By Brian Manley
Student Brian Parker with the Geo Metro. |
One of my Advanced Auto students, Brian, works part time at a service station to gain experience and make a little money. Brian and the other technician at the station had been working on an overheating and no-start issue on a 1992 Geo Metro for several weeks, when he told me the owner was going to send it to another shop for diagnosis. “What? Have it towed here!” was my response to that. I didn't want to pass up a perfect car to use for demo in my Engine Performance unit. I immediately called the owner and talked him into towing it to our shop at the school, guaranteeing him that we'd diagnose it and, because I'd be using his vehicle for troubleshooting demonstration, I wouldn't charge him anything. He couldn't lose!
Spark tester and noid light. |
Once in our shop, I popped the hood to see some shiny parts: the radiator, water pump and ignition coil all looked new. Brian further explained the thermostat had been replaced and they had replaced the head gasket - all because of the overheating condition. The ignition system had been tested, and a rebuilt distributor had been installed, without any effect on the intermittent no-start status. The intermittent running-and-dying condition had also recently developed into a full-blown no-start, which we confirmed.
The two-wire distributor with frayed insulation - a good visual check. |
Brian checked the coolant and oil levels, which were fine, then hooked up the spark tester and noid light. There was no trigger to either one. He then plugged in the scan tool to find a code 42: Crank Angle Sensor Circuit - No Signal For Three Seconds. As we studied the schematic for this code, we saw one distributor pickup assembly, two wires connecting it to the PCM, and the PCM itself. The distributor on this Geo has just two wires, so this was a great time to show my students how to test a magnetic reluctance sensor, and then move downstream in our testing to isolate the root cause.
The combustion leak test. |
When the distributor shaft rotates, a fluctuating magnetic field is generated due to changes in the air gap between the crank angle sensor and the distributor shaft reluctor. As a result, a small AC voltage is induced in the crank angle sensor, which then travels to the PCM through terminals B1 and B9. The PCM monitors this signal, and if it falters for more than three seconds, it sets the code 42. No signal, no ignition coil or injector trigger. The students thought it all made sense so far.
This stray rubber plug was found in the lower radiator tank. |
We first tested the resistance of the pickup coil, which was 178 ohms, right between the 140 to 180 ohm spec. I explained to my classes that it still needs to produce an AC voltage of at least 100 millivolts before we can call it “good.” It made just over .1 volt AC while cranking, within the .1 to .4 volt spec. The next step was to confirm the signal at the PCM, which was still 100 mv. Well, the final box in our diagnostic chart now called for a replacement PCM, so we tested all of the powers and grounds that feed the PCM, and the solenoids and injector that were controlled by the PCM. They all tested well, so Brian went off to find a PCM. He returned a week later with one of only two Geo PCMs in the state of Colorado. It cost $100; the new one cost $700.
The original and the used PCM. |
With a used PCM plugged in we turned the key and the engine fired right up! Now the next step was to run the engine and monitor the cooling system, which promptly overheated after less than 10 minutes of running. OK, let's begin listing the possibilities for an overheating condition:
- Low or diluted coolant
- Water pump and belt tension
- Restricted radiator
- Restricted airflow through the radiator
- Inoperative cooling fan
- Defective thermostat
- Ignition timing maladjusted
- Poor circulation
- Combustion gases in the coolant
- Collapsed lower hose
We assumed the water pump was good and the t-stat was working, so we searched for other possibilities. As the engine ran, the coolant circulated, but not that well. I wondered about the head gasket; could it have been installed backward, blocking the coolant holes? Brian assured me the gasket only fits one way, so I started wondering about combustion gases. Did the engine overheat enough to crack a head?
The new ignition coil and the fuel injector resistor. |
Even though the head had been leak-checked at the machine shop, I asked Brian to perform a combustion leak test. The result was negative. We refocused on the fact that the coolant flow in the radiator seemed a little weak, so Brian pulled both radiator hoses and the t-stat for examination. We clamped a garden hose to the upper radiator hose, which was still clamped to the radiator top neck, twisted the cap on, and turned on the water. The pressure pushed water into the reserve tank, but very little came out the bottom of the radiator! What Brian pulled out is shown below: a stray plug that was pushed into the lower tank.
Touching the PCM wiring harness caused the engine to die. |
After reassembling the cooling system, the engine purred along and ran mid-range on the temp gauge, until ...
With the car running, it was time to button up the PCM and all of the dash parts, so under the dash I went.
The fault is isolated to two wires - the two that come from the distributor. |
The moment I moved the PCM, the car faltered and died! I was grounded to the chassis, so it couldn't be static electricity; what's goin' on here? I plugged in the original PCM and the engine still didn't fire, so I plugged the “new” one in and it started up!
Removing the female PCM terminals for test fitting. |
After quite a bit of trial-and-error, I found that the car would run if I moved the harness to the left, but not if I moved it to the right. I knew that many General Motors PCMs have surface-soldered connections that were questionable, so I tried my best to just move the circuit board and nothing else. The car didn't die. I finally found that moving just two wires would cause that car to stumble and die. Which two wires were they? Terminals B1 and B9 - the same two that come from the distributor pickup coil. It was all making sense to me now ...
The test fit! |
I removed the two terminals from the PCM connector so I could test fit them on the PCM terminals. Both of them had no drag whatsoever. I searched through my assortment of GM connectors, but I didn't have any that matched (because this was a Japanese PCM). I ended up carefully reshaping the two terminals so they would fit snugly, and all was well.
What is the lesson in all of this? As I always tell my students, it's always something, and most of the time it's something simple. I often preach that if you think the PCM is the problem, you're probably not testing the system thoroughly enough (with the exception of many '90s model PCMs). Turns out that sometimes the PCM is the problem.
Systematically walking through this diagnosis with Brian and other students showcased the fact that if techs understand the system they're working on, have a good diagnostic process, and can use their test equipment to their fullest potential, then virtually any customer concern can be diagnosed and repaired. We all just have to keep fine-tuning those skills - and that applies to me, too.
