Functional power solutions for powertrain

For the PDF version of this article, click here.

Power semiconductors for today's automotive powertrain and engine control applications often include some level of signal-processing capability and intelligence while operating in energy absorption modes rarely experienced in other power designs. Thermal management issues arise as these devices are located on or near the engine or transmission to facilitate system integration and testing.

Partitioning of power and signal processing functions as well as packaging becomes critical to achieve optimal reliability and performance in powertrain applications. Through the use of state-of-the-art power silicon, high-performance BiCMOS control functions and new power packaging approaches, new higher- power lower-dissipation complex functional power solutions are being developed to address automotive design challenges.

PHYSICAL ENVIRONMENT

Power semiconductors used in automotive powertrain controls must endure harsh environments. Because the need to have fully tested systems, such as engines and transmission, requires the electronic control units to be mounted on or near the deliverable tested system, the thermal environment for these power semiconductors can reach ambient temperatures of 150 °C (Figure 1).

For a semiconductor, Tj max is the critical factor since blocking capability, gate threshold voltages, as well as other vital characteristics, are all bounded by this parameter. Exceeding Tj max is the cause of most failures. Couple this with the fact that in many automotive applications the power device must operate in energy absorption modes rarely experienced in other power designs, and it becomes clear that an understanding of thermal limits of power semiconductors and consideration of thermal management details is absolutely necessary for insuring that designs will continue to provide the reliability required by the automotive market.

ELECTRICAL ENVIRONMENT

The electrical environment for automotive applications is also quite different from that of most power systems. Unlike other transient environments where external influences have the greatest impact, the transient environment of the automobile is one of the best understood. The most severe transients result from either a load dump condition or a jumpstart over-voltage condition. Other transients may also result from relays and solenoids switching on and off and from fuses opening.

The circuit designer must ensure reliable circuit operation in this severe transient environment. The transients on the automobile power supply range from the severe, high-energy transients generated by the alternator/regulator system, to the low-level “noise” generated by the ignition system and various accessories. Figure 2 shows some of the voltage transients that must be considered in automotive systems design.

FUNCTIONAL POWER SOLUTIONS

Functional power monolithic technologies are progressing down the feature size path by adding more and more capabilities. They are reaching their limits in the basic power and analog requirements primary to any power application. Several of the advantages for a multiple die functional power solution include:

use of the lowest power loss switch for a given area;
use of the latest technology advances in power discrete technology;
improved isolation between the power and sensitive analog blocks;
improved modularity of the design by having one control die used with various power discrete devices;
ability to combine all types of discrete power technologies with high-performance analog control blocks; and
improved optimization of the switching function for on resistance or switching losses.

A functional power device is first and foremost a power device. It normally has limited data-processing responsibilities compared to other IC devices. Therefore, it is important to approach functional power from a power perspective. For a power device, the primary concerns are voltage-blocking capability, current- handling capability and thermal performance. Heat-related power losses in the device must be removed to keep the junction temperature below the point at which potential damage due to extremely localized heating can occur.

For devices that are primarily handling power with limited signal processing, the power process alone will often work. Figure 3 shows a simple drain-to-source voltage feedback signal and a gate drive disable control function implemented on a monolithic power MOSFET die. This type of architecture is often needed in automotive injectors and solenoid drivers. This example shows a device called the FDSS2407 by Fairchild Semiconductor, a 62 V, 132 mΩ, 5 V logic-level gate dual MOSFET in SO-8. This device has a 5 V logic-level feedback signal of the drain to source voltage. In this example, multiple devices can be wired “OR'd” to a single monitoring circuit input. The gate-disable function allows the FDSS2407 to be turned off independently of the drive signal on the gate. This function provides a second control circuit with the ability to deactivate the load if necessary. It can also be wired “OR'd” allowing multiple devices to be controlled by a single open collector/drain control transistor.

INTEGRATED POWER DEVICES

With functional power systems, there is a need to process both power and data. In some cases, the data-processing function is so complex that it is more cost-effective to use a silicon process optimized for signal processing for the smart functions of the device, and to use an entirely different silicon process optimized for power devices for the power functions of the device. To achieve this dual goal, the power and control functions could be placed in separate packages, but separate packages consume board area.

The need for smaller electronics requires integration of the separate, optimized silicon processes into a single, smaller package. These more compact devices must also provide the power handling, die interconnect, power and signal connections — and possibly die substrate isolation — along with physical support and environmental protection.

An example of this type of integration is the FDMS2380, also by Fairchild, shown in Figure 4. This device is a dual, intelligent low-side driver with a built-in recirculation and demagnetization circuits designed specifically for driving inductive loads. Its inputs are CMOS compatible. The diagnostic output on the device provides an indication of open load and demagnetization mode. Built-in over-current, over-voltage and over-temperature circuits protect the device and, in case of over-current or over-temperature, this product will automatically operate in freewheeling recirculation mode for inductive loads.

With a multichip functional power technology, excellent electrical isolation between power and control silicon is provided. The thermal communication between power and control silicon is dampened. This improves product ruggedness and reliability particularly in harsh electrical environments.

With a multichip smart power technology it makes sense to use the most efficient power silicon for the smart power functions that the footprint (board space) will allow. The lowest R_DS(ON) MOSFETs can be used as needed for the lowest power loss.

PACKAGING

Packaging has moved well beyond being a chip carrying and chip-to-board interface element and has become a powerful tool in solving problems associated with the automotive environment. For functional power devices, proper packaging allows integration of the power and control silicon while maintaining excellent isolation between power and control functions. The isolation improves product ruggedness and reliability.

Package technology provides functional power solutions allowing integration of optimized silicon processes in a single, small package while providing the electrical, thermal, and environmental performance required for automotive electronics. In addition to leadless QFN packaging, other package styles can be used for packaging multiple die. For many years, IGBTs and diodes have been packaged in standard discrete packages like TO-247, TO-263 and TO-220. Functional power devices typically require more than the standard three leads to communicate with the surrounding circuitry.

Other packages, including unique modules, can be designed to combine power and control functions. This results in better performance, easier protection, and higher reliability. Figure 5 shows an IPM intelligent power module designed for an EPS application that contains:

power device = IGBT with FRD (or MOSFET);
driver IC = HVIC & LVIC; and
package = transfer molding with ceramic or DBC.

Conventional power steering systems use a method of magnification to reduce the amount of effort used to turn the steering wheel. This method is operated with a hydraulic power steering pump system. Power steering with a fixed magnification ratio has major drawbacks. If the system is designed to reduce the force needed to turn the wheel when the vehicle is stationary, the steering can seem loose at greater speeds. A conventional power-steering system has a hydraulic pump that is mechanically connected to the engine and is running all the time. It uses energy even when the automobile is going in a straight line.

This constant load on the engine could be eliminated with full electronic power steering (EPS) systems, which place electric motors on the rack-and-pinion steering unit or, for small cars, on the steering column. These electric motors are efficiently controlled with low R_DS(ON) power MOSFETs employed in an IPM as previously described. These systems drain less power from the engine, thus providing more energy and fuel efficiency. In addition to less operational energy, these systems are between $5 to $10 less expensive than conventional hydraulic power steering systems. Full EPS systems are also lighter than conventional hydraulic power-steering systems. This reduced weight provides an additional contribution to fuel efficiency.

FLEXIBILITY

Most automotive designs are somewhat customized, yet automotive system designers are under constant pressure to reduce design cycle time. Traditional “smart power” monolithic technologies have complicated fabrication processes, which reduce the speed and flexibility of developing new devices. Since a monolithic development needs to make both power and signal processing on the same fabrication process, iterative designs can be costly and slow. With a multichip smart power technology the power and signal device development can occur in parallel. This enables faster development of new products optimized for a customer application. Often, an IC process is optimized for driver functions only (not power); and while the IC portion may stay the same, the power-handling requirements often vary. Use of the latest-generation power technologies can expand system life by migrating the power section to the latest power technologies.

PERFORMANCE

Lower R_DS(ON) products are needed for less power loss. Low R_DS(ON) means less voltage drop across the switch to measure at a given current. Any noise in the measurement method makes accurate measurement difficult. Reduction of noise and accurate sensing of load conditions are important considerations in automotive module design. Multi-die smart power devices allow for more accurate measurement of low voltage drops and low currents. With a multichip smart power technology excellent isolation between the power and the control silicon is provided. This isolation improves product ruggedness and reliability particularly in harsh electrical environments.

CONCLUSION

Semiconductors for automotive powertrains must operate in one of the harshest physical and electrical environments used in high volume production. Products are needed that provide less power loss, higher degrees of flexibility, and the ability to process small signals and large amounts of power. The solutions need to be cost effective. Through the combined use of innovative packaging and different semiconductor processes designed for specific functions like power or signal processing, new solutions will allow for less expensive more efficient and reliable power train systems.