In this article, I’d like to continue to discuss some electrical circuit basics. Without a thorough understanding of the fundamentals, we struggle to understand the more complex areas of diagnosis. In particular, I am going to discuss temperature sensors, how they work and how best to test them.
In its most basic form, a temperature sensor for many automotive applications is a variable resistor, whose resistance varies with temperature. The most common type we deal with have what is known as a Negative Temperature Co-efficient (NTC), which means as the temperature of the sensor rises its resistance decreases. Typical resistances would be around 10,000 Ohms at 200C and dropping to 200 Ohms at 1000C, but check the manufacturer’s data for the model you are working on.
This all sounds pretty straightforward but an engine control unit cannot directly measure resistance, so it needs some circuitry to convert the varying resistance into a format it can understand. Figure 1 shows a typical ECU temperature input circuit. It converts the variable resistance into a variable voltage, which in turn is converted to a digital reading via an Analogue to Digital Converter (ADC) within the control unit.
For better clarity, I have redrawn the circuit in Figure 2. We can see that the fixed resistor and the temperature sensor are in series and form what is known as a voltage divider circuit. We have 5 Volts at the top and ground at the bottom. The voltage at the centre will vary, depending on the resistance of the temperature sensor.
Now, here comes the maths bit. The voltage at the centre can
be worked out using Ohm’s Law. We first need to know the total resistance of the circuit. R1+R2 = Rt (10,000 + 1,000 Ohms). From this we can work out the current flowing from 5V to ground. Ohm’s Law again gives us A = V / Rt. So, at 20C, the calculation would be A = 5 / 11,000 or 0.00045 Amps.
Now we know the current flowing through the circuit, we can work out the voltage drop across the temp sensor. So, using Ohm’s Law again, (we use just the resistance of the temp sensor in this calculation), V = A x R, (V = 0.00045 x 10,000) or 4.55 Volts. At 20C, the ECU would see approx. 4.5 Volts.
The calculations at 1000C would be; Rt = 1,200 Ohms. Current through the circuit A = 5 / 1200 = 0.00417 Amps. Voltage across the Temp Sensor V = 0.00417 x 200 = 0.84 Volts.
HOW DO YOU TEST IT?
So now we know how the circuit works, what is the best method to test it?
The simplest and easiest way is via live data. By measuring the temperature close to the sensor with a thermometer and comparing the reading with live data, we can assess the whole circuit. We are looking for the readings to be within a few degrees of each other. If they are not in agreement, then we have a problem that needs further investigation.
If the reading isn’t correct then the live data should give you a clue as to what is happening. The minimum reading is usually around -400C and the max around 1500C. If you are seeing the minimum reading then the lower part of the voltage divider circuit has no resistance (short to ground); either the sensor has an internal short or the wiring has a short to ground. If you see the maximum reading then the lower part of the voltage divider has infinite resistance (open circuit); the sensor or wiring is open circuit.
If the reading is just inaccurate, the next step would be to carry out resistance checks on the sensor and compare those to the manufacturer’s data. They normally give several readings at different temperatures for you to be able to check and compare. Any difference to the data would obviously mean the sensor is faulty and requires replacement.
What other things could you test? Well you could measure the voltage at the sensor. With the cable disconnected, one wire will go ground through the control unit and the other will go to the fixed resistor. As I mentioned last month, without current flow there will be no voltage drop; the circuit is now not under load, so there will be no voltage drop across the resistor that forms the top half of the voltage divider. What you will measure is a voltage fairly close to the regulated 5V supply.
If you want to really go to town, you can back-probe the sensor and use a Scope to monitor the voltage over a long time base of say 200s per division, giving around 30 mins on the screen. I have caught a few anomalies like this, such as a Peugeot 206 that would stall at certain temperatures. Live data was too slow to show up the issue but the scope showed the sensor intermittently shorting internally.
Although it may seem like a fairly simple circuit, there is a little more to it than first meets the eye and the principles can be applied to other circuits with the car. I hope this information has helped to increase confidence in your test procedure.