Automobiles typically have dozens of electronic systems operating interactively in a relatively compact space. These systems must operate reliably in a wide range of environments over extended periods of time. As a growing number of these systems play an ever expanding role in protecting the safety of a vehicle’s occupants, there is an increasing need to ensure that the integrity of these systems will not be compromised by electromagnetic interference.
The traditional design, build and test approach to automotive EMC compliance will not be sufficient to ensure the safety or reliability of tomorrow’s automobiles. A Design for Guaranteed Compliance approach promises to ensure that automotive components will meet all EMC requirements the first time they are tested, and that compatibility will not depend on the specific vehicle or system in which the components are installed. More work needs to be done before this concept reaches its full potential, but electronic system designers can already derive significant benefit by applying this approach to products currently under development.
In a graduate automotive electronics course at Clemson University (AuE 835), electronic systems are defined as systems that are managed or controlled by computers. By that definition, the number of electronic systems in automobiles has been zero for the majority of the automobile’s illustrious 120-plus year history. It was only in the 1980s that microprocessors started to see widespread use in automobiles, primarily for engine control. By the 1990s, they were required in automobiles for on-board diagnostics. Today, a typical automobile has dozens of embedded processors managing everything from passenger comfort and entertainment to critical vehicle safety systems. In fact it is no longer possible to build vehicles that meet federal requirements for passenger safety without microprocessors that control systems such as the anti-lock brakes, tire pressure monitoring and electronic stability control.
Cars and trucks are safer than they have ever been , due in-part to the implementation of various electronic systems that can compensate for driver error or unexpected obstacles. Examples of some of these systems include electronic stability control, adaptive cruise control, automatic braking, lane departure warning and blind-spot detection. However, as automobiles begin to rely more heavily on electronic systems for passenger safety, it becomes increasingly important to ensure that these systems work reliably, and as intended, 100% of the time.
Less than a decade ago, none of the electronic systems commonly found in production of cars were designed to apply the brakes or turn the steering without explicit driver input. Today, a car may have several systems with these capabilities plus nearly a dozen that can actuate the throttle , and the number of systems that control safety-critical actuators increases with every new model year. As electronic systems play a bigger role in automotive safety, ensuring the electromagnetic compatibility (EMC) of these systems has never been more important.
The aerospace electronics industry is also concerned with reliability and electromagnetic compatibility. A failure in a critical system can ruin a multi-billion dollar space mission. Likewise a critical failure in a commercial aircraft can jeopardize the lives of its passengers and crew. For these reasons, aerospace systems are generally designed so that the failure of any one component won’t compromise a safety-critical (or mission-critical) system. Automobiles have not traditionally been held to the same high reliability standards as aerospace electronics, but this is changing. Unlike aerospace systems, an automotive electronic system that appears in a popular vehicle may experience hundreds of millions of operating hours every year. In a safety-critical system, a failure rate as low as one for every million operating hours, could affect hundreds of people every year.
The new automotive safety standard, ISO 26262, is designed to ensure that vehicle electronics meet certain criteria for reliability and safe operation. The standard employs a risk-based approach for assigning components to various risk classes and defines procedures for the development, operation and service of components depending on their class. Modeling and testing (including EMC modeling and testing) play a key role in these procedures.
Prof. Hubing and Clemson graduate student, Dexin Zhang, making measurements on a pickup truck with interior trim removed.
Prior to the widespread use of electronic systems in automobiles, electromagnetic compatibility was a relatively minor concern in the automotive industry. Any EMC problems that arose in prototype vehicles were generally dealt with in an ad hoc fashion. Was the engine spark being picked up by the AM radio? Use resistive spark plug wires. Was relay noise causing spurious transients that burned out ignition components? Apply snubber circuits or employ better isolation.
As electronics began to be used for more automotive systems, individual manufacturers began developing their own EMC tests to ensure that the components or systems from their Tier 1 suppliers would be unlikely to cause interference problems when integrated in a vehicle. Eventually, there was an effort to consolidate these tests to make EMC testing more meaningful and cost-effective across the industry. Standards organizations such as SAE and CISPR formed committees of interested parties, such as automobile manufacturers, Tier 1 electronics suppliers and EMC test houses. These committees developed standards and test practices for evaluating conducted and radiated EM emissions, conducted and radiated EM susceptibility, transient voltages, power bus noise, electrostatic discharge and other forms of electromagnetic interference.
Clemson graduate student, Li Niu, setting up a radiated emissions test in the Clemson Vehicular Electronics Laboratory.
Today, automotive electronic systems are subjected to a wide range of EMC tests, both at the component level (before being installed in a vehicle) and vehicle level. These tests give the manufacturer confidence that electromagnetic compatibility problems will not be a significant source of vehicle failures or customer complaints over the life of the product. However, the rate at which EMC standards are written and revised has been slow compared to the exponential growth in the number of automotive electronic systems and their role in ensuring vehicle safety. It is not realistic to expect to test for all possible sources of interference or anticipate all possible interactions between systems that various Tier 1 suppliers develop for a given vehicle. Also, current EMC test procedures are performed on components and systems that are new and in good working order. There is generally very little testing performed on systems that have undetected failures or components that are out of specification (e.g. due to age or excessive heat).
In order to meet the requirements of ISO 26262 and ensure that electronic components and systems do not present an unacceptable risk, we must change the current development paradigm. Rather than designing the best we can and testing to ensure compliance, we need to use design procedures that ensure compliance and test procedures that validate compliance with the design criteria.
Prof. Hubing and Clemson graduate student, Chentian Zhu, measuring the currents in a three-phase motor driver.
Design for Guaranteed Performance
Is it possible to design automotive components in a way that ensures compliance with electromagnetic compatibility requirements? Even the simplest electronic control units (ECUs) in an automobile typically have hundreds of circuits. The component designers rarely have complete information regarding the attached wire harness routing or other key parameters that affect EMC performance. In addition, EMC test set-ups can vary depending on who is conducting the test, and results obtained for EMC tests conducted at different test sites can vary significantly [3, 4]. So is it really possible to guarantee that a particular component design will meet all automotive EMC requirements the first time it is tested? In most cases, the answer is yes.
A Design for Guaranteed Performance approach requires that an ECU design be reviewed circuit-by-circuit to ensure that the combination of electrical and geometric parameters cannot couple sufficient energy (in or out) to cause an EMC failure under any circumstances. For example, if a given radiated emissions specification can be exceeded only if an ECU radiates at least 4 nW of power at a given frequency; then the designer must ensure each circuit on the board is incapable of supplying 4 nW of power at that frequency, either alone or in conjunction with other circuits. In the event that a particular circuit is required to supply more than 4 nW of power at that frequency in order to meet functional criteria, then that signal path must be design in a manner that ensures that all power in excess of 4 nW is accounted for and dissipated within the ECU.
The following section demonstrates how Design for Guaranteed Performance can be implemented in order to meet radiated emission requirements. Note that this section deals with actual radiated emissions, as opposed to electric- or magnetic-field coupled emissions (i.e. coupling in the near field). Since many automotive “radiated” EMC tests actually measure near-field coupling, a similar procedure must be employed for ensuring near-field coupling in a particular test environment does not exceed the corresponding limit.
Design to Meet Radiated Emissions Criteria
Step 1: Know Your Limits
The first step in the process of designing for guaranteed performance is to know your limits. For example, if your product specification requires that the radiated field strength at a distance of 10 meters is no greater than 30 μV/m at 80 MHz, then the power radiated must be less than:
where r is the distance from the source, η0 is the intrinsic impedance of free space and Dmax is the maximum directivity of the radiating structure. (In this case, we’ll assume the radiating structure is not electrically large at 80 MHz, so the directivity should not exceed 1.6.) Equation 1 indicates that a circuit that is incapable of supplying more than 1 nW of power at 80 MHz will not be capable of being the sole source of a radiated emissions failure at this frequency. For analog circuits, the maximum power supplied by the source can be readily determined, and is probably already known. For digital circuits, the maximum power at any given frequency is probably not part of any functional specification. Designing for guaranteed performance requires that the maximum power these circuits are capable of generating is known and/or controlled.
Step 2: Control Risetimes
Any digital circuit source that cannot deliver at least 1 nW of power at 80 MHz cannot be the sole source of radiated emissions that exceed the specification in the example above. This is true no matter how poorly the circuit is routed and no matter what other circuits or structures the signal may be coupled to. A primary objective in designing for guaranteed compliance is to ensure that no source produces power capable of exceeding the specification unless that power is necessary for the circuit to serve its intended function.
For every digital circuit in the system that is operating at clock speeds of 16 MHz or less, the best way to control the emissions at 80 MHz and higher is to control the transition times. Slowing transition times to 20% of the bit width significantly reduces the energy in the signal at frequencies above the fifth harmonic. For example, suppose our signal source is an automotive microcontroller output with a 3.3-volt signal amplitude and a 165-ohm source resistance (i.e. 20 mA max current). If the operating frequency is 100 kHz and we’re driving a 5-pF load, the bit width is 10 μsec and the transition time is 1.8 ns. The signal voltage at each odd nth harmonic of the clock frequency (assuming a trapezoidal waveform) would be given by:
where A is the peak-to-peak signal voltage, tr is the signal transition time, and T is the signal period. Now let’s assume that all of the power produced by the source is radiated (i.e. the source drives a lossless matched antenna).
Figure 1 shows a plot of the radiated emissions compared to an FCC radiated emissions test limit. Note that this component is capable of producing radiated emissions about 15 dB above the limit at 80 MHz.
Figure 1: Maximum possible radiated emissions from a 100-kbps, 3.3-volt, 20-mA source with a 1.8-ns transition time
Now suppose we insert a 20-kΩ resistor in series with the output of this controller. For the same 5-pF load, the transition time would now be increased to 220 ns. Since the transition time is about 2% of the bit width, the integrity of the digital signal is preserved. Figure 2 shows the radiated emission that would result if all of the power produced by this source was radiated. Note that this circuit is incapable of producing enough power to exceed the FCC radiated emission specification at any frequency above 30 MHz. From a design standpoint, this means that this signal can be routed anywhere on the board or on any attached wiring harness. No matter how poorly the routing is done, or how many other circuits this signal couples to, it cannot be the source of a radiated emission failure.
Figure 2: Maximum possible radiated emissions from a 100-kbps, 3.3-volt, 20-mA source with a 220-ns transition time
For the purposes of meeting a radiated emissions specification, most of the digital circuits in most automotive ECUs can be eliminated from further consideration simply by controlling their transition times. Once the transition time has been slowed to between 10% and 20% of a bit width, only the faster circuits will remain viable sources of radiated emissions. Series resistors are one way to control transition times. The use of controlled slew-rate sources can also be effective.
Faster digital circuits cannot be eliminated as possible sources simply by limiting their transition times, but controlling these times reduces the amount of power available to be radiated. It also reduces the range of frequencies that must be considered. Figure 3 shows the maximum possible radiated emissions from the same automotive controller in the previous example when it is operating at 1 MHz. Note that it is capable of exceeding the limit by about 35 dB at 80 MHz, and is not guaranteed to meet the requirement at any frequency below about 600 MHz. Slowing the transition time down to 90 ns using an 8-kΩ resistor results in the maximum emissions plotted in Figure 4. Although the circuit is still not guaranteed to be okay at 80 MHz, it is within 5 dB of the limit, and the controlled transition time guarantees that the circuit will not exceed the limit at frequencies above 100 MHz.
Figure 3: Maximum possible radiated emissions from a 1-Mbps, 3.3-volt, 20-mA source with a 1.8-ns transition time
Figure 4: Maximum possible radiated emissions from a 1-Mbps, 3.3-volt, 20-mA source with a 90-ns transition time
Step 3: Control Current Paths and Coupling
After controlling all transition times, the number of circuits capable of being a radiated emissions source in a typical automotive ECU will be relatively small. It should be noted that circuits capable of radiating a few dB above the limit are unlikely to do so. The calculations in the previous section assumed that the source was connected to a lossless impedance-matched antenna. This is unlikely to happen in a real device, so the actual emissions will generally be lower than the maximum possible emissions. Nevertheless, each circuit capable of radiating above the specification limit should be examined to ensure that most of the power being supplied at the offending frequencies reaches the load. The best way to do this is to trace the signal current paths from the source to the load and back again.
During the design review, it is important to recognize any structures or coupling paths that could steal power from the intended signal path and direct it elsewhere. Traces over a solid ground plane (microstrip traces) are relatively easy to examine, because they present fewer opportunities for unwanted coupling. Levels of crosstalk between adjacent microstrip traces can easily be kept below −20 dB simply by following general design guidelines for trace routing. If your maximum coupling is −20 dB, then the maximum radiated power will be 20 dB lower than the maximum available power calculated in the previous section.
Step 4: Recognize and Control Coupling to Antennas
In many automotive ECUs, applying Steps 1-3 above will eliminate the vast majority of the circuits from consideration as possible sources of a radiated emissions problem. At this point, attention can be focused on the few remaining circuits/signals capable of being a problem source. By recognizing that all radiated emissions problems involve both a source and an antenna, and further recognizing that any antenna that is small relative to a wavelength will not be efficient, it is possible to examine every possible antenna structure in a system and evaluate the maximum possible coupling between every possible source and every possible antenna. Typically, in an automotive ECU the possible antennas are limited to one or more cable harnesses, one or more large heatsinks, and perhaps a metal enclosure or connection to the vehicle chassis. Evaluating 3 to 5 possible antennas against 1 to 10 possible circuit sources is a very manageable task.
While it is possible to gain some insight relative to the maximum possible coupling between a circuit and antenna structure using electromagnetic modeling tools, a more effective approach is to apply closed-form equations designed to calculate the maximum possible emissions from specific source-structure geometries typically found in electronic systems. Examples of these closed-form equations can be found in [5-14]. Several of these are implemented in a web-based calculator on the Clemson Vehicular Electronics Laboratory website . Maximum emissions calculations basically assume that everything not known about a particular source/antenna geometry is worst case. For example, if a source drives a cable of unknown length, the calculator assumes that the cable is a lossless, resonant antenna at every frequency of interest. If a transition time is unknown, it is assumed to be zero. While these calculations will generally overestimate the actual measured radiated emissions, they provide an upper bound. For EMC work, an upper bound is generally preferable to an exact calculation of emissions from a specific configuration. Designing a system to be compliant with a radiated emissions specification based on maximum emissions calculations ensures that the specification will be met regardless of where it is tested or what it is connected to.
While the procedure outlined above is relatively time consuming, it is not as difficult as it might first appear. It is true that every circuit on the board must be evaluated, and a board with average complexity might have hundreds of circuits. However with a little experience, a designer will learn to recognize many of the circuits that are incapable of being the source of a problem without explicitly doing a calculation. Also, simple alterations to many types of circuits that change them from being possible sources to negligible contributors will become evident to the designer over time. This will make it easier to adjust the design rather than calculate the possible emissions. Since many of the calculations required for this procedure can be automated, these calculations can eventually be integrated with existing computer-aided design tools. This will further reduce the burden on the design engineer and reduce the level of EMC design experience required to implement the procedure.
It is tempting to think that an electronic system designed for guaranteed compliance will be over-designed and expensive. However, most of the EMC design decisions made early in a product’s development cycle have little or no impact on the cost of the product. The costs associated with building a compliant product based on the procedure outlined above are often limited to a few cents worth of components. On the other hand, the costs associated with fixing an EMC problem in a product that has already been prototyped can be relatively high. Routing and placement options that would have been easy to implement early, may no longer be available. Fixing problems that are discovered when testing a prototype is much more likely to involve relatively expensive filters or shields and can cause delays to the product development schedule.
The previous section describes a design procedure for guaranteeing compliance with radiated emissions requirements. Similar procedures can be followed to ensure compliance with other EMC requirements, although it is important to recognize that “radiated” automotive EMC test procedures are often measuring near-field coupling over much of their designated frequency range. Also “conducted” emissions tests measure both conducted and field-coupled emissions. Procedures for guaranteeing compliance with susceptibility tests can guarantee a maximum coupled signal, but they can’t guarantee that every part of the system will tolerate that signal unless that has been previously determined for each component.
There are many advantages to a design for guaranteed compliance approach. Although the technique requires more time and effort than a design-rule based approach, it is much more likely to result in a low-cost, compliant design on the first pass. The design for guaranteed compliance approach also highlights the most vulnerable aspects of a design, making it easier to diagnose any problems that may occur in the field. On the other hand, the rigorous, but arbitrary enforcement of EMC design rules can add cost and complexity to a design without necessarily improving its EMC performance. In fact, the misapplication of a well-intended design rule can significantly degrade EMC performance .
Another advantage of the design for guaranteed compliance approach is that it does not rely on testing to establish the electromagnetic compatibility of the design. Testing alone can never be thorough enough to anticipate all the possible problems/interactions that can arise in modern automotive systems. With the design for guaranteed compliance approach, the compatibility is designed in and the testing simply confirms that the design performs as expected. Although there is much more work to be done before this approach reaches its full potential, it is likely that the design-for-compliance-and-test-for-validation concept will ultimately define the methods used to ensure that increasingly complex automotive electronic systems meet all EMC performance requirements.
- 2011 Safety Numbers Show Decline In Traffic Fatalities—But DOT’s Work Continues, Fast Lane: The Official Blog of the U.S. Secretary of Transportation, http://fastlane.dot.gov, Dec. 10, 2012.
- Automotive Electronic Systems website, Clemson Vehicular Electronics Laboratory, http://www.cvel.clemson.edu/auto/systems/auto-systems.html.
- Tae-Weon Kang and Hyo-Tae Kim, “Reproducibility and Uncertainty in Radiated Emission Measurements at Open Area Test Sites and in Semianechoic Chambers Using a Spherical Dipole Radiator,” IEEE Trans. on EMC, vol. 43, no. 4, Nov. 2001, pp. 677 – 685.
- A. Kriz and W. Muellner, “Analysis of the CISPR 25 Component Test Setup,” Proc. Of the 2003 IEEE Int. Symp. on EMC, May 2003, pp. 229-232.
- Y. Fu and T. Hubing, “Analysis of radiated emissions from a printed circuit board using expert system algorithms,” IEEE Trans. on Electromagnetic Compatibility, vol. 49, no. 1, Feb. 2007, pp. 68-75.
- X. He and T. Hubing, “A closed-form expression for estimating the maximum radiated emissions from a heatsink on a printed circuit board,” IEEE Trans. on Electromagnetic Compatibility, vol. 54, no. 1, Feb. 2012, pp. 205-211.
- C. Su and T. Hubing, “Calculating radiated emissions due to I/O line coupling on printed circuit boards using the imbalance difference method,” IEEE Trans. on Electromagnetic Compatibility, vol. 54, no. 1, Feb. 2012, pp. 212-217.
- C. Su and T. Hubing, “Improvements to a method for estimating the maximum radiated emissions from PCBs with cables,” IEEE Trans. on Electromagnetic Compatibility, vol. 53, no. 4,
- Nov. 2011, pp. 1087-1091.
- X. Dong, H. Weng, D. G. Beetner, T. Hubing, “Approximation of worst-case crosstalk at high frequencies,” IEEE Trans. on Electromagnetic Compatibility, vol. 53, no. 1, Feb. 2011, pp. 202-208.
- H. Zeng, H. Ke, G. Burbui and T. Hubing, “Determining the maximum allowable power bus voltage to ensure compliance with a given radiated emissions specification,” IEEE Trans. on Electromagnetic Compatibility, vol. 51, no. 3, Aug. 2009, pp. 868-872.
- S. Deng, T. Hubing and D. 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.
- H. Shim and T. Hubing, “A closed-form expression for estimating radiated emissions from the power planes in a populated printed circuit board,” IEEE Transactions on Electromagnetic Compatibility, vol. 48, no. 1, Feb. 2006, pp. 74-81.
- H. Shim and T. Hubing, “Model for estimating radiated emissions from a printed circuit board with attached cables due to voltage-driven sources,” IEEE Transactions on Electromagnetic Compatibility, vol. 47, no. 4, Nov. 2005, pp. 899-907.
- Maximum Radiated Emissions Calculator (MREMC), Clemson Vehicular Electronics Laboratory website, http://www.clemson.edu/ces/cvel/modeling/EMAG/MaxEMCalc.html, March 2013.
- EMC Design Guideline Collection, LearnEMC website, http://www.learnemc.com/tutorials/guidelines.html.
|Dr. Todd Hubing
is the Michelin Professor of Vehicle Electronic Systems Integration at Clemson University and Director of the Clemson Vehicular Electronics Laboratory (CVEL). His research focuses on the design of electronic components and systems that work safely and reliably in automotive environments. Dr. Hubing holds a BSEE degree from MIT, an MSEE degree from Purdue University and a Ph.D. from North Carolina State University. He was an engineer with IBM for 7 years and a faculty member at the University of Missouri-Rolla for 17 years before joining Clemson University in 2006. At Clemson, he teaches classes in automotive electronics integration to graduate students in the automotive engineering department. He also teaches classes in vehicle electronics, electromagnetic compatibility and digital signal integrity to graduates and undergraduates in the electrical engineering department. His company, LearnEMC, provides EMC instruction and design assistance to practicing engineers. He is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), a Fellow of the Applied Computational Electromagnetics Society, and a Past-President of the IEEE Electromagnetic Compatibility Society.