If you have been a mechanic for say, ten years or more, you have no doubt seen your fair share of weird and wonderful faults (and even weirder symptoms) on a wide variety of vehicles. While such faults and defects often resist all initial attempts to diagnose them, we usually resolve them in the end, but there is another category of faults that sometimes call for special measures such as laboratory tests if we are to resolve them.
This writer has encountered many such exotic faults in his more than four decades of fixing cars, and he has learned a few things about exotic faults and symptoms in that time. However, looking back over his career and some of the mistakes he had made, this writer now wishes that he had known some of the things he had learned thirty years or more ago.
You may wonder where this is going, but this article is not so much about going anywhere, as it is about where it is coming from. Let's explain that- this writer was recently invited to consult on a no-start condition on a three-year-old diesel vehicle with an SCR (Selective Catalytic Reduction) exhaust after-treatment system.
The details of the consultation are not relevant to this article, but what is important about it is that the workshop that was dealing with the vehicle immediately threw its whole arsenal of diagnostic equipment at the problem, which immediately buried the relevant details of the issue under a mountain of irrelevant test data. Judging by the number and variety of exhaust and other parts the shop had already removed from the vehicle when this writer arrived, it was clear that the young mechanics in this workshop had dug themselves into a deep diagnostic hole by likely breaking at last three of the most basic rules of modern automotive diagnostics.
So, against this backdrop, this article will discuss two case studies of weird faults on different cars this writer was involved with in various ways. Before we get to specifics, though, it is perhaps worth mentioning that the case studies in this article revolve around the diagnostic mistakes the young mechanics had made in their efforts to diagnose the no-start condition, so let us begin with the mistake of-
Image source: https://www.vehicleservicepros.com/service-repair/diagnostics-and-drivability/article/21295100/all-in-a-diagnosticdays-work
Our first case study involves a misfire on a 2019 Hyundai Tucson as reported by the scan report shown above, only the engine was most certainly not misfiring- at least, not in a way that was detectable just by looking and listening to the engine running at idling speed. In fact, the engine was running as smoothly as the proverbial Swiss clock, and it was only when the engine speed exceeded about 1000RPM that a barely detectable vibration or roughness in the engine became evident, but this could not be described as a “misfire” in any way- or so we thought at the time.
The funny thing about this misfire was that the vehicle did not come in for the misfire issue; we has serviced the vehicle two weeks before, and there were no active fault codes on the vehicle when its owner collected it from us after the service, which included an oil change and replacement of the spark plugs. The only reason the vehicle was with us was that its owner and his family were going on a long road trip in a few days, and he wanted us to look the vehicle over to make sure nothing untoward had cropped up since the service.
For us, doing full diagnostic scans was a part of looking vehicles over, so there we were; a diagnostic scan had turned up two misfire codes without the engine actually misfiring, so we did the right thing- we cleared the codes and started the engine again. Both codes came right back, which meant that the codes might be real, after all.
Thus, and on the assumption that the codes were real, we started looking for likely and/or probable causes of the misfire codes. We were not too bothered by the random misfire code because even if a cylinder-specific misfire code happens randomly or intermittently, but many misfire detection systems will report such a misfire as a random misfire along with the cylinder-specific code.
Also, since we replaced the spark plugs with the recommended NGK plugs only two weeks before, we did not consider it likely that a new sparkplug would be the cause, but just on the off-chance that the spark plug on cylinder #2 might be implicated, we removed all the plugs to compare them to each other. Consider the image below-
Image source: https://www.vehicleservicepros.com/service-repair/diagnostics-and-drivability/article/21295100/all-in-a-diagnosticdays-work
This image shows the actual spark plug we removed from cylinder #2, but pay attention to the brown staining on the porcelain. The staining was present on only this plug but because staining like this is caused by the intense magnetic field around the metal body of sparkplugs “baking” dust and oil residue that might be trapped in the rubber isolator onto the porcelain, we did not think too much about the fact that only one spark plug was stained. Satisfied that the plug was good, we reinstalled all the plugs and started the engine. Both trouble codes reappeared immediately.
We then began investigating all the most likely causes of misfires on petrol engines. We let the engine warm up and took temperature readings of the coils; all were within a degree or so from each other, so the coils were good. We then hooked up an oscilloscope and a pressure transducer and took captures of the ignition waveforms referenced to the ignition timing on each cylinder, but this test told us nothing, except that a) cylinder compression was roughly equal across all the cylinders over twenty engine cycles, and b) that there were no leaking valves on the engine.
The fuel rail pressure was within specifications, and a fuel volume test showed that the injectors were delivering equal volumes of fuel to the cylinders. We then used a smoke machine to check for vacuum leaks, but we could not find any, which largely obviated the need to do a volumetric efficiency test, especially since the available freeze frame data showed that the misfire occurred at only 1180 RPM.
By this time, two technicians had spent well over four hours on the vehicle, and we were fast running out of options or things to test, but then the issue of the single stained spark plug arose again. Somehow, it seemed strange that only one spark plug, and the spark plug on cylinder #2 at that, showed any discolouration, so on a hunch, we replaced that spark plug with a new one.
Changing the discoloured spark plug solved the problem; the codes did not come back even after we took the vehicle on a long test drive and started it multiple times in rapid succession after the test drive.
It seems that we had overlooked the obvious, which was the fact that coronal staining (discolouration) rarely if ever, occurs only on one spark plug. While we could argue that coronal staining does not typically affect the operation of spark plugs, we disregarded the possibility that the discolouration on the spark plug was not caused by intense magnetism, but by something else, such as compression pressure and combustion gases leaking out of the spark plug.
It is said that hindsight always brings perfect clarity, so to obtain that clarity we sent the spark plug to its maker, who ran some tests on the plug, and which tests confirmed that the plug was indeed leaking minute quantities of combustion gases. They also said that since the ignition spark was directly in the leak path, the leak was probably disrupting the growth and propagation of the combustion flame front, thereby causing poor or incomplete combustion in the affected cylinder at some engine speeds.
Although the test report explained the misfire (that was not an actual misfire), one question remained, and it is this: given that the misfire counter did not record the misfire, why did the misfire detection system report the problem as a misfire, and not as a cylinder power contribution issue?
We could not quite figure this out, so the best explanation we could come up with was that some questions do not have answers, but since we received a free replacement spark plug along with the test report on the leaking spark plug we did not pursue the issue any further, which brings to the time we-
Image source: https://www.vehicleservicepros.com/service-repair/diagnostics-and-drivability/article/21295100/all-in-a-diagnosticdays-work
This image shows a multimeter reading of the charging output of a new alternator (arrowed) on a 2008 Toyota Camry fitted with a 2.4L engine that had done just under 220 000 km.
This customer brought the car in for us to check the battery condition because the engine cranked over very slowly and the headlights barely lit up when they were turned on with the engine not running. The alternator warning light was also glowing dimly with the engine running. According to the customer, the battery was only a few months old, and they wanted to claim a replacement under warranty conditions if the battery turned out to be defective.
We confirmed the customer’s concerns and ran some tests on the battery with our high-end battery test equipment. The battery passed all tests, but since we needed a fully charged battery to test the operation of the starter motor, we brought the battery up to a full charge before reinstalling it in the vehicle.
The engine turned over at normal cranking speed, and the headlights seemed to work OK with the fully charged battery, but just to be sure that the charging system was working properly, we measured the alternator’s output. The result of this test is shown in the image at the top of this section.
The low output voltage would explain the chronically discharged battery, so we obtained authorisation from the customer to replace the alternator with a reputable aftermarket unit. However, the alternator replacement did not resolve the problem, because a) the alternator warning light was still glowing dimly, and b) the new alternator’s output was also just under 13 volts.
The only thing that could explain this was that the new alternator was also defective, but since we did not have an electrical test bench to test this theory, we did the next best thing, which was to return the alternator to the supplier and ask for another replacement.
Since we spend a lot of money with this supplier, the new replacement was duly delivered, but this unit also seemed to be defective, since its output voltage was also just under 13 volts, and the warning light remained on. At this point, we realised that there must be something else going on, simply because the odds of two successive new alternators of the same brand from the same supplier being defective must be staggeringly large.
Based on this realisation, we started looking for the problem. A full diagnostic scan turned up some trouble codes relating to low system voltage, but nothing else worth mentioning. We already knew the vehicle had a charging problem, so the codes were hardly surprising. Next, we checked the idling speed, which was smooth and within specs, and went on to verify that all the primary grounds were present and in good condition. For good measure, we cleaned the battery posts and cable terminals, before checking for signs of overheated fuses and/or relays in both fuse boxes, but everything turned out to be in near-perfect condition.
As a final screening test, we made up and ran a suitable jumper cable between the alternator's output point and the battery-positive terminal, but the output voltage stubbornly remained just under 13 volts.
Now there remained just two things we could do; we could obtain authorization from the customer to perform a deep dive into the Camry’s wiring, or we could spend some time researching the problem. We chose to take the second option since the customer, who was stuck in our waiting area, was getting restive about the inordinate amount of time we were taking to do some simple battery tests.
Our research turned up several TSBs and technical articles about charging system issues on various Toyota vehicles, but none seemed to quite fit our situation because we have already done the things the TSBs recommended without resolving the problem.
Eventually, though, we turned up a TSBN dating from 2011 that described exactly the situation we were faced with. More specifically, this TSB addressed this exact problem on high-mileage 2.4L Camry’s fitted with electronic throttle control systems, which systems it turned out, could directly affect how the alternator worked.
The TSB recommended running a suitable jumper cable between the alternator and the battery, but this made no difference. The TSDB recommended that we also check the idly quality and speed, but we already did that, and that also made no difference.
Finally, the TSB recommended that we check the condition of the throttle body, and clean both the throttle plate and its immediate area if any carbon or oily deposits were present. Now you may wonder what the throttle body’s state of cleanliness might have to do with the alternator’s output voltage, but it turned out that the alternator’s output voltage depends as much on the actual idling speed, as it does on the throttle opening, which is what determines the idling speed. Consider the image of the Camry’s throttle body below-
Image source:https://www.vehicleservicepros.com/service-repair/diagnostics-and-drivability/article/21295100/all-in-a-diagnosticdays-work
We have all seen throttle bodies in far worse condition than this. However, the TSB went on to explain that instead of using the actual idling speed as the primary input to control the alternator’s output voltage at low engine speeds, the engine management system uses the correlation between the throttle plates’ position and the position of the throttle pedal as the principal input to calculate an appropriate throttle opening to allow the engine to idle.
Many drive-by-wire systems work like this, but in the vase of 2.4L Camry engines, the engine management system does not use the actual idling speed to verify the correlation between the throttle pedal and the throttle plate's positions or to compare the throttle plate's actual position to its desired position. As a practical matter, the upshot of the above is this-
Since only the correlation between the throttle pedal and the throttle plate’s inferred positions counts in calculating an appropriate idling speed, excessive carbon deposits around the throttle plate that impede the flow of air through the already small throttle opening, could make the ECU think that the engine is idling at a much lower speed than it actually is.
This explanation kind of made perfect sense, so we cleaned off the carbon deposits from around the throttle plate with an approved solvent and performed an idle relearn procedure. In addition, to test this explanation, we also reinstalled the original alternator, which immediately extinguished the warning light and produced a charging current of 14.3 volts at idling speed.
In this case, failing to read the instructions, aka researching system strategies, could have saved our customer a five-hour wait and us a lot of paperwork in trying to get refunded for the second replacement alternator, which was supplied correctly. As it happened, we never were refunded, which cost we had to add to the fact that we could not charge the customer for anything more than cleaning the throttle body, which took less than 30 minutes.
Both of the above cases happened many years ago, but that is beside the point. The point is that we lost money in both cases; in the first case we only lost valuable time, but in the second case, we lost both time spent on unnecessary/ wasted labour, and actual money on an expensive part and labour we could not charge the customer for, which leaves us with this-
We all make mistakes, and for the most part, we make honest mistakes. However, when a mistake results from us not knowing what we don’t know, can we still say that it was an honest mistake? Perhaps, since we can’t ever know everything about everything and anything, but at the same time, there are some things we should know.
There are many examples of things we should know, but at the top of the list is the need to know that if we are unfamiliar with a particular system, we should always research that system’s technical implementation and operational strategies before we dive headlong into trying to diagnose issues in those systems.
On the other hand, how were we to know that minute quantities of combustion gases leaking from a sparkplug can sometimes look just like harmless coronal staining? Now that was a perfect example of not knowing what one does not know, but we did get a free replacement sparkplug out of that episode, which was a small consolation, but still…
