For the PDF version of this article, click here.
Diesel-powered cars accounted for only 3.2% of the 2005 U.S. light vehicle sales, but J.D. Powers forecasts growth to 9% in 2012. Diesels already constitute about 50% of the Western Europe car market share where fuel prices are much higher than the U.S.’s record prices of 2005. Why? Diesel fuel is 15% to 20% cheaper than gasoline and the mileage is 10% to 20% higher than a vehicle with a comparable spark-ignition (SI) engine. And, contrary to the popular belief of most Americans, the diesel does not smell, is not noisy, and has excellent acceleration. Electronics has played a role in the diesel’s acceptance in Europe. The electrical components and electronic controls for diesels have several striking differences from those used in SI engines but there are areas with similarities.
DIESEL VS. SI ENGINE
Instead of spark plugs that ignite gasoline in the combustion chamber and a throttle control that increases the air flow to provide higher torque and horsepower, a diesel engine uses a combination of glow plugs to provide heat and high compression to self ignite fuel oil. Varying the amount of fuel injected into the cylinder controls the speed and torque. With this basic operation, diesel engines provided increased efficiency and high torque compared to comparable SI engines, since they were first produced in the early 1900s. Consequently, diesel engines in the United States and other regions provided the power for heavy trucks and buses. These large vehicles and systems gave diesels the reputation of slow, noisy, soot-generating machines. Then electronic control changed the diesel even more dramatically than it changed the SI engine. Improved diesel fuel reduced the objectionable odor and sulfur level allowing advanced control techniques to address the remaining shortcomings. Figure 1 shows some of the key electronic elements of today’s diesel engine management. In addition to the high-pressure pump, injector, glow plug and EGR actuator, the figure identifies 16 sensor measurements that could be made as part of the control system including the after-treatment techniques. Direct fuel injection solved many of the diesel’s problems including soot and noise.
Electronics is an integral part of the solution to make user-friendly diesels.
Comparing the diesel system to an SI engine, Stephan Lehmann, strategic marketing manager for the global automotive business at Freescale Semiconductor said, “The system concept is not widely different. You still use a microcontroller at roughly the same complexity.” A simple gasoline four-cylinder engine would actually need less performance. But a gasoline direct-injection SI engine requires essentially the same computing power as the diesel—a mid to high-end 32-bit microcontroller (MCU) to run the engine fast enough. “You might need a little bit more on the sensor side especially to handle advanced Euro 4 or Euro 5 requirements or requirements in the states for California,” noted Lehmann.
THE DIRECT INJECTION DIFFERENCE
Initially, diesel engine management systems used electromagnetic (solenoid) injectors. These systems operated at 300 bar (4,350 psi). More recently, with common rail direct injection (CDI) technology, it is possible to inject at 1500 to 2000 bar (21,750 to 43,499 psi). CDI still uses solenoid injectors but adds the capability of piezoelectric or simply piezo injectors. The main manufacturers of common rail diesel injectors (Bosch, Denso, Delphi and Siemens) are moving toward CDI.
“One of the key things that was a big step forward is piezo injector technology,” noted Freescale’s Lehmann. “The piezo injector acts faster than traditional injectors and that gives them the opportunity to have more injections per cycle,” This allows system designers to improve exhaust quality and performance and to reduce noise. To have multiple injections, the MCU needs to have the response time. Most suppliers use two voltages, 12 V and 40 V or 50 V, and some even have 12 V and 70 V for solenoid injector operation. The higher voltage is generated by a dc-to-dc converter and uses a peak and hold waveform. Piezo injectors require 200 V to 300 V and frequently use IGBTs—the same technology used to switch the coil in SI engines. Moving to piezo injectors provides faster, more reliable pulses from one injection to the next. Piezo injectors can generate multiple pulses easier than solenoids and manage variable lift of the injector, which is not possible with solenoids. Variable lift means lifting the needle in the injector in a programmable manner. Obtaining five pulses from a solenoid is challenging and six or seven pulses could be common in the future. “To manage the heat release in the chamber the fuel is broken into several pulses with phases known as pilot injection, pre-injection, main injection, post-injection and others,” observed Patrick Leteinturier, senior principal automotive systems, Infineon.
In addition to multiple pulses, variable nozzle, and variable lift, future diesel engines will use a higher pressure rail. Solenoid injectors can meet Euro 4 and even Euro 5 but the after treatment could be more difficult. The system tradeoffs in choosing the right technology depends on cost in the injection vs. cost of the after treatment. Figure 4 shows the cutaway of a piezo injector that identifies the piezo actuator.
Controlling piezo actuators requires additional software and in- creased application development cost. As a result, Siemens introduced an architecture that addresses both gasoline and diesel engines. The EMS 2 layered architecture is independent from components, hardware and the microcontroller. As shown in Figure 3, the system designer has the choice of blocks developed specifically for diesel or gasoline engines with the option to implement hybrid technology. By organizing the EMS 2 software into modules, system designers can access the modules required for a specific development project from a database.
The top level in the architecture addresses the entire vehicle. The next level down focuses on various systems, including the model-based functional strategies describing engine operation. These upper two levels are essentially independent of the system components and hardware requirements. The next level down specifically addresses the component-dependent aspects. Eighteen groups, such as fuel, have a total of 65 individual function packages like injector driver control. Hardware control that relates specifically to controlling the processor is handled separately. This software approach provides hardware independence and meets the requirements of the AUTOSAR initiative.
MICROCONTROLLERS IN DIESEL ENGINE MANAGEMENT
Leteinturier noted that today’s advanced systems require a high- performance microcontroller to provide in-cylinder signal conditioning and, in terms of complexity, is equivalent to making a knock measurement for each cylinder. In a V-engine today, there are two knock sensors. Because a pressure measurement for each of the four cylinders in the diesel has to be engine synchronized, the complexity will be greater than the way today’s knock sensing is handled. This requires acquiring data every 0.5 of crank angle. To handle this performance, a 150 MHz or higher frequency processor is required. Today’s performance is in the 80 MHz range.
Axel Hahn, director-NAFTA Microcontroller Marketing Automotive & Industrial Business Group, Infineon, said, “It is important to understand the requirement is not only for performance, the clock frequency is going up. There will be an increasing requirement in code size, which leads to a microcontroller with small on-board flash.” Today, 2 MB of flash memory is used but Leteinturier expects that 4 MB will be required for Euro 5. Some control could be distributed in these systems especially where high-voltage control is required. Today, the high-voltage piezo actuation is controlled from a separate module. The maturity level of these systems does not currently allow optimized partitioning, so the separate vs. integrated (or centralized vs. distributed control) issue remains to be resolved.
The MCU requires more complexity than a four-cylinder SI engine, but a processor for a gasoline direct injection engine would probably be comparable—an MCU in the mid to high-end range of 32-bit products. “The system block diagrams are rather similar,” noted Freescale’s Lehmann. Today’s diesels that use Freescale technology typically have a 40 MHz to 66 MHz processor. To satisfy the requirements of future diesel systems, Lehmann noted that the next-generation systems require further improvement in response times, up to 150 MHz, to allow more complex algorithms with faster sensor data.
“We are not using a single core any more,” said Infineon’s Hahn. “We have a kind of small co-processor.” Infineon’s TriCore is a combination of microprocessor, microcontroller and DSP that has a PCP, a peripheral event controller, which off-loads the TriCore. The TriCore handles the higher-level tasks and the PCP handles the lower-level drivers and lower-level tasks. Partitioning is accomplished by soft-ware. In the diesel, low-level drivers handle hardware-specific tasks, such as analog information from sensors, to prepare the data in a format that makes it easier for the main processor to control. The TriCore handles the applications software and strategy to manage the system.
Not managing the right amount of fuel at the precise time creates excessive NOx requiring post treatment by techniques such as a NOx trap or selective catalyst reduction (SCR) system. In the lean NOx trap shown in Figure 4, the ECU controls an EGR throttle valve based on inputs from six sensors including pressure temperature, air flow and possibly a NOx measurement. Onboard diagnostics (OBD) for the particulate trap requires a pressure sensor for upstream and downstream pressure measurements.
“One of our customers measures the NOx in the exhaust to verify when the filter is full and then it would feed back into the engine management system and they would change the mix to clean up the filter,” said Freescale’s Lehmann. In at least one system in development, a separate ECU performs this function and communicates with the main engine management ECU via CAN. Similar functionality is used in gasoline engines to address NOx control.
Another approach to reduce NOx emissions injects urea or AdBlue, as it is commonly known, into the exhaust stream after the particulate filter, which is then followed by a selective catalytic reduction (SCR) element. About two liters of mixture of urea and water is required with every refuel. The closed loop SCR catalyst system shown in Figure 5 adds the SCR catalyst, a urea tank, injector and NOx sensor. However, there are other options.
SMART GLOW PLUGS
Adding electronics to control the glow plug operation virtually eliminated the wait time prior to starting diesel engines. Beru combined its instant start system (ISS) technology with a pressure sensor to provide feedback for advanced NOx control. To achieve very low NOx emissions, an alternate combustion process called homogeneous charge compression ignition (HCCI) in the research phase of development uses closed loop in-cylinder pressure measurements. “The technology developed for diesel is highly suitable for HCCI engines to use in direct injection,” said Leteinturier. The way to burn gasoline in the future could be close to identical to the way to burn diesel.” This approach may be able to eliminate the NOx treatment.
As shown in Figure 6, the piezoresistive pressure sensor (developed in cooperation with Texas Instruments) is placed on a diaphragm located away from the cylinder pressure and the ISS heating element, which only glows at the tip. The heating element mounts on flexible bearings that transmit pressure to the membrane. The sensor’s location also minimizes its exposure to vibration.
According to Infineon’s Leteinturier, high engine position accuracy of less than 0.1 (required to get good synchronization data) means that the signal from the in-cylinder pressure sensor must be processed close to the sensor. Providing sampling frequencies in the range of 100 kHz per sensor and about 12-bit accuracy, requires an application processor with DSP performance capability and a high-speed digital communication (CAN bus).
THE DIESEL FUTURE
In Europe, the diesel engine is the engine of the present and the future for fuel economy and performance. With the many approaches to solve the tougher requirements of Euro 5 and SULEV, electronics will also solve its emissions’ handicap. In the United States, the diesel or spark ignition engine battle will heat up as lower sulfur diesel fuel is available nationwide this fall thanks to a federal mandate. The SI engine’s best case 33% efficiency occurs only at full load and with the throttle closed, the efficiency decreases to as low as 5% or less. “The diesel has the advantage that you can run with high efficiency quite a lot of the time,” emphasized Leteinturier. “It can reach 40 % to 45% efficiency. That is going to be very good.”
. 2. Patrick Leteinturier’s presentation, “Diesel Engine Management,” Infineon Technologies, Feb. 9, 2006.
3. Jean-Marc André’s presentation “Platform Strategies for Engine Management Systems of the Future,” Siemens.
4. Fiat-GM Powertrain presentation at International Symposium Diesel Engine: The NOx & PM Emissions Challenge, Oct. 13-15, 2004, Monopoli (Bari), Italy.
ABOUT THE AUTHOR
Randy Frank is president of Randy Frank & Associates Ltd., a technical marketing consulting firm based in Scottsdale, Ariz. He is an SAE and IEEE Fellow and has been involved in automotive electronics for more than 25 years. He can be reached at [email protected]