A Power Shortage Is Driving Automotive Applications To 42 V

Changes in the electrical system alter the landscape under the hood as well as foster the development of power semiconductors.

Looming under the hood, a power shortage is about to transform the automotive electrical system. The traditional 14-V power system is being taxed to its limits as car makers design-in a host of electronic features to improve the rider's comfort and convenience, while also boosting the car's performance in terms of handling, safety, and energy efficiency. Inside the cabin, sophisticated entertainment systems, climate controls, and other devices raise electrical power consumption. Meanwhile, the shift from mechanically controlled functions to electronic versions within the engine compartment poses an even greater threat to existing 14-V power budgets.

Currently, functions like antilock brakes and engine controls rely on power electronics tied into the car's electrical grid. Plus, drive-by-wire systems, which trade mechanical linkages for electronic controllers, sensors, and actuators, are now starting to implement power steering, braking, and throttle. Electronically controlled suspension systems that sense and adapt to drive conditions are in the works as well.

These new controls not only offer new ways of enhancing fuel efficiency and handling, but also eliminate some of the mechanical constraints of conventional car designs. Electronic actuators can be positioned more easily than mechanical assemblies that must be placed near the engine, where they derive power from the front-end accessory drive. Power-steering hydraulic pumps are an example of such assemblies.

With all of these systems coming on board, however, power demands are expected to rise sharply. Average power levels are presently around 1 kW with peak power demands in the vicinity of 2 kW. But, these levels are expected to rise significantly in the near future. Within five years, we may witness peak power levels of 12 kW.1 These levels are beyond what current 14-V systems can bear. Batteries and alternators won't be able to keep up. Even existing wiring will be inadequate as rising current levels produce unacceptable voltage drops.

The solution is to move to a higher system voltage, which boosts the power generating and handling ability of alternators, batteries, and cabling. That higher voltage is 42 V—a value selected by the MIT Consortium on Advanced Automotive Electrical/Electronic Systems and Components. Comprised primarily of automakers and their suppliers, this association is working to develop standards for a 42-V automotive electrical system. The consortium calls this system the 42-V PowerNet.

The consortium aims to define system requirements in a way that allows maximum flexibility. As Thomas Keim, director of the consortium, says, "There's no interest in standardizing the architecture or layout of components," because car manufacturers want to decide on this individually. Instead, the goal is to standardize critical system parameters, such as voltage range and battery termination. So, component vendors can develop a pool of components from which manufacturers will order.

Also, the consortium is addressing issues relating to safety. For instance, there must be some industry consensus on jump-starting requirements, particularly when the introduction of 42 V initially means a dual 14- and 42-V electrical system within cars.

Progress On A Voltage Range
So far, the greatest progress toward standardization has been in defining the voltage range. A German organization known as Forum Bordnetz, consisting of German and other European car makers and suppliers, has developed a specification that establishes undervoltage and overvoltage limits, maximum allowable ripple, etc. (see the table). The MIT Consortium is endorsing this specification. The voltage limits determined by this proposed standard will have a direct impact on the development of the power components for emerging automotive applications.

Although there's still discussion on various approaches, it appears, at least in the beginning, that the 42-V system will coexist with 14 V. This will be a dual-voltage system where high-power loads operate off of the higher voltage and medium- to low-power loads run off of the lower voltage. Among the architectures being discussed are dual- and single-battery systems. In both systems, power that's generated by a 42-V alternator is rectified and stored in a 36-V battery (42 V when fully charged). A dc-dc converter is then used to generate the 14-V power bus.

In a dual-battery system, a 12-V battery (14 V when fully charged) follows the dc-dc converter, providing energy storage for the 14-V loads. In the single-battery approach, the 14-V loads run off of the dc-dc converter. Each approach has pros and cons. With a dual-battery system, 14-V loads that are active when the engine is turned off won't drain the battery used to start the car. The dc-dc converter also can be sized smaller because the 14-V loads can rely on the battery to provide peak power.

On the other hand, the 12-V battery adds weight and cost to the vehicle. Also, it's possible that isolation of 14-V loads from the 36-V battery can be accomplished through power management circuits rather than a second battery. When the 12-V battery is eliminated, there's an additonal issue: whether to generate 14 V with a single, centralized dc-dc converter or multiple point-of-use converters configured in a distributed power architecture.

Another consideration is the changing of 14-V load requirements over time. The 14-V headlights may eventually be supplanted by 300-V high-intensity-discharge (HID) lamps, while other, lower intensity parking, braking, and signaling lamps may be replaced with LED arrays.

In addition, other controls may either be modified to operate off of 42 V or to run off dc-dc converters. The end result could be a changeover to a single 42-V bus that reduces cabling requirements. The interconnect problem isn't only an issue for power wiring. Improvements in control bus structures will be needed to reduce the size and cost of other cables and connectors that interface with all of the car's electronic and electromechanical functions.

Aside from concerns about 14-V power requirements, there's the pressing issue of how to maximize the benefits of 42-V power. A major advantage of 42 V is that it allows the alternator and starter motor to be merged into one machine commonly referred to as an integrated starter alternator (ISA). The ISA would take up residence in the drive train between the engine and transmission. One proposal places the ISA within a dual-battery, dual-voltage architecture (Fig. 1).

The ISA approach has many advantages, most of which lead to better fuel efficiency. It's likely, though, that implementation of the ISA will vary somewhat from market to market. These variations should reflect differences in priority when it comes to minimizing fuel economy.

Improving miles per liter is a top priority in Europe and Japan, thanks to legislation and economics, so we might expect that ISAs will help develop small cars with super efficiency. But, there's a strong demand in the U.S. for large luxury cars and SUVs. The ISA's first priority will be to generate the high levels of power demanded by these cars.

Nevertheless, U.S. car makers will still need to achieve some improvement in the fuel efficiency of these cars. This is because of the corporate average fuel economy (CAFE) standards. To meet these requirements, manufacturers can do one of two things. They may improve the fuel economy of their luxury vehicles, which is where they really make their profits. Or, they can just make smaller, more efficient cars.

These factors aside, the benefits of designing with an ISA should entice all. Using an ISA permits start-stop operation, whereby the engine is turned off each time the car comes to a stop. When driving is resumed, the electric motor is used to start the car moving while restarting the engine.

This balancing of electric and gas power makes this a "hybrid" approach that boosts fuel efficiency. In its operation, it makes a seamless transition from electric motor power to engine power. That leads to another benefit—the elimination of the "clunk" that typically accompanies the startup of a combustion engine. In addition, it's possible to use the ISA to recover braking energy and store it electrically in the battery. There also is a fuel savings associated with reducing the weight of the car by eliminating the starter motor as a separate unit.

To make the ISA work efficiently as both a power plant for the car and as a starter, an efficient power circuit is required. It must invert dc from the battery to generate the three-phase ac drive to the starter motor. Conversely, it must rectify the alternator's output to charge the 36-V battery and run the assortment of power-hungry 42-V and 14-V loads in the car. These functions can be realized with a power module built of a basic three-phase bridge. Although the architecture may be relatively simple and familiar, the actual task of designing such a module is no easy assignment.

According to Steve Clemente, vice president of the Electronic Motion Systems division at International Rectifier Corp., the toughest part of developing an ISA power module is dealing with the heat, which is a custom job. "From my perspective, the biggest problem is thermal management," he says. The inverter must be able to operate over an ambient temperature range of −40°C to 125°C (Fig. 2). At the same time, the power module has to be rated for peak power levels of 25 kW, although in normal operation it will be operating at around 5 or 6 kW. The packaging and cooling design will be critical.

To construct the power module, it's probable that multiple MOSFET dies will need to be paralleled in order to achieve the required power ratings. Consequently, some reliable form of current sharing must be used to prevent MOSFETs from burning up. All aspects of construction come under scrutiny, including the sizing of wirebonds to handle the high current levels, the selection of materials with properly matched thermal coefficients of expansion, and the design of fluid-cooled heatsinking.

In the end, to pass qualification, the power module must pass a rigorous test in which it spends hundreds of hours cycling between full-engine speed and full torque, then hundreds more hours running at full speed. Designers also have to worry about EMI. The switching frequency for the MOSFETs should be high enough to reduce switching losses, yet low enough to minimize EMI problems.

Despite the challenges, semiconductor and module vendors have strong motivation to develop products for the ISA application. Within a few years, the ISA could go into high-volume production. This will keep module builders increasingly busy. Semiconductor vendors will have their hands full, too.

For each power module built of a standard three-phase bridge, there will be six MOSFET switches. Each of these six switches may actually be comprised of four or more MOSFET dies configured in parallel to achieve the required current and power ratings of the module. As a result, the ISA application may tax fabrication capacity as much as it tests process technology.

Power MOSFETs
Any power semiconductors that operate off the 42-V supply voltage must be rated to withstand its worst-case overvoltage. According to the proposed 42-V PowerNet specification, the maximum transient overvoltage allowed on the bus is 58 V. In theory, a 60-V transistor could do the job, but vendors are generally developing more conservatively rated MOSFETs for this application. Typically the drain-to-source breakdown voltage rating is 75 V.

The higher rating affords a greater margin of safety for designers who must consider reliability as a critical concern. After all, as Clemente observes, the automotive industry is "risk aversive." Proof of this is the long qualification time for a new piece of electronics hardware—possibly two years or more. On the other hand, reducing the breakdown voltage rating lowers on-resistance for a given die size.

Vishay Siliconix is one of the semiconductor vendors developing components for the ISA power module application. Klaus Pietrczak, senior manager of automotive marketing, believes that over time, the 75-V rating will come down to about 65 V. That change could either lead to parts with lower RDS(ON) or to smaller, cheaper dies. (Unfortunately, though, as the die gets smaller, heatsinking becomes more difficult.) These gains will be coupled with the usual evolutionary improvements in cell density.

But for now, 75-V MOSFETs are the norm in the ISA application. For example, Vishay-Siliconix has developed a 75-V n-channel enhancement-mode MOSFET in 6- by 6-mm die form. Maximum RDS(ON) for this die (part number SUC85N08-04) in a TO-247 package is 4 mΩ at 25°C and VGS = 10 V. This part is fabricated in a trench-cell process which, according to the company, affords a much greater cell density than a conventional planar process. Moreover, the trench process is far from its limits in terms of achievable cell density. Much denser devices will be coming out of this technology, Pietrczak claims.

Other vendors, though, have taken the planar route in MOSFET development. International Rectifier chose a striped planar design to produce its 75-V MOSFET (part number IRFC2907). The size 6 die features an RDS(ON) of 2.7-mΩ maximum at 25°C and VGS = 10 V. In the TO-247 version, RDS(ON) is 4.5 mW max. Gordon Gray, technical marketing manager for discrete components, says that IR decided to produce this transistor in a striped planar process despite the ability that the trench process has to provide lower RDS(ON).

Gray notes that at present, striped planar is a more rugged process in regards to avalanche voltage. One of the benefits of the IRFC2907 is that it guarantees a repetitive avalanche rating up to TJ maximum. In the future, however, the company will likely transition 75-V parts to the trench process as its performance on avalanche rating is improved.

In producing its 75-V MOSFET for ISA, Infineon Technologies also went with a planar process on account of its ruggedness. There are no immediate plans to develop similar parts using a trench process. The company is currently developing a 75-V die produced in its OptiMOS planar process. The part's RDS(ON) is 2.8 mΩ at 25°C and VGS = 10 V. This product also will be offered in a D2PAK with 7.5-mΩ RDS(ON). Both parts are due to be released to production by the end of this calendar year.

Although the ISA is a primary application of 75-V MOS technology, the 42-V electrical system will afford many other opportunities for the development of power components. Infineon, for example, indicates that it's working on a triple-output power-supply chip that will convert the 42-V supply to 5, 3.3, and 2.5 V. At first the IC will step down 42 V to 7 V, and then use LDOs to generate the three clean outputs. In addition, the company has a number of 75-V high- and low-side protected switches in production. These can be driven directly by logic level inputs from a microcontroller for controlling electronic actuators.

The 75-V MOSFET technology developed for the ISA and other 42-V applications will be critical in making 42-V system power a reality. As the electronic content of the vehicle rises, these semiconductors will play a greater role in determining both the performance and price of the car. To start, this will be for luxury vehicles only, but eventually will extend to cars with more modest price tags. Down the road, these components will be at the forefront of efforts to develop even more advanced automotive technologies based on hybrid or fully electric designs.

References:

  1. "Automotive Electronics Power Up," John G. Kassakian, et al. IEEE Spectrum, May 2000.
  2. "Road Vehicles—Environmental Conditions for Electrical and Electronic Equipment for a 42-V PowerNet—Part 2: Electrical Loads." AG "Normung," Forum Bordnetzarchitektur, Stand. 31.01.2000. Preliminary, still in preparation. See www.sican.de.

For further information, check the MIT Consortium's web site at www.auto.mit.edu/consortium.

Manufacturers Providing Information For This Article
Infineon Technologies
(248) 374-2522
shawn.slusser@infineon.com
www.infineon.com

International Rectifier
(310) 252-7019
jengle1@irf.com
www.irf.com

On Semiconductor
(602) 244-5654
randy.frank@onsemi.com
www.onsemi.com

Sican GmbH
(facilitator for Forum Bordnetz)
hartman@sican.de
www.sican.de

Vishay Siliconix
(800) 554-5565, ext. 25
or (408) 567-8220
www.siliconix.com

Hide comments

Comments

  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Publish