AfterMarketNews Brake&Frontend BodyShopBusiness Counterman EngineBuilder Fleet Equipment ImportCar Motorcycle & Powersports News Servicio Automotriz Shop Owner Tire Review Tech Shop Tomorrow's Tech Underhood Service Speedville

How Automotive Computers Work and How They Fail

Print Print Email Email

I’ll begin this month’s Mechanic Connection with a case study. Because of warranty concerns surrounding a newly-installed fuel pump on his wife’s ’98 Jeep Wrangler, a customer was referred to me by my local parts jobber. The man’s original concern was that the Jeep was starting rather slowly because “it isn’t getting enough fuel.” Because his small contracting business was seasonally busy, the man had a local repair shop replace the fuel pump.

A few days after the fuel pump replacement, the wife discovered that her Jeep had a dead battery. So the contractor places the battery on a slow charge and returns at lunch time only to find that the battery is still dead. In frustration, he turns his commercial-grade battery charger on a high-boost charging setting with no result. After replacing the battery, he still can’t get the Jeep to start.

Scan tool data is vital for diagnosing PCMs. This data indicates that short-term fuel trim or fuel control is operating within normal parameters.  His next step is to remove and inspect the new fuel pump. He then decides the fuel pump is defective and wants the jobber to warranty the fuel pump. At that point, my jobber referred his customer and his problem to me.

To many amateur mechanics, an automotive computer is a “black box” that mysteriously controls all of the engine functions. In reality, an automotive computer or powertrain control module (PCM) is an information processor that turns electrical inputs from system sensors into electrical outputs into system actuators. As with any computer, the PCM uses a binary code to execute complicated math formulas programmed into its read-only memory. 

When the ignition is turned on, the PCM uses ambient temperature, coolant temperature, and barometric pressure sensor inputs to determine the correct air/fuel mixture ratio for starting the engine. In some applications, critical data like barometric pressure or altitude might be stored in an adaptive memory built into the PCM so the engine will immediately start.

The PCM also activates the in-tank fuel pump for a few seconds during cranking to pressurize the fuel system. As the engine is cranked, the PCM receives a signal from the crankshaft position sensor indicating cranking speed. When the cranking speed exceeds approximately 200 rpm, the PCM re-activates the fuel pump to supply fuel to the engine. At this point, the PCM turns on actuators like the ignition coils and fuel injectors to start the engine and begin operating in an open-loop mode.

As soon as the oxygen sensors located in the exhaust system start producing a signal, the PCM switches into the closed-loop mode, which means that it is now adjusting fuel delivery, spark advance, and idle speed to meet various operating conditions. The PCM also begins running a series of tests on emissions system actuator components to see if they are working. If the PCM finds an emissions-related failure in any powertrain system, it turns on the “check engine” light.

Returning to our case study, it didn’t take much testing to suspect that the PCM on the Jeep Wrangler wasn’t working. First, the orange “check engine” light didn’t momentarily illuminate as it should when I turned on the ignition. Second, my scan tool wouldn’t communicate with the PCM.

Automotive computers are generally tested by the process of elimination. The first step is to eliminate the possibility of faulty power supply and ground connections. Most PCMs have at least one terminal that’s constantly connected to B+ or battery positive and another that receives key-on voltage from the ignition switch. Most also have two redundant ground connections to B- or battery negative.

To find corroded connections, blown fuses or burned fusible links in the electrical system, I began testing each fuse with a volt meter. During testing, I found that the PCM fuse supplying the PCM key-on voltage was burned out. Unfortunately, replacing the fuse didn’t restore scan tool communications with the PCM, so I confirmed B+ voltage at the PCM’s battery terminal and key-on voltage at the PCM’s ignition terminal. I also load-tested the PCM’s two ground wires just to verify a good connection with battery negative.

Most automotive electronic parts rebuilders claim that the majority of PCMs returned under warranty are in perfectly good operating condition. The fact is that, while they’re blamed for many “mysterious” system malfunctions, most modern, post-1996 on-board diagnostics II (OBD II) PCMs are very reliable.

Going back to the original complaint of slow starting, I suspected that the original battery had developed a weak cell, which caused hard starting by erasing the adaptive memory in the PCM during cranking. Of course, low voltage during cranking might also shut off the PCM. So the initial slow starting complaint could have been caused, not by a bad fuel pump, but by a bad battery.

Why did the PCM fail? Let’s go back to the fact that the Jeep’s owner also runs a small construction company and that his battery charger probably is a high-amperage unit designed to start heavy-duty diesel trucks. It might also be an older battery charger that will charge well over the 17-volt limit for modern vehicles with on-board electronics.

When the owner turned his battery charger on high boost, the charging voltage probably spiked well beyond the 17-volt rule of thumb. While most modern PCMs automatically deactivate the alternator in an over-voltage situation, none can withstand exposure to more than 18-20 volts from an over-boost on a battery charger. Could the owner also have ruined the PCM by reversing the polarity on the battery charger? Perhaps, and that’s why many auto manufacturers recommend removing the battery for recharging.

I also tested the lighting and the electrical and electronic components in the system for damage. Fortunately, the scan tool would communicate with the air bag module, which was the only other electronic control module included on this otherwise Plain Jane Jeep. All other accessories, including lights, radio, and windshield wipers worked properly.

It’s always important to determine why the old PCM failed. In some cases, shorted actuators like fuel injectors or ignition coils can burn out the injector or coil driver on some older OBD I computer designs. Although most OBD II PCMs protect their circuits from shorted actuators and other electrical overloads, the PCM can also be ruined by something like a leaking windshield seal dripping water into its case. Consequently, most rebuilders want their installers to explain why the unit is being replaced so the returned core can be accurately diagnosed and repaired.

When replacing a PCM, it’s the installer’s responsibility to inspect the harness connectors for signs of corrosion or mechanical damage. It’s also important not to touch the PCM’s connector pins because the human body can transmit enough static electricity to degrade an interior component. In most cases, it’s also recommended that the battery negative cable be disconnected before removing and replacing the PCM. Last, it’s important to install the PCM in its original mounting to eliminate vibration as a future cause of failure.

When selling an OBD computer, don’t forget that most post-1996 PCMs require “reflashing” to allow the PCM to communicate with other on-board computers and modules. Most aftermarket reman PCMs can be programmed by the rebuilder or stocking warehouse to conform to the vehicle identification number (VIN) application.

 After the new PCM is installed, it’s always important to test engine functions such as fuel delivery and ignition to ensure that all of the sensors and actuators are operating as designed. Some scan tools, for example, feature bi-directional controls that allow the technician to activate the ignition, fuel injection, fuel pump and other actuators for testing purposes. Another quick check is to verify that the vehicle has the correct PCM by using the scan tool to read either the VIN or the calibration number of the PCM. 

In the case of the 1998 Jeep Wrangler, all parameters checked out as they should, so the Jeep with the “failing fuel pump” was ready for delivery.


About Author

  • Chuck Hawley

    Hi my name is Chuck, my son has an 2000 chevy Silverado 5.3 eng. that has starting issues. turn key engine rows good,but does not try to start until the keys is released,then no start.,to give you a clue, he was changing starter, got aliitle rough, broke crank shaft sensor, I rewired this, and it started. Now it has intermittent no start. we have fuel. acts like the old Chrysler with duel ballast resistors

  • thatguyof671

    Hi, I have a 2010 toyota corolla thats cranking but not starting. Before it not starting, I was driving it with a weak battery. It would take two tries to start the car for the first use of the day. And if I was to turn it off, I would have a hard time getting it started. Then one day, the car felt like it was about to run out of gas (loses power continuously). Felt like I wasnt gonna make it home, I then parked it for about 3 weeks. Before that, I was told that it was my ignition coils. So I purchased a new set and installed it. And the problem seemed solved. In less then one week, I parked at a family members house and left the car on. In less then five minutes I returned to my car realizing that it was off. When I started the car, it was bogging like I was out of gas and then it would shut off. I was able to get it going again by continuously trying and made it not so far. Then after trying and trying again to start it, I noticed that the check engine light and everything else was off. Only the open car door light would turn on. (I was able to change the battery at my family members house). Now my car cranks but cant start, and those check engine lights dont turn on but the open car door does. Everything else in the car works though. Would you happen to have any advice on what could be the problem? I googled and learned that it could be my ECU…

  • cj hamann

    I have had a question about high voltage (14.64) to interior fuse block ( all fuses ), and lower voltage (11.74 ) to the PDC box under the hood on a 95 Ford Explorer Sport 4.0 OHV vinx X. Today I removed the Alternator and had AutoZone test the unit on their machine. It passed the ripple test, voltage, rectifier test, lamp/in, but failed the lamp/out test. The AutoZone employee said that the diode must have failed. I then came home and decided to disassemble the alternator and visually inspect it. I then spray cleaned the unit with electrical cleaner and reassembled the unit and reinstall. When I restarted the vehicle the first test at the PDC fuses was 13.42 volts, I then increased rpms and let run a little longer and turned unit off and restarted engine and then my test showed 14.64 volts at all the PDC fuses. I then went for a test drive noticed right away that the vehicle felt like it had more horsepower or response and shifted quicker or with better response. Only did the slightly sluggish response return briefly for less than a minute ( 30 secs). Now I am going to purchase a Alternator and hopefully this never returns. I did attach my 15/2 amp charger to the battery cables at 2Amps for the entire time I worked with the battery out of the system and this may have helped in the system not losing all its memory. Still not sure if a problem can be created by doing this, but the relearning curve on this unit is a little slower than honey on a 0 degree day…thanks ..cj

  • cj hamann

    One other note about the alternator. After reassembly I checked the diodes by testing the meter test leads to the supply post ( threaded post ) and the case for ground and I had a ol ( no reading when tested one way and a reading of 1274 when tested with the leads reversed. According to postings on the internet ,this is suppose to mean that the diodes are good . Not 100% what good relates to, such as all positive diodes are ok and maybe negative diodes are failing ?…Does Lamp/in refer to AC is being produced and diode sends but does not convert to Lamp/out DC ?…Can a faulty diode work intermittently ?….Hoping to see a improvement on a old car, but it is still in excellent shape , and crushing or parting out a vehicle seems like a waste…But buying more gas than a hemi dragster can use seems a waste also..Hope she returns to what someone had in mind when they built her….cj

  • cj hamann

    So far the low voltage at idle to the PDC box has not returned. Even at a cold start in the dark I have tested the PDC fuses and all read the same 14.64 volts. Car responds much better. New Alternator arrives Monday and the TPS a little later. I believe the erratic shift behavior with the new vale body in the 4R55E trans with the Ford Update according to the TSB is all related to the changing voltage from the alternator. The GEM module has also worked a little erratic and this would be confirmed through the Bluetooth Diagnostic device from EasyDiag Plus 2 tester. Even though this is a OBD1 system with a Obd 2 connector inside the car you can still test the components through the GEM module if you open the EasyDiag vehicle window and look up manually for a FORD Explorer, and NOT Explorer Sport ( even though it is). Explorer gives you the options of engines and Sport only gives you SOHC engine version. You can fully test the unit and even let the GEM module run it’s own test.This can be very helpful when you want a fault code or if you create one to verify the operation of the Module and it’s functions. It really is amazing that after all the time and effort put into all the testing and replacements of parts, that if I would have first would have tested the Voltage when engine running and tested a cold and hot voltage I may have found this earlier and saved alot of gas and trouble. There was a few times when testing that the voltage was with in or very close to specs, but not low enough to say it was faulty. Keep in mind all I have is a digital volt meter and a vacuum gauge and compression tester with a few wrenches, not your ideal way of repair but enough to test the items as needed. Not every one can dissemble a Alternator, but if needed you can verify voltage at the fuse block in the PDC and the interior fuse block and verify the voltage, and then happily go to your parts guy and ask I hope you have the part I need, and you know if they do not they can probably find it for you….thanks for the help….cj

  • cj hamann

    1 note about the Alternator before reassembly is that the Inner Brush location on the Armature has a small groove cut in it from the regulator brush. According to a site on YouTube this is the Negative side of the charging system ( may refer to the Lamp/out test, not clear on this).The outer brush location would have to be the Positive side, if this is the case then maybe the groove is making the brush act erratic and overheat the brush holder for the brush to the regulator and create a variable that affects the voltage to the PDC box. This would explain the different engine performance and also the way the PCM learns what relays and or sensors are sending it. The PCM would then have different learning curves that would constantly change, even though a engine code and or light is never produced. Remember to test the voltage at each side of the fuse in the PDC box and interior box ” WITH THE ENGINE RUNNING ” to verify if your Alternator may have a problem. I have also taken each ground and cleaned them earlier in the past spring and reconnected and still this problem was still there. Maybe it all started with bad grounds, but all the connections were clean and only a notice of surface corrosion was observed…Hope this helps…cj

  • cj hamann

    PS the voltage now at PDC box under hood is a 11.94 volts.

  • Shanan West

    I have a 2002 ford escape 6 cyl 4wd…we replaced coils and plugs, oxygen sensor map sensor a burnt up Egr Sensor and hose. We also reset the trouble codes… then we followed some youtube video instructions to clear the pcm and drive the vehicle to get it to remember . The vehicle seemed to do great for a day or two… and now on two occasions it has failed to take off at a green light.. it gradually started to move and then when I seen another vehicle approaching as I slowly drifted onto the road, the vehicle would not stop. I seem to recall hearing a zing sort of noise upon acceleration before working on the vehicle.. kinda noise like a homes outdoor heat pump’s fan hitting the cage. I don’t know where to go from here. HELP Please ..before I get killed in this thing.

    • Okc Dave

      That makes no sense unless you have multiple problems. First you never clearly described the initial problem. Second you don’t describe what failed to take off at a green light means. Does it mean engine seems to run fine but transmission isn’t locking up? Is it even auto or manual? What does “slowly drifted onto the road” mean? Is it important or just info we don’t need?

      What does “vehicle would not stop” mean? The prior has nothing to do with the state of your brake system, except if engine dies then you have diminished vac assist and a hard brake pedal to get it to stop. “Zing sort of noise” is nothing I’ve ever heard described before, would randomly guess that’s transmission related, except you mention a fan hitting a cage which makes me wonder if an additional problem is your fan really is hitting something.

      If you feel it is dangerous, don’t drive it, for your sake and that of fellow motorists.

  • Hello, I have been getting ripped off by this mechanic (I think) every time I take my 2006 Pt cruiser in for something not working it is always something really expensive. The last time I had it in my transmission wouldn’t change gears it is automatic it would hesitate and sometime not change at all He told me that it was my computer. That the transmission wasn’t talking to the computer. So all together I paid around 1200.00 that time. He also put a thermostat in cause my when my daughter pulled over it was over heating so I guess that is why he did. But he said he could not get it to over heat. He did some other adjustments. But anyway its been about a month now the car has been running fine but my daughter was backing out and the car just stopped. The lights and dash are all on everything is working. Only thing it acts like it isn’t getting any gas. It really tries to start but it will not. Now my question is we don’t know if he programmed the computer. Another thing and I don’t know my son n law said it was getting gas how he knew that I don’t know. but my question is will the car run if the computer is not programmed. because we took it to Firestone first and they couldn’t figure anything out because they said the computer was reading a lot of codes. What that means I do not know. so now the car is back at the guy who charges me an arm and a leg. He has the car so he don’t over charge me again. And this is another question if and I say if because he said a wire must of come loosed. I am a dummy when it comes to a car I just know it has four wheels and you push the gas peddle and it goes. But if I have to have a gas pump about how much should it be with labor just a ball park figure. So I really had two questions sorry could you help me so I don’t get robbed again. Thank you!!

    broke lady.

  • Levon

    I bought a 2006 Jeep commander on a note and after 3 weeks I noticed the transmission seemed to be slipping and rpm ran high into 3,4 thousand rpm at 30mph. I took it back to the dealer for a warrenty coverd repair, and 3 weeks later, I am still driving a tental car. Turns out that when uninstaling the trans, the somehow burned out the engine wiring harness and the pcm among other countless items that have yet to be discovered. I was told that they had to send the Jeep to a different dealer to teprogram the pcm. Once it got there though, the new pcm was found to be fried as well! I dont want the truck back, i only want my deposit back and the months truck payments refunded, and to supply me with a similar vehicle in the same price range! I believe due to the type of damage the caused to the pcm, and engine wiring harness, that its a total loss ! They keep promise me twice a week that they can get thing back to normal, but every 3rd day they find a new malfunction caused by the burnout that they caused removing the transmission. Why did they not disconnect the battery? Why did the attempt to start the car without a transmission? Why does the pcm keep failing after installing a new one? What is it gonna take to get them to admit that they totaled the truck, instead of promising me that ill get it back in good working condition? What steps need to be taken to get me out of that truck and into a similar suv and transfer my note to that new vehicle? Should the owe me the cost of the truck, plus the lost time of one monthe and car note and insurance that was paied during the time that they tried to cover up their major screw up ! ? Pleas help me !

  • Christine

    2001 Mazda tribute v6….engine reved at 3 only in park….got new computer….was fine for 2 months…did it again….returned on warranty got new computer…stopped for a few weeks then ignition problems car woyld not start…pvc hose got changed…computer unhooked and looked at again by different mechanic…when put back together car started fine…mechanic mentioned something about an ehc module or something like that was frying the computer.he said car could last a week or 6 months if I didnt fix it……put two mechanics on phone together and they argued that neither one knows what they are talking about…well decided not to do it…ran fine for 5 months ….now in the morning when it hasnt ran its fine but throughout the day after driven when put in park it revs back up to 3 and wont come down….very aggravating. …need this truck for another 6 months desperately. …UGH…HELP