Tech Feature: Tune-Up and Emissions

Fixes From The Edge

Sure, you’re not a technician, but when a fix goes bad, you’re likely to get at least part of the blame. At the very least, the technician may just want to dump on you out of sheer frustration.

At times like these, it is best if you understand some of the frustrations that can come with even a seemingly simple fix, and the hair-pulling that accompanies problems that defy logic.

Here are some examples from the real world.

Take, for example, an ailing Honda. A technician is faced with an old Honda Accord, a 1985 model with fuel injection. Now, he doesn’t want to end up spending more on the car than it’s worth, of course, but it is a devil car.

Towed in with a no-start after it ran out of gas, but it wouldn’t start after gas was added. Good spark, engine cranks, only fires up when the tech sprays carb cleaner in the intake. Plugs are good, voltage okay to the injectors, noid light check is okay, ECU checks out on the lab scope, injector grounds too, and the power to the fuel pump is good. After sitting awhile, the car started and ran for an hour. Fuel pressure was 32 to 40 psi, which is a good range. Perhaps the fuel filter? Perhaps bad gas, which was dark and stank, was the problem. Perhaps the fuel pump was weak on volume. Replaced them all, then the no-start returns. The car runs fine for a few minutes at a time, with no codes, and then starts to crap out. About this time the tech is about ready to sell his soul for the solution (I told you it was a devil car), but if he called you, what would you say?

Would you suggest that something else must be the problem, that perhaps he tripped a switch, or suffered from brain fade and missed something?

That would surely get you in dutch with the customer, but in this case it was exactly what happened. The tech had done a few checks of the system that were a waste of time because of the voltages it uses, but the real problem was that he had replaced the proper fuel filter bolt with a slightly longer one that blocked fuel flow. Bad gas and a failing fuel pump, it turns out, were the problem, but his bolt error had created a new one that seemed like the same old problem.

Or how about this one involving a 1998 Chevy Lumina with a misfire. With 160,000 km on the 3.1L six, and codes stored, the technician understandably assumed that ignition components were at the end of their useful life, and started replacing items: new plugs, wires, coil pack, injectors, and so on. No change. With known-good injectors and fuel rail, the vehicle would run fine for about 10 minutes, then start misfiring. A new coil pack and module had no effect. At wit’s end, the shop techs give the customer an oil change and send him on his way. Voila! Problem solved. Apparently, the oil was a bit old and for some reason caused the engine to misfire. In any case, the misfire went away and has not returned. Weird and wonderful at the same time.

In closing, here is one right out of the twilight zone, which shouldn’t surprise considering it comes all the way from California, land of the weird and wonderful.

Consider for a moment the world’s fastest Dodge Grand Caravan. Bucking and stalling, it arrives at the shop of an unsuspecting tech. Confidently he plugs in his scan tool and goes for a ride which, according to the snapshot readings, reveals his velocity at 647 mph, and barometric pressure of zero. Returning to subsonic speeds, he surmises that these values are probably incorrect.

Here is his account from his flight log: “I’ve checked the speed sensor waveform, nice sine wave, I’ve gone through the trans computer and found no codes and normal values. I’ve gone through the ATM, and everything worked fine. When back probing the TPS and MAP there is a .2 VDC signal on the reference voltage wire, key off. The MAP and TPS both test well dynamically KOEO and KOER.”

Two test computers later, the problem is still there. The engine never lost fuel pressure or spark when it was bucking, and on both the computers there was a .2 to .4 VDC signal on the MAP and TPS signal circuit with the key off. So where to now? Back, it turns out, for another computer.

Though the tech didn’t know it at the outset, Chrysler had made several updates to the computer for this model–to deal with stalling issues and odd readings no less. The two computers he had tried did not contain the updates. The moral, from the technician in the twilight zone, was not to trust that a dealership’s test computer, or a bargain basement-priced aftermarket unit, would have the appropriate updates.

There are certainly many cases where even the best technicians can be stumped. You should, however, be on the lookout for technicians who seem to have the same diagnostic problems time and time again. They are candidates for training in proper diagnostic procedures, and perhaps may need some of the right tools.

In an time where a single oxygen sensor can cost hundreds of dollars, it is not acceptable to keep throwing parts at the problem, in or out of the twilight zone.

Oil Leaks and the Oxygen Sensor

Oil can play a role less intriguing than the Chevy Lumina’s problem, but can be just as effective at killing an oxygen sensor.

The position of the sensor near the likely source of oil collection, and the way it is oriented, can cause oil to collect around the sensor.

At this point it is useful to remember that an oxygen sensor requires an outside air source in order to provide the comparison to the oxygen content in the exhaust stream. Block the passage with grimy oil residue, and the sensor will fail to provide the proper voltage signal back to the fuel control, and you will end up with excessive emissions.

In cases where a vehicle is suffering from excessive oil leakage, it is not inconceivable that a sensor can show signs of failure in only a few months. Technicians should fix the source of the oil leak if a failed sensor is found to be oil-soaked.

Oxygen Sensors May Fail Silently

One of the issues surrounding oxygen sensors is their typical failure mode.

While many other components on a modern emissions system will trigger a fault code, and in turn cause a warning light to illuminate on the instrument panel, this is not always the case with an oxygen sensor.

Often, the first clue that a consumer will have that there is a problem is when their car fails an emission test. In fact, according to Robert Bosch Inc., one study even found that 99.7% of consumers did not even know their vehicle had an oxygen sensor. Afterward, they may be counselled on the fact that their fuel mileage was off, but the oxygen sensor may have degraded so gradually–often they become sluggish and slow to react rather than failing outright–that the car owner fails to notice.

In fact a failed oxygen sensor is, according to a study quoted, once again, by Robert Bosch Inc., “the single greatest source of excessive emission for fuel-injected vehicles.”

A quick glance at the chart reveals why. Since the oxygen sensor is the primary input for the air-fuel (A/F) ratio, when it degrades or fails entirely, combustion conditions aren’t optimized and combustion by-product levels can skyrocket.