25.3.1 Static Voltage Ranges

In most present-day automobiles, a Lundell-type alternator provides dc electrical power with a lead-acid battery for energy storage and buffering. The nominal battery voltage is 12.6 V, which the alternator regulates to 14.2 V when the engine is on in order to maintain a high state of charge on the battery. In practice, the regulation voltage is adjusted for temperature to match the battery characteristics. For example, in [1], a 25°C regulation voltage of 14.5V is specified with a − 10 mV/°C adjustment. Under normal operating conditions, the bus voltage will be maintained in the range of 11–16V [3]. Safety-critical equipment is typically expected to be operable even under battery discharge down to 9 V, and equipment operating during starting may see a bus voltage as low as 4.5–6 V under certain conditions.

In addition to the normal operating voltage range, a wider range of conditions is sometimes considered in the design of automotive electronics [3]. One possible condition is reverse-polarity battery installation, resulting in a bus voltage of approximately –12 V Another static overvoltage condition can occur during jump starting from a 24-V system such as on a tow truck. Other static overvoltage conditions can occur due to failure of the alternator voltage regulator. This can result in a bus voltage as high as 18 V, followed by battery electrolyte boil-off and a subsequent unregulated bus voltage as high as 130 V Typically, it is not practical to design the electronics for operation under such an extreme fault condition, but it should be noted that such conditions can occur. Table 25.1 summarizes the range of static voltages that can be expected in the automotive electrical system.

TABLE 25.1 Static voltage range for the automotive electrical system [3]

25.3.2 Transients and Electromagnetic Immunity

Power electronic circuits designed for automotive applications must exhibit electromagnetic compatibility, i.e. the conducted and radiated emissions generated by the circuit must not interfere with other equipment on board the vehicle, and the circuit must exhibit immunity to radiated and conducted disturbances. The Society of Automotive Engineers (SAE) has laid out standards and recommended practices for the electromagnetic compatibility of automotive electronics in a set of technical reports [4]. These reports are listed in Table 25.2. Here we will focus on two of the basic requirements of automotive power electronics: immunity to power lead transients and limitation of conducted emissions.

TABLE 25.2 SAE J1113 electromagnetic compatibility technical reports

A major consideration in the design of an automotive power electronic system is its immunity to the transients that can appear on its power leads. A number of transient sources exist in the vehicle [5] and procedures for validating immunity to these transients have been established in documents such as SAE J1113/11 [4, 6] and DIN 40389 [1]. Table 25.3 illustrates the transient test pulses specified in SAE J1113/11. Each test pulse corresponds to a different type of transient. The vehicle manufacturer determines which test pulses apply to a specific device.

TABLE 25.3 Transient pulse waveforms specified in SAE J1113/11

Transients occur when inductive loads such as solenoids, motors, and clutches are turned on and off. The transients can be especially severe when the bus is disconnected from the battery, as is the case for the accessory loads when the ignition is switched off. Test pulse 1 in Table 25.3 simulates the transient generated when an inductive load is disconnected from the battery and the device under test remains in parallel with it. When the inductive load is a dc motor, it may briefly act as a generator after disconnection. This transient is simulated by test pulse 2b. Test pulse 2a models the transient when current in an inductive element in series with the device under test is interrupted. Test pulses 3a and 3b model switching spikes that appear on the bus during normal operation. Test pulse 4 models the voltage transient that occurs on starting.

Perhaps the best-known electrical disturbance is the so-called load dump transient that occurs when the alternator load current drops sharply and the battery is unable to properly buffer the change. This can occur when the battery becomes disconnected while drawing a large amount of current. To understand why a major transient can occur under this situation, consider that the Lundell-alternator has a very large leakage reactance. The high commutating reactance interacting with the diode rectifier results in a high degree of load regulation, necessitating the use of a large back emf to source rated current at high speed [7]. Back voltages as high as 120 V may be needed to generate rated current into a 14 V output at top speed. Analytical modeling of such systems is addressed in [8]. Two effects occur when the load on the alternator suddenly steps down. First, as the machine current drops, the energy in the alternator leakage reactances is immediately delivered to the alternator output, causing a voltage spike. The peak voltage reached depends on the electrical system impedance, and may be limited by suppression devices. Second, once the alternator current is reduced, the voltage drops across the leakage (com-mutating) reactances are reduced, and a much larger fraction of the machine back-emf is impressed across the dc output. The proper output voltage is only re-established as the voltage regulator reduces the field current appropriately. With conventional regulator circuits, this takes place on the time scale of the field winding time constant (typically 100 ms), and results in a major transient event. In systems without centralized protection, a load dump can generate a transient with a peak voltage in excess of 100 V lasting hundreds of milliseconds. Test pulse 5 in Table 25.3 (expressed as a current waveform in parallel with an output resistance) is designed to simulate such a load-dump transient; other load-dump tests are even more severe [1, 3].

25.3.3 Electromagnetic Interference

Strict limits also exist for the amount of electromagnetic interference (EMI) that an automotive electronic component can generate. Limits for both conducted and radiated emissions are specified in SAE standards J1113/41 and J1113/42 [4, 9, 10]. Here we will consider the conducted EMI specifications for power leads, since they directly impact the design of EMI filters for automotive power electronics. Meeting the conducted specifications is a major step towards achieving overall compliance.

The conducted EMI specifications in SAE J1113/41 limit the ripple that an electronic circuit can inject onto the voltage bus over the frequency range from 150 KHz to 108 MHz. The amount of ripple injected by a circuit usually depends on the bus impedance. To eliminate any variability due to this, EMI compliance testing is done using a line impedance stabilization network (LISN) between the bus and the device under test, as illustrated in Fig. 25.2. The LISN is also sometimes referred to as an artificial mains network (AN). Essentially, the LISN ensures that the equipment under test receives the proper dc voltage and current levels and also sees a controlled impedance for the ripple frequencies of interest. Figure 25.3 shows the magnitude of the LISN output impedance for a low-impedance input source; the effective impedance is 50 Ω over most of the frequency range of interest. The 50-Ω termination impedance of the LISN is typically provided by the measurement equipment. The EMI specifications are stated in terms of the allowable voltage ripple (in dB μV) appearing across the 50-Ω LISN resistance as a function of frequency.

FIGURE 25.2 Conducted EMI test set up with LISN. L_LISN = 5 μH, C_LISN = 0.1 μ-F, and R_LISN = 50 Ω.

FIGURE 25.3 The LISN output impedance magnitude for a low impedance input source.

There are a wide range of other technical considerations for EMI testing, including the arrangement of the equipment over a ground plane and the types and settings of the measuring devices. One characteristic to consider is that the EMI measurements are done across frequency with a spectrum analyzer having a prespecified receiver bandwidth (RBW). For frequencies between 150 kHz and 30 MHz, the receiver bandwidth is 9 kHz, resulting in spectral components within 9 kHz of one another being lumped together for purposes of the test. A full test procedure is defined in the SAE specifications, beginning with narrowband measurements and moving to wideband measurements if necessary. Figure 25.4 illustrates the narrowband conducted EMI limits for power leads in SAE 11113/41. It is interesting to note that for the commonly used Class 5 limits, the allowable ripple current into the LISN at 150 kHz is less than 100 μA!

FIGURE 25.4 SAE J1113/41 narrowband conducted EMI limits for power leads. The specification covers the frequency range from 150 kHz to 108 MHz.

As seen in the previous section, the transient disturbances generated by electrical and electronic equipment are an important consideration in automotive applications. Because power electronic circuits typically contain switches and magnetic elements, they are potential sources for such transients, especially when powered from the switched ignition line. SAE J1113/42 specifies methods for testing and evaluating the transients generated by automotive electrical components, and proposes transient waveform limits for different severity levels. The equipment under test is set up in a configuration similar to that in Fig. 25.2, but with a switching device on one side or the other of the LISN, depending on the application. The equipment under test is then evaluated for transient behavior at turn on, turn off, and across its operating range. The voltage transients at the input of the equipment are measured and evaluated with respect to magnitude, duration, and rise and fall times. Specific limits for such transients are specified by the vehicle manufacturer, but SAE J1113/42 proposes a representative set of limits for four different transient severity levels.

Due to the tight conducted emissions limits, input EMI filter design is an important consideration in automotive power electronics. Single or multistage low-pass filters are typically used to attenuate converter ripple to acceptable levels [11–13]. When designing such filters, the parasitic behavior of the filter components, such as capacitor equivalent series resistance and inductance, and suitable filter damping are important considerations [14]. One must also ensure that the filter design yields acceptable transients at switch on and off, and does not result in undesired dynamic interactions with the power circuit [13]. Attention to appropriate filter design, coupled with proper circuit layout, grounding, and shielding goes a long way towards meeting electromagnetic interference specifications [14].

25.3.4 Environmental Considerations

The automobile is a very challenging environment for electronics. Environmental factors influencing the design of automotive electronics include temperature, humidity, mechanical shock, vibration, immersion, salt spray, and exposure to sand, gravel, oil, and other chemicals. In 1978, the SAE developed a recommended practice for electronic equipment design to address these environmental considerations [3, 4]. This document, SAE J1211, provides quantitative information about the automotive environment to aid the designer in developing environmental design goals for electronic equipment. Here, we briefly summarize a few of the most important factors affecting the design of power electronics for automotive applications. For more detailed guidelines, the reader is referred to [3] and the documents cited therein.

Perhaps the most challenging environmental characteristic is the extreme range of temperatures that can occur in the automobile. Table 25.4 summarizes some of the temperature extremes listed in SAE J1211 for different locations in the automobile. Ambient temperatures as low as –40°C maybe found during operation, and storage temperatures as low as –50°C may be found for components shipped in unheated aircraft. Maximum ambient temperatures vary widely depending on vehicle location, even for small differences in position. Because ambient temperature has a strong impact on the design of a power electronic system it is important to work closely with the vehicle manufacturer to establish temperature specifications for a particular application. For equipment that is air-cooled, one must also consider that the equipment maybe operated at altitudes up to 12,000 feet above sea level. This results in low ambient pressure (down to 9 psia), which can reduce the heat transfer efficiency [3]. For equipment utilizing the radiator-cooling loop, maximum coolant temperatures in the range of 105–120°C at a pressure of 1.4 bar are possible [15].

TABLE 25.4 Automotive temperature extremes by location [3]

Vehicle location	Min temp. (°C)	Max temp. (°C)
Exterior	− 40	85
Chassis
Isolated	− 40	85
Near heat source	− 40	121
Drive train high temperature location	− 40	177
Interior
Floor	− 40	85
Rear deck	− 40	104
Instrument panel	− 40	85
Instrument panel top	− 40	177
Trunk	− 40	85
Under hood
Near radiator support structure	− 40	100
Intake manifold	− 40	121
Near alternator	− 40	131
Exhaust manifold	− 40	649
Dash panel (normal)	− 40	121
Dash panel (extreme)	− 40	141

In addition to the temperature extremes in the automobile, thermal cycling and shock are also important considerations due to their effect on component reliability. Thermal cycling refers to the cumulative effects of many transitions between temperature extremes, while thermal shock refers to rapid transitions between temperature extremes, as may happen when a component operating at high temperature is suddenly cooled by water splash. The damaging effects of thermal cycling and shock include failures caused by thermal expansion mismatches between materials. Test methods have been developed which are designed to expose such weaknesses [3, 16]. The thermal environment in the automobile, including the temperature extremes, cycling, and shock, are challenging issues that must be addressed in the design of automotive power electronics.

A number of other important environmental factors exist in the automobile. Humidity levels as high as 98% at 38°C can exist in some areas of the automobile, and frost can occur in situations where the temperature drops rapidly. Salt atmosphere, spray, water splash, and immersion are also important factors for exterior, chassis, and underhood components. Failure mechanisms resulting from these factors include corrosion and circuit bridging. Dust, sand, and gravel bombardment can also be significant effects depending on equipment location. Mechanical vibration and shock are also important considerations in the design of automotive power electronic equipment. Details about the effects of these environmental factors, sample recorded data, and recommended test procedures can be found in [3].

25.4.1 High Intensity Discharge Lamps

High intensity discharge (HID) lamps have started to appear in automobiles as low-beam headlights and fog lights. The HID lamps offer higher luminous efficacy, higher reliability, longer life, and greater styling flexibility than the traditional halogen lamps [17, 18]. The luminous efficacy of an HID lamp is over three times that of a halogen lamp and its life is about 2000 hours, compared to 300–700 hours for a halogen lamp. Therefore, HID lamps provide substantially higher road illumination while consuming the same amount of electrical power and, in most cases, should last the life of the automobile. The HID lamps also produce a whiter light than halogen lamps since their color spectrum is closer to that of the sun.

High intensity discharge lamps do not have a filament. Instead, light is generated by discharging an arc through a pressurized mixture of mercury, xenon, and vaporized metal halides –mercury produces most of the light, the metal halides determine the color spectrum, and xenon helps reduce the start-up time of the lamp [17, 19]. Unlike halogen lamps that can be powered directly from the 12-V electrical system, HID lamps require power electronic ballasts for their operation. Initially, a high voltage pulse of 10–30 kV is needed to ignite the arc between the electrodes and a voltage of about 85 V is needed to sustain the arc [43]. Figure 25.5 shows a simplified power electronic circuit that can be used to start and drive an HID lamp. A step-up dc–dc converter is used to boost the voltage from 12 V to the voltage needed for the steady-state operation of the HID lamp. Any dc–dc converter that can step up the voltage, such as the boost or flyback converter, can be used for this application. An H-bridge is then used to create the ac voltage that drives the lamp in steady state. The circuit to initiate the arc can be as simple as a circuit that provides an inductive voltage kick, as shown in Fig. 25.5.

FIGURE 25.5 Simplified power electronic circuit for an HID lamp ballast.

25.4.2 Pulse-width Modulated Incandescent Lighting

Future automobiles may utilize a 42 V electrical system in place of today's 14 V electrical system (see Section 25.7). Because HID lamps are driven through a power electronic ballast, HID lighting systems operable from a 42 V bus can be easily developed. However, the high cost of HID lighting –as much as an order of magnitude more expensive than incandescent lighting –largely limits its usefulness to headlight applications. Incandescent lamps compatible with 42 V systems can also be implemented. However, because a much longer, thinner filament must be employed at the higher voltage, lamp lifetime suffers greatly. An alternative to this approach is to use pulse-width modulation to operate 12 V incandescent lamps from a 42 V bus [20].

In a pulse-width modulated (PWM) lighting system, a semiconductor switch is modulated to apply a periodic pulsed voltage to the lamp filament. Because of its resistive nature, the power delivered to the filament depends on the rms of the applied voltage waveform. The thermal mass of the system filters the power pulsations so that the filament temperature and light production are similar to that generated by a dc voltage with the same rms value. The PWM frequency is selected low enough to avoid lamp mechanical resonances and the need for EMI filtering, while being high enough to limit visible flicker; PWM frequencies in the range of 90–250 Hz are typical [20].

Ideally, a 11.1% duty ratio is needed to generate 14 V rms across a lamp from a 42 V nominal voltage source. In practice, deviations from this duty ratio are needed to adjust for input voltage variations and device drops. In some proposed systems, multiple lamps are operated within a single lighting module with phase staggered (interleaved) PWM waveforms to reduce the input rms current of the module.

Another issue with PWM lighting relates to startup. Even with operation from a 12 V dc source, incandescent lamps have an inrush current that is 6–8 times higher than the steady-state value, because of how filament resistance changes with temperature; this inrush impacts lamp durability. The additional increase in peak inrush current due to operating from a 42 V source can be sufficient to cause destruction of the filament, even when using conventional PWM soft-start techniques (a ramping up of duty ratio). Means for limiting the peak inrush current –such as operating the controlling MOS–FET in current limiting mode during startup –are needed to make practical use of PWM lighting control.

While PWM incandescent lighting technology is still in the early stages of development, it offers a number of promising advantages in future 42 V vehicles. These include low-cost adaptation of incandescent lighting to high-voltage systems, control of lighting intensity independent of bus voltage, the ability to implement multiple intensities, flashing, dimming, etc. through PWM control, and the potential improvement of lamp durability through more precise inrush and operating control [20].

25.4.3 Piezoelectric Ultrasonic Actuators

Piezoelectric ultrasonic motors are being considered as actuators for window lifts, seat positioning, and head restraints in automobiles [21, 22]. These motors work on the principle of converting piezoelectrically induced ultrasonic vibrations in an elastic body into unidirectional motion of a moving part. Unidirectional motion is achieved by allowing the vibrating body to make contact with the moving part only during a half-cycle of its oscillation, and power is transferred from the vibrating body to the moving part through frictional contact. Ultrasonic motors have a number of attractive features, including high-torque density, large holding torque even without input power, low speed without gears, quiet operation, no magnetic fields, and high dynamics [21, 23]. These characteristics make ultrasonic motors an attractive alternative to electromagnetic motors for low-power high-torque applications.

Various types of ultrasonic motors have been developed. However, because of its compact design, the traveling wave type is the most popular ultrasonic motor [24]. Figure 25.6a shows the basic structure of such a motor. It consists of a metal stator and rotor, which are pushed against each other by a spring. The rotor is coated with a special lining material to increase friction and reduce wear at the contacting surfaces. A layer of piezoelectric material, such as lead zirconate titanate (PZT), is bonded to the underside of the stator. Silver electrodes are printed on both sides of the piezoceramic ring. The top electrode is segmented and the piezoceramic is polarized as shown in Fig. 25.6b. The number of segments is twice the order of the excited vibration mode.

FIGURE 25.6 (a) Basic structure of a traveling wave piezoelectric ultrasonic motor and (b) structure of the piezoceramic ring and electrode for a four-wavelength motor. Arrows indicate direction of polarization. Dashed lines indicate segments etched in the electrode for poling but electrically connected during motor operation.

When a positive voltage is applied between terminals A and C, the downwards poled segment elongates and the upwards poled segments contract. This causes the stator to undulate, waving down at the elongated section and up at the contracted one. When the polarity of the voltage is inverted, the undulations are also inverted. Hence, when an ac voltage is applied a flexural standing wave is created in the stator. To get a large wave amplitude, the stator is driven at the resonance frequency of the flexural mode. An ac voltage between terminals B and C similarly produces another standing wave. However, because of the spacer segments in the piezoceramic ring, the second standing wave is 90° spatially out of phase from the first one. If the two standing waves are excited by ac voltages that are out of phase in time by 90°, a traveling wave is generated. As the traveling wave passes through a point along the neutral plane, that points simply exhibits axial (z-axis) motion. However, off-neutral plane points also have an azimuthal (ϕ-axis) component of motion. This azimuthal motion of the surface points propels the rotor. Ultrasonic motors require a power electronic drive. A power electronic circuit suitable for driving an ultrasonic motor is shown in Fig. 25.7. The two H-bridges are controlled to generate waveforms that are 90° out of phase with each other.

FIGURE 25.7 Drive circuit for an ultrasonic motor.

25.4.4 Electromechanical Engine Valves

Electromagnetic actuators are finding increasing application in automotive systems. These actuators are more desirable than the other types of actuators, such as the hydraulic and pneumatic actuators, because they can be more easily controlled by a microprocessor to provide more precise control. An application of electromagnetic actuators that is of particular interest is the replacement of the camshaft and tappet valve assembly by electromechanically driven engine valves [25]. The opening and closing of the intake and exhaust valves can be controlled to achieve optimum engine performance and improved fuel economy over a wide range of conditions determined by variables such as the speed, load, altitude, and temperature. The present cam system provides a valve profile that can give optimum engine performance and improved fuel economy only under certain conditions.

Two power electronic circuits suitable for driving the solenoids for valve actuation are shown in Fig. 25.8. The circuit of Fig. 25.8a is suitable for solenoids that require unidirectional currents through their coils, while the circuit of Fig. 25.8b is suitable for solenoids that require bidirectional currents through their coils.

FIGURE 25.8 Power electronic circuits for driving solenoids.

25.4.5 Electric Air Conditioner

It is desirable to replace some of the engine-driven functions of a vehicle with electrically driven counterparts. The benefits of driving these functions electrically include the elimination of belts and pulleys, improved design and control due to independence from engine speed, and resulting increased efficiency and improved fuel economy. Furthermore, there is the opportunity for operation of the function in the engine-off condition.

The air conditioner is an example of an engine-driven function that could benefit from electrification. The engine drives the compressor of the air conditioner. Consequently, the speed of the compressor varies over a wide range and the compressor has to be over-sized to provide the desired performance at engine idle. Also, since the compressor speed is dependent on the engine speed, excessive cooling occurs at highway speeds requiring the cool air to be blended with the hot air to keep the temperature at the desired level. Furthermore, shaft seals and rubber hoses can lead to the loss of refrigerant (CFC) and pose an environmental challenge.

In an electric air conditioner, an electric motor is used to drive the compressor [26]. The motor is usually a three-phase brushless dc motor driven by a three-phase MOSFET bridge. The speed of the compressor in an electric air conditioner is independent of the engine speed. As a result, the compressor does not have to be over-sized and excessive cooling does not occur. Also, shaft seals and hoses can be replaced with a hermetically sealed system. Another benefit of an electric air conditioner is the flexibility in its location, since it does not have to be driven by the engine.

25.4.6 Electric and Electrohydraulic Power Steering Systems

The hydraulic power steering system of a vehicle is another example of an engine-driven accessory. This system can be replaced with an electric power steering (EPS) system in which a brushless dc motor is used to provide the steering power assist [27]. The electric power steering system is more efficient than the hydraulic power steering system because, unlike the engine-driven hydraulic steering pump, which is driven by the engine all the time, the motor operates only on demand. Another system that can replace the hydraulic power steering system is the electrohydraulic power steering (EHPS) system. In this case, a brushless dc motor and inverter can be employed to drive the hydraulic steering pump. The ability of the EPHS system to drive the pump only on demand leads to energy savings of as much as 80% as compared with the conventional hydraulic system. Challenges in implementing EPS and EPHS systems include meeting the required levels of cost and reliability for this critical vehicle subsystem.

25.4.7 Motor Speed Control

Some of the motors used in a vehicle require variable speed control. Consider, as an example, the blower motor used to provide air flow to the passenger compartment. This motor is typically a permanent magnet dc motor with a squirrel-cage fan. The speed of the motor is usually controlled by varying the resistance connected in series with the motor winding. This method of speed control leads to a significant power loss. A low-loss method of speed control employs semiconductor devices as shown in Fig. 25.9. In this case, the speed of the motor is controlled via PWM –that is, by switching the MOSFET on and off with different duty-ratios for different speed settings. An input filter is needed to reduce the EMI generated by the switching of the MOSFET. This method of speed control is equivalent to supplying power to the motor through a variable-output dc-to-dc converter. The converter is located close to the motor and no filter is required between the converter output and motor winding.

FIGURE 25.9 Low-loss circuit to control the speed of a motor.

Another low-loss method that can be used to control the speed of a motor employs a three-phase brushless dc motor. The speed in this case is controlled by controlling the MOS-FETs in the dc-to-three-phase-ac converter that drives the motor.

25.6.1 The Lundell Alternator

The Lundell, or claw-pole, alternator is a three-phase wound-field synchronous machine that is almost universally used for power generation in present-day vehicles [1]. As illustrated in Fig. 25.11, the rotor is made of a pair of stamped pole pieces (“claw poles”) fixed around a cylindrical field coil. The field winding is driven from the stator via a pair of slip rings and brushes, and causes the two pole pieces to become opposing magnetic poles. A full-bridge diode rectifier is traditionally used at the machine output, and a fan mounted on the rotor is typically used to cool the whole assembly.

FIGURE 25.11 Structure and circuitry of the conventional Lundell alternator.

The dc output voltage of the alternator system is regulated by controlling the field current. A switching field regulator applies a pulse width modulated voltage across the field. The steady-state field current is determined by the field-winding resistance and the average voltage applied by the regulator. Changes in the field current occur with an L/R field-winding time constant in the order of 100 ms or more. This long field-winding time constant and a large stator leakage reactance are characteristic of this type of alternator and tend to dominate its performance.

The alternator is driven by means of a belt, and is designed to operate over a wide speed ratio of about 10:1, though much of its operating lifetime is spent within a narrower 3:1 or 4:1 range. The gearing ratio provided by the belt is a design variable for the alternator; an alternator mechanical speed range from 1800 to 18,000 rpm for a 12-pole machine is typical.

A simple electrical model for the Lundell alternator is shown in Fig. 25.12. The armature of the alternator is modeled as a Y-connected set of leakage inductances L_s and back voltages v_sa, v_sb, and v_sc. The fundamental electrical frequency ω of the back-emfs is one-half of the product of the number of machine poles p and the mechanical speed ω_m. The magnitude of the back-emfs is proportional to the electrical frequency and the field current. For the sinusoidal case, the line-to-neutral voltage back-emf magnitude can be calculated as:

(25.1)

where k is the machine constant and i_f is the field current. The diode bridge feeds a constant voltage V_o representing the battery and other loads. This simple model captures many of the important characteristics of the Lundell alternator, while remaining analytically tractable. Other effects, such as stator resistance, mutual coupling, magnetic saturation, and waveform harmonic content, can be incorporated into this model at the expense of simplicity. The constant-voltage battery load on the alternator makes the analysis of this system different from the classic case of a diode rectifier with a current-source load. Nevertheless, with reasonable approximations, the behavior of this system can be described analytically [8]. Using the results presented in [8], alternator output power vs operating point can be calculated as:

(25.2)

where V_o is the output voltage, V_s is the back-emf magnitude, ω is the electrical frequency, and L_s is the armature leakage inductance. Extensions of Eq. (25.2) that also include the effect of the stator resistance are given in [8].

FIGURE 25.12 A simple Lundell alternator model.

As can be inferred from Eq. (25.2), alternator output power varies with speed, and is maximized when the back-emf magnitude of the machine is substantially larger than the output voltage. In a typical Lundell alternator, back voltages in excess of 80 V may be necessary to source-rated current into a 14 V output at high speed. Furthermore, as can be seen from Eq. (25.2), the armature leakage reactance limits the output power capability of the alternator. These characteristics are a result of the fact that significant voltage drops occur across the leakage reactances when current is drawn from the machine. These drops increase with speed and current, and cause the alternator to exhibit significant drop in output voltage with increasing current. Thus, an appropriate dc-side model for a Lundell alternator is a large open-circuit voltage (related to the back-emf magnitude) in series with a large current- and speed-dependent output impedance. This characteristic, coupled with the long field time constant, is the source of the undesirable load-dump transient characteristic of the Lundell alternator. In this transient, the large open-circuit voltage is transiently impressed across the alternator output when the load is suddenly reduced.

The efficiency of the conventional Lundell alternator is relatively poor. Typical efficiency values are in the order of 40–60%, depending on the operating point [1, 36, 37]. At low and medium speeds, losses tend to be dominated by stator copper losses. Iron losses become dominant only at very high speeds [1].

25.6.2 Advanced Lundell Alternator Design Techniques

The conventional diode-rectified Lundell alternator, though inefficient, has so far met vehicle electrical power requirements in a cost-effective manner. However, continuing increase in electrical power demand and growing interest in improved fuel economy is pushing the limits of conventional Lundell alternator technology. This section describes some established and emerging technologies that can be used to improve the performance of the Lundell alternator.

25.6.3 Alternative Machines and Power Electronics

The demand for increased alternator power levels, efficiency and performance also motivates the consideration of alternative electrical machines, power electronics, and design approaches. While no alternative machine has yet displaced the Lundell alternator in production vehicles, primarily due to cost considerations, some potential candidates are reviewed in this section. These include machines that are mounted directly on the engine rather than driven from a belt. These direct-driven machines become important as power levels rise. This section also addresses the more general case of the combined starter/alternators. While the use of a single machine to do both starting and generation functions is clearly possible, a separate (transient-rated) dc machine is presently used for starting. This is because the large mismatch in starting and generating requirements has made the combined starter/alternator approach unattractive. However, as alternator power ratings increase, the mismatch is reduced, and a single starter/alternator system becomes more practical. A combined system has the potential to eliminate the need for a separate flywheel, starter, solenoid switch, and pinion engaging drive. It also has the potential to allow regenerative braking and “light hybrid” operation, and to provide idle-stop capability (i.e. the ability to turn off the engine when the vehicle is stopped and seamlessly restart when the vehicle needs to move) for reduced fuel consumption. A move to this more sophisticated approach relies upon advanced electrical machines and power electronics.

25.7.1 Trends Driving System Evolution

The conventional 12-V automotive electrical power system has many defects, including a widely varying steady-state system voltage and large transients, which force the electrical functions to be over-designed. However, these limitations alone have not been a strong enough driver for automotive companies to seriously evaluate advanced alternatives. Now a number of new factors are changing this situation. The most important of these are future load requirements that cannot be met by the present 12-V architecture.

25.7.2 Voltage Specifications

A major issue when implementing a high or dual voltage system is the nominal voltage of the high-voltage bus, and the operating limits of both buses. While there are many possibilities, there is a growing consensus in the automotive industry for a nominal voltage of 42 V for the high-voltage bus (corresponding to a 36 V lead-acid storage battery) [39, 56, 60]. This voltage is gaining acceptance because it is as high as possible while remaining within acceptable safety limits for open wiring systems (once headroom is added for transients) and it provides substantial benefits in the power semiconductors and wiring harness [61]. Furthermore, this voltage is sufficient to implement starter/alternator systems and “light” hybrid vehicle designs [62, 63]. While no vehicles equipped at 42 V are in production at present, availability of 42 V components is rapidly increasing and 42 V equipped vehicles may be expected early in this decade.

The permissible static and transient voltage ranges in an electrical system are important design considerations for power electronic equipment. At present, no universally accepted specification exists for high or dual voltage automotive electrical systems. However, the preliminary specifications proposed by the European automotive working group, Forum Bordnetz, are under wide consideration by the automotive industry [56]. These specifications, summarized in Table 25.7, impose tight static and transient limits on both the 42 and 14 V buses. The upper voltage limit on the 14 V bus is far lower than in the conventional 12-V system. The allowed upper limit on the 42 V bus is also proportionally tight. These strict limits facilitate the use of power semiconductor devices such as power MOSFETs and lower the cost of the protection circuitry needed in individual functions. However, they also require much more sophisticated means for limiting transients (such as load dump) than is found in conventional systems, which imposes a significant cost. Appropriate voltage range specifications for dual/high voltage electrical systems are thus a subject of ongoing investigation by vehicle manufacturers, and will likely continue to evolve for some time.

TABLE 25.7 Voltage limits for 14 and 42 V buses proposed in [56]

25.7.3 Dual-voltage Architectures

Conventional automotive electrical systems have a single alternator and battery. Dual-voltage electrical systems have two voltage buses and typically two batteries. Single-battery configurations are possible, but tend to be less cost effective [61]. A variety of different methods for generating and supplying energy to the two buses are under investigation in the automotive community. Many of these have power electronic circuits at their core. This section describes three dual-voltage electrical system architectures that have received broad attention. In all three cases the loads are assumed to be partitioned between the two buses with the starter and many of the other high-power loads on the 42 V bus and most of the lamps and electronics on the 14 V bus.

The dc/dc converter-based implementation of Fig. 25.19 is perhaps the most widely considered dual-voltage architecture. In this implementation, an alternator and associated battery provide energy to one bus (typically the 42 V bus), while the other bus is supplied via a dc/dc converter. If a battery is used at the dc/dc converter output, the converter needs to be rated for slightly above average power. Otherwise, the converter needs to be rated a factor of two to three higher to meet peak power requirements [61]. The architecture of Fig. 25.19 has a number of advantages. The dc/dc converter provides high-bandwidth control of energy flow between the two buses, thus enabling better transient control on the 14 V bus than is available in present-day systems or in most other dual-voltage architectures. Furthermore, in systems with batteries on both buses, the dc/dc converter can be used to implement an energy management system so that generated energy is always put to best use. If the converter is bidirectional it can even be used to recharge the high-voltage (starter) battery from the low-voltage battery, thus providing a self jump start capability. The major challenge presented by this architecture is the implementation of dc/dc converters having the proper functionality within the tight cost constraints dictated by the automotive industry. Some aspects of design and optimization of converters for this application are addressed in [64].

FIGURE 25.19 Dual-voltage architecture based on a dc/dc converter.

The dual-stator alternator architecture of Fig. 25.20 is also often considered for dual-voltage automotive electrical systems [65, 66]. In this case, an alternator with two armature windings is used along with two rectifiers to provide energy to the buses and their respective batteries. Control of the bus voltages is achieved via a combination of controlled rectification and field control. Typically, field control is used to regulate one output, while the other output is regulated using a controlled rectifier. Figure 25.21 shows one possible implementation of this architecture. It should be noted that to achieve sufficient output power and power steering from the dual-wound alternator, the winding ratio between the two outputs must be carefully selected. For 42/14V systems, a winding ratio of 2.5:1 is typical [66]. Advantages of this electrical architecture include low cost. However, it does not provide the bidirectional energy control that is possible in the dc/dc converter architecture. Furthermore, there are substantial issues of cross-regulation and transient control with this architecture that remain to be fully explored.

FIGURE 25.20 Dual-voltage architecture based on a dual-wound alternator.

FIGURE 25.21 Model for a dual-wound alternator system. The two output voltages are regulated through field control and phase control. For a 42/14V system, a winding ratio between the two stator windings of 2.5:1 is typical.

In a third architecture, a single-output alternator with a dual-output rectifier is employed. This approach is shown schematically in Fig. 25.22. As with the dual-stator alternator configuration, this architecture has the potential for low cost. One widely considered implementation of the dual-rectified alternator is shown in Fig. 25.23 [65, 67–69]. Despite its simplicity, this implementation approach provides less functionality than the dc/dc converter-based architecture, generates substantial low-frequency ripple which must be filtered, and has serious output power and control limitations [66]. An alternative implementation, proposed in [37] and shown in Fig. 25.24, seems to overcome these limitations, and may potentially provide the same capabilities as the dc/dc converter-based architecture at lower cost. Clearly, this architecture has promise for dual-voltage electrical systems, but remains to be fully explored.

FIGURE 25.22 Dual-voltage architecture based on a dual-rectified alternator.

FIGURE 25.23 A dual-rectified alternator with a phase-controlled rectifier.

FIGURE 25.24 A dual-rectified alternator with a switched-mode rectifier.