|
About half of the hardware design effort which this project demands falls under the
category of "Signal Interface". We need to design circuits which will allow the microcontroller
to digitally sample the incoming engine signals, as well as generate compatible output signals to
send on to the stock ECU. The nature and complexity of these circuits is dictated by the form
of the input signals. Some of these signals are simple and well-known; for example the MAP
sensor signal is analog, 0-5 volts, and the O2 sensor signal is analog, 0-1v. But other signals
aren't so simple. We'll be researching the characteristics of these signals from factory manuals
and initial datalogging, and posting results here as they become available.
According to the electrical diagram in the Haynes manual for a 1998 Integra, the TDC sensor is simply a coil. Both electrical connections are routed to the ECU. If this is correct, then the signal should be an analog pulse of varying amplitude and duration (proportional to engine speed), and unknown maximum amplitude. The exact TDC position should be at the peak pulse amplitude. An active derivator may be used to create an edge at this time. At 8000rpm this signal will have a frequency of 133Hz. The ignition input signal (which may be used to observe and control spark timing) is labeled "IGN Input Signal" in the Haynes manual. This signal drives a circuit which charges the coil. It is unclear if this signal is merely a switch or if it actually provides the charging current. Thus it may be either a 5v logic-level signal, or it may be a high current 12v line. Also the duty cycle of this signal remains unknown. It may be 50%, or have a small, fixed, ON time or OFF time. Further research will resolve these issues and allow for interface design. This signal originates at the Mass Air Flow sensor on DSMs, and provides feedback on the rate of air mass flow coming in to the engine. The signal is coded as a varying frequency proportional to air flow, which suggests but does not guarantee a fixed voltage. A simple digital input may suffice for this signal. The duty cycle / ON time / OFF time remain unknown for this signal. This is one of the simpler signals to which we will interface. The MAP sensor output is a varying 0-5 volt analog signal. 0v represents 0psi absolute, while atmospheric pressure (at sealevel) is reported to be 3.0v. The following information was taken from HostBoard regarding the MAP sensor signal.
Successfully interfacing both an input and output for this signal should be easy. The stock ECU certainly has high-impedance inputs for this signal, so the D/A outputs will be able to drive the modified MAP signal. And the MC68HC16Z1 has high-impedance A/D inputs, which may be directly connected to the MAP sensor wire (coming from the motor). The throttle position signal, like the MAP sensor signal, is an analog voltage probably ranging from 0-5 volts. This signal may also be directly connected to the A/D inputs and D/A outputs for easy interfacing. The voltage ranges for this signal are currently unknown. It has been said that the computer requires a TPS voltage below 0.5 volts for idling. The voltage level for Wide-Open-Throttle is currently unknown. The intake air temperature sensor signal is the odd man out. The temperature sensor itself is a variable resistor, and one side is tied to ground. Presumably, the other side is tied to a certain voltage (perhaps 5v?) by the ECU, and the current through the circuit is measured. Interface strategies for this signal may be complex and require further research before development is possible. The oxygen sensor behaves like a variable voltage source, with up to 1 volt in a rich condition, or down to 0 volts in a lean condition. Interfacing this signal is simple; since only an input is necessary (there's no need to modify the O2 sensor signal), it may be Tee'd and connected directly to an A/D input line. The knock sensor is a piezo transducer, which produces high-frequency voltage signals roughly proportional to the movement of the crystal. This sensor output may be digitally analyzed to determine the level of 'knock' or detonation present. The maximum output voltage of this sensor is unknown. A high-impedance A/D input, along with possibly an op-amp for scaling, should be sufficient to sample this signal. Since this signal will not be modified, an output interface is not necessary. |
| After all the necessary research has been completed regarding the charachteristics of the engine signals, interface documentation (explanation and schematics) will be posted here. |
