Here is an interesting diagnostic case from my own practice.
I am not mentioning the make and model of the vehicle I have been working on and I am doing it on purpose. An idea here is to show the way of thinking that would be applicable to many different engines and control systems. Also it shows how the diagnostic can be completed without previous experience working on the same model or system and where not all of the required technical information is available. By the way, pretty much the same can be said about every case of diagnostics – you never have all the experience and all the information you’d like to have and must somehow make up for it… By the way, previous experience might do bad service sometimes, directing your thinking in the wrong way, away from where the problem is!
So the symptoms on the vehicle in question are… well, the engine runs like shit :-). Pretty much it has it all: poor idle, stumbling, cutting out under load, occasionally stalling on idle, lack of power, severe misfire under load, hesitation, bucking, popping into exhaust… It is still possible to drive it as long as you are gentle on gas pedal, but surely there is neither pleasure or safety in driving the vehicle like that.
Scan tool shows codes for crankshaft position sensor circuit and cylinder 5 injector circuit. Later on, after couple of test drives codes for ignition coils circuits also came up. Customer had already replaced crankshaft position sensor.
Normally if it comes to no start, poor start or all sorts of stalling / cutting out issues crankshaft position sensor (CKP) or camshaft position sensor (CMP) would be on my Usual Suspects list :-). But when the concerns are rough idle, loss of power, hesitation, misfiring, popping into exhaust – well, I would strongly suspect there is something else that is no good. Is it bad fuel pump that doesn’t supply enough fuel pressure? Bad mass airflow sensor? Blocked exhaust? Or maybe there is actually bad ignition coil and / or injector as codes suggest?
As critical as they are CKP and CMP sensors are needed just to get engine running. Without correctly timed signals from them control module would not know if the crankshaft and camshaft(s) are actually spinning, what is the RPM and at what moment of time to ignite the spark in cylinders to get engine running. But once it got fired up and running then it is mostly up to multitude of different other sensors and actuators to determine how much air is entering into the engine and how much fuel needs to be mixed with it to have proper air-to-fuel mixture, correct burning of that mixture and therefore proper engine output. I say “proper output” because driver doesn’t need the full output all the time but rather some portion of it depending on driving conditions as commanded by pressing on gas pedal.
Most of the people and (which is the worst part!) plenty of mechanics think that next step after reading codes is to replace parts or components mentioned in code description. Sometimes that works, indeed (I would say 2 out of 3 times it works), and that’s why so many people think that “computer” actually tells for mechanic what part to replace. LOL, good luck with that when all you got is P0171 Fuel Mixture Too Lean Bank 1 code – its hard to find new fuel mixture for your car in parts stores ;-).
In this case crankshaft position sensor was already replaced so do I have to install new cylinder 5 injector and ignition coils module now? No, I don’t think so. And that is why my position was called Diagnostic Technician rather then a Mechanic for the last 10 years working in different dealerships pretty much around the globe. My job was (and still is) not just replace parts – it was to diagnose what is causing the problem and then fix it right, first time oh, don’t tell me that stupid “on time” line again!
The right next step is to look at the live data while driving the vehicle and experiencing the problem. This is relatively quick and easy way to see what kind of input from various sensors “sees” the control module that controls diagnosed system – in this case it is obviously Engine Control Module (or Powertrain Control Module, whatever way manufacturer likes to call it). My expectation was to see that fuel trims and oxygen sensor signals would change somehow when I am having that lack of power / hesitation / popping issue. For example, if the fuel pump is bad then under load it would not deliver enough fuel and fuel mixture would go lean resulting on oxygen sensor signal going low and short term fuel trim going high (as the system tries to compensate for lack of fuel delivery).
I was a bit surprised because data readings indicated that fuel mixture actually goes rich when I am experiencing lack of power. Well, this is unusual… Somehow system delivers too much fuel and mixture go so rich that it can not burn? But what would be causing that?
I managed to make another useful observation though. As among other data I recorded RPM signal (normally I am dong it just for the reference, as to see later what was the engine speed and load at any particular moment) I noticed there were occasional spikes in RPM reading that could not be quite explained by slippery road conditions (on slippery roads as the wheel spin occurs there might be momentary raise in engine RPM too) as there seemed to be too sharp and too short. Looking at the data stream represented as a graph by my scan tool helped me to spot this with sharp spikes on RPM graph (top one):
Definitely I should have a closer look at the crankshaft and camshaft position sensors signals. At least I should make sure that they both operating properly and are in correct sync before I would proceed with any further diagnostics. As I knew already, crankshaft position sensor was already replaced – as customer told me and also visual inspection confirmed that the sensor looks new. I have to notice here that whatever customer says should be taken with grain of salt and must be verified without any doubts. Not because customers are lying (although some of them do!) but because customer does not have to know the vehicle that well and might be unable to describe the problem or do some work properly and completely. Relying diagnostic process only on customer’s description of the concern is the straight way to costly diagnostic mistakes. Would be there a problem with camshaft position sensor (sometimes the problem with one would result in fault code to be set for another, because none of the self-diagnostic software embedded in control modules is perfect)? Would they require some specific synchronization procedure or adjustment after replacement that was not performed? Or would the new crankshaft position sensor be defective?
So I am proceeding to the next step and setting up diagnostic technician’s best friend – lab scope. By the way in my 10 years working in the different dealerships none of them had decent scope – such a shame, as this is literally an eye opening tool when it comes to electric signals! Unfortunately there is no quick way to connect lab scope to the vehicle as there is for diagnostic scan tool. The best way is to use break-out boxes with compatible connectors for each and every control module on each and every vehicle, but the cost for those incredible accessories is from few hundreds bucks and up – and you need literally have hundreds of them. This will never pay back unless you have very narrow specialization in particular make or even system of the vehicle. This kind of narrow specialization is definitely not viable when living in remote area with barely 100 000 people around.
Therefore I have to go slow way: disassemble control module connector that contains required circuits (locate them first with wiring diagrams or otherwise there is even slower way of mapping and tracing connector pinout), strip down protective tape and conduits and loosen up the wires around that connector so then it is possible to backprobe them with appropriate needle probes. Quite a tedious and time consuming process, especially considering the need to ensure good and reliable contact with each circuit to be tested while not damaging wire terminals or connector pins!
So first I am reading the signals from both the crankshaft (white line) and camshaft (orange line) position sensors:
Right away, I am wondering why some of the crankshaft position sensor signal pulses are longer then others (circled in green). It is not necessarily an indication of the problem. Some systems are designed to have longer pulses from crankshaft or camshaft position sensors to indicate top dead centre or crankshaft rotation angle. Although I can’t see any logic in this sequence of long and short pulses and suspect there might be some problem here. For sure, the best way to tell if there is a problem or not is to have a sample signal waveform from a known good vehicle of the same model, year and engine option. But now I have only known bad vehicle. I have service manual for this vehicle but, as usual, information in it is quite limited and doesn’t provide any waveforms at all. Quick search on the internet did not bring any good enough results either.
Therefore my next best option was to have a look at the sensor signal ring and see what kind of shape it has and what kind of signal it supposed to produce. Not always a simple thing to do as this signal ring is usually attached somewhere to the crankshaft, so it is inside the engine. Luckily on this vehicle I did not have to strip down the whole engine to have a look at it :-). I used small and cheap USB camera that I put in place of the crankshaft position sensor. All I wanted to see was this:
This is one of the holes on the side of the flywheel that triggers crankshaft position sensor pulses. Notice the scratch marks – for a while I thought sensor tip came into contact with rotating flywheel due to incorrectly set sensor air gap (or even because of the radial play on the flywheel for whatever reason) and this was the reason for bad signal. In fact, as it was found from reading service manual there was no problem here and sensor had a special pad on its tip that supposed to contact flywheel upon installation and this is how air gap is being set.
Then I rotated the crankshaft, counted and inspected holes and made reference marks on the front pulley for their position. Turns out all of them were of the same size and positioned in groups of four every 120 degrees of the crankshaft rotation, only 12 holes altogether. Now it is also clear how the camshaft position sensor works – it is just switches twice per one revolution of the camshaft (which spins with half the speed of the crankshaft on any four-stroke engine. Yes, there are two-stroke engines with camshaft too, although they are incredibly rare.
Now I am pretty sure that crankshaft position sensor, despite being brand new, is faulty. Do I have to know what might be wrong with it? If I want to be a good Diagnostic Technician I should have at least some general understanding how each and every sensor works. On this vehicle I have Hall effect crankshaft position sensor as opposed to inductive pickup sensor. How do I know? Because there are three wires at the sensor connector. No, not every Hall effect sensor has three wires! ABS sensors on modern vehicles are also Hall effect sensors but they have only two wires. Although all the inductive pickup sensors have two wires. Another sign is the square wave form of the signal, although to see that you have to connect a scope first. Hall effect sensors always would have some internal circuitry to shape and amplify output signal. If the elements of this circuit are incorrectly chosen or failed it is quite likely to cause signal to be “stuck” after it it was triggered. More then likely this is what is going on here.
Although, would that be enough to cause engine running with such a broad range of different symptoms. How this bad signal from the sensor would cause oxygen sensor and fuel trims indicate rich mixture?
To get an answer for this I used recording feature of my scope (probably the best feature of it!). I had some sporadic roughness in engine operation at idle, like a small shakes from time to time. I decided to read the control signal for the ignition coils. This six cylinder engine has three ignition coils that are firing once every revolution of the crankshaft – this is called waste spark system as every second spark event at each spark plug occurs at the end of exhaust stroke as opposed to at the end of compression stroke – so it does nothing, hence “wasted spark”. At idle ignition timing sets the spark event right at the end of each four pulses group (pointed by red arrows on the picture below). I only read signals from two out of three coils so for clarity I just draw an approximation of how the third signal should look like (blue line on a graph below, don’t be hard on me for them sloppy lines, drawing on computer screen is not that easy )
Less then two seconds of recording the signal was enough to catch the bad guy – and here is his ugly face. After few longer pulses that should not be there at all Engine Control Module loses the track of crankshaft position and speed and fires the spark at the wrong moment, then it “realises” making a mistake and fires it again at the same stroke (two closely positioned spark events represented by the peaks on the green line). It also fires the spark at the wrong moment on another cylinder, twice (yellow line). As I haven’t recorded the third coil signal I can only guess if it was also fired twice or not (it should have been somewhere at the end of the next group of four pulses). Then while the control module tries to count pulses and figure out where it should send the spark next time it misses the spark event where it should be. And this is where that small engine shake comes from!
My guess is that injector was also operated twice on one or more cylinders and therefore injected twice more fuel per stroke – and this is why mixture was going rich. I might have got more interesting results if I would have driven the vehicle and have engine operating under load and at higher RPM. Unfortunately even in good road conditions driving the vehicle while trying to record some signals with diagnostic scope is not safe, not even mention doing the same thing on very slippery winter roads. Also diagnostic time is paid by customer so I did not want to spend more of it that it was necessary.
Now is the final step. New, high quality crankshaft position sensor gets installed and signals checked again while the scope is connected. All the signals are clean and lined up in logical fashion now (again, don’t be hard on me for blue graph drawing):
Engine idles nice and smooth now. After disconnecting the scope and wrapping up all the wiring back the way it was I do final test drive, verify engine works fine under load and at higher RPM (oh, yeah, there is a huge difference now!), clear fault codes and call it case closed.
By the way I spent more time writing this post and preparing illustrations then working on that vehicle :-).