Toyota Wide Range Air:Fuel Sensor, John Thornton, Underhood Service, January 2002

This month’s article is not going to focus on a specific driveability problem, but rather a specific testing technique. Have you had the opportunity to see a Toyota Wide Range Air:Fuel sensor yet? You may have seen one and not even noticed it. Why? Because this new type of "oxygen" sensor looks very similar to a conventional oxygen sensor. Photo 1 shows a Wide Range Air:Fuel sensor in a 2001 Camry with a 2.2L engine.

At first glance, it looks like a standard 4-wire heated O2 sensor. Well, it is definitely not a standard 4-wire O2 sensor. Toyota first started using Wide Range Air:Fuel sensors with some 1997 models in California. By 1999, their use had spread to other Toyota and Lexus models.

Wide Range Air:Fuel sensors are sometimes called Linear Air:Fuel sensors. Both Honda (in some limited Civics) and Cadillac (1999 Catera) have used this unique type of air:fuel measuring device. But what is the benefit?

Let’s answer that question with a chart shown in Figure 1.

This chart from Toyota shows the relationship between air:fuel ratio and Wide Range Air:Fuel sensor output (relative) voltage. However, some explanation is in order.

We know that a conventional O2 sensor toggles or switches around the stoichiometric point (14.7:1). The PCM never knows the exact air:fuel ratio, only that it is richer or leaner than 14.7:1. When O2 voltage is above the typical target voltage of 0.45 volts, short-term fuel trim tells the PCM to slightly narrow injector pulsewidth. When O2 voltage is below this target voltage, short-term fuel trim tells the PCM to slightly widen injector pulsewidth. The result is the familiar O2 sensor voltage pattern that we see.

The Wide Range Air:Fuel sensor does not operate on this principle. A Wide Range Air:Fuel sensor produces an output that accurately identifies the current air:fuel ratio over a wide lean-to-rich range. The PCM not only knows what the air:fuel ratio really is, but it can command to a specific ratio. How good is it? Well, only an engineer could tell us for sure. Nonetheless, I decided to perform some emissions testing to satisfy my own curiosity.

My first step was to bring a 2001 Camry with a 2.2L 5S-FE engine to an I/M 240 lane for emissions testing. This vehicle is ULEV certified per the decal, so I was expecting some low HC and CO numbers. Figure 2 shows the results of this 240-second dyno run.

The numbers were excellent! All three measured gases (HC, CO and NOx) showed 0.00 grams/mile over the 240 seconds. The tested was performed four times; all with similar results.

While the dyno runs were being performed, I put my gas analyzer probe in the tailpipe. Generally, measuring HC in parts per million (ppm) and CO in percentage (%), on the road (or on the dyno) can be good predictors of I/M 240 results. Figure 3 shows part of the results. The values shown in Figure 3 represent a moment in time. Please take my word for this; those were typical of the values seen over the 240 seconds. The HC never (and I mean never) exceeded 0 ppm. The CO bounced around up to 0.03%. Peak NOx was 144 ppm, and that lasted for just a few seconds.

I also recorded some road tests with my gas analyzer and saw similar results. Understand that emissions control is a system; it is not just a component. But I’m pretty sure the Wide Range Air:Fuel sensor plays an important role in this system.

With that introduction info behind us, let’s get into the testing of this sensor. Please review the chart in Figure 1. You’ll notice on the vertical scale there is a voltage that corresponds with a specific air:fuel ratio.

The chart implies that this Wide Range Air:Fuel sensor produces a voltage in direct relation to the air:fuel ratio. This is not true. The voltage shown on the chart is what one would see if one were using the factory Toyota scan tool to measure the air:fuel sensor parameter. Toyota states that the output of this sensor can only be measured with a scan tool. (It appears that the Toyota factory scan tool and the Vetronix Mastertech with the Toyota OE software are the only tools that support this parameter currently.) The sensor’s output is not a changing analog voltage, but rather a small (< 0.020 amps) bi-directional current. Internal circuitry within the PCM converts the analog current output into a voltage. It is this converted voltage that can be seen on the scan tool. To me, this is a diagnostic challenge. There always is a way to check something. Maybe the technique is direct, or maybe it is indirect.

Let’s start off with what Toyota shows on their scan tool, and then we will look at two alternative testing methods.

Figure 4 shows a 30-second snapshot taken from the Toyota factory scan tool that is manufactured by Vetronix. The Wide Range Air:Fuel sensor output is shown as AFS B1 S1. The second parameter shown is fuel trim (AF FT B1 S1). The engine is at idle, fully warmed up and in closed loop. As you probably have already noticed, the voltage shown for the Wide Range Air:Fuel sensor doesn’t change much.

As per the graph shown in Figure 1, 3.3 volts corresponds to the 14.7:1 air:fuel ratio. As the voltage displayed on the scan tool decreases, the oxygen content of the exhaust decreases (rich exhaust). As the voltage displayed on the scan tool increases, the oxygen content of the exhaust increases (lean exhaust). Keep in mind that the oxygen content of the exhaust corresponds to a specific voltage. Therefore, the PCM knows the exact air:fuel ratio based on the sensor’s output signal.

Note that the scanner Wide Range Air:Fuel sensor is 3.29 volts with a correction of 0.99 (1% negative). This is almost ideal. The PCM is maintaining an air:fuel ratio that is very close to stoichiometric.

As the scanner voltage decreases (from 3.30 to 2.80 volts), exhaust gases are said to be rich (exhaust oxygen deficient). As the scanner voltage increases (from 3.30 to 3.80 volts), exhaust gases are said to be lean (excess exhaust oxygen). This voltage will not oscillate like that of a conventional oxygen sensor. The Wide Range Air:Fuel sensor voltage is relatively stable. The scanner voltage will change due to extreme rich or lean conditions in the exhaust.

Next, study the example shown in Figure 5. Is the exhaust rich or lean? What is the fuel trim indicating? Well, at 2.63 volts the exhaust is on the rich side. Remember, voltages less than 3.3 volts indicate a rich exhaust while voltages greater than 3.3 volts indicate a lean exhaust. The fuel trim at 0.81 (or 81%) equates to a 19% rich bias. The air:fuel ratio is 19% rich of stoichiometric. I was flooding this engine with propane. So, one way to check this sensor is with a scan tool. The only catch is that the scan tool we use must support this parameter.

Before we go any farther, another important point needs to be made regarding the operating temperature of this Wide Range Air:Fuel sensor. For this sensor to function properly it must operate at about 1,200° F vs. the typical 600-700° F that most O2 sensors operate at. Check out the heater current shown in Figure 6 on page 30.

Channel 1 is connected to a current probe clamped around one heater wire, and Channel 2 is connected to the PCM (control) side of the heater circuit. Heater current is duty cycled by the PCM. Peak current is just over 6 amps. I have not seen heater currents like this before. Certainly, there are manufacturers who are pulsing (duty cycling) the heater control line. Six amps is a fair amount of heater current.

I want to emphasize one more time that these sensors work at a much higher operating temperature than we tend to be familiar with. If there is a heater circuit problem, there will be a Wide Range Air:Fuel sensor problem.

Let’s now examine circuit voltages. Figure 7 is a simplified drawing of the Wide Range Air:Fuel sensor. The two heater wires have not been shown. The sensor has two signal lines. One line has 3.3 volts on it, and the other has 3.0 volts on it (relative to engine ground).

These two voltages do not change. If this is true, then the voltage (300 millivolts) across the sensor’s two signal lines does not change. I had a hard time with this when I first started checking these sensors. No matter what I did with the throttle, with propane or with controlled vacuum leaks, I could not get those voltages to change. They are fixed.

What changes is the current flow through the sensor. Again, the voltages are fixed, therefore, we have to test in a different fashion. We’ll discuss two methods.

Please keep in mind for this sensor to operate it must be close to 1,200° F. This says the heater must be functional. The heater circuit must always be checked for proper current flow. Additionally, any tests we perform will require heater operation.

The first Wide Range Air:Fuel sensor test can be done with a lab scope or a multimeter. Credit for this test must be given to Snap-on. I don’t know who originated this test, but it is described in a help menu found in the Snap-on Vantage. Disconnect the sensor’s 4-wire connector and jumper the two heater wires so as to complete the heater circuit. The heater must be functional for the sensor to work.

Do not jumper the sensor’s signal lines, but connect a scope or multimeter to them. Connect the positive lead of the scope to the 3.3-volt line, and connect the negative lead of the scope to the 3.0-volt line. This is shown in Figure 8.

The scope has been connected to the two disconnected signal lines from the Wide Range Air:Fuel sensor. The heater is still connected and operational. Figure 9 shows the sensor’s response to propane. While idling, propane is being added to the engine’s intake. The sensor produces a voltage similar to that of a conventional oxygen sensor. About 1 volt indicates a rich exhaust.

This is a good method for testing one portion of the Wide Range Air:Fuel sensor. Without going into the theory of operation, the Wide Range Air:Fuel sensor is made up of two cells. One of those cells is similar to a typical O2 sensor. The other cell is used to pump oxygen into or out of a reference chamber. The pumping action and, ultimately, the exhaust oxygen content are determined by a very small current measurement. Per my research, the current levels are less than 0.020 amps. Note: SAE papers #930232 and #930233 contain detailed information about sensor operation.

The final method for testing involves the use of an ammeter. While the above method is a good technique, measuring current puts us closer to what is really happening with this Wide Range Air:Fuel sensor.

Figure 10 shows the setup. A digital multimeter has been connected in series in the 3.3 volt signal line. Set the DMM to the milliamp scale. Use jumpers so as to complete the heater circuit and the 3.0 volt circuit of the air:fuel sensor. Note the meter orientation to the circuit. The meter’s red lead is connected to the sensor and the meter’s black lead is connected to the PCM. The meter is in the 3.3-volt circuit. Photos 2 and 3 show what my Fluke 87 displays with a rich exhaust and a lean exhaust.

In this example, a rich exhaust produced a positive 4.89 milliamps, and a lean exhaust produced a negative 1.53 milliamps. If the meter leads had been reversed, so would have been the polarities seen on the meter.

This test can be taken one step further. Instead of using a conventional DMM, a graphical multimeter can be used. The Fluke 867 graphical multimeter was used to acquire the upcoming patterns. As before, the meter is connected in series with the 3.3-volt circuit. Use jumpers to complete the heater circuit and the 3.0-volt circuit of the air:fuel sensor.

Figure 11 shows how the Fluke 867 was connected for the following recordings. Please note, in this next group of tests the positive lead of the meter was going to the 3.3-volt feed line of the PCM. Figure 12 shows the relationship.

The next five figures are snaphots taken from the Fluke 867. Short of the factory scan tool, I believe this is an effective and accurate way to test the Toyota Wide Range Air:Fuel sensor.

In Figure 13, the vertical scale is ±14.2 milliamps, and the horizontal scale is from 0 to 12.8 seconds. This was taken at idle. I am snapping the throttle open and close. Wide Range Air:Fuel sensor current drops to -6.9 milliamps (rich exhaust). The sensor responds to changes in the air:fuel ratio. This is good.

In Figure 14, the vertical scale is ±14.2 milliamps, and the horizontal scale is from 0 to 12.8 seconds. Fuel injector #1 has been disconnected. An open injector should bias the exhaust lean. Average current appears to be about 2.8 milliamps. Per the Toyota scanner, the air:fuel ratio sensor voltage was 3.5 volts. Remember that voltage above the base of 3.3 volts indicates a lean exhaust.

In Figure 15, the vertical scale is ±14.2 milliamps, and the horizontal scale is from 0 to 64 seconds. With the engine at an elevated rpm, I am adding and removing propane to the intake. The Wide Range Air:Fuel sensor responds appropriately.

In Figure 16, the vertical scale is ±14.2 milliamps, and the horizontal scale is from 0 to 25.6 seconds. I am performing a brake torque in reverse. Sensor current drops to -11 milliamps. As you now know, this indicates a rich exhaust.

In Figure 17, the vertical scale is ±14.2 milliamps, and the horizontal scale is from 0 to 25.6 seconds. While cruising at about 30 mph, I did an aggressive WOT acceleration and then a long deceleration. The current is dropping (going rich) for about the first 15 seconds, before it starts to increase (lean decel).

I hope these five figures demonstrate how effectively this Wide Range Air:Fuel sensor can be checked with an ammeter. A graphing DMM is handy, but not necessary, a typical DMM will work. If our scan tool cannot support this parameter on a Toyota, we have an effective alternate test for this unique sensor.

To summarize, we looked at three methods for testing the Toyota Wide Range Air:Fuel sensor. The first, and probably the easiest, was through the use of a scan tool. The only issue was that not all scan tools support this parameter (yet).

The second method was to disconnect the two signal lines and connect them to a lab scope. The heater circuit had to remain intact. The sensor was then tested in a similar fashion to a conventional O2 sensor.

The third approach was to connect a DMM in series with the 3.3-volt signal line, and jumper the remaining three wires back to their respective circuits. With this technique, we could actually monitor the current the PCM was measuring as the exhaust oxygen content changed. This test brought us closest to actual circuit operation.

When one of these unique "oxygen" sensors appears in your bay, I hope this article has given you the knowledge and confidence to properly evaluate it.