Automotive EMC: A moving target

With the electronic content of automobiles growing, so is the concern for electromagnetic compatibility (EMC) of automotive systems. The methods designers have been using, as well as the specifications, are being challenged by such changes in automotive technology as telematics, electric power, and the use of electronic control systems in place of mechanical ones. Research teams are developing new methods, tools, and standards for automotive EMC testing.

The growth of electronics in automobiles has been staggering. Many industry analysts expect that by 2010, nearly 30% of an automobile's cost will be its electronic component. Vehicles will contain noise-canceling entertainment systems, wireless voice and digital communications, active suspension, antilock brakes, and satellite navigation just for starters. Electronic locks, keyless entry, power windows and mirrors, and power seating will combine with intelligence that recognizes a driver from the entry code and adjusts the car to his or her personal settings. Complex wire harnesses will be replaced with serial communications buses and wireless signaling for sensors, and traditional mechanical systems such as steering and braking will become electrically controlled.

In the face of such a vast array of electronic functions, the need for automotive EMC test has become critical. In 1995, the European Union (EU) created directive 95/54/EC to specify the test methods and legislated limits for radiated emissions and immunity of both vehicles and electronic subassemblies (ESA) used in vehicles. No legislation defining such limits exists within the US, although the Society of Automotive Engineers (SAE; www.sae.org) has had standards since the early 1990s for testing, and manufacturers have developed their own internal performance standards in order to meet the requirements of their geographical markets.

Those specifications, however, have needed frequent revision, and new standards have emerged. In 1995, the International Electrotechnical Commission (IEC; www.iec.ch) released CISPR 12 (covering the limits and measurement methods for emissions) and CISPR 25 (covering immunity of vehicles and ESAs). In addition, the International Standards Organization (ISO; www.iso.ch) has revised and expanded its standards covering ESA immunity to transients (ISO 7637), electrostatic discharge (ISO 10605), and radiated energy (ISO 11452). As a result of these organizations' efforts, the CISPR and ISO standards have essentially replaced the SAE and manufacturer standards. (A summary of public standards and their applicability appears in Table 1.)

But the changes are far from over. EU directive 95/54/EC is undergoing revision, and the SAE has released a 2004 edition of its Surface Vehicle EMC Standards Manual, reflecting the addition of new RF sources in vehicles. The automotive industry is also looking at the impact alternative-power vehicles will have on EMC. Alternative power, such as batteries, fuel cells, and battery-gasoline hybrid engines, can significantly add to the EMC problems in automobiles.

Some of the recent investigations into such EMC problems were presented at the IEEE Symposium on Electromagnetic Compatibility held in August 2003 in Boston, MA. The presentations indicate that both standards and test procedures will need further change to properly characterize the behavior of electrically driven vehicles (Ref. 1). And one group of researchers found that new test instruments may also be needed.

In their paper entitled, "Investigation of Electromagnetic Emissions Measurements Practices for Alternative Powertrain Road Vehicles," Alastair Ruddle et al. note that standard broadband emissions tests call for the vehicle's engine to be disengaged, idling at a constant speed. Narrowband emissions tests require that the engine be off but the electrical power be on to test the ESA emissions.

With alternative-power vehicles, however, these conditions do not exercise the drive system. Electric motors draw current based on their load and draw no current when not engaged, so the standard test setup would not stimulate emissions from the motor or its power system. Further, with the motor off, some ESA components would not be active, so the standard test setup would miss any emissions from these modules.

When the authors tested a variety of electric vehicles under steady driving conditions, by running the vehicle at a constant speed on a dynamometer, the vehicles showed excessive emissions from 30 MHz to 127 MHz. These emissions would have been missed under the standard test conditions. The conclusion is that existing test conditions and methods will fail to adequately characterize alternative-power vehicle emissions. Automotive EMC test methods will have to change as alternative-power vehicles become more common. The nature of the change is uncertain, however, and Ruddle and his co-authors recommend testing alternative-power vehicles under load.

Chingchi Chen, of Ford Motor Co., offers a different approach in his paper, "Characterization of Power Electronics EMI Emission." Chen investigated the relationship between the switching pattern of a power system and its radiated emissions. Using a two-pulse test technique to reproduce switching transitions, he recorded system transient responses under different load conditions. He was then able to accurately predict the power system's emissions by modeling the pulse width modulation (PWM) patterns the system would generate under load.

Whether or not such modeling is acceptable as part of a specifications test, it is a useful design diagnostic tool. An ability to accurately predict the EMC behavior of system components helps designers implement corrective action early in the design phase. That was the goal of a European research project described in the paper "Continuous Simulation of System-Level Automotive EMC Problems" by a team of university and automotive researchers from Europe.

The researchers began their modeling process by using SPICE models to characterize the EMC behavior of the control module's processor chip. The characterization then moved to board-level and on to the vehicle level, as the team used SPICE to model the CAN bus as a transmission line network. Even the vehicle chassis itself was modeled using the CAD data for the car frame; a field solver was used to calculate scattering parameters and develop the SPICE model for bus-to-antenna coupling in the vehicle. With all the models in place, the team could simulate system-level EMC.

The researchers tested their models against measurements and found good agreement. This level of simulation will not replace final validation tests, but it does serve as an early detection system for potential EMC problems. It also implies that future automotive EMC control efforts could involve sharing of models from the chip level on up in the design process.

Along with new modeling schemes, test methods, and specifications, research into automotive EMC is leading to the development of new test tools. Researchers M. Kull et al. created a remote probe system with optical links, which they described in "A New Test System for Measurements of Fast Transients in Passenger Cars." The probe system was used to measure the spikes caused in cable harnesses by the switching of high-power devices such as active suspension systems and antilock brakes. The optical links to the probes were necessary to avoid introducing ground loops in probe shields when making simultaneous measurements at several points in a vehicle. In addition, the links protected the measurement from picking up signals during high-field-strength immunity tests.

This creation of new test instrumentation and methods demonstrates that automotive system EMC test is far from a settled discipline. With the continual introduction of new electronic systems, including electric drive, EMC specifications and standards are trying to hit a moving target. EMC test engineers can thus look forward to continual growth and change in automotive EMC test for years to come.