Understanding Engine Management - Part 1
For educational purposes only. Take on an empty stomach. See your mechanic if symptoms persist.
Engine Management Systems Series.
Author/s: Jacques Gordon
Issue: June, 1999
The first in a series describing the components, how they work and how they work together.
Whether it uses pistons, rotors or turbine wheels, 2-strokes, 4-strokes or continuous process, every internal combustion engine you've ever seen draws in air and fuel, pressurizes the mixture, bums it and dumps the exhaust: suck, squeeze, bang, blow. In the piston engines that we see and use every day, it's possible for every one of these basic functions to be controlled by a computer. Many of the latest models don't have a throttle cable or linkage, and the computer operates the throttle valve (suck). Miller-cycle engines control the throttle or delay the closing of the intake valve to control compression ratio (squeeze). The most common equipment is that which controls fuel quantity and spark timing (bang). And of course many engines with variable valve timing control both intake and exhaust valve opening (blow).
While we're not aware of any piston engine in production (yet) that electronically controls all four functions, today's ECM at least keeps track of each degree of crankshaft rotation. You've probably already had some experience with many of the sensors and output devices that make all this possible, but a true understanding of each component is the key to knowing what that tiny electronic brain is really trying to do. In this article, we'll look at a few of the most critical sensors found on 1990 and later Japanese and Korean engine management systems and describe the signals they produce. As the series continues in future issues, we'll eventually cover every sensor, actuator and control strategy in the most popular systems from Asia, Europe and the U.S.A.
Every automotive type of sensor is either passive or active and produces one of three possible signals: voltage, frequency or pulse width modulation. A passive sensor produces its own signal without power or command from the ECM. Piezeo-electric knock sensors, wheel speed sensors and some rpm and position sensors are passive because they produce their own voltage. They may have only one wire connection to the ECM, the signal wire. Active sensors must receive a reference voltage or query from the ECM before they can return a signal. These include throttle position, air flow and some speed or position sensors, because they receive a reference voltage from the ECM and somehow modify that reference to generate a return signal. They connect to the ECM with at least three wires; power, ground and signal. For very low power signals, or if the wire runs past a likely source of interference, there may also be a shield around the signal wire that's grounded at the ECM. In future systems, these sensors will all feed their signal to a data bus, a single sensor circuit that carries all the signals in a specific sequence, encoded so the ECM can tell them apart. For now though, each sensor has its own discrete wiring connection to the ECM.
Throttle Position Sensor
The ECM uses this signal as a direct reading for engine load, needed for calculating ignition timing, fuel injection and automatic transmission functions. The TPS is a potentiometer, a rotating resistor that produces an analog signal, meaning infinitely variable voltage (up to the maximum). Typically the computer supplies a 5V DC reference, and with the throttle closed, the resistance of the TPS reduces the output signal to about 0.5 volts. As the throttle opens, signal voltage increases, typically to a maximum of 4.5 volts. If the sensor returns the full reference voltage or no voltage at all, the computer assumes a malfunction and will ignore the signal, set a code and turn on the MIL. More common is a partial failure, where the signal drops out through a small range of the sensor's motion. This malfunction causes driveability problems with symptoms that don't always make you think of TPS failure, like, harsh shifting or ignition timing problems. On most Japanese systems, total TPS failure may not stop the engine, but the ECM will turn on the MIL and probably switch into a limp-home mode. Many vehicles also have separate switches to tell the computer when the throttle is closed or wide open. The TPS voltage signals can be seen with a DVOM, but the only way to find a dead spot in the rotating resistor is with a scope.
MAP and BARO sensors
These are pressure sensors. Picture a diaphragm with a semi-conductor chip mounted on one side. Pushing on the diaphragm distorts the shape of the chip, causing a relatively linear change in resistance across the chip. This is the principle behind most electronic pressure sensors. The difference between a MAP and a BARO sensor is the location of the measurement and how the data is used. The BARO sensor is often built into the ECM and reads barometric pressure, the static pressure of the air around us that correlates directly to air density. The Manifold Absolute Pressure sensor measures atmospheric pressure in the intake manifold. Zero pressure exists only where there is no atmosphere, so here on the ground, manifold pressure is always more than zero. Any time the engine is running, manifold pressure is less than atmospheric and increases as the throttle opens. The data from these two sensors is used along with rpm to calculate the amount of air entering the engine, a speed/density calculation. Systems with an actual air flow sensor may still use a MAP sensor for load calculation. If the BARO sensor fails, the computer may use a reading from the MAP sensor taken in the instant between key-on and engine-start. If the MAP sensor fails, the engine may start but it won't rev above idle because the computer can't calculate the air quantity. The MAP signal can be seen by back-probing with a DVOM while the engine is running. With the wiring disconnected, you can see the resistance across the sensor terminals change while working the sensor with a hand vacuum pump.
Air Flow sensor
Actually measuring the air flowing into the engine is much faster and more efficient than calculating it. There are two basic air flow measurements: volume and mass. Most Japanese systems measure mass, which is the final value the computer works with anyway. In a hot-wire Mass Air Flow (MAF) sensor, a platinum wire or a thick film is maintained at a temperature of about 120 [degrees] C above ambient. The ECM monitors the power needed to maintain the temperature; more air carrying away the heat requires more power to hold the temperature constant. While it's actually current that makes the sensor element hot, the ECM or the sensor's own internal circuitry controls current by either adjusting the voltage level or by pulsing the voltage. With a hot-wire element, the wire is momentarily heated when the ignition is switched off to bum off any dirt deposits. This is not needed with hot-film types because it uses a temperature sensing resistor mounted on the back of a microchip, where it's protected from deposits. This chip also includes much of the control and signal conditioning circuitry. Thick-film is slower than the hot-wire type but both react to changing air-mass in milliseconds. On both types, only a calibrated fraction of the total intake air flow moves past the sensing element. The output signal of either type will be either analog voltage or frequency, which can be seen with the appropriate DVOM or on a scope. The ECM interprets that signal directly as grams per second (g/s) of air flow.
The Volume Air Flow (VAF) sensor measures air volume and temperature, then the ECM calculates the mass. The electronics are similar to a TPS: a vane pivots a shaft that turns a potentiometer to modify a reference voltage. Air flow through the housing pushes the vane open against a balancing chamber, with faster air flow moving the vane farther. The IAT sensor is usually mounted in the VAF housing, and sometimes a BARO sensor is also used for fine tuning the g/s calculation. The output is an analog voltage signal and by back-probing the correct terminals, the signal will increase as the vane is moved by hand.
The Karman-Vortex air flow sensor was used by Mitsubishi in the early 1990s. As air flows through the sensor housing, an ultra-sonic generator transmits sound waves at right angles to the flow of air. Obstructions built into the housing cause eddy currents in the flow, and an acoustic detector reads the frequency of these currents as they change with increasing air flow. It's extremely accurate and automatically accounts for differences in air temperature and density (barometric pressure), but it can be easily `confused' by pulsations in the air flow. The output signal is frequency, which can be seen with the appropriate DVOM or a scope.
These sensors report shaft speed and position to the ECM. There are three different types of CKP and CMP sensors. The inductive sensor is a permanent magnet with a coil of wire around it, positioned close to a toothed wheel on the shaft. As each moth passes the magnet, it disturbs the magnetic field and a sine-wave voltage is induced in the coil. The computer interprets the frequency of the sine wave as rpm. A specific position, such as TDC, can be referenced by making the gap between two of the teeth different from the rest, making the sine wave signal different at that location. The distance of the sensor from the toothed pick-up wheel is critical and often adjustable. The inductive sensor is simple and tolerant of harsh environments, but the other two types of sensor are usually inside the distributor or dedicated housing. The Hall effect sensor has a permanent magnet mounted next to a semi-conductor chip that has a current passing through it. A metal wheel with tabs is mounted so that the tabs pass between the magnet and the chip. As the wheel rotates, the interruption of the magnetic field changes the current flow through the chip. The resulting frequency signal can represent either speed or position, and is commonly used for both.
The other active sensor type is a photo diode, using an LED to produce light, another type of diode that detects light, and a wheel with slots or holes passing between them. The openings in the wheel can represent degrees of rotation or any other desired reference. On certain Nissan systems, the distributor contains one wheel with three sets of slots, LEDs and detectors that read rpm, camshaft position and cylinder number one. On many Asian vehicles, the loss of a cylinder reference signal will not stop the engine if it is already running, but the ECM needs both reference and rpm signals for starting.
On most Asian OBDII systems, there is a dedicated sensor that reads crankshaft speed for misfire detection, usually an inductive type mounted right on the block. The resistance of inductive sensors can usually be tested with a DVOM if you can find the specification. For the other sensors, the presence of a signal can often be detected with a DVOM while the engine is running or cranking, but the quality of the signal can only be tested `live' by back-probing the signal wire with a scope.
The sensors described here are all an engine needs for basic fuel quantity and spark timing; air mass, crankshaft speed and camshaft position. In future editions of this series, we'll look at other devices used to control fuel trim, ignition advance, valve timing, intake manifold tuning valves, traction control, emission control and anything else managed by the ECM. Then you'll know everything the ECM knows, and more.
Line drawing illustrations courtesy of Autodata Publications Inc., who also provided much of the research material for this article.
COPYRIGHT 1999 Cahners Publishing Company
COPYRIGHT 2000 Gale Group
93 Probe GT - for sale! $5,300.
95 Probe GT - the 007 version - smooth and refined.