– Single connection,
multiple measurements
Source 2 Source 2
Test port 1 Output 1 Output 2 Test port
• Mixer and Converter Tests 2

• Applications DUT

• Amplifier Measurements
• Summary

Page 3
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Single Connection Multiple Measurement System

Page 4
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 SCMM System to PNA-X

Page 5
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
The Need for Component Test
• Components are underlying building blocks of RF systems

• Magnitude and phase information crucial for simulation during

design stage
• Ensure devices meet specifications
during manufacturing

Page 6
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
S-parameters: Core of Linear Characterization
• S-parameters: complex (magnitude and phase) reflection and
transmission in forward and reverse directions Transmission

DUT
• Fully describe linear behavior of RF components Reflection

• Distortion caused by non-flat amplitude and deviation from linear

phase/constant group delay
• Necessary, but not sufficient for full system simulation

Frequency
Magnitude

Phase

Frequency

Page 7
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Non-Linear Distortion of Active Devices
• System impairments also result from nonlinear device behavior
• Important to include nonlinear characteristics in simulation
• Nonlinearities usually dependent on power presented to device
• Common measurements include gain compression, AM-to-PM
conversion, and harmonic and intermodulation distortion

Page 8
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Vector Network Analyzers Key to Component Test
Vector network analyzers (VNAs)…
• Are stimulus-response test systems
• Can sweep frequency or power
• Characterize linear (S-parameters) S21
and nonlinear (compression, IMD, etc.) performance
S11 S22
• Are very fast for swept measurements
• Provide the highest level of measurement accuracy S12

In-band

DUT
f1 f2 2f1-f2 f1 f2 2f2-f1 ...

Page 9
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Agenda
• Overview of Component Testing Using a VNA
• Modern VNA Architectures
• Nonlinear Amplifier Tests LO

– Intermodulation distortion Source 2

To receivers

– Phase versus drive OUT 1

Source 1
OUT 2
OUT 1 OUT 2

– Hot S22
R1 R2
– True-mode stimulus A B

– Single connection,
multiple measurements Source 2 Source 2
Test port 1 Output 1 Output 2 Test port
2

• Mixer and Converter Tests DUT

• Summary

Page 10
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Traditional VNA Architectures
• Traditional VNAs have one RF source, two test ports
• Four-port analyzers are common now

LO
RF Source
RF Source R

LO LO LO LO
R1 R2 LO

A B A B C D

Test port 1 Test port 2 Test port 1 Test port 2 Test port 3 Test port 4

Reference receivers for each test port Reference receiver shared among all test ports

Page 11
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Nonlinear Testing With One Source – Power Sweeps
• Sweep power at a fixed frequency (CW)
• Measure gain compression and phase compression (AM-to-PM
conversion)
Saturated Output Power
Output Power (dBm)

Compression region

Linear region
(slope = small-signal gain)
Input Power (dBm)

Page 12
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Nonlinear Testing With One Source – Harmonics
• Sweep fundamental frequency
• Measure harmonics generated by DUT
• VNA must be able to tune receivers independent of stimulus
(frequency-offset mode)

2nd harmonic example

Tune receiver to 2xf1

Stimulus

DUT
f1 2xf1

Page 13
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 • Second internal source
Modern Two-Port
• Internal signal combiner
VNA Architecture • Flexible signal routing
• Internal modulators and pulse generators
rear panel

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 Pulse generators R2
A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2

Page 14
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 • Second internal source
Modern Four-Port • Internal signal combiner
VNA Architecture • Flexible signal routing
• Internal modulators and pulse generators
rear panel

LO
Source 2
Source 1
Pulse generators
OUT 1 OUT 2
OUT 1 OUT 2 To receivers

R1 R3 R4 R2
A C D B

Test port 1 Test port 3 Test port 4 Test port 2

Page 15
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Easily Switch Between One- and Two-Source Tests

Page 16
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Desirable Source Attributes
• High port power to drive amplifiers into compression and to
ensure measurable distortion products
• High port power for driving mixer LO ports
• Low source harmonics for more accurate harmonic and IMD
measurements
• Wide power-sweep range to cover linear-to-nonlinear transition
• Built-in pulse modulators for simple pulsed S-parameter setups
+16 dBm

-60 dBc
P OUT 1 OUT 2

Page 17
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Benefit of Low Source Harmonics for IMD Tests
In-band
Error Due to Interfering Signal
100

10 -
+ DUT
Error (dB, deg)

phase error f1 f2 2f1-f2 f1 f2 2f2-f1

1

magn error
0.1
• Source harmonics will interfere with device-
generated intermodulation products
0.01
• Quantify error by applying f1+2f2 or 2f1+f2
at input to DUT where power of 2f1 or 2f2
0.001 is same as source 2nd harmonic power level
0 -5 -10 -15 -20 -25 -30 -35 -40 -45 -50 -55 -60 -65 -70
Interfering signal (dB)

Page 18
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Desirable Receiver Characteristics
• Excellent linearity with a high compression point
• Narrowband IF path for improved IMD and pulsed measurements
• Internal IF gates for narrowband pulsed measurement
• External IF inputs for antenna-system remote-mixing situations

A Narrowband filter
Anti-alias Digital FIR
filter ADC IF filter
IF gate

External
IF input

Page 19
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
PNA-X Provides Industry-Leading Performance

N5242A PNA-X Performance

• Frequency Range 10 MHz to 26.5 GHz
• IF Bandwidths 1 Hz to 600 kHz
• Dynamic Range 130 dB at 24 GHz
• Trace Noise (1 kHz IF BW) 0.0006 dB at 22.5 GHz
• Output Power +16 dBm at 24 GHz
• Source Harmonics -60 dBc at 24 GHz
• 0.1 dB Receiver Compression +12 dBm at 20 GHz
• Power Sweep Range (ALC) 40 dB at 24 GHz

Page 20
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Additional Access Loops Expand Range of Measurements
Example 1: switch between normal Example 2: switch between network
path and high-power path analyzer and external source and SA/VSA

Booster amp

To rear access loops

To rear access loops

DUT
DUT

Page 21
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Two-Port PNA-X High-Power Amplifier Measurements
Booster amplifier

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 R2
A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2

DUT

Page 22
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Single Connection, Multiple Measurements
Alternate connection point for more signal power

Sig gen

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 R2 SA / VSA

A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2

DUT

Page 23
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Pulsed S-parameters PNA-X 8510

PNA
• Ease of Setup
– Internal pulse modulators (one or two)
– Internal pulse generators (four)
– Very fast pulse-profile measurements
(20-30x improvement over PNA) 100

Dynamic Range (dB)

Narrowband
• Wide- or narrow-band detection Detection PNA-X
80
– Wideband down to 2 us (< 1 us soon)
– Narrowband down to 33 ns 60 Wideband
Detection Narrowband
• Dynamic range improvements 8510 Detection PNA
40
for narrowband detection Narrowband
Detection
8510
– Crystal-filter path with increased gain 20

– Patented software-gating technique

Duty Cycle (%)
(especially helpful for small duty cycles) 0
100 10.0 1.0 0.1 0.01

Page 24
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Narrowband Example with Low Duty Cycle (.001%)
0
-10
-20 PNA
-30
-40 40 dB improvement!
-50
dB

-60 PNA-X
-70
-80 PNA pulse
-90 PNA-X pulse (SW gate off)

-100 PNA-X pulse

PNA-X CW
-110
PRF = 200 Hz
-120 PRI = 5 ms
PW = 1 us
7.74 8.24 8.74 9.24 9.74 10.24 10.74 11.24 11.74 12.24 12.74 IF gate = 50 ns
PNA IF = 44 Hz
Freq (GHz) PNA-X IF = 50 Hz

Page 25
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Agenda
• Overview of Component Testing Using a VNA
• Modern VNA Architectures
• Nonlinear Amplifier Tests LO

– Intermodulation distortion Source 2

To receivers

– Phase versus drive OUT 1

Source 1
OUT 2
OUT 1 OUT 2

– Hot S22
R1 R2
– True-mode stimulus A B

– Single connection,
multiple measurements Source 2 Source 2
Test port 1 Output 1 Output 2 Test port
2

• Mixer and Converter Tests DUT

• Summary

Page 26
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Two-Port PNA-X Two-Tone IMD Measurements

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 R2
A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2

DUT

Page 27
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
IMD Examples

Swept-frequency IMD Swept-power IMD

Page 28
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Optimizing IMD Measurements
• Use narrowband (crystal-filtered) IF path
• Set IF bandwidth as narrow as possible
(limited by measurement speed)
• Set attenuation in receiver path to
optimize noise floor and receiver IMD
• Avoid source spurs
– single tone spurs generated by synthesis process
– avoid multiples of 10 MHz by 3 to 10 times the IF bandwidth
• Avoid source cross modulation (due to ALC loop interaction)
– use combiner with good isolation between sources
– use open-loop leveling
Isolation

Page 29
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Looking for Source Spurs
Log mag
non-zero span ratio = 1.41
(spur)

ratio = 1.02
(no spur)

Lin mag
zero span

Mean/(Std. Dev.) ≥ 1.25

indicates a spur (for 201 points)

Polar
zero span

Page 30
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Phase Versus Drive
• Measure phase change on fixed signal F2 (low power) induced by
amplitude change on signal F1 (high power)
• Combine signals with same hardware setup as for measuring IMD
• Perform a power sweep with F1
• Tune receivers to F2 and measure S21 phase
• Calculate delta phase divided by delta amplitude (e.g., deg/dB)

Perform power Measure

sweep S21 phase

F1 F2 F1 F2
DUT

Page 31
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Two-Port PNA-X Phase Versus Drive
Jumpers changed to
reverse the combiner

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 R2
A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2

DUT

Page 32
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Two-Port PNA-X Hot S-Parameters
Jumpers changed to
reverse the combiner

LO

Source 2

To receivers
Source 1
OUT 1 OUT 2
OUT 1 OUT 2

R1 R2
A B

Source 2 Source 2
Test port 1 Output 1 Output 2 Test port 2
10.5 GHz, -5 dBm *

10 GHz Hot S22 example * Example values

+15 dBm * DUT

Page 33
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
True-Mode Stimulus
• Accurately characterize balanced amplifiers
in nonlinear region of operation
• Apply true-differential or true-common-mode Differential
stimulus (forward and reverse direction) o
(180 out-
• Apply arbitrary phase offset (CW only) of-phase)

• Measure differential-mode, common-mode,

and cross-mode S-parameters
3 4
• Correct amplitude/phase imbalance due to
mismatch with single-ended 4-port calibration
Common
(in-phase)

1 2

Page 34
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Agenda
• Overview of Component Testing Using a VNA
• Modern VNA Architectures
• Nonlinear Amplifier Tests LO

– Intermodulation distortion Source 2

To receivers

– Phase versus drive OUT 1

Source 1
OUT 2
OUT 1 OUT 2

– Hot S22
R1 R2
– True-mode stimulus A B

– Single connection,
multiple measurements Source 2 Source 2
Test port 1 Output 1 Output 2 Test port
2

• Mixer and Converter Tests DUT

• Summary

Page 35
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Four-Port PNA-X Scalar Mixer/Converter Measurements

LO
Source 2
Source 1
OUT 1 OUT 2
OUT 1 OUT 2 To receivers

R1 R3 R4 R2
A C D B

Test port Test port Test port Test port

1 3 4 2

DUT

Page 36
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Four-Port PNA-X Vector Mixer/Converter Measurements

LO
Source 2
Source 1
OUT 1 OUT 2
OUT 1 OUT 2 To receivers

R1 R3 R4 R2
A C D B

Test port Test port Test port Test port

1 3 4 2
Ref
mixer DUT

Page 37
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Vector Mixer/Converter Measurements – High LO Power
Booster amplifier

LO
Source 2
Source 1
OUT 1 OUT 2
OUT 1 OUT 2 To receivers

R1 R3 R4 R2
A C D B

Test port Test port Test port Test port

1 3 4 2
Ref
mixer DUT

Page 38
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Example of Fixed and Swept LO Measurements

Swept LO
Fixed IF

Fixed LO
Swept IF

Page 39
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Speed Improvements with Internal Second Source
SMC Cycle Time (Swept LO/Fixed IF)

1000
199x

100
37x
162x

10x
113x PSG (SW trig)
Seconds

28x PSG (HW trig)

10
MXG (HW trig)
Internal 2nd source
19x 9x
1x

6x
1

1x

1x

0.1
51 201 1601 Correction ON
Trace points IF bandwidth = 100 kHz

Page 40
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Advanced Calibration Techniques for Mixers/Converters
••Simple
Simplesetup
setupwith
withno
no
Conversion loss external signal source Conversion loss,
and match external signal source delay, and match
••Match
Matchcorrection
correction
ELIMINATES
ELIMINATESATTENUATORS!
ATTENUATORS!

GPIB
Reference
mixer

Calibration
Power meter mixer/filter DUT
DUT

Scalar Mixer Calibration (SMC) Vector Mixer Calibration (VMC)

• Highest accuracy conversion-loss • Most accurate measurements of phase and
measurements with simple setup and cal absolute group delay
• Removes mismatch errors during calibration • Removes magnitude and phase errors for
and measurements by combining one-port transmission and reflection measurements by
and power-meter calibrations calibrating with characterized through mixer

Page 41
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Example Conversion Loss Measurements

8720 without attenuators 8720 with attenuators

PNA without attenuators using SMC

Page 42
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Agilent's Patented Vector Mixer Calibration
Technique Calibration mixer/filter pair
OPEN

IF-
Three step calibration: SHORT

• Step 1: measure one-port error terms LOAD

at input and output frequencies RF IF+ IF- = RF-LO

• Step 2: completely characterize a LO
calibration mixer/filter pair using
reflection measurements Internal switch
(acquire S11, S22, and “C21”)
• Step 3: calibrate transmission tracking b2
term of test system using
characterized mixer/filter as a SHORT

through standard OPEN

LOAD

Reference mixer
Calibration mixer/filter

Page 43
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Example Group Delay Measurements

PNA with VMC (no attenuators)

VNA with
attenuators

Page 44
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Agenda
• Overview of Component Testing Using a VNA
• Modern VNA Architectures
• Nonlinear Amplifier Tests LO

– Intermodulation distortion Source 2

To receivers

– Phase versus drive OUT 1

Source 1
OUT 2
OUT 1 OUT 2

– Hot S22
R1 R2
– True-mode stimulus A B

– Single connection,
multiple measurements Source 2 Source 2
Test port 1 Output 1 Output 2 Test port
2

• Mixer and Converter Tests DUT

• Applications
• Summary

Page 45
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Frequency Converters w/Embedded LO
S.W Option = 084

LO LO

Embedded LO

Page 46
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Embedded LO Option 084:
1. Allows measurement of relative phase and absolute group delay
without accessing to embedded LO source.

2. Is an extension of the Vector Mixer Calibration (VMC, opt. 083).

• PNA requires option 014, 080, 081, UNL and 083.
• PNA-X requires option 080, 083.

Embedded LO Option 084 is NOT:

1. Fixed output frequency.
2. DUT LO frequency > 10MHz of reference mixer LO frequency.

Page 47
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Introducing the N5242A Option 029
Source-corrected noise figure
option extends single-connection
multiple-measurement capability
of the PNA-X
• Measure key amplifier parameters up
to 26.5 GHz with a single connection
(e.g. S-parameters, noise figure,
Agilent's unique noise-figure-calibration compression, IMD, harmonics)
technique uses an ECal module as an
impedance tuner to remove the effects of • Achieve the highest measurement
imperfect system source match accuracy of any solution on the
market

Page 48
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
PNA-X’s Unique Source-Corrected Technique
• PNA-X varies source match around 50 ohms using an ECal module
(source-pull technique)
• With resulting impedance/noise-figure pairs and vector error terms,
very accurate 50-ohm noise figure (NF50) can be calculated
• Each impedance state is measured versus frequency

(Z1, F1), (Z2, F2), …

frequency

Page 49
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Noise Figure Uncertainty Example (ATE Setup)
4.500
Amplifier:
Gain = 15 dB Y-factor with noise source
Input/output match = 10 dB
NF = 3 dB
connected to DUT via switch matrix
Gamma opt = 0.27 ∠ 0o
4.000
Fmin = 2.7 dB
Rn = 12 – 33
0.75 dB

3.500
0.5 dB
NF (dB)

0.2 dB
PNA-X Y-factor with noise source
3.000
directly at DUT input

2.500

2.000
0.5 2.5 4.5 6.5 8.5 10.5 12.5 14.5 16.5 18.5 20.5 22.5 24.5 26.5
GHz
* Note: this example ignores drift contribution

Page 50
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Uncertainty Breakdown (ATE Setup)
Coaxial amp, Noise figure = 3 dB, Gain = 15 dB (@ 4.5 GHz)
Fmin = 2.8 dB, Gopt = 0.27 +j0, Rn = 37.4

0.9
0.8
0.7
0.6
0.5
97% confidence (dB)
0.4
0.3
0.2
0.1
0 Y-factor (Embedded)
Y-Factor (Not Embedded)
Total Uncertainty

ENR Uncertainty

Mismatch
Wave (Embedded)
Noise Parameters

S-parameters

Jitter
Wave (Embedded)
Y-Factor (Not Embedded)
Y-factor (Embedded)
Uncertainty Contributor

Page 51
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Comparing Accuracy of Two Methods
• Noise parameter effect present for both methods
• Y-factor
– Noise source directly to DUT: good source match
– Noise source in ATE or probe situation: poor source match
• Cold source
– Without source correction: poor source match
– With source correction: excellent effective source match

Page 52
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
N5242A Option 086 (GCA)
Gain Compression Application
GCA builds upon PNA-X’s strength in
single-connection active-device measurements,
providing gain compression data fast and
accurately,
at multiple frequencies, with a simple setup.
Measure a key amplifier spec multiple times faster
compared to current methods, with GCA’s SMART sweep.

Achieve the highest measurement accuracy of any solution

in the market with mismatch correction.

Frequency converters not supported at intro. Future

enhancement.

A firmware upgrade for existing PNA-Xs.

Free 14-day trial license for PNA-X owners to try out Option 086.

Page 53
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
 Nonlinear Testing GCA – Power Sweeps

Page 54
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Easily Switch Between One- and Two-Source Tests

Page 55
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Swept
Harm

Page 56
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Swept
Third

Page 57
Beyond S-Parameters
© Agilent Technologies, Inc. 2007
Summary
Modern VNA architectures provide:
• Two internal signal sources with high output power and low harmonics
– Simplifies measurements of IMD, phase vs. drive, hot S22 and more
– Provides convenient and fast LO signal for mixer and converters
• Flexible signal routing
– Combine sources with internal signal combiner for variety of measurements
– Add signal-conditioning hardware and external test
equipment via front and rear-panel RF access loops
• Complete set of pulsed S-parameter hardware
– Modulators, pulse generators, IF gates,
wide and narrowband IF filters
– Detection choices offer tradeoff between speed, dynamic range, and resolution
Characterize linear and nonlinear behavior of amplifiers,
mixers and converters with simple test setups and fast test
times

Page 58
Beyond S-Parameters
© Agilent Technologies, Inc. 2007