Auto Electronics

Building ECUs for Electric Power-Assisted Steering Systems

Packaging considerations for building electronic control units (eCus) for driving electric motors used in electric power steering applications.

Electric power-assisted steering has well established itself in the automotive vehicle market. In principle there are two types: electro-hydraulic power steering (EHPS), where an electric motor drives a hydraulic pump (a similar pump is used in classical power steering) and electric power steering (EPS), where the electric motor directly assists the steering motion.

While current market research sees penetration of EHPS at around 8% and staying constant over the next few years, EPS is seen as growing very strongly. Currently, the market share for EPS is 25% and is expected to be equipped in every second vehicle sold in the next 10 years.

EHPS is the system with more components compared to EPS, since next to the electric motor, usually a brushless DC (BLDC), and the necessary electronic control unit (ECU) it still retains the hydraulic pump. The advantage of EHPS over conventional power steering is that the drive of the hydraulic pump can be controlled and becomes more energy efficient. The pump does not need to build up pressure if there is only a little or no steering required. This cannot be achieved with a conventional power steering system (or only with a lot of effort and cost), because the pump is driven directly via the combustion engine.

For an EPS there is no hydraulic system anymore. The electric motor is assisting the steering motion directly. The absence of the hydraulic system means there is less cost. In case of failure there is, however, an increase in risk by the electric motor acting on the steering column directly. This has led to a delay in introduction of these systems and has opened a market opportunity for EHPS, which has less risk associated with it in case of failure. However, through redundancies for critical sub-components (similar to the aircraft industry) and extensive validation testing, this risk has been reduced.

The next evolution for power steering will be steer-by-wire systems, where there is no mechanical link between steering wheel and steering axle anymore. The electric motor steers directly and the movement of the steering wheel by the driver is only detected by sensors and communicated to the motor.

An ECU unit, as shown in Figure 1, is needed to drive the electric motor. It consist of three principle functional blocks:

  1. The control module.
  2. The power module.
  3. The control software.

The control module drives the power module by turning the power switches on and off and performing certain diagnostic functions. The power module consists of three half-bridges. Each half-bridge powers one phase of the three-phase electric motor. The power switches are field-effect transistors (FETs). And the control software consists of algorithms to control speed and torque of the electric motor.


This module consists of a microprocessor that is mounted on a PCB along with peripheral components. The module requires several inputs e.g., signals from the angle sensors that determine the exact angular position of the electric motor's rotor (there is a sensorless drive also, where the angular position is inferred from the response to pulses superimposed on the phase signals). It monitors the temperature of the power module, protects against overvoltages and currents and carries out emergency strategies in case of failure.

The control module drives the power module via its output signals. The turn-on/turn-off sequences are dependent on the speed and torque requirements of the steering system.


This module delivers the power to the three-phase electric motor (mostly brushless DC). The power switches are driven typically by sinusoidal pulse width modulation that leads to a sinusoidal signal at the phase outputs of the power module.

The packaging and interconnection technology of the module has to be optimized in terms of thermal management since the module produces significant power dissipation through conduction and switching losses of the power switches. This is done optimizing the junction temperatures of the FET power switches by means of thermal simulation (Figure 2) and variation of the package. It is next verified through prototyping and measurement by an infrared camera (Figure 3).

Turning the FETs on and off also requires special design. For instance, because the parasitic diode of the commutating low-side FET is in flyback, the turn-on of a high-side FET can result in a high current pulse. If the high-side FET now takes over the phase current too rapidly it causes the diode to be still conducting even during its reverse recovery time (Figure 4). Conversely, a turn-off of a FET switch can lead to overvoltages due to stray inductances of the conducting tracks and the speed of turn-off. Because the latter is usually dictated by the system (and turning off more slowly would increase the switching losses) the focus is on reducing stray inductances as far as possible.

In case of a substantial overvoltage (Figure 5) the FET can go into avalanche. This clamps the voltage and the current in the stray inductance can decay. Avalanche is a reversible and valid operation for a FET. However, the losses during an avalanche event are considerable and it needs to be checked to ensure that the junction temperature stays below its specified maximum (typically 175 °C for a module).

The control module requires a measurement of the bridge current in order to determine the torque. This current is measured inside the power module by means of a shunt resistor. This resistor needs to be sufficiently low in resistance in order to keep power losses low or distort the phase signals. However, it needs to be high enough in resistance in order to allow an accurate voltage measurement.

A thermistor is also located inside the power module for thermal protection. The thermistor is placed in close vicinity of a FET in order to measure the junction temperature as well as possible.

For EHPS power steering in normal operation, the motor is spinning all the time to keep the pump pressurized. This means that the con-ducting FET switches are alternating all the time and provide the phases with current in line with the rotational direction of the motor. This reduces the average current for a particular FET compared to the phase current and the junction temperature reaches some averaged equilibrium (thermal averaging).

An EPS system can also have operational modes without rotating electric motors.

This is the case when the motor is asked to provide a static torque only (as will be the case when a car is parked right next to a curb). In this case, the current will flow constantly through specific FETs and the current per FET is considerably higher. The power module has to be dimensioned for this extreme case and requirements for EPS are therefore usually higher than for EHPS.


The necessary software to control the electric motor can be subdivided into two blocks: the basic motor control, e.g., by means of space vector control (this generates the necessary pulse width modulation (PWM) for all six FET switches of the power module) and the overriding system software with communication protocols, fail-safe operation and diagnostic functions. Additional com-puting power is required for sensorless operation since the software has to convert the response of superimposed pulses into angular information of the rotor.


The packaging technology for an EHPS and an EPS system can be similar. Due to cabling and space restrictions the ECU unit usually is integrated somewhere into the steering system. One possibility is to integrate the module into the mid-flange of the steering system. It is made from aluminum and provides sufficient space for the unit.

The power module needs to be attached to a plane aluminum surface in order to have sufficient thermal transfer and therefore cooling (see sidebar). The control module consists of a populated PCB. It can be attached either directly to the aluminum plane of the mid-flange, or can be placed on risers above the power module in case the PCB has components on either side. Power module and control module are interconnected by through-hole contact leads or other appropriate technology (bond wires, flex foil, press-fit pens). A plastic carrier with integrated lead-frame can be used to hold larger components like electrolytic capacitors and interconnect them.


The required ECU unit for electric steering systems has reached production status for some time now. It consists of a compact unit control module, which is made up of a control module and a power module. Control software drives all necessary functions of the electric motor. The ECUs for EHPS and EPS follow largely similar design principles, but differences in specific modes of operation have to be looked at in detail.


  1. Infineon Application Note, V 1.1, July 2006, p. 7.

  2. Semikron Application manual, 2004, p. 42.

  3. Vishay Semiconductor Application Note (Document No. 88842), July 18, 2002, p. 2.

  4. Maxim Application Note 848, Nov. 12, 2001, p. 4.

  5. Vishay Siliconix AN601 (Document No.: 70752), Feb. 15, 1994, p. 2.


Peter Sommerfeld is Electronic Motion System's director for program management and marketing in Willich, Germany.

The Packaging Technology for the Power Module

Because of space restrictions and power density a design of the power stage with discrete components often does not meet the requirements. In this case, application-specific power modules are being used. A common construction of such a power module is using a direct-bonded copper (DBC) substrate, which consists of a ceramic tile plated on both sides with copper. This substrate can be structured and FETs can be soldered to it in the form of bare silicon die. A plastic insert molded lead-frame provides external contacts (Figure 6).

An alternative approach is the CoolPAK design. Here the bare silicon die are soldered directly to the insert molded lead-frame. An additional substrate is not required as shown in Figure 7. The lead-frame facing the cooling surface is exposed of plastic in order to conduct heat. By means of small nubs in the plastic a short circuit between lead-frame and cooling surface is prevented (Figure 8).

Hide 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.