Design for Guaranteed EMC Compliance · 2013. 4. 30. · Calculator (MREMC) Development of...
Transcript of Design for Guaranteed EMC Compliance · 2013. 4. 30. · Calculator (MREMC) Development of...
Automotive EMC Workshop
Clemson Vehicular Electronics Laboratory
Reliable Automotive Electronics
Design for Guaranteed EMC Compliance
April 29, 2013
Todd Hubing Clemson University
EMC Requirements and Key Design Considerations
Radiated Emissions Radiated Susceptibility
Transient Immunity
Electrostatic Discharge
Bulk Current Injection
• 1 HF GND • Risetime Control • Filtered I/O • Adequate
Decoupling • Balance Control
• 1 HF GND • Filtered I/O • Adequate
Decoupling • Balance Control
• LF Current Path Control
• Chassis GND on board
• Filtered I/O • Adequate
Decoupling
• LF Current Path Control
• Chassis GND on board
• Filtered I/O • Adequate
Decoupling
• 1 HF GND • Chassis GND on
board • Filtered I/O • Adequate
Decoupling • Balance Control
2
In 2011, CVEL began to guarantee that the automotive products they reviewed/designed would meet all automotive EMC
requirements the first time they were tested.
Clemson Vehicular Electronics Laboratory - 2013
What we are NOT doing
3
NOT Relying on EMC Design Guidelines
Clemson Vehicular Electronics Laboratory - 2013
What we are NOT doing
4
NOT Modeling Products with Numerical EM Modeling Codes
Numerical EM modeling codes give precise answers to precisely defined problems. EMC geometries are not well-defined.
We don’t want to know how much a given configuration will radiate. The answer to that question depends on a lot of factors that we have no control over. We want to know if our product will meet its requirements.
Clemson Vehicular Electronics Laboratory - 2013
What we ARE doing
5
Identifying all possible sources, victims and coupling paths
SOURCE ANTENNA
Clemson Vehicular Electronics Laboratory - 2013
History
6
1989 2013 1995 2006
Begin development of
numerical modeling
software for EMC analysis at
UMR
Formation of EMC Expert
System Consortium
First Maximum Radiated Emission Calculator (MREMC)
Development of Performance-Based Design
for EMC Process
Investigation of fundamental EMI source mechanisms driving common-mode radiation from printed circuit boards with attached cables, IEEE Trans. on EMC, Nov. 1996.
Calculating radiated emissions due to I/O line coupling on printed circuit boards using the imbalance difference method,” IEEE Trans. on EMC, Feb. 2012.
Estimating maximum radiated emissions from printed circuit boards with an attached cable, IEEE Trans. on EMC, Feb. 2008.
Development of algorithms for calculating radiated emissions and EM coupling for various PCB structures
~40 publications relating to ability of various PCB structures to radiate
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Concept
7
315 MHz RF Transmitter (5 watts)
Connector for Antenna
What is the maximum 3-meter radiated field strength at 315 MHz?
a. impossible to predict without knowing what antenna is connected b. impossible to predict even if the antenna is known c. 15 V/m d. none of the above
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Concept
8
315 MHz RF Transmitter (5 watts)
Connector for Antenna
What is the maximum 3-meter radiated field strength at 315 MHz?
2
02
14 2
radrec
EPP D
rπ η= = 022max
radPE D
rη
π=
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Concept
9
315 MHz RF Transmitter (5 watts)
Connector for Antenna
What is the maximum 3-meter radiated field strength at 315 MHz?
02 2
377 5 6 4 14 62 2 3max
( )( W)( . ) . V/m
( m)radP
E Dr
ηπ π
Ω= = =
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Concept
10
What is the maximum 3-meter radiated field strength at 200 MHz?
We can put an upper bound on the radiated emissions at any given frequency! The more we know about the product design, the lower this upper bound becomes.
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Concept
11
Possible Antenna? Possible Source?
(Processor with 200-MHz clock)
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Calculation
12
~
Cable Equivalent
voltage
VCM
heatsink
board
=CM DMCV VC
boardmax
heatsink board cable
0.2234= × ×DMC rV E
C F F
To Floor
To Floor
~ VDM Noise voltage
Heatsink
References [1] H. Shim and T. Hubing, “Model for Estimating Radiated Emissions from a Printed Circuit Board with Attached Cables Driven by Voltage-Driven Sources,” IEEE Transactions on Electromagnetic Compatibility, vol. 47, no. 4, Nov. 2005, pp. 899-907. [2] Shaowei Deng, Todd Hubing, and Daryl Beetner, "Estimating Maximum Radiated Emissions From Printed Circuit Boards With an Attached Cable,“ IEEE Trans. on Electromagnetic Compatibility, vol. 50, no. 1, Feb. 2008, pp. 215-218.
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Calculation
13
heatsink
board
0.43 pF5.14 pF
==
CC
20cm 20cm
100 cm
5cm 5cm
1cm
Spacing between heatsink and board is 1 cm
Clemson Vehicular Electronics Laboratory - 2013
Maximum Radiated Emissions Calculator
14
Clemson Vehicular Electronics Laboratory - 2013
Performance-Based EMC Design Procedure
15
1. Determine worst-case signal characteristics
2. Calculate maximum possible emissions from signal driving matched antenna
3. If > limit at any frequency, control risetime with series resistor
4. Recalculate maximum possible emissions from signal driving matched antenna
5. Proceed to Step II.
Step I: For each net on each board:
Clemson Vehicular Electronics Laboratory - 2013
MS Excel Spreadsheet Calculation
16
Clemson Vehicular Electronics Laboratory - 2013
Design Review Procedure
17
1. Determine worst-case emissions due to each of the 5 MREMC algorithms that apply to your design
2. For any net that does not meet the specification at every frequency as determined by a given algorithm, adjust the design until the net is compliant.
Step II: For each net at each frequency over the limit:
Clemson Vehicular Electronics Laboratory - 2013
Example 1: Microcontroller Output Driver
18
Vsource = 3.3 V
Imax = 20 mA
Cin = 5 pF
Rseries = 0 Ω
CLK Freq = 100 kHz
Available Information
Rsource = 165 Ω
T = 10 µs
tr = 1.82 ns
Calculated Parameters
Suppose we connected an output of this microcontroller directly up to an impedance-matched antenna…
Automotive microcontroller in typical application:
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
10 100 1000
3-M
ETER
E-F
IELD
IN D
B(u
V/M
)
FREQUENCY IN MHZ
Maximum Radiated Field
FCC Limit
Absolute maximum possible emissions!
Clemson Vehicular Electronics Laboratory - 2013
Example 1: Microcontroller Output Driver
19
Vsource = 3.3 V
Imax = 20 mA
Cin = 5 pF
Rseries = 20 kΩ
CLK Freq = 100 kHz
Available Information
Rsource = 8165 Ω
T = 10 µs
tr = 220.0 ns
Calculated Parameters
Suppose we connected an output of this microcontroller directly up to an impedance-matched antenna…
Same output with 20-kΩ series resistor:
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
10 100 1000
3-M
ETER
E-F
IELD
IN D
B(u
V/M
)
FREQUENCY IN MHZ
Maximum RadiatedFieldFCC Limit
Clemson Vehicular Electronics Laboratory - 2013
Series Resistors
20
Optimal control
Minimal cost / Minimal footprint
Predictable behavior
Easy to adjust without affecting layout
Reduces power bus noise
Why use series resistors to control transition times?
Clemson Vehicular Electronics Laboratory - 2013
Example 2: Microcontroller Output Driver
21
Vsource = 3.3 V
Imax = 20 mA
Cin = 5 pF
Rseries = 0 kΩ
CLK Freq = 1 MHz
Available Information
Rsource = 165 Ω
T = 1 µs
tr = 1.82 ns
Calculated Parameters
Suppose we connected an output of this microcontroller directly up to an impedance-matched antenna…
Same output with 1 MHz output:
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
10 100 1000
3-M
ETER
E-F
IELD
IN D
B(u
V/M
)
FREQUENCY IN MHZ
Maximum Radiated Field
FCC Limit
Clemson Vehicular Electronics Laboratory - 2013
Example 2: Microcontroller Output Driver
22
Vsource = 3.3 V
Imax = 20 mA
Cin = 5 pF
Rseries = 8 kΩ
CLK Freq = 1 MHz
Available Information
Rsource = 8165 Ω
T = 1 µs
tr = 90 ns
Calculated Parameters
Suppose we connected an output of this microcontroller directly up to an impedance-matched antenna…
Same output with 1 MHz output and 8-kΩ series resistor:
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
10 100 1000
3-M
ETER
E-F
IELD
IN D
B(u
V/M
)
FREQUENCY IN MHZ
Maximum Radiated Field
FCC Limit
Clemson Vehicular Electronics Laboratory - 2013
Example 3: Xilinx Vertex-6 FPGA SelectIOTM
23
Vsource = 2.5 V
Imax = 240 mA*
Cin = 5 pF
Rseries = 0 kΩ
CLK Freq = 1 MHz
Available Information
Rsource = 50 Ω
T = 1 µs
tr = 0.55 ns
Calculated Parameters
Suppose we connected an output of this FPGA directly up to an impedance-matched antenna…
With 1 MHz output :
Clemson Vehicular Electronics Laboratory - 2013
Example 3: Xilinx Vertex-6 FPGA SelectIOTM
24
Vsource = 2.5 V
Imax = 240 mA*
Cin = 5 pF
Rseries = 20 kΩ
CLK Freq = 1 MHz
Available Information
Rsource = 50 Ω
T = 1 µs
tr = 221 ns
Calculated Parameters
With 1 MHz output and 20-kΩ series resistor: Suppose we connected an output of this FPGA directly up to an impedance-matched antenna…
Clemson Vehicular Electronics Laboratory - 2013
Example 4: Xilinx Vertex-6 FPGA SelectIOTM
25
Vsource = 2.5 V
Imax = 240 mA*
Cin = 5 pF
Rseries = 0 kΩ
CLK Freq = 32 MHz
Available Information
Rsource = 50 Ω
T = 31 ns
tr = 0.55 ns
Calculated Parameters
With 32 MHz output and 0-Ω series resistor: Suppose we connected an output of this FPGA directly up to an impedance-matched antenna…
Clemson Vehicular Electronics Laboratory - 2013
Example 4: Xilinx Vertex-6 FPGA SelectIOTM
26
Vsource = 2.5 V
Imax = 240 mA*
Cin = 5 pF
Rseries = 500 Ω
CLK Freq = 32 MHz
Available Information
Rsource = 50 Ω
T = 31 ns
tr = 6.0 ns
Calculated Parameters
With 32 MHz output and 500-Ω series resistor : Suppose we connected an output of this FPGA directly up to an impedance-matched antenna…
Clemson Vehicular Electronics Laboratory - 2013
3 Elements of a Radiated Emissions Problem
27
SOURCE ANTENNA
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Nets)
28
Direct Radiation from Trace
Trace Drives an Attached Cable and/or Heatsink
Trace Couples to another Trace that Drives an Attached Cable
Trace Drives the Power Bus
Need to know: net dimensions
Need to know: net dimensions, net placement, connector placement, and board dimensions
Need to know: net dimensions, net placement, connector placement, and board dimensions
Need to know: board dimensions
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Nets)
29
Direct Radiation from 10-cm trace, 1-mm above plane
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Nets)
30
Trace Couples to another Trace that Drives an Attached Cable
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Nets)
31
Trace Drives an Attached Cable and/or Heatsink
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Nets)
32
Trace Drives an Attached Cable and/or Heatsink
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Components)
33
Direct Radiation from Component
Component Drives an Attached Cable and/or Heatsink
Component Couples to another Trace that Drives an Attached Cable
Component Drives the Power Bus
Negligible
Measure or model the equivalent dipole source for the component and use the trace algorithm
Measure or model the equivalent dipole source for the component and use the trace algorithm
Need to know: board dimensions, CPD and load Cs, component datasheet information
Clemson Vehicular Electronics Laboratory - 2013
MREMC Algorithms (Shielded Products)
34
Board analysis should be done as if there were no shield
E-field coupling problems can be mitigated with E-field shielding
Common-mode currents on cables can be mitigated enclosure to cable filtering
Wiring Harness
Digital Return PlaneChassis Ground Plane
Capacitors connectingchassis ground to thedigital return plane
Chassis connectionto chassis ground
Chassis connectionto chassis ground
Clemson Vehicular Electronics Laboratory - 2013
h = 0.5 - no plane h = small - w/ plane
MREMC Algorithms (Differential Signals)
35
Use Imbalance Difference Model to convert all differential signals to equivalent common-mode sources
Then apply the same algorithms used for single-ended signals
Clemson Vehicular Electronics Laboratory - 2013
Susceptibility Calculations
36
PORT 1 PORT 2
Calculate Maximum Possible S21 Maximum Radiated Emissions Calculator (MREMC)
Clemson Vehicular Electronics Laboratory - 2013
Application to Infotainment System
37
5 Circuit Boards, mixed-signal RF, audio, video
Internal ribbon cable connections
Unshielded external connections
AM/FM Radio
3 Camera Interfaces
GPS
DVD Player
USB
Fold-out Display
Clemson Vehicular Electronics Laboratory - 2013
EMC Testing
38
Clemson Vehicular Electronics Laboratory - 2013
Identify Antennas
39
Clemson Vehicular Electronics Laboratory - 2013
Design Guidelines
40
e.g. No ground traces. No shared ground vias.
Many design “rules” were violated in the final design. Attempting to comply with a complete list of design rules would have made the product unnecessarily expensive.
Nevertheless, some rules make too much sense to ignore. (Even if they are not explicitly required.)
Clemson Vehicular Electronics Laboratory - 2013
Design Flexibility
41
e.g. grounding option
It’s a good idea to leave options open to deal with unexpected issues.
Clemson Vehicular Electronics Laboratory - 2013
Design Standards
42
e.g. GPS antenna interface
Many circuit geometries were based on known success with prior products.
Clemson Vehicular Electronics Laboratory - 2013
For This Product Design (Nets)
43
Direct Radiation from Trace
Trace Drives an Attached Cable and/or Heatsink
Trace Couples to another Trace that Drives an Attached Cable
Trace Drives the Power Bus
No calculations made. Provided HF current return for all nets not eliminated after Step 1 (critical nets).
Optimized each critical net, but relied on filtering to chassis to guarantee compliance.
No calculations made. Visually highlighted all I/O and kept several trace heights away from critical nets.
No calculations made. Focused on providing excellent HF decoupling.
Clemson Vehicular Electronics Laboratory - 2013
For This Product Design (Components)
44
Direct Radiation from Component
Component Drives an Attached Cable and/or Heatsink
Component Couples to another Trace that Drives an Attached Cable
Component Drives the Power Bus
No calculations made. Negligible.
No calculations made. Judged to be a non-issue.
No calculations made. Visually highlighted all I/O and kept critical components away.
No calculations made. Focused on providing excellent HF decoupling.
Clemson Vehicular Electronics Laboratory - 2013
Current Project Status
45
Documenting MREMC algorithms
Increasing awareness
Looking for software partner
Formulating radiated susceptibility algorithms
Clemson Vehicular Electronics Laboratory - 2013
Performance-Based EMC Design of Electronic Systems
46
In Compliance Magazine, May 2013.
Automotive Testing Technology International, Nov. 2012.
Clemson Vehicular Electronics Laboratory - 2013
Expected Outcomes
47
Software tools will make this technique easier to implement and accessible to non-expert design engineers
Will increase consumer demand for EMC-specific component information
Will not replace EMC engineers, but will allow more sophisticated designs
Will help engineers to use numerical EM modeling tools more effectively