Electromagnetic Compatibility of an Experimental Electric Vehicle
Developing electric vehicles presents vehicle OEMs and their electronic component and semiconductor suppliers with a variety of opportunities for vehicle designs. It also provides system designers with an entirely new set of technological hurdles. The most formidable challenge is the vehicle’s susceptibility to large levels of electromagnetic interference (EMI) produced by its own high-power switching electric motors.
These motors, connected to high-voltage, high-current battery packs, can have switching times as short as 100 ns. The electromagnetic fields generated by this high-current switching can affect any component in the vehicle—from the engine controller or the anti-lock braking system to optional features such as an AM radio or a CD player. For electric vehicles to become a practical reality, these EMI issues must be overcome.
Background
A cross-functional research team was established by the Motorola Semiconductor Products Sector to research emerging automotive technologies. One particularly promising technology is a serial multiplex communications network to replace complex and costly wiring harnesses in vehicles. A serial multiplex communications network is created by connecting each module or system that must collect, calculate, exchange or display data to a single serial communications bus.
To determine whether the serial communications technology currently available could operate in the EMI-intensive environment of an electric vehicle, a Society of Automotive Engineers (SAE) J1850 serial communications system was designed, installed and tested in Motorola’s electric vehicle research platform, a 1993 Dodge Dakota pickup truck converted to electric power. The network was simple enough to be monitored easily but complex enough to ensure that any EMI-related problems became evident quickly.
The SAE J1850 allows the use of a single-wire bus at 10.7 kb/s or a dual- wire bus at 41.7 kb/s. Since it is a balanced system, the dual-wire bus would have been more tolerant of common-mode EMI. However, the single-wire bus was chosen to prove that a multiplex system could function in the EMI-intensive environment.
Figure 1 shows a block diagram of the electric vehicle multiplex system. The DC-to-AC inverter uses insulated gate bipolar transistors (IGBTs) that are controlled by the centered pulse-width modulated vector-control technique.
The AC motor control node consists of an MC68HC705V8 connected via the serial peripheral interface port to a Motorola MC68332 32-bit microcontroller unit (MCU). The MC68332 performs the vector-control algorithms that drive the AC induction motor. Noise spikes are created as the IGBTs switch on and off at a rate of 8 kHz. These spikes are the by-product of the rising and falling edges of the IGBTs.
The switching edge time is determined by the IGBT characteristics. In this case, the turn-on time is 200 ns and the turn-off time is 500 ns. These spikes radiate from the input DC wires and the output AC wires, and then couple onto other wires and components in the system.
The radiated levels were so high that the AM radio in this vehicle never functioned properly. Also, the unshielded J1850 bus wire had common-mode spikes induced at levels up to 20 V.
Initially, the MCU did not function properly because of these high-level spikes. The network communications and MCU functioned properly in the presence of these spikes after using a combination of design alterations and signal processing by the MCU, digital filtering in the SAE J1850 communications modules and opto-isolation between the MCU and the sensing circuits.1
Class-B Data Communications Network Interface Bus
Voltage measurements were made on the multiplexed bus signal and ground lines at the point where they entered the sensor node. The measurements were performed with a digital storage oscilloscope (DSO) using various digital sampling rates to ascertain whether the DSO was being alliased by the voltage spikes. The DSO was powered through an isolation transformer to prevent ground loops.
The test vehicle was bonded to a shield-room floor with a ground strap attached to the point where the return of the accessory 12-VDC battery was grounded to the vehicle chassis. The high-voltage battery pack, consisting of 24 12-VDC car batteries wired in series, was isolated from the chassis ground to meet safety regulations. The test vehicle was placed on jack stands to simulate the vehicle traveling at 25 mph.
Figure 2 shows a plot of the multiplexed data signal with respect to the chassis ground (top trace). With no throttle, the signal appears clean. The middle trace shows the multiplexed signal ground measured with respect to the chassis ground. The bottom trace is the mathematical difference between the top and middle traces and it represents the differential signal between the data and its signal ground.
The DSO sample rate in Figure 3 was increased to 1 GS/s. The levels of the spikes were up to 20-V peak on the top and middle traces. These traces represented a common-mode coupling since each line was measured with respect to chassis ground. The bottom trace shows that the spikes were much lower when measured differentially; that is, about a 6-V peak.
The multiplexed data signal and signal ground were not twisted or shielded but they ran side by side between the various nodes. If the multiplexed cable was shielded and used a twisted pair, then both the common-mode and differential-mode spikes would be reduced.
Radiated Magnetic Field Emissions
The broadband radiated magnetic field emissions were measured in accordance with SAE J551/5 (Draft) Performance Levels and Methods of Measurement of Magnetic and Electric Field Strength from Electric Vehicles, Broadband, 9 kHz to 30 MHz.2 The test method involved placing the center of a 60-cm loop antenna a distance of one meter above the ground level and one meter away from the nearest part of the vehicle. All four sides of the vehicle were measured for magnetic-field emissions in all three orthogonal polarizations of the loop antenna with the car elevated on jack stands and the rear wheels spinning at 25 mph.
The maximum magnetic-field emissions emanated from the front of the vehicle. In accordance with SAE J551/5, the side of the vehicle with the highest levels also was tested at 10 mph and 40 mph to determine the speed that produced the maximum radiation. A speed of 10 mph provided the highest levels.
Radiated Electric-Field Emissions
The broadband radiated electric-field emissions also were measured in accordance with SAE J551/5. The test method involved placing the base of a one-meter monopole antenna at ground level and three meters away from the nearest part of the vehicle. The monopole base was grounded to the metal shield-room floor to form a counterpoise. All four sides of the vehicle were measured for electric-field emissions with the car elevated on jack stands and the rear wheels spinning at 25 mph.
The maximum electric-field emissions emanated from the passenger side of the vehicle. These levels are shown in Figure 4. In accordance with SAE J551/5, the vehicle side with the highest levels also was tested at 10 mph and 40 mph to determine the speed that produced the maximum radiation. A speed of 25 mph produced the highest levels.
The receiver bandwidths and scan times were the same as those used for the magnetic-field radiated emissions tests. The data and test limits in Figure 4 are shown in broadband units of dBµV/m/MHz. The values can be converted from dBµV/m/MHz to dBµV/m/kHz by subtracting 60 dB.
All of the calibrated electric-field measurements were made at three meters from the vehicle in accordance with SAE J551/5. Additional electric- and magnetic-field measurements were made with radiation hazard probes to determine location and approximate levels. These broadband probes measured average rather than peak field strength.
A reading of several volts/meter emanated from the vehicle’s radio antenna. The antenna apparently was acting as a re-radiator of electromagnetic energy. No wonder that the AM radio had never functioned properly. The highest readings were found in close proximity, such as about 20 cm, to the high-voltage battery pack where levels were measured up to 0.3 A/m and 300 V/m.
Conclusions
The Motorola electric vehicle platform had substantially high levels of radiated electric and magnetic fields. The levels were up to 34 dB above the SAE J551/5 limits. Also, conducted spike levels were measured up to 20-V peak on the SAE J1850 serial communications bus.
The electric vehicle platform was built without EMI control features since it was not a production vehicle and the purpose was to identify next-generation semiconductors. The intent of the multiplex project was to verify whether the serial communications technology currently available could operate in a worst-case EMI-intensive environment of an electric vehicle.
The answer is a qualified yes because isolation and signal filtering were necessary to ensure reliable operation of the MCUs and the multiplexed communications system. Other vehicle circuits and components such as the AM radio, however, require further reduction of the high-current/high-voltage switching noise of the AC inverter.
Indeed, if you installed EMI control features to reduce the radiated emissions by 34 dB, then the vehicle would comply with SAE J551/5. The conducted spike levels on the multiplex bus also would be reduced to 0.4-V peak or less.
As a result, compatible operation would be ensured in the vehicle, perhaps without the need for extensive signal filtering and isolation. Of course, the EMI control measures would increase the cost of the vehicle systems.
References
1. Powers, C. and Huettl, T., “Using Existing Multiplex Communication Technology to Implement an Electric Vehicle Communication Network,” SAE Future Transportation Technology Conference, Los Angeles, CA, Aug. 1995.
2. SAE J551/5 (Draft), Performance Levels and Methods of Measurement of Magnetic and Electric Field Strength from Electric Vehicles, Broadband, 9 kHz to 30 MHz, Feb. 2, 1995.
About the Authors
Harry Gaul, principle EMC engineer, is currently working on EMC compatibility of military products. He received his B.S.M.E. degree from the University of Colorado and has 18 years EMC experience for space and military systems. Motorola, Space and Systems Technology Group, 8201 E. McDowell Rd., Scottsdale, AZ 85252, (602) 441-5321.
Thomas Huettl, systems engineer, is a member of Motorola’s GENESIS advanced vehicular electronics team. He has 14 years experience with power electronics for industrial, computer and automotive systems. His current assignments include electric vehicle systems design and automotive multiplex networking. Mr. Huettl received a B.S.E.E. degree from Arizona State University. Motorola, Semiconductor Products Sector, 7755 S. Research Dr., Suite 110, Tempe, AZ 85284, (602) 755-2518.
Chuck Powers is a principle staff engineer at Motorola’s Microcontroller Technologies Group currently assigned to strategic system development. He has 10 years experience with microcontrollers and communications for industrial, computer and automotive systems. Mr. Powers received his B.S.E.E. degree from Texas Tech University. Motorola, Semiconductor Products Sector, 6501 William Cannon Dr. W., Austin, TX 78735, (512) 891-4594.
Copyright 1997 Nelson Publishing Inc.
February 1997