Intermittent faults with no definable pattern can be the toughest to repair. But sometimes you get lucky.
This poor customer had been in twice before with the same complaint. When coming to a stop, the engine would die like she had turned the key off, but only occasionally. The car would not restart when this happened, and she would have to have the car towed into our shop. On the last two visits, the car would start without fail when it arrived, leaving our techs scratching their heads, trying to figure out what was causing the problem. You've never experienced a situation like that, now have you?
INFORMATION, PLEASE!
I was fortunate enough to be in the service aisle when the car arrived for the third time, and I took the opportunity to talk directly with the customer. Had she noticed that the problem occurred only when first starting the car (when cold), or was it only after she had been driving for some time (when hot)? Did it ever fail when driving or only when she came to a stop? These were just a few of the questions I had for her in an attempt to see if there was any pattern to the failure I might be able to duplicate.But not in this case. According to the customer, it happened without rhyme or reason. It had been more than a week since the car was last in the shop, and she had driven more than 1,000 miles without a problem and without any changes to her normal driving habits. The last work done to the car was the installation of a new ignition control module and replacement of the crankshaft position sensor. She thought the problem had been solved – until she stopped to make a left turn and it died again, that is.
I took the keys and went out to the car. When I tried to fire the engine, all I got for my trouble was the characteristic "click" of a dead battery. Apparently, in her frustration, the customer had run the battery down in her attempts to restart the car. Or could it be a new problem in the charging system that led to the stall? I got some help and brought it into the shop, connected the battery charger, and charged it while I attended to other business.
IT COULDN'T BE THAT EASY
That afternoon, I went back to the Pontiac. The battery tested OK after charging, so it was time to try again. Based on this car's history, I expected the engine to crank right up.Fortunately, it wouldn't start. I say fortunately because at least now I had a fighting chance of finding this problem. The engine cranking speed sounded normal with no audible evidence of a mechanical failure being the cause. Next, I connected my fuel pressure gauge to the fuel rail fitting, and verified that fuel availability was also not an issue. OK, I have fuel in the rail and the engine has sufficient compression to at least attempt to start ... how about spark? Connecting my spark tester to an ignition lead, I cranked it over and got nothing.
THIRD TIME'S A CHARM
VEHICLE: 1998 Pontiac Grand Am DRIVETRAIN: 3.8 liter V6, automatic transmission MILEAGE: 83,547 miles COMPLAINT: Intermittent stall when stopping, unable to restart when failed.Now we're on to something. An engine that stalls just like "someone turned the key off" is a characteristic symptom of ignition-related failures. While I knew I was going to take a closer look at the ignition system, I wasn't quite done with my basic "fact gathering."
Next on the basics list is injector operation. This test can be done with a lab scope, a "noid" light or even a stethoscope, listening for the "click" of the injector opening and closing. I used the last as a quick check and heard nothing. I verified the lack of injector action with my scope and began to feel I was dealing with an electrical issue and not necessarily a component failure.I didn't hold out much hope of finding any stored codes because of the low battery level, but it was worth a shot. I connected my scan tool, and entered the powertrain control module (PCM) using the General Motors (GM) enhanced data mode. Of course, there were no codes.
Thinking about how this particular system worked, I knew that the PCM had to see a signal during cranking from both the camshaft position (CMP) sensor and the crankshaft position (CKP) sensor in order to know where the engine was in terms of top dead center (TDC) and cylinder synchronization. Without that information, both the injectors and ignition would not fire.
I also knew from past experience that a lost CKP signal would often not be logged as a code because the PCM would see the RPM drop at the same time the engine died and think that the driver had turned the engine off. But that occurred after the sensor had gotten hot, not on an initial cold start.
I switched to live data on my scanner and pulled up the parameter identifications (PIDs) for the CKP and CMP. On the Pontiac, the CKP actually provides two signals, referred to as the 18x and 3x. I watched the scan data while cranking the engine, looking for a normal reading.YOU DON'T ALWAYS GET WHAT YOU WANT
What I got was peak RPM readings of more than 12,000 and cranking readings that were erratic. Of course, there was no way this engine was spinning at that kind of RPM, and my search area was getting narrower. I obviously had a problem – well, the PCM did anyway – with the CKP and CMP signals the PCM was receiving. With out that information, the PCM wasn't going to let that car go anywhere.
My next question was, "Is the information from the sensors bad, or is the right information just not getting to the PCM?" Here is where a lab scope comes into play. It allows me to see the signal that the PCM would see and monitor its action.
On this particular system, all three signals pass through the ignition control module (ICM) on their way to the PCM and connect via a subharness to the main harness at the top of the engine. Because I was looking for a corrupted signal, I decided to start here, roughly in the middle.I watched my scope as a buddy cranked over the engine. The signals I saw looked like the sensors were certainly trying their best, but couldn't quite hold on. They would start out OK, but almost immediately collapse before trying again. The signal faults were happening so fast I didn't have a chance to freeze the frame. I repositioned the leads and tried again – and the engine started right up.
Wasn't expecting that. I guess that explains the intermittent part.
With the engine running, I took a look at the signals. The CMP and CKP 18x signals looked perfectly normal, as did the CMP and CKP 3x. I tried to get the "no-start" to repeat, but the car restarted every time I tried. Tugging on the harnesses and connectors had no effect, either. It just kept running with no change in what I was seeing on my scope.
LISTEN TO YOUR GUT
Let's review: With the car dead, it had no spark or injector pulse, most likely caused by corrupted or lost signals from the CMP and/or CKP sensors. The signals with the car running all looked normal, with full amplitude and clean cut-offs and pull down all the way to ground.
These signals go through the ICM, so it may be getting lost there; however, that part was new and from the original equipment manufacturer (OEM). I've seen my share of new parts fail, so it was a possibility. But it was not likely, considering the problem existed before the replacement as well. The CKP was new, too. What ever had caused the problem was now gone.
My gut was telling me that this was a wiring or connection issue. These circuits are low current, and it wouldn't take much to block them. Hindsight is always 20/20, and I should have done a voltage drop on the sensors' grounds with the first sign of a corrupted signal. However, with the car now running and the clean drop to ground my scope was showing, I didn't think it would help me much now.I decided to check the connectors before I dove into the harness by checking the connector integrity with a "drag" test. This test is done by using the correct mating pin for the connector you want to test. Using the wrong size can actually damage the connector. I felt for a slight "drag" in each socket of the connector to make sure the tension on the terminals was OK. I also inspected for damaged connectors, corrosion and pins that were recessed in the connector block. They all looked OK.
Now I started looking hard at the harness, and I realized why my gut was telling me "electrical." There had been work done on the top end of this car, and the subharness from the sensors and ICM to the main harness had been misrouted. It was very tight in spots, so much so that the mating connector to the main harness was out of position by at least 2 inches.
Luckily, this harness was easy to remove, and after stripping away the insulation, I found the ground splice from the ICM to the sensors had literally only one strand of wire holding it together. Vibration, harness stress or some similar factor must have been causing this connection to "make or break" intermittently, resulting in a loss of signal from the sensors, and the "stall."
Because of corrosion in other sections of the harness and some rubbed through damage, I elected to replace the harness rather than repair it. Sometimes, careful testing and observation will lead to discovering the cause of an intermittent fault. And then there are times like this one, when gut feeling and a little luck provide the answers.PETE MEIER is an ASE CMAT, a member of iATN, and a full-time tech with CarMax in Tampa, FL. He started doing oil changes and minor repairs more than 30 years ago and brings a variety of experience to bear. His current job involves all manufacturers' lines, and, as Pete says, "provides me constant opportunity to learn something new." Diagnosing electrical and driveability problems are his favorite challenges.
