Very-Near-Field Solutions
for
Far-Field Problems
Local representation by Electronic Instrument Associates
Frank Krozel – (630) 924-1600
frank@electronicinstrument.com
http://www.electronicinstrument.com
Agenda
Corporate Information
Introduction to Near-Field
Very-Near-Field Implementation
EMxpert Validation
EMxpert Test Applications
EMxpert Demonstration
Customer Case Studies
Conclusion
Corporate Introduction
An Established Company
Private Canadian corporation since 1989
– Coverage in 49 countries
Unique and patented products for RF and EMI
– Very-near-field magnetic measurements
Recognized innovative products
A Leader
World Leading Developer of Visual Real-Time
EM and RF Diagnostic Solutions
Antenna and PCB Designers
Product Integration and Verification Engineers
Pre-Compliance Not Compliance
Exciting Value Proposition
Substantially Reduce Project Development Costs
Dramatically Increase Designer Productivity
Significantly Accelerate Time-to-Market
1 Hour in a Chamber or 1 Second with EMSCAN?
Chamber on your Desktop
EMxpert
–
EMC/EMI diagnostic tool enabling
designers to rapidly diagnose and
solve EM problems in a single design
cycle in their own lab environment
RFxpert
–
APM tool enabling engineers to
quickly evaluate and optimize their
designs with real-time antenna
performance characterization at their
desk
Fundamentals
High-density planar antenna array
High-speed electronic switching
Very near-field measurements
Far-field predictions
Real-time real-fast
No chamber
Near-Field Introduction
Very-Near-Field Techniques
Near-Field Scanning
Near-field scanning for antenna measurement
Near-field scanning for EMC applications
What is Near-Field?
Anything not in
the far-field
Far-field is where
the pattern is not
changing with the
distance
Common definitions
Usually stay out of the
reactive region
Where is Near-Field?
Emissions approximate
a plane wave when the
maximum phase difference
between any point
on the source is 22.5°
The equates to ⁄
Fields from Infinitesimal Dipole
0
1 Distances should be
measured form the
edge of the reactive
region
For low frequencies the
dimensions most devices
are not≫ C.A.Balanis – Antenna Theory : Analysis and Design 3rd Ed.
More Realistic Approximation
For a case of D = 1m
No radiating
near-field
for low frequencies
Situation for many EMC
problems
Reactive region
extend beyond
the measurement
distance of EMC test
J.C.Bolomey - Engineering Applications of the Modulated Scatterer Technique
EMI/EMC Test Solution
Compliance parameters in far-field
Debugging via near-field
Far-Field Measurements
Far-field site far and demanding a large area
Open-air-test-site (OATS) avoids reflections
Difficult nowadays because of EM noise pollution
Image: www.noiseblog.com
Far-Field Measurements cont.
Controlled environment
Anechoic or semi-anechoic chamber
Image: www.geosig.com
Very-Near-Field Alternative
Origin of all emissions
Insight into root causes
Calculate far-field emissions from very-near-field
Very-Near-Field Scanning Solutions
Very close so coupling has to be managed
Data collection is very fast
Small area can collect the majority of the
emissions
Very-Near-Field Implementation
EMxpert
Real-time measurements
( <1 sec)
Compact tabletop instrument
Cost effective solution
Hardware
1218 probes in a 29 x 42 array
–
2436 loops in X
Antennas
–
–
Sensitive down to -135 dBm
Inefficient for EMI isolation
3.75mm resolution
Scan area 21.8cm x 31.6cm
Plastic case for safe testing
Functionality
Spectral scan
–
Problem frequencies
Real-time near-field spatial
scan
–
Sources of radiated emissions
Far-field prediction
–
Third party software
50 kHz to 4 GHz
Automated report generator
Technical Specifications
System Configuration
LAN/USB
USB
EMSCAN Application
External Trigger
EMSCAN Adapter
Spectrum Analyzer
Control
RF
EMSCAN Scanner
Spectral Scan: Scanner
Spectral Scan: Far-Field Antenna
Device Emission: Shows up in near-field scan and far-field position scans but not in far-field
ambient
Device Emission at an Ambient Frequency: Shows as ambient in the far-field ambient scan
and as a marked peak in the near-field scan; suspected to be a legitimate peak that happens
to occur very close to an ambient signal
Suspected Device Emission: Signal that is not in the far-field ambient or near-field device scan
but appears in the position scans; it is a suspected cable emission
Ambient Frequency: A signal that shows up in the far-field ambient scan and nowhere else
Spatial Scan: Scanner
Spatial Scan: Hand-Held Probe
Up to 1218 measurements at
the maximum possible
resolution of the probe
Spatial scan scaled up to
maximum resolution
Far-Field Application
Software to predict Open Area Test Site (OATS) or
free space radiated EMI of PCB
– Compensated EMxpert very-near-field data
– PCB structure and design models
Absorber mat 2 mm
EMxpert Validation
Simulation
Simulation
EMC Analysis for a PCB mounted switching regulator using
Electromagnetic Simulator
Mitsuharu Umekawa
EDA Application Engineering
Electronic Measurement Group
Agilent Technologies
Microwave Workshop and Exhibition
December 1st , 2011
Simulation
Validation of ADS Momentum simulation models
EMC Toyo corporation (EMSCAN Representative)
Tokyo, Japan
November 1, 2011
EMxpert Validation
Far-Field Measurements
Case Study
PCBs containing split planes on ground plane fail EMI
requirements more often than those without
Measure at DVT Solutions in Calgary Canada
Why Split Planes on Ground Plane?
In digital designs, current always travel in a loop
– Return via ground connection
Full ground plane is the usual design
– Inadequate for multiple power supplies with
varying VDC
Plane splitting permits more than one voltage
ground return on a single layer
– Each voltage assigned an isolated copper
region
Negative Effects of Split Planes
Reflections due to impedance mismatch
– Signal and/or clock distortions
Increase in common ground impedance in ground path for
adjacent signals and/or clocks
– High cross-talk
Signal and/or clock return currents departing from
indented path
– Whole plane an effective broadband radiator with
multiple resonances when current flows along the
edges of a split plane
Split Plane EMI Mitigation
Mitigation with shorted jumpers
Mitigation with de-coupling capacitors
Case Study Parameters
Impact of Open-Ended Micro-Strip Line Designs
– Very-Near-Field Levels Analysis
• Effects of joining split planes at discrete locations
with simple short along ground plane split
• Effects of placing 470pF chip capacitors between split
planes at discrete locations along ground plane split
– Far-Field Levels Verification
• Changes in the far-field levels from ground plane
split with shorts and chip capacitors between split
planes
Test PCBs
Test Case 1
Ground Plane
w/o Split
Test Cases 3 - 7
Split Plane
c/w Modifications
Test Case 2
Ground Plane
c/w Split
Case 1: Micro-strip Line
Case 2: Ground Split
Cases 3 to 4: Shorting Jumper
Test Case 3
Center
Test Case 4
Edge
Cases 5 to 7: Chip Capacitors
Test Case 5
1 x 470pF in Center
Test Case 6
1 x 470pF at Edge
Test Case 7
3 x 470pF in Center
C
(pF)
Frequency
(MHz)
Z
(Ohm)
470
220
1.54
470
780
0.43
1410
220
0.51
1410
780
0.14
Very-Near-Field Setup of EMxpert
Very-Near-Field Measurements
220 MHz Spatial Scan
Very-Near-Field Effects of Split
Solid Ground
Split Ground
Shorting Jumper
Split Ground
Solid Ground
Jumper Center
Jumper Edge
Capacitor
470pF Center
Solid Ground
1410pF Center
470pF Edge
220 MHz Analysis
Intelligent analysis still requires an engineers mind
– Combining measured very-near-field with knowledge
of the PCB layout can identify root layout mistakes
Near-field and layout analysis indicates split ground
plane fault
Split ground plane is a problem because of disturbed
return current paths
Using this knowledge an optimum solution will recover
the most uniform return current distribution
Prioritize uniform return currents over actual level
220 MHz Analysis
Analysis of current distributions with most uniform being
prioritized over actual level
– Case 3 Jumper centre
– Case 7 Three capacitors centre
– Case 5 Single capacitor centre
– Case 6 Single capacitor edge (no mitigation)
– Case 4 Jumper edge (no mitigation)
Very-Near-Field Measurements
780 MHz Spatial Scan
Very-Near-Field Effects of split
Solid Ground
Split Ground
Shorting Jumper
Split Ground
Solid Ground
Jumper Center
Jumper Edge
Capacitor
470pF Center
Solid Ground
1410pF Center
470pF Edge
780 MHz Analysis
Visualization of hot spots for quick identification of
mitigation techniques from best to worst
– Case 7 Three capacitors centre
– Case 5 Single capacitor centre
– Case 3 Jumper centre
– Case 6 Single capacitor edge
– Case 4 Jumper edge
Very-Near-Field Conclusion
3-capacitor centre preferred as solution across all bands
Far-Field Measurements
Far-Field Setup
Receiving Setup on Antenna Mast
Transmitter (DUT) Setup on Turntable
Far-Field Analysis
Frequency
(MHz)
Microstrip
Line
(dBuV/m)
Split
(dBuV/m)
Test Case 1
Test Case 2
Test Case 3
Test Case 4
Test Case 5
Test Case 6
Test Case 7
Full Ground
Split
Center
Edge
One Cap
Center
One Cap
Edge
Three Cap
Center
220 MHz
27.2
56.2
24.3
53.5
48.4
40.2
34.1
780 MHz
35.1
59.4
53.3
62.8
49.4
68.0
49.3
Label
Split+Jumper
(dBuV/m)
Split+470pF
(dBuV/m)
Split+1.41nF
(dBuV/m)
Summary
Case
220MHz
FF Level
220MHz
FF order
Full Ground
27.2
Split Ground
220MHz
NF order
780MHz
FF level
780MHz
FF order
780MHz
NF order
-
35.1
-
56.2
-
59.4
-
Jumper Centre
24.3
1
1
53.3
3
3
Jumper Edge
53.5
5
5
62.8
4
5
470pF Centre
48.4
4
3
49.4
2
2
470pF Edge
40.2
3
4
68.0
5
4
1410pF Centre
34.1
2
2
49.3
1
1
Far-Field Conclusions
3-capacitor centre or jumper centre preferred
Depending on the frequency of higher concern
Mitigating Split Plane EMI Effects
Substantial emission increase from split-planes
– Very-near-field results indicate reduction in emission
with decoupling capacitance at both 220MHz and 780
MHz
– Far-field results confirm significant reduction in levels
from decoupling capacitors
High correlation between very-near-field and far field
results when fault type taken into account
Purely looking very-near-field emission levels without
considering distribution can be misleading.
– Spatial results are critical
Test Duration for all 7 Test Cases
EMxpert
– 1 ¾ hours for 2 frequencies
Automated probe
– 21 hours for 2 frequencies
Chamber
– 28 hours for 2 frequencies
Conclusion
High correlation between very-near-field and far-field
results
Dramatically reduced test times in very-near-field
EMxpert is thus an effective tool to quickly optimize
number of decoupling capacitors, their capacitance values
and locations to mitigate split ground plane effect at low
and high frequencies
Test Applications
Functional Testing
Changes in real-time
–
–
Sources of emission
Current loop
A/B Comparison
Obsolescence management
Production unit versus gold
standard
Repair
Conducted Immunity Insights
Injected signal path
Component susceptibility
Common Mode
How signals couple onto
connectors creating common
mode problems
Broadband Emissions
Broadband emissions created by switching mode power supply
Amplitude(dBuV) Auto
30.0
20.0
10.0
0.0
0.00
100.00
200.00
300.00
400.00
500.00
600.00
700.00
Frequency(MHz) Auto
800.00
900.00
1000.00
Filtering
Effectiveness of filters
Shielding
Effectiveness of shields
In-Situ Testing
High Resolution Testing
Fast identification of the
emitting component with
EMxpert scanner
Third-party hand-held probe
for pin or trace level testing
– Data integrated in EMxpert
software application
– Automatic report generator
Pre-Compliance Testing
Far-field measurements without chamber or OATS
Far-field prediction FCC/ANSI, CISPR and user-defined limit lines
Break
Demonstration
Demonstration
Customer Case Studies
Currents Distribution
Taximeter PCB test with no capacitor – noise
circled in red
Taximeter PCB test with capacitor reduces
noise by at least 8 dBuV
Filtering - Before
Satellite Receiver Board positioned on
EMxpert. Unfiltered connection to power
supply circled in red.
Frequency = 96.9 MHz Noise radiating from
the power supply
Filtering - After
The same board and test setup as before. A
ferrite filter was added and the board retested, demonstrating that the filter eliminated
the noise
After applying the filter the noise completely
disappears
Shielding - Before
Small demodulator board
shown with no shielding.
Frequency = 269.996MHz noise
emanating from the board.
Shielding - After
Demodulator board with foil
tape shielding.
EMxpert spatial scan shows noise
eliminated after adding shielding.
Conducted Immunity
Satellite Receiver Board positioned on EMxpert
Baseline spatial scan at 50MHz
Conducted Immunity – Signal
External signal at 50MHz injected at the
video connectors on the top
Spatial scan at 50MHz showing the
current paths of the injected signal
Customer Burn-In Tests Case Study
Medical Industry
PCB with Pre-Programmed Tests
Wi-Fi module
High speed RAM
HDD
Microcontroller
HDD Tests
Noise from the HDD is being coupled on to the Wi-Fi antenna cable.
HDD Tests with Ferrite
One solution is a ferrite place on the antenna cable.
RAM Tests
Burn-in test of the RAM shows coupling to HDD and Wi-Fi cable.
Customer SERDES Case Study
Automotive Industry
Objectives
Quantify the EMI emissions profile by comparing half-duplex
deserializer (SERDES) to next generation full-duplex design
Determine whether full-duplex design impacts EMI profile and, if so,
quantify the difference
Results courtesy of National Semiconductor / Texas Instruments
Test Method
Design team placed original half-duplex board on the EMxpert
scanner to generate a baseline measurement.
Powered the device under test (DUT) and activated the scan
Test Results
Baseline results
Full-duplex results
Conclusion
No spikes and very similar peak emissions
Better EMI profile (more blue in the spatial scan)
No appreciable change occurred in full-duplex mode
– Implementation of the full-duplex feature with no additional mitigation measures.
Design team conducted the scans on the EMxpert system in their
offices
Results in minutes
To test the new design in a third party chamber would have required
that an engineer travel to an off-site test facility for the better part of
a day
Customer SSCG Case Study
Automotive Industry
Objective
Generate compelling and quantified evidence to present to
automotive industry customers that SSCG feature reduces EMI
emissions .
Results courtesy of National Semiconductor / Texas Instruments
Test Setup
Place device under test (DUT) on the EMxpert scanner with SSCG
turned “OFF”, power and capture emissions profile
To demonstrate the effectiveness of the feature, run identical test, but
with SSCG function turned “ON.”
Test Results
SSCG OFF
SSCG ON
Conclusion
The design team was able to compare 4 different methods to
implement SSCG
They were able to carefully compare results that rather dramatically
highlighted the benefit of the SSCG feature
Feature drastically reduces overall electromagnetic radiation
Whenever the customer support team presented the above
comparison, it universally resulted in a customer response of “Wow”!
Reason: Automotive engineers’ biggest challenge is reducing EMI
Any features that reduce EMI result in faster time-to-market,
less shielding, and lower costs
Conclusion
Summary
Advantages
– Continuous peak hold scan for spurious events
– Real-time view of emission sources and currents
– Fast pre-compliance regulatory data
– Low acquisition cost and zero operational cost
Benefits
– Test time reduction
> 100 x
– Rapid design iteration, prototyping & optimization
– Reduced chamber investment or third party testing
– Cost effective preparation to compliance
Business Case
In a single product life cycle, avoiding one board re-spin
and retest can save $$$
Third party testing
–
–
–
–
8 hours driving
Night in a hotel
$3000 for chamber time
4 days for debugging
Thank You
www.emscan.com