ECU Inputs: Air and Water Temperature Sensors

For proper operation, the ECU needs to know the temperature of the air going into the engine as well as how warm the engine is. It uses a pair of identical Nippon Denso 179700-0110 temperature sensors to supply that information to the ECU. One sensor is used to measure the intake air temperature and the other is used to measure the temperature of the engine coolant. They look like this:

The sensing element is mounted inside the tip of the metallic protrusion at the bottom of the sensor. The metallic jacket protects and isolates the sensing element from whatever nasty substance it might be immersed in.

Temperature Effects

Temperature information can be used to make both gross adjustments and fine adjustments to the air-fuel ratio. The gross adjustments are familiar to a rider of old (i.e. back in the carbureted era) when they were faced with starting their bike for the first time on a cold morning. In that case, the human rider acted as the temperature sensor, and they responded to starting a cold engine engine by putting the choke on. Choking the intake tract would make a carburetor produce an artificially rich mixture, which would make it easier for the engine to start. Once started, the rider would have to make more decisions about regarding the position of the choke as the bike warmed up until they could finally disable the choke completely.

The computer on a fuel-injected machine makes the same decisions by measuring the temperature of the coolant. If the engine is stopped and the coolant is cold, the computer will make sure to enrich the mixture when the rider hits the starter switch. The computer will see the temperature rise as the engine warms and will respond by reducing the extra fuel back to normal levels.

The computer can also make some fine adjustments to the air-fuel ratio based on the air temperature. Hot air is less dense than cold air, and less dense means less molecules of air per unit of volume. To maintain a correct mixture ratio, a less dense volume of air requires less fuel. In the old days, changing carburetion to compensate for temperature would mean swapping jets. Depending on the bike, swapping jets could range from being a complete pain to being merely annoying. The ability of the fuel injection computer to tweak its operation based on the operating environment help improve we would call the 'rideability' of the fuel-injection system. These little adjustments make the bike feel like it runs better no matter how it is being ridden or what conditions it is being ridden in. It takes the manufacturer a long time and plenty of experimentation to get things tweaked just right.

Sensor Operation

These sensors are based on thermistor technology. A thermistor is a resistor whose resistance changes according to its temperature. If the resistance of a thermistor gets higher as its temperature rises, the thermistor is said to have a positive temperature coefficient. If the thermistor resistance falls as the temperature rises, thermistor is said to have a negative temperature coefficient. The other thing to know about thermistors is that their change in resistance for a given change in temperature is non-linear.

I got a spare sensor off eBay and did some tests to see how it responded to temperature changes. The experimental setup involved a pot of boiling water on the stove containing the sensor and a thermometer. The sensor was connected to a multimeter so I could measure the resistance at any point in time. The plan was to map out the sensor's response curve by bringing the water to a boil, shutting off the heat, and then measuring the change in resistance against the changing temperature of the water as it slowly cooled down. In the picture below, you can see a measurement in a cup of water taken at a room temperature of 20 degrees C showing a resistance of 2598 Ohms.

Using the test setup, I measured the following resistance vs. temperature data for the sensor:

To get the measurement for 0C, I filled an insulated cup with ice and water and let the sensor soak for a while.

Looking at the measurement data, it is easy to see that this sensor is an NTC (Negative Temperature Coefficient) since the resistance rose as the temperature fell. One other thing to note is that thermistors are manufactured with a specific resistance measured at 25 degrees C. In this case, the thermistor has a value of basically 2000 Ohms (2K Ohms) at 25 degrees C. This is a good sign since a 2K value is one of the standard thermister values.

The nonlinearity of the response is also apparent. For example, the change from 22C to 24C caused the resistance to change by 180 Ohms. At the high end of the scale, the change from 88C to 90C caused the resistance to change by only 14 Ohms, or less than 1/10 as much as the change from 22C to 24C.

Fortunately for our Aprilia ECU, even though the change is non-linear, it is quite predictable. Thermistor manufacturers print tables or formulas that can be used to calculate a temperature given a specific resistance. Here is a table from one manufacturer describing the response curves for two of their thermistor families:

The way to interpret this table is that at 25C, the ratio column says that the thermistor will have a value of 1.00 times its nominal 25C resistance. At 70C, a thermistor from the curve 2 family will have a resistance of .2172 times its nominal 25C value. For the Aprilia sensor with its measured value of 1992 ohms at 25C, the resistance at 70C should be 1992 * 0.2172, or 433 Ohms. This corresponds almost exactly with the 432 ohms measured at 70C. If the Aprilia sensor had a curve-1 response, its value would have been (1992 * 0.1753) or 349 ohms, which is nowhere near the measured value at 70C. Comparing other curve-2 values from the table for temperatures that had been measured, we get the following:

It would appear that we have an excellent match to the curve-2 family, at least from 30C through 90C. The resistance down near 0C is 10% off, but I figure that it is close enough that we could use the curve-2 characteristics to convert the Aprilia/Denso sensor resistance information back to a temperature, especially since you are not likely to find me on my bike when the temp is 0C.