Truglia is the owner of Car Clinic, a state-of-the-art repair facility in Mahopac, N.Y. He is ASE certified with a M.A. from Columbia University. In the automotive world he has been trained by Technicians Service Training and Automotive Technician Training Services. Car Clinic’s facility is fully equipped with factory-level equipment and services American, European and Asian vehicles, including diesels and hybrids. (All vehicles were diagnosed by G. “Jerry” Truglia, Alex Portillo and Craig Truglia.)
Back in the day when electronic fuel injection and computers in vehicles were first coming out, oxygen sensors were new. Technicians freaked out wondering how they worked and had to figure out new ways to diagnose vehicles. Eventually, they got used to the pattern failures (i.e. oxygen sensors stuck lean, bad heater circuits, etc.) and understood how they worked: the richer the fuel mixture, the higher the voltage and the leaner the fuel mixture, the lower the voltage.
Then, air/fuel sensors came out and also confused technicians worldwide. The sensors worked in the reverse fashion, and instead of switching between high and low voltages, they used a steady, unspecified voltage. As recently as 2010, a young, ASE L1 with 12 years in the field told me, “There is no way to properly diagnose air/fuel sensors.”
However, as more time passes, technicians are getting increasingly familiar with diagnosing these sensors. Nonetheless, there are still some changes afoot that are worth tracking.
‘New-school’ oxygen sensors
Many service techs thought that the days of having to learn anything about oxygen sensors were over. Think again. Chrysler vehicles still exclusively use oxygen sensors, as opposed to air/fuel ratio sensors, to give the powertrain control module (PCM) feedback in order to calculate whether combustion is stoichiometric, or in plain English, if the exhaust is rich or lean.
However, intent on not making the lives of technicians easy, vehicle manufacturers are beginning to employ oxygen sensors that read between voltages higher than 1V. In fact, the front and rear oxygen sensors on Chryslers for the last few years, including 2014, are 5V sensors. On an example using a 2008 Jeep Grand Cherokee 5.7L (see Figure 1), the oxygen sensors read between 2.6 and 3.4V when we use enhanced OBD II on an OTC Genisys. Why is the Bank 2 sensor stuck at 3.3V? We’ll see in a moment.
Why is this? Instead of being a zirconium dioxide oxygen sensor, these 5V sensors use titanium dioxide.
It is probable that this fact has skipped many technicians’ notice, simply because they do not look at Chrysler oxygen sensors very often. Chryslers do not like throwing system lean DTCs, even when they have massive vacuum leaks.
The way titanium dioxide oxygen sensors on Chryslers work is that they receive 5V of power from the PCM from what is called an “O2 return” wire. The Bank 1 and Bank 2 sensors share the same O2 return wires and each individual sensor has a “signal wire” which in reality is the ground wire (refer to Figure 2 for a Chrysler wiring diagram, courtesy of Mitchell 1 ProDemand).
This shows how the power/return wire runs through the variable capacitor internal to the sensor and exits as the signal wire to the PCM.
The changes in oxygen in the exhaust create a voltage drop in the sensor, which the signal/ground wire sends as a voltage to the PCM.
The PCM then interprets these voltages to know what air/fuel mixture or catalytic efficiency is. Just like on a standard oxygen sensor, these sensors “switch” up and down in an identical fashion.
Instead of switching between 0V and 1V, Chrysler specifies that they switch between 2.5V and 3.4V. When using OEM Enhanced data on a scan tool, or a labscope, this is how it will appear.
However, in generic OBD II the sensors simply read between 0V and 1V and work just like a classic oxygen sensor. As long as the technician uses generic OBD II to diagnose these sensors, he won’t even notice the difference. See Figure 3 for an example on a 2014 Dodge Charger 3.6L, where the front oxygen sensor worked just like a traditional one, but the voltages reflected in generic OBD II, between 0V to 1V are inaccurate.
It is worth noting that the rear oxygen sensors on these newer Chryslers are also of the 5V variety, but they likewise work identically. In Figure 4, you’ll see that the rear oxygen sensor on the same vehicle had a steady voltage of around 685mV according to generic OBD II. We know this is inaccurate, but we know that on traditional oxygen sensors, a steady reading of such a voltage would reflect high catalytic efficiency. After going WOT, the voltage drops precipitously as the fuel mixture is leaned out during the throttle snap. (Figure 4a shows some new vehicles, such as the 2013 Ford Edge 3.5L.)
After testing two different 2014 vehicles, both with the Chrysler 3.6L motor, we found that the update rate on the oxygen sensor was practically instant. The only time testing these sensors might be confusing is when using OEM enhanced data on the scan tool as opposed to generic OBD II.
Oftentimes technicians would rather use generic OBD II because when looking at air/fuel sensors, generic OBD II once gave us numbers that were not accurate for these sensors. While this may be important when looking at an air/fuel sensor on an older vehicle where it would be tough to know whether it is shifted lean or not without having the right specifications, for these Chryslers it is not terribly important. The 0V to 1V scale works fine and is much more familiar to the average technician.
Furthermore, newer vehicles now give the correct specification for the front air/fuel sensor in generic OBD II, or even better yet, give us the mA PID to go by, which is all one really needs.
Being that this is the case, it is best to get into the habit of using generic OBD II when diagnosing all check engine light issues.
2008 Jeep Grand Cherokee 5.7L P2098 and P0153 DTCs
A Jeep rolls into the bay with no driveability problems and two DTCs. A P2098 (Downstream Fuel Trim System 2 Lean) and P0153 Bank 2 Side 1 oxygen sensor slow response appear related. They share a power wire, so an issue with this wire can compromise both sensors. Further, an exhaust leak on this bank can also conceivably throw these DTCs.
However, in Figure 1, the front oxygen sensor appears to be switching normally while the rear one is stuck at 3.3V. No exhaust leaks were found and as a quick down and front oxygen sensor working normally seems to exclude a common power issue.
In order to quickly rule out the PCM, the oxygen sensor was removed to see if the PID went to 5V, which tells us that there is an open circuit.
If the PID is correct, then it is the sensor giving the incorrect feedback, not the PCM misinterpreting it.
Before going crazy looking for common power issues, especially when the front sensor works so normally, it is wise to replace the rear oxygen sensor. See Figure 5: Due to the connectors for the oxygen sensors, this vehicle was tested again after the exhaust cooled down a bit.
The rear oxygen sensor reads 5V as it should unplugged, ruling out the PCM, while the front oxygen sensor gives good data as it begins to warm up and fall into the 2.5V to 3.4V range specified by Chrysler.
Air/fuel sensors and the mA PID
The air/fuel sensor is the most common fuel-control sensor there is and it is probably the only fuel control sensor most technicians come across these days.
Air/fuel sensors work off the principles of Lambda, covered in much detail in my recent article Lambda diagnostics: Solve those lean problems fast, which appeared in the May/June 2014 issue of Auto Service Professional. Here’s a recap:
For all intents and purposes, if you have a Lambda of 1.0 you have a perfect running engine. If you go below 1, you begin running rich. If you go above 1, you run lean. Anything within 0.97 to 1.03 is normal, but if you go above these numbers and the vehicle has a code for fuel trim or a converter issue, it is worth taking a closer look. However, don’t be hyper sensitive. If the vehicle is running fine and has a Lambda 1.08 or 0.95, that could be “good enough.”
Just remember how it works: above 1 is lean and below 1 is rich.
Now, also remember that instead of the Lambda number 1, it is important to know the air fuel sensor’s voltage specification. Anything above that voltage specification is lean and below it is rich.
One of the toughest things about air fuel sensors is that no one tells you what a known good voltage is. Without knowing what your PID should be, it is very difficult to diagnose an air/fuel sensor.
The following are known good voltages for air/fuel sensors compiled over the last few years: 3.3V (Toyota), 2.8V (Honda), 1.9V (Hyundai), 2.44V (Subaru), 1.47V (Nissan), 1.00 Lambda (all European manufacturers). Companies are not always forthcoming with this information, so you will have to compare voltages with known good vehicles. Occasionally, service information will contain such data.
More importantly, remember the number zero in place of 1 Lambda. This is perfect air/fuel mixture when using the mA specification which is now available on almost every newer and new vehicle. For every single air/fuel sensor to date a perfect air/fuel mixture will measure 0 mA. Each tenth of a milliamp above zero is about a tenth of Lambda in the lean direction (i.e. 0.1 mA equals 1.1 Lambda, 0.5 mA equal 1.5 Lambda, 1.0 mA equals 2.0 Lambda, etc) and each fraction of a milliamp below zero is a fraction rich (i.e. -0.1 mA equals 0.9 Lambda, -0.15 mA equals 0.85 Lambda, and etc.) Hence, this works fundamentally the same as emissions analysis.
In the old days, in order to test for mA the technician had to connect his meter in series with the air/fuel sensor in amps mode.
Now, almost all newer vehicles contain a mA specification in generic OBD II. This makes diagnosing air/fuel ratio sensors a breeze, as long as the principles of Lambda are understood and the means to test the sensor, such as a bottle of propane, are available. See Figure 6: The air/fuel sensor in a 2014 Kia Soul didn’t reveal much of anything when graphed. However, it reached 2.6 mA when in wide-open-throttle, which showed a movement in the lean direction, before settling at an ideal lambda of 0 mA.
Figure 7 shows an example of a 2013 Ford Edge perfect Lambda air/fuel sensor reading of -0.016 mA, which is good for a 1.004 Lambda.
Refer to Figure 8, which shows a perfect Lambda as displayed on an ANSED emissions analyzer. The mA reading technically should be rich, but we are talking about hundredths of a mA, which is for too small for a computer to display accurately.
Only worry about tenths of a mA, smaller fractions have no bearing on Lambda. ●