WIXLIB

DEALING WITH TOYOTA A/F SENSORSThis jumper harness is inserted between the vehicle wiringand the A/F sensor. Its 1-ohm resistors allow voltage dropmeasurements exactly proportional to current flow.Fig. 1 This low-amp probe scope capture shows a normalduty-cycled heater pattern on channel A and no currentflow through the inoperative heater on the other bank onchannel B. The probes on each channel were set to 100mV/A.celeration enrichment. On a sudden contamination or downstream air leaks. erase freeze-frame memory and resetclosed-throttle deceleration, voltage Since Toyota allows the rear O2 sensor a any monitor information, so you’ll wantshould rise at least as high as 4.0 volts, in- surprisingly large authority over fuel to make sure you’ve read and recordeddicating a very lean fuel-cut condition. trim, you’ll want to test it carefully as well. all of that before taking this step.You’ll note that these voltage trends are (Many generic interfaces call out the rearexactly the opposite of those we expect O2 sensor’s contribution to fuel trim sepa- Testing, Testingfrom a conventional O2 sensor. As you rately under a heading such as FTB1S2.) If your scan tool allows you access tocan see from Fig. 2, your actual readingsIf you believe a rear O2 sensor issue bidirectional control of injector volume,may well go far beyond this narrow volt- underlies an excessive fuel trim correc- you can use that function to verify theage band. I recorded voltages all the way tion, you can remove the PCM battery performance of a suspect air/fuel sensor.from 2.312 to 4.997 volts. If your read- fuse to clear the KAM, temporarily re- Enabling this test function temporarilyings seem to be nearly an order of mag- setting the rear O2 sensor’s trim to zero. suspends normal closed-loop operation,nitude smaller, read the sidebar However, this will also clear codes, allowing you to monitor the A/F sensor’s“Through the Looking Glass.”reaction to increased or deIf you didn’t reach the voltcreased injection quantity. Asage thresholds on your roadyou alternately command atest, start by checking the ba25% increase and a 12.5%sic circuit integrity. Undecrease in injection quantiplugged KOEO, you shouldty, look for the A/F sensorhave 3.0 volts on the (usuallyPID to fluctuate from somewhite) AF wire and 3.3where below 3.0 volts (a richvolts on the (usually blue)indication) to somewhereAF wire at the vehicle sideabove 3.35 volts (a lean indiof the harness plug as meacation). On a test involving ansured with a voltmeter con80,000-mile Highlander, Inected to ground. If theserecorded responses of 3.68voltages are not found, checkvolts at 12.5% lean and 2.71the wiring, then replace the Fig. 2 This “flight recording” from a known-good car shows volts at 25% rich.typical air/fuel sensor voltage and fuel trim data plotted withfaulty PCM.Perform this test, togglingWhat’s the norm? On light rpm. Note the values of the various parameters at the cur- between adding and subto moderate throttle, report- sor-indicated time of 107 seconds. This is the response to a tracting fuel, a few times ined voltages will generally fall sustained full-throttle acceleration. The A/F sensor has quick sucession. The A/Finto a narrow range (perhaps dropped to a low of 2.312 volts, indicating a very rich mixture, sensor should respond withinshown as a calculated equivalency ratio of .896, or an air/fuel .3 volt) centered around ratio of roughly 13.2:1. By time-stamp 116, on closed-throttle 1.1 seconds of each com3.3 volts. Normal fuel trims decel, when fuel cut-off has occurred, the A/F sensor voltage manded fuel shift. Using theshould be 0 10%. If fuel has peaked at 4.997 volts (actually, the available reference Toyota dealer-level interface,trims are out of line, be sure voltage for this sensor circuit leg) at a calculated equivalency the best choice for this test isto check the MAF sensor for ratio of an ultra-lean 1.234 (an air/fuel ratio of about 18.1:1).the A/F Control test, since it32July 2010

DEALING WITH TOYOTA A/F SENSORSallows a quick toggle fromexhaust gas measurementsone extreme to the other,verify their accuracy.whereas the Injector VolumeThe second test is sometest scrolls in 1% increments.what more complex, butIf your scan tool does notmakes for a good cross-checksupport the Injector Volumeof a suspect sensor. Unplugor A/F Control tests, you canthe sensor connector andalso try forcing the mixturecarefully jumper the heaterlean or rich by creating a largepower and ground leads tovacuum leak or by addingtheir appropriate counterpropane. Before jumping toparts in the sensor pigtail.any conclusion, however, beMake sure these leads aresure you’re taking closedproperly separated and doloop responses into accountnot touch one another. Nowby monitoring short-term fu- Fig. 3 Forming a ten-turn wrap of wire around a socket jumper the (usually blue)el trims as well as the A/F or other suitable mandrel allows your low-amp probe to AFS , 3.3-volt lead to itssensor data. Since a good A/F pick up small current variations, as shown in this screen corresponding wire in thesensor responds quickly to capture. Clamp the jaws around the resulting coil and read sensor, being sure to keep iteven very abrupt changes in the output. Remember to divide your results by ten!well insulated and protectedmixture composition, look forfrom inadvertent contact withsignificant variation in STFT, rather the datastream shows closed-loop oper- any of the other wires. Use a longer leadthan in the air/fuel sensor PID itself.ation on all banks. If it does, but lambda for the final AFS jumper, winding it exand equivalence values remain mis- actly ten times around a fairly large socketHow It Worksmatched, you can be pretty sure that at or other circular object and taping theThe A/F sensor produces a small elec- least one A/F sensor is bad.resultant coil in several places. Removetromotive force (EMF) which, in turn,If Equivalence PIDs aren’t shown in the socket, then place your low-ampmay augment or oppose current flow your scan tool, there are still at least two probe around a portion of the circumferacross the nominal .3-volt potential be- ways to check up on an A/F sensor’s per- ence of the ten-turn coil you just made.tween the AF and AF terminals. formance. Unfortunately, both are a bitSet your probe to its maximum sensiThe current flow is positive when the cumbersome and not as informative as tivity (usually 100mV/A) and hook it upmixture is lean and negative when rich. plotting the A/F sensor data directly. to your scope or DMM. Bring the enThere’s no current flow in the detection The first check is to examine STFT re- gine up to temperature, then observe thecircuit when the mixture is at stoi- sponse to a variety of conditions, such as amp readings. At idle, there should bechiometry. Maximum current flow is on artificially manipulating the air/fuel bal- little or no current flow. On snap-throttlethe order of 15 milliamps (mA), but is ance, as outlined above. Fuel trim cor- acceleration, current should flow in oneusually much less.rections that appear to be responsive direction, while closed-throttle fuel cutI said I’d show you some alternatives across a range of conditions are usually a off decel should result in current flow inif you couldn’t reach the A/F sensor in good indication of overall system perfor- the opposite direction. A sample wavedatastream. As you can see from Fig. 2, mance, especially when actual tailpipe form is shown in Fig. 3 above.the Equivalence PID effectiveThe exact magnitude ofly mirrors the data from thethe current flow is not imA/F sensor. This means thatportant here, although youyou can infer the A/F sensorshould read a maximum ofdata by examining the equivaabout 140mA, correspondlency data, provided there areing to a sensor output curno current DTCs that mightrent of one-tenth as much,suspend closed-loop operation.or 14mA. (The purpose ofIt’s also good practice to verifythe ten-coil jumper is tothe accuracy of the equivalencymultiply the sensor’s smalldata calculation by checkingcurrent by a factor of tenthe lambda value of the actualso your probe can pick ittailpipe exhaust gas composiup reliably even at its lowertion. The equivalency datalevels.) It’s this current flowfrom a good A/F sensor shouldwhich the PCM translatesmatch up well with actual Fig. 4 The waveform in this screen capture provides a into its A/F sensor voltagetailpipe lambda measurements. general feel for the air/fuel sensor’s performance. But is it PID, so the purpose of thisIf it does not, check to be sure enough on which to base your diagnosis?test is simply to verify that34July 2010

DEALING WITH TOYOTA A/F SENSORSthe sensor is functioning more or less asintended, and that current magnitudeand direction change appropriately.An alternative approach is illustratedin the top left photo on page 32. I waslucky enough to scavenge the matingconnector from another application, although, as alert readers will notice, the36July 2010white and blue wires change color in theharness in order to match up to the A/Fsensor wire positions. This allowed meto build an extension harness, into whichI then added two 1-ohm resistors—onein the blue wire and one in the white.This in turn allowed me to hook up a labscope to measure the voltage dropacross either resistor while a jumper bypassed the other one. The voltage dropacross a fixed resistor is exactly proportional to the current. Where the resistoris 1 ohm, a voltage drop of 1mV corresponds to a current flow of 1mA.I expected to find a recognizablewaveform when I toggled the air/fuel ratio via my scanner, but was disappointedto capture what looked mostly like random noise when scoping the resistor inthe blue wire while the other resistor wasbypassed with a jumper lead. But when Ibypassed the resistor in the blue wireand connected my scope leads across theresistor in the white wire, I obtainedsome useful information, as illustrated inthe waveform shown in Fig. 4. This corresponds to the changing current flow asI toggled the mixture rich and then lean.How does the information from theseexperiments help us? First, we nowknow that the air/fuel sensor’s two leadsdo not function as they would in a classicseries circuit. Second, we can use thenear-verticality of the waveform’s rise asan indicator of its ability to respondquickly 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 Icommand a fuel reduction, is lessabrupt. This is consistent with thephysics of the situation, as the total available fuel volume cannot be decreased asquickly as it can be increased.An odd mishap led me to discov

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 app