Most of us have become pretty comfortable with the conventional oxygen sensor, and feel like we have a pretty good handle on telling the good ones from the bad ones. So it’s time to move out of our comfort zone and on to the oxygen sensor’s more sophisticated younger cousin, the air/fuel (A/F) sensor. These come in a few different varieties, but I’ll concentrate on the Toyota sensor exclusively because the company was an early adapter of this technology and it’s found on many of its vehicles. These sensors are used only upstream of the catalytic converter; downstream sensors are always of the conventional O2 sensor type.
How do you tell whether the vehicle has an O2 sensor or an A/F sensor? Not all Toyotas use an air/fuel sensor, but they’re in increasingly widespread use throughout the product line. The first place to look is on the underhood Vehicle Emission Control Information (VECI) label. Of course, as we all know, sometimes the hood isn’t original, or the label is missing. In such a case, check the VIN with your local dealer. Sometimes, however, even a VIN check is inconclusive. At that point, careful examination of the wire colors at the harness connector plug should prove conclusive.
Let’s first look at a few of the more common DTCs associated with a bad air/fuel sensor, then take a deeper look at advanced diagnostic techniques and datastream analysis.
Far and away the most common air/fuel sensor fault codes in the northern part of the country are P1135 and P1155, indicating heater circuit failures in the sensors for bank 1 or bank 2, respectively. These are two-trip codes. While conventional oxygen sensors function at 650° to 850°F, the Toyota A/F sensor requires 1200°F for reliable accuracy.
Diagnosis here holds few surprises, as it largely mirrors the procedures for diagnosing heater circuit faults in oxygen sensors. Some applications feature separate dedicated A/F sensor heater fuses, commonly on V6 models, although there are a few dual-bank 4-cylinders as well. Most 4-cylinder single-bank applications do not utilize a dedicated heater fuse, so if the car runs, you can be sure the fuse is good, since it also powers the injectors.
Toyota A/F sensor heaters are nominally in the 75-watt range (approximately 5 to 7 amps), though actual current consumption may vary with the applied duty cycle. A low-amp probe can quickly verify heater performance. Typical heater control programming applies full power for about 20 seconds after start-up, then maintains heat with an applied duty cycle that varies as needed. When exhaust flow is low, as at idle, more heat is commanded. There are also some one-trip heater codes that occur in the relatively rarer event of a completely open or shorted heater circuit.
Accompanying the P1135/P1155 codes may appear code P1130 or P1150—air/fuel sensor range or performance fault codes, also two-trip codes. If there are heater codes as well, they must be dealt with first. In fact, even if a range or performance code appears alone, check the associated heater circuit and actual heater performance first, since these are critical to correct A/F sensor functioning. The Toyota air/fuel sensor typically operates at about 1200°F, so relying on exhaust gases alone to bring the sensor up to temp simply doesn’t work reliably.
Whatever DTCs you retrieve, be sure to check for relevant TSBs at the outset. For example, a code P0031 and/or P2238 on some applications, such as a 2.4L Camry, may require both a new sensor and a recalibration of the PCM—Toyota-speak for a reflash.
Rounding out the DTC list specific to Toyota A/F sensors are codes P1133 and P1153, indicating an unacceptable slowdown in the sensor’s response rate. The monitor for these DTCs executes an active test and looks for the sensor to respond within a maximum of 1.1 seconds. Once again, these are two-trip codes.
In addition, there’s a very common one-trip code that you may encounter. It’s P0125, rather misleadingly titled “insufficient coolant temperature for closed-loop operation.” Don’t be fooled! Here’s what really happens: The A/F sensor is monitored for voltage change activity under specified conditions. If there’s insufficient change in the sensor’s output voltage during a 90-second interval of off-idle operation at or above 1500 rpm with a road-speed of 25 to 62 mph, and at least 140 seconds have elapsed since start-up (before the 90-second interval begins), DTC P0125 will set.
Be sure to check for other pending codes as well, since they may offer important clues. (Toyota products equipped with conventional O2 sensors can also set this code, for analogous reasons.) In many cases, the fault will lie not in the sensor itself, but elsewhere in the air/fuel induction system, most likely because of a vacuum leak.
Another surprisingly common problem involves familiar lean codes P0171 and/or P0174. Expect to find combined short-term and long-term fuel trims of 50% or more in the freeze-frame data. Once again, these codes frequently stem from an underlying heater circuit fault. But if the customer complaint or the repair history reveals recent work, check to make sure someone hasn’t swapped in a conventional four-wire heated oxygen sensor by mistake. Although the connector latch is slightly offset and doesn’t line up, it’s actually possible to interchange the two types of sensors! When in doubt, verify that you have the correct part number. All Toyota air/fuel sensor part numbers start with the prefix “89467”; OE oxygen sensors, in contrast, start with prefix “89465.”
In the case of any of these DTCs, once you’ve ruled out heater malfunctions (be sure to check for pending heater codes) and incorrect parts, it’s time to dig in for more detailed diagnostics. As always, be sure to read and record all available freeze-frame data and to check monitor completion status at the outset. If someone has recently cleared DTCs or interrupted the keep-alive memory (KAM) circuit, some monitors may still show incomplete, depriving you of potentially useful information. But remember that monitor completion does not signify a passing grade; it merely means that the test sequence has been executed one time. Fortunately, the monitors for the A/F sensors and heaters run fairly quickly and easily in most instances, so a little driving may be all that’s required to bring their status to completion so you can check for pending codes.
Now it’s time to look into datastream. The relevant PIDs here are AFS1, STFT1 and Equivalency Ratio. Of course, in multibank engines, you’ll need to substitute or add the bank 2 PIDs as well. The first problem you’re likely to encounter is accessing the A/F sensor PIDs at all. These may show up in certain generic interfaces, but they may not look quite right. See “Through the Looking Glass” on page 57 for an explanation of why.
If you have access to the actual A/F sensor PIDs, it’s probably in a manufacturer-specific interface. Look for them under the User Specified or the Extended data lists. Still not found? Depending on your scan tool, you might try headings such as A/F O2 Data or Fuel and Emissions. The relevant PIDs may show up in some generic interfaces, often labeled as WRAF (wide range air/fuel sensor), but are not accessible with all scan tools.
To achieve a fast enough update rate on your scanner, once you’ve reached the A/F sensor PIDs, you’ll want to deselect all other parameters except the ones I mentioned above. If you can’t access them with your scanner, don’t give up just yet. I’ll be showing you some alternatives later on.
You might be tempted to hook up your scope at this point to monitor the A/F sensor voltage output. Don’t bother; it won’t work! The “voltage” is not a value that can be measured at the sensor. Instead, in response to both the amount and direction of current flow through the A/F sensor, this voltage is generated internally, inside the PCM, where it’s encoded into the datastream.
Since our test-drive sequence is going to involve some hard accelerations and decelerations, have an assistant do the driving while you monitor the data. Or, perhaps better yet, set up your scanner to record the data as you drive. Fig. 2 on page 32 will give you an idea of what you’re after.
On hard acceleration, look for the reported voltage to drop to somewhere below 2.8 volts, reflecting the effects of acceleration enrichment. On a sudden closed-throttle deceleration, voltage should rise at least as high as 4.0 volts, indicating a very lean fuel-cut condition. You’ll note that these voltage trends are exactly the opposite of those we expect from a conventional O2 sensor. As you can see from Fig. 2, your actual readings may well go far beyond this narrow voltage band. I recorded voltages all the way from 2.312 to 4.997 volts. If your readings seem to be nearly an order of magnitude smaller, read the sidebar “Through the Looking Glass.”
If you didn’t reach the voltage thresholds on your road test, start by checking the basic circuit integrity. Unplugged KOEO, you should have 3.0 volts on the (usually white) AF1 wire and 3.3 volts on the (usually blue) AF2 wire at the vehicle side of the harness plug as measured with a voltmeter connected to ground. If these voltages are not found, check the wiring, then replace the faulty PCM.
What’s the norm? On light to moderate throttle, reported voltages will generally fall into a narrow range (perhaps 6.3 volt) centered around 3.3 volts. Normal fuel trims should be 0 610%. If fuel trims are out of line, be sure to check the MAF sensor for contamination or downstream air leaks. Since Toyota allows the rear O2 sensor a surprisingly large authority over fuel trim, you’ll want to test it carefully as well. (Many generic interfaces call out the rear O2 sensor’s contribution to fuel trim separately under a heading such as FTB1S2.)
If you believe a rear O2 sensor issue underlies an excessive fuel trim correction, you can remove the PCM battery fuse to clear the KAM, temporarily resetting the rear O2 sensor’s trim to zero. However, this will also clear codes, erase freeze-frame memory and reset any monitor information, so you’ll want to make sure you’ve read and recorded all of that before taking this step.
Testing, Testing
If your scan tool allows you access to bidirectional control of injector volume, you can use that function to verify the performance of a suspect air/fuel sensor. Enabling this test function temporarily suspends normal closed-loop operation, allowing you to monitor the A/F sensor’s reaction to increased or decreased injection quantity. As you alternately command a 25% increase and a 12.5% decrease in injection quantity, look for the A/F sensor PID to fluctuate from somewhere below 3.0 volts (a rich indication) to somewhere above 3.35 volts (a lean indication). On a test involving an 80,000-mile Highlander, I recorded responses of 3.68 volts at 12.5% lean and 2.71 volts at 25% rich.
Perform this test, toggling between adding and subtracting fuel, a few times in quick sucession. The A/F sensor should respond within 1.1 seconds of each commanded fuel shift. Using the Toyota dealer-level interface, the best choice for this test is the A/F Control test, since it allows a quick toggle from one extreme to the other, whereas the Injector Volume test scrolls in 1% increments.
If your scan tool does not support the Injector Volume or A/F Control tests, you can also try forcing the mixture lean or rich by creating a large vacuum leak or by adding propane. Before jumping to any conclusion, however, be sure you’re taking closed-loop responses into account by monitoring short-term fuel trims as well as the A/F sensor data. Since a good A/F sensor responds quickly to even very abrupt changes in mixture composition, look for significant variation in STFT, rather than in the air/fuel sensor PID itself.
How It Works
The A/F sensor produces a small electromotive force (EMF) which, in turn, may augment or oppose current flow across the nominal .3-volt potential between the AF1 and AF2 terminals. The current flow is positive when the mixture is lean and negative when rich. There’s no current flow in the detection circuit when the mixture is at stoichiometry. Maximum current flow is on the order of 15 milliamps (mA), but is usually much less.
I said I’d show you some alternatives if you couldn’t reach the A/F sensor in datastream. As you can see from Fig. 2, the Equivalence PID effectively mirrors the data from the A/F sensor. This means that you can infer the A/F sensor data by examining the equivalency data, provided there are no current DTCs that might suspend closed-loop operation. It’s also good practice to verify the accuracy of the equivalency data calculation by checking the lambda value of the actual tailpipe exhaust gas composition. The equivalency data from a good A/F sensor should match up well with actual tailpipe lambda measurements. If it does not, check to be sure the datastream shows closed-loop operation on all banks. If it does, but lambda and equivalence values remain mismatched, you can be pretty sure that at least one A/F sensor is bad.
If Equivalence PIDs aren’t shown in your scan tool, there are still at least two ways to check up on an A/F sensor’s performance. Unfortunately, both are a bit cumbersome and not as informative as plotting the A/F sensor data directly. The first check is to examine STFT response to a variety of conditions, such as artificially manipulating the air/fuel balance, as outlined above. Fuel trim corrections that appear to be responsive across a range of conditions are usually a good indication of overall system performance, especially when actual tailpipe exhaust gas measurements verify their accuracy.
The second test is somewhat more complex, but makes for a good cross-check of a suspect sensor. Unplug the sensor connector and carefully jumper the heater power and ground leads to their appropriate counterparts in the sensor pigtail. Make sure these leads are properly separated and do not touch one another. Now jumper the (usually blue) AFS1, 3.3-volt lead to its corresponding wire in the sensor, being sure to keep it well insulated and protected from inadvertent contact with any of the other wires. Use a longer lead for the final AFS2 jumper, winding it exactly ten times around a fairly large socket or other circular object and taping the resultant coil in several places. Remove the socket, then place your low-amp probe around a portion of the circumference of the ten-turn coil you just made.
Set your probe to its maximum sensitivity (usually 100mV/A) and hook it up to your scope or DMM. Bring the engine up to temperature, then observe the amp readings. At idle, there should be little or no current flow. On snap-throttle acceleration, current should flow in one direction, while closed-throttle fuel cut-off decel should result in current flow in the opposite direction. A sample waveform is shown in Fig. 3 above.
The exact magnitude of the current flow is not important here, although you should read a maximum of about 140mA, corresponding to a sensor output current of one-tenth as much, or 14mA. (The purpose of the ten-coil jumper is to multiply the sensor’s small current by a factor of ten so your probe can pick it up reliably even at its lower levels.) It’s this current flow which the PCM translates into its A/F sensor voltage PID, so the purpose of this test is simply to verify that the sensor is functioning more or less as intended, and that current magnitude and direction change appropriately.
An alternative approach is illustrated in the top left photo on page 32. I was lucky enough to scavenge the mating connector from another application, although, as alert readers will notice, the white and blue wires change color in the harness in order to match up to the A/F sensor wire positions. This allowed me to build an extension harness, into which I then added two 1-ohm resistors—one in the blue wire and one in the white. This in turn allowed me to hook up a lab scope to measure the voltage drop across either resistor while a jumper bypassed the other one. The voltage drop across a fixed resistor is exactly proportional to the current. Where the resistor is 1 ohm, a voltage drop of 1mV corresponds to a current flow of 1mA.
I expected to find a recognizable waveform when I toggled the air/fuel ratio via my scanner, but was disappointed to capture what looked mostly like random noise when scoping the resistor in the blue wire while the other resistor was bypassed with a jumper lead. But when I bypassed the resistor in the blue wire and connected my scope leads across the resistor in the white wire, I obtained some useful information, as illustrated in the waveform shown in Fig. 4. This corresponds to the changing current flow as I toggled the mixture rich and then lean.
How does the information from these experiments help us? First, we now know that the air/fuel sensor’s two leads do not function as they would in a classic series circuit. Second, we can use the near-verticality of the waveform’s rise as an indicator of its ability to respond quickly to a change in mixture. (Remember, the PCM sets a DTC for a response slower than 1.2 seconds.) Note, however, that the subsequent fall, as I command a fuel reduction, is less abrupt. This is consistent with the physics of the situation, as the total available fuel volume cannot be decreased as quickly as it can be increased.
An odd mishap led me to discover an unexpected waveform that turns out actually to be diagnostically valuable and repeatable (see Fig. 5 on page 57). The nearly vertical rise corresponds to the 125% air/fuel enrichment command from the scanner, while the subsequent fall memorializes the 212.5% enleanment command. The negative lead of the scope was attached to ground, while the positive lead was connected to the PCM, with the blue wire disconnected. This configuration set a DTC P2238 (Air Fuel Sensor Pumping Circuit Low), but the near-vertical rise and fall indicate a quick sensor response to the commands.
Then there’s the Mode 6 option, which offers a couple of windows through which to view and evaluate the A/F sensor’s performance. Manufacturers have wide discretion over when, how and even if they choose to present Mode 6 data. Toyota allows scan tool access to Mode 6, but there are a couple of important caveats. To begin with, it’s vital that the monitor status shows “complete.” An incomplete monitor may store incorrect data in Mode 6, primarily to serve as a kind of placeholder for the correct data once it arrives. Placeholder data values always show a passing grade because failure data may suspend certain monitors, preventing them from ever running to completion.
Depending on year and model, some Toyotas may not show Mode 6 data from the current trip, even if the monitor in question is not yet completed, until the key is cycled off. These data values are then erased from Mode 6 until new data is generated as part of the next trip and the monitor has completed.
Finally, not all scan tools allow access to all Mode 6 information, even if the monitors involved have run to completion. For example, the Highlander data below was available from only two of the five scan tools I tried, with one additional tool providing data for only one of the two A/F sensors and denying even the existence of the hexadecimal address (MID$10, TID$06) of the Bank 2 sensor response test!
On this Highlander with a constellation of codes, including a pending P1133, I recorded the following information for the two A/F sensors:
MID$ 01 MID$ 10
TID$ 06 TID$ 06
(Bank 1 Monitor Complete) (Bank 2 Monitor Complete)
Value 4.429576 Value 3.462604
Limit 3.997696 Limit 3.997696
TLT 0 TLT 0
Result: FAIL Result: PASS
After replacing both A/F sensors and running the monitors, the recorded Mode 6 values changed to .693692 and .511668, respectively, indicating a substantially faster response time. For those keeping Mode 6 scores at home, the TLT 0 designation indicates a test whose result must not exceed a specified maximum. TLT 1 indicates a test that requires a certain minimum to pass.
Depending on the interface used, you may also find results such as these: MID$ 01 TID$ 06 0B1B ≤4000 or MID$ 10 TID$ 06 0831 ≤4000. These apparent alphanumeric jumbles are, in fact, simply hexadecimal numerals indicating passing grades on the same two tests.
What makes an A/F sensor better than a conventional oxygen sensor? In a word, accuracy. When an A/F sensor functions correctly, it doesn’t report merely rich or lean; rather, it reports how rich or how lean. This information allows much tighter control over the resulting exhaust gas composition. This translates into not only a better balance of fuel economy and performance, but also potentially longer catalytic converter life. This latter consideration can be a major cost containment benefit under the 8-year/80,000-mile federal emissions warranty.
What ails Toyota air/fuel sensors? Like a conventional oxygen sensor, the Toyota A/F sensor can be damaged by contamination or blockage of the sensing surface caused by carbon deposits or exposure to antifreeze, burning oil or leaded fuels. Certain volatile solvents, including those associated with some nonautomotive (high-volatility) RTV sealants, can cause the sensor to fail completely or to behave erratically. And it should come as no surprise that these sensors do not withstand sharp or direct impacts very well. In my part of the country (the Midwest), A/F sensors seem to have an average life expectancy of about 80,000 to 110,000 miles, with the majority of failures stemming from heater circuit faults.
Like any exhaust sensor, the A/F sensor can be misled by leaks in the exhaust system. Sometimes very careful examination may be required to locate small exhaust leaks at the spring-loaded joints between the exhaust manifold and the front pipe/catalytic converter assembly on certain vehicles. Such leaks typically occur on moderate to hard acceleration and may result in erroneous fuel trim commands and related driveability concerns, though small leaks are not likely to result in any DTCs.
Some repair notes are in order here. Replacing a Toyota air/fuel sensor is usually straightforward enough, although the female threads of many a manifold have accompanied the original sensor when it was removed. Fortunately, a standard M18x1.5 threaded insert is widely available if retapping is unsuccessful. And don’t cheap out on the antiseize compound, either; use the high-temperature nickel-based stuff. Remember, the heater will kick out 1200°F or more!
If the connector latch doesn’t seem to engage quite right, or if excessive force is required to plug together the connector shells, stop and recheck the part numbers and the application. It’s possible you may have been sent a conventional oxygen sensor when you needed an A/F sensor, or vice versa.
Make sure the sensor pigtail is correctly routed and that any retainers are properly installed, to prevent damage to the wiring or induced radio frequency interference from other components.
Torque is important; for all applications I found, it was 30 ft.-lbs. Overtorquing can damage the sensor, so remember two important points here: 1. The spec applies to cleaned threads with a light coat of antiseize compound, not to dry or dirty threads. 2. Unless you’re using a deep-broached center-driven socket, you’ll have to calculate the correct setting for your torque wrench, taking into account the offset of your socket.
Conclusions
While a rapidly varying voltage is an excellent diagnostic predictor of a good oxygen sensor, the opposite is true with the Toyota air/fuel sensor. Scan tool interfaces may not offer sufficient diagnostically relevant information in some instances. Using actual tailpipe emissions to calculate lambda values provides a crucial cross-check of sensor accuracy, especially when coupled with equivalency data. A lab scope can be used to successfully measure sensor response time. Mode 6 data can also provide valuable insight into the air/fuel sensor’s state of health.