Is it just me who thinks that when most mechanics are presented with a misfire on a petrol engine, they go straight for the ignition system as it’s the ‘most common’ system to cause the problem?
Then, depending on the arrangement, they usually replace the ignition coils, spark plugs and either strike it lucky and get a fix or are still met with a misfire fault. Only then do they start doing some digging. I recently had to deal with three different vehicles, all petrol, and all with misfires. I want to show that there is more to it than just spark plugs, coils and ignition leads.
My place of work includes car sales. With stock running low, my boss decided to buy some more cars. This included three vehicles that were described as having “warning lamps illuminated and running issues.” Being a blind sale online you have to take a risk as you cannot view before you buy. You then have to hope the damage is not catastrophic or that it has not been around the houses with no joy and sent to auction to become someone else’s problem. The vehicles purchased with running faults were a Citroen DS3, Dacia Duster and a Nissan Qashqai, all of which had petrol engines.
After a week or two, the vehicles arrived and I set to work on diagnosing each of them. I chose the Citroen first which had the 1.6 EP6C engine fitted, simply because it was the first off the lorry. The car started and ran but had a misfire accompanied by warning lights and messages for a fault. Scanning the car for faults showed a cylinder 4 misfire code. This code doesn’t tell you why it is misfiring, only which cylinder is misfiring. We now a had direction in which to head.
For a petrol engine, the key elements required for correct running are the fuel system, ignition system and mechanical integrity, i.e. good compression. This also must all be happening at the correct time. If we have a misfire one or more of these must be off.
A good habit I have gotten into over the years with misfires is to check the mechanical condition of the engine first in order to ensure it can create compression. If all is ok, only then do I move onto the ignition/fuel side of things.
To do this I first prevented the engine from starting, which can be done in several different ways, and cranked the engine, listening for any irregularities as it was turning over. A trained ear can pick up an issue before any tools need to be used. In this instance, all sounded well but to confirm this I carried out a relative compression test using an oscilloscope. While this can’t physically measure the compression, it can compare each cylinder to the others which confirms if they are even or relative to each other. This is done by placing a current clamp around a battery cable and measuring the current draw from the starter motor in amps. In this way we can visually see the work done by the starter to crank the engine over.
A key thing to remember though is that the scope trace may look good and have even peaks throughout, but if all cylinders are low on compression the waveform will also look good. As I said, we are testing them relative to each other. A good thing to do while carrying out this test is to check the amperage at the top of the peaks and compare it to the battery AH (amp hour). A rough guide is to expect three times the AH of the battery. So, if we have a 60AH battery I would want to see a peak of around 180 amps drawn by the starter upon cranking. This way I know the compression of the engine is in the right ball park. As shown in Fig.1, I have put a thin black line across the top of the peaks where you would measure with the scope software and arrows pointing to two individual compression events to show what a good trace looks like. So, for a 4-cylinder you would count four and for a 6-cylinder, six peaks etc. You can then add in a reference point to work out which cylinder is 1 or 4 or whatever one has an issue. I will explain more about this later in the article, so read on.
Back to the Citroen; After confirming mechanical integrity in the form of compression, I then decided to make sure we had a good functioning ignition system. As we had a fault code for cylinder 4, I decided to test that first by carrying out a stress test of the ignition coil. Upon pulling the coil from the engine I noticed it was soaked in engine oil and the rubber coating of the coil had been softened by the oil and was damaged and badly mis-shaped. The coils on this vehicle have three wires, which include a power supply, a ground and a control/trigger/turn-on signal. This vehicle uses four individual pencil coils. Upon testing the coil, it was noted there was no spark whatsoever so it appeared to be faulty. To be certain the wiring side was ok, I removed cylinder 3’s ignition coil and plugged it into the connector of cylinder 4. Having four separate coils allows me to swap their positions to prove system integrity quickly. Attempting a coil test again on cylinder 4 showed no spark and swapping the coil back to cylinder 3 and repeating the test showed the coil to fire. This meant we had proven the fault was somewhere in the supply, ground or control of the coil. Using the oscilloscope, I proceeded to test the power, ground and turn-on signal. The power and ground were good; However the turn-on signal was missing. Next, I tested the wire from the connector on cylinder 4 coil back to the engine control unit. This proved the wire to be good so the fault had to lay within the engine control unit. The evidence would suggest that the coil was damaged due to the oil leak. The damage to the rubber insulation had allowed the high voltage intended for the spark plug to be instead sent to the control unit causing internal damage. A replacement control unit was sourced, cloned and fitted along with new spark plugs, a new ignition coil and gasket for the cam cover. The vehicle then started and ran without fault.
Logging a fault
Next was the Dacia Duster which also had a 1.6 petrol engine and was logging a fault for misfires on cylinder 3. Again, it didn’t tell us why, only which cylinder. This is commonly done by the crankshaft position sensor measuring the variations in the speed of the crankshaft. Following the same process as before, I checked compression first which was good followed by ignition which also was good. However, upon testing the fuel side of things I found injector 3 was not working correctly. It still didn’t look correct when looking at its waveform using an oscilloscope. Carrying out some more checks, I found its internal resistance to be wrong compared to the 16 ohms of the other three. As with the coils on the Citroen, when you have multiple components which are the same you can use them for a known good and then use that data to test your component in question. On this injector, the resistance was close to 0 ohms, indicating it was shorted internally. A new injector was fitted, the system retested and we now had a fully functioning engine.
Cause for concern
Finally, the Nissan Qashqai was looked at which had a 1.2 GDI petrol engine. Unlike the other two vehicles, this uses direct injection. This presents a further cause for concern as these engines are becoming more and more common for misfires caused by excessive carbon build-up in the inlet ports and on the back of the inlet valves. This is due to the fuel now being injected directly into the cylinder. Previously on manifold injection the fuel would clean the valves on its way into the cylinder. The carbon build-up restricts the path for air to enter the cylinder and this is very important for GDI with its multiple different running modes. However, I will leave the subject of GDI at that as it could make its own article.
This vehicle also had a cylinder 3 misfire fault code stored. Upon carrying out my misfire process, the engine could be heard to have an inconsistent cranking speed, indicating a problem.
Carrying out a relative compression test showed we did indeed have a problem and no compression was present on a cylinder. To verify it was actually cylinder 3 I connected a second channel on the scope to the ignition coil trigger for cylinder 1. Checking technical data showed the firing order to be 1-3-4-2 so if the missing peak was the next after the firing line for cylinder 1 it would indeed pin down cylinder 3 as the cause of the misfire. As can be seen in Fig.2, the green line indicated when cylinder 1 coil fires and the next peak of compression is missing so it was indeed cylinder 3. The next question was why was there no compression on that cylinder? At this point we had two choices; We either stripped the engine down until we identified the cause or we use a pressure pulse sensor so that we could measure pressure pulses in the inlet, exhaust and crankcase to find where the cylinder pressure was escaping to. This would be much like a cylinder leak-down test only we can do it cranking the engine and without having to set the valves at specific positions and introduce compressed air to find where it is leaking. So, it is much faster. In fact, if we had multiple pulse sensors we could evaluate the entire engine in one cranking capture with an oscilloscope, but I am not so fortunate.
I went for the latter and connected my one and only pulse sensor to a third channel and one by one captured an inlet, exhaust and crank case pressure waveform against relative compression and the coil trigger for cylinder 1. Then using a cylinder over-lay chart, I could see what each valve should be doing at what time and compare it to my waveforms to find where the problem was. Now to be brutally honest this takes a lot of time and practice to master, and I still find myself needing some assistance. So, after reaching out to a few fellow techs for their opinion we concluded that there was an exhaust valve issue on cylinder 3. Again, this subject could fill its own article so I won’t get into the nitty gritty of how and why but we had an answer where the compression was going.
Yes, a cylinder leak-down test would have shown this, but not all running issues are obvious and some require the engine to turn to reveal themselves which a leak-down test will just not show. Now this job so far was still in-house, but if it was a retail job I could tell the customer exactly what the problem was without removing much more than an engine cover or dipstick/pipe to put my pulse sensor in to measure if required. This would allow them to decide if they wanted to continue with the repair or for them to take the vehicle away. This would not be so easy if the cylinder head was removed from the engine.
The cylinder head was then removed and one of the exhaust valves was found to be incomplete. Please refer to Fig.3. This backed up the diagnosis made before the engine was even dismantled, proving how accurate and powerful this form of testing can be. All three cars in fact prove how complex the causes of a misfire can be, and how fitting a new coil or spark plugs/leads would not have fixed any of the vehicles in this article. If you carry out logical testing and gather data which drives your next move, you will always find the cause of the fault.