 | |||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||
|
Codeless in the Shop: How Do You Begin to Troubleshoot?Posted 3/1/2007By Jeff Bach
Being a specialist can be both a blessing and a curse in this industry. When you've seen as many weird problems and difficult-to-diagnose issues with some of the cars that I have, it's hard to imagine not having a certain feel for a problem that points you in the direction of the shortest route to the fix. The majority of the problems I find have made their way through the doors because of a malfunction indicator lamp (MIL), an obvious symptom or a "barely made it here" statement from the customer. These problems can sometimes be difficult to pinpoint without the aid of the proper tools, including both an information source and a technical database. I consult these for technical bulletins and known similar problems. I sometimes find myself in the position of playing technical adviser to a number of technicians, most of whom are very skilled technicians in their own areas of expertise but like me, aren't afraid to tackle problems outside their familiarity zone. Trying to get a feel for a problem over the phone can be difficult without having the advantage of that all-important test drive. It's that feel that sometimes will let you know almost by second nature what a problem is without doing any testing at all. Take, for instance, that flat spot you felt on that mid '70s car that had a carburetor. I have to admit I can remember when the best test I had to determine the difference between a bad accelerator pump and a pole piece wire that was separating due to the movement from the vacuum advance, had you risk getting a nonOSHA-approved haircut and eyebrow trimming! The feel was nearly identical though, and it was easier to take off the lid of the air cleaner and watch the accelerator squirt than to pull the vacuum from the advance and go for another test drive. That feel is something that's hard to teach. It comes from years of flat rating and trying to decide the quickest way to diagnose and repair one vehicle while moving on to the next one. I can't remember ever getting paid for diagnosis. Knowing what was wrong with a car was just expected. If you were a good mechanic, you just knew. Diagnosis wasn't a service you could charge for. Not until they started coming out with those dinosaur diagnostic machines that could tell you everything wrong with your car just by hooking up all those leads, did diagnosis become a commodity. That was before the days of self-diagnostic computers that analyzed the system for you. They would spit out a code for the bad component with the touch of a few buttons, right? These days, it's a little more like playing "20 questions." The computers on today's cars contain a wealth of helpful information that can greatly increase the odds of finding an elusive driveability problem without having it eat your lunch - if you have the right resources. When I encounter an OBD-II vehicle with a driveability issue and no MIL, I like to know some things about the vehicle before I begin doing my testing. One of the important questions I like to have answered is: What repairs, if any, were made to the vehicle since the problem first appeared? If repairs were made, is the car currently doing what it was when the symptoms first appeared or has it developed new symptoms? I call this "Wasnowits syndrome" (that is, it was doing A, now it's doing B and C). If so, what took place prior to the new problems? "Wasnowits" can sometimes be cured just by resetting the long-term memory or clearing the powertrain control module (PCM) learned fuel strategy. I've seen an oil change clear up driveability issues (fuel in the oil) that had me chasing my tail. The fuel in the oil was entering the engine through the positive crankcase ventilation (PCV) system, creating a rich condition. This is fairly common on a vehicle that has had extensive testing or cranking with no spark. You might be surprised at how difficult it can be to trace a driveability issue that stems from a bad battery or an alternator problem.
By adding fuel, then creating a slight vacuum leak, you can see how well the O2 is responding. This is especially helpful when doing bank-to-bank comparisons. I had a customer's '96 Cadillac Deville with a 4.6 Northstar engine that came in with a slightly rough idle complaint. No codes and very little else was noticeably affected. One bank did, however, have a higher LTFT. After doing all the usual testing for a bank specific problem, we finally found that one of the four cams had been timed one tooth off. One of the less common but difficult to find problems I run across is a coolant sensor circuit that goes askew intermittently. This problem usually results in hard starting because of over-fueling cold; however, as the engine warms up, the sensor begins reading normally. Connection issues and corrosion problems have a tendency to become worse as the cold affects them. As I write this article, I happen to have two very similar Cadillacs, both with 4.6 liter Northstar systems, on the lot with similar symptoms and different causes. My favorite test for this type of situation is what I call the first pulse test. To do this, you need to hook a current probe to the injector circuit and set the scope to trigger a single shot on the first pulse. Depending on coolant temp, this pulse can be fairly wide (see Figure 1). This pulse width at 52 mS is normal for this temp (20 F). A lot of fuel injection systems I've seen use a spritz pulse, which opens all the injectors for a wide first pulse then drops back or off for the next few cranking ignition pulses based on coolant temperature. The Sias car in Figure 1 is showing 52.75 mS. This one had a bad fuel pressure regulator, which caused the hard start. The Ayers (37mS) car also in Figure 1 shows a 34 F temperature, which is 14 degrees warmer than it's been all day, and the car hasn't been running since it was towed in last night. Neither car had a code stored and both ran well when warm. Without having the ability to compare them this way, the injector pulse alone would probably not have tipped me off that there was an issue with the coolant sensor. This time of the year, most of the driveability issues we're seeing are cold engine related, and much of the time the problems seem to go away as the engine gets warmer. The last few weeks here in Ohio, the temperature hasn't even reached the 20s, and some nights we're seeing single digits. In these conditions, flooding occurs easily and any testing that requires engine cranking is best kept at a minimum. The conditions are great for stimulating quick test ideas. Some of the questions I find myself asking techs when I'm doing phone support on fuel-related driveability issues are:
I've had this happen a time or two, usually from an O2 sensor that was shorting due to oil contamination on the wires, allowing the circuit for the heater feed to contact the O2 signal circuit and feed the PCM a steady, rich signal. Here's another clue that a skewed sensor may be feeding false information to the PCM. The complaint is that the car develops its driveability issue after the system goes into closed loop. Here the PCM seems to be controlling the mixture fine on preset values until it goes closed loop and starts taking in and using information from the sensors. If disconnecting a sensor that puts it in backup mode improves the driveability a great deal, then start looking closely at sensor inputs for one that is skewed. The easiest way to find some of these driveability issues is to simply wait until they get bad enough to set codes. Sometimes I wish we could just be like doctors and say "drop one of these pills in the tank every day for two weeks and call me if the symptoms get worse." I had a mentor years ago when I was at the Cadillac dealership that had a son in medical school. He said his son was taught not to waste time doing a lot of testing on patients with mild symptoms. Just prescribe some type of medication to stall them for a couple of weeks. He said 90 percent of all ailments will either get better on their own or bad enough to make them more diagnosable within 10 days. Either way, the doctor gets paid and still gets the credit.
|
||||||||||||||||||||||||||||||||
Copyright (c) 1996-2009. Automotive Service Association®. All rights reserved.
| |||||||||||||||||||||||||||||||||
How recently was the PCM cleared? Once cleared, it can take a good bit of driving under different conditions to establish long-term fuel trim (LTFT) numbers. These are helpful in determining whether the PCM is adding or subtracting fuel to compensate for an external air or fuel leak. High LTFT numbers at idle that drop close to normal at say 1,500-2,000 rpms can indicate a slight vacuum leak, which is quite common on many cars, especially where plastic intakes are involved. An overall high LTFT may be an indication of a mass air flow (MAF) sensor gone awry or possibly a weak fuel pump. A weak fuel pump will usually cause LTFT to start to increase with engine load and may drop toward normal at lower speeds. When an engine has dual O2s and the ability to control each bank individually, LTFT can be a big help in finding the difference between an issue that affects the engine overall such as MAF fuel supply, engine temperature, air intake or coolant sensors, air injection reactor (AIR), exhaust gas recirculation (EGR), etc. versus a bank specific problem that affects one side more than the other such as a vacuum leak injector problem, bad O2 sensor, etc. I have fought hesitation problems that resulted from lean conditions on engines that were running low fuel trim numbers due to the O2 sensor signal circuit leaching current from the O2 heater circuit, basically telling the PCM it had too much fuel. The way I found it was by having a current probe hooked to the injector feed circuit on one channel while monitoring the O2 signal with the other and watching the interaction between them.
