2004 to 2020 Mazda 3 Forum and Mazdaspeed 3 Forums banner

1 - 2 of 2 Posts

1,788 Posts
Discussion Starter · #1 ·
I’ve been building up to this one for a while. This will be my last generalized guide in this series. It is very important to read the previous guides before this one, especially zero 103: engine flow process. It is absolutely vital that you understand at least the basics of how the engine functions and where the high and low pressures/temperatures are before trying your hand at diagnosis.

Diagnosing car related problems is a little like playing Sherlock Holmes. Or maybe Dr. House if you want. To quote Arthur Conan Doyle, “Once you eliminate the impossible, whatever remains, no matter how improbable, must be the truth”. This is the cornerstone of diagnosis.

In this guide we will look at a general way of thinking when diagnosing a problem and a few case studies to show how these principles can be applied to any problem. This is not about X problem is caused by Y fault because that’s what you’ve been taught, its about why that fault causes that problem.

First and foremost, if you want to repair any vehicle properly there is one thing that you must absolutely stop immediately: STOP REPLACING PARTS AT RANDOM! I don’t care if you’ve been fixing cars for years, you’re a qualified mechanic, or your buddy ‘reckons’ what the problem is. It doesn’t matter. You’re doing the job wrong by replacing parts that you think may cause the problem without anything to back that theory up.

Do not fall into the trap of the usual suspects. For example, the MC accent has an issue with blowing its coilpacks. That should always be on the list to check when diagnosing for a possible ignition problem (ie. Misfires) but it is not always going to be the problem. Check it and don’t forget everything else.

The result of this behavior is spending a lot of money on parts and not solving the problem. Sometimes you might get lucky and fix the problem, but I’m interested in 100% accuracy with my repairs. Not guesses.

I will admit that sometimes it comes down to random replacement of a part or two simply because there is no practical way to test or eliminate a cause. This should always be left as a last resort.

The short version here is stop replacing parts randomly or on gut feelings. Forget gut feelings, they’re useless. Forget what anyone says if they are not going through a logical process of elimination to determine the problem. And finally, never scratch a single possibility off your list until its PROVEN! Don’t be afraid to revisit old theories as you get new information either.

So in general the process for diagnosis will go like this:

Step 1: what is the problem and what are the possible causes?

This is where my engine flow process article can come in handy. Some things may not seem obvious until you understand how the engine works. Take all the clues available, no matter how small. Every little symptom should be considered even if they seem unrelated at first. I don’t believe in coincidences and I’m rarely proven wrong with that. Noting when certain symptoms appear can be helpful too, maybe only at a certain speed or when the engine is warm, etc.

The next and probably the most important clue you need is to get that code scanner onto the diagnostics port and check for any codes! Those trouble codes can be the most useful item in your diagnosis. They can tell you things the ECU is detecting that you may not be able to, such as an electrical fault or failed sensor.

Sometimes the code is indicating an obvious fault, like say an o2 sensor. Other times it may not be a fault of an individual part but reporting an adverse condition, such as a random cylinder misfire.

Step 2: eliminate the possibilities

At this point you probably have a theory. Resist the urge to replace parts based on that theory. Instead, it’s time to prove it. Consider all your clues in the previous step. What doesn’t fit? For example if you have a slight oil leak and the level is fine, but have a random cylinder misfire trouble code I think it would be safe to say that the oil leak is not responsible. Cross that one off. Keep going down your list until you reach the point where you’re left with the possibilities you can’t cross off. Sometimes you get lucky and there is only one possibility left. Now its time to order some parts! But what if you have a few possibilities? Read on…

Step 3: test your theory

You’ve got a few possibilities left to cross off your list now and the answer is not so obvious at this point. Time to start testing! I’m not going to go through all the tests here. If I did this article would turn into a novel! There are many available and many are cheap to buy the equipment for. Some you can even rig up on your own. For example, I was once concerned that I may have a leak in my EVAP system. My test was to use my small air compressor and a secondary pressure regulator to reduce pressure to about 4psi….or maybe it was 6, can’t remember point is but putting air into the system I was able to test for leaks. My pressure did not drop over the course of a few hours. Pretty leak tight I’d say.

Your homework now would be to research what tests are available for any given item you want to test. Forums and google are great resources. Knowing the differences between various tests is good knowledge as well. For example, do you know what a leak-down test can show you over a compression test?

Some test tools are very overlooked and can be very useful. Example is a mechanic’s stethoscope. Oh sure you can use a stick to your ear….but if you believe that putting that stick up to your ear is going to give you the sound detail that a device designed for this task will….well maybe you’re not ready to stop replacing parts randomly. Did I mention that I picked mine up for $10 about 7 years ago? Yeah still works guys.

The point of this entire thought process is to save you both time and money. It stops replacing parts needlessly. That takes money to buy parts and your time. Time is money, so that’s a lot of cash you’re wasting! Read on to the case studies that I’ve given below. This should give you a better idea on how to apply these principals to your repair job. These are either problems that have happened to myself and how I solved them, or other examples from forum posts. No names will be used.


This is a rather simple one. Not that long ago, I found a post where someone had coolant spilling out the overflow tank. Immediately, I was able to jump to either a blown head gasket or cracked head. Cracked block is a possibility, but rather unlikely. Further information from the OP revealed that there was a previous overheating problem and he/she then removed the thermostat.

So what happened here and why was I able to jump to that conclusion?

Let’s start with the diagnosis itself and deal with the ‘what happened’ later. Firstly, consider the pressures of systems involved and let’s make a few assumptions based on the system flow paths:

- There must be a motive force to displace the coolant in the system and push it out the overflow. It does not come out on its own.

- Consider the systems that given the failure of a gasket or metal itself would be able to interface with the coolant system. This put power steering, A/C and the brake system (as well as clutch if it exists here) out. That leaves us with engine gasses, engine oil, and transmission oil (if automatic).

- Next, for enough volume and force to displace a noticeable amount of coolant we can cross both the trans oil and engine oil off the list. There simply is not enough volume. Furthermore, the OP was indicating that this was pushing out in what sounded like bursts. Oil systems won’t do that, they provide a constant stead flow of oil. Finally, the OP also indicated that there was no oil contamination found in the coolant. That leaves us with one possibility: combustion chamber gasses. Nothing else has the kind of pressure and volume to displace coolant like that.

Remember Arthur Conan Doyle’s quote here. There is no other possibility than gasses from the engine combustion chamber(s) entering the coolant system displacing the coolant. Only place for it to go is out the overflow tank. Only way for that to happen is if there is a break from a cylinder to the cooling jacket either through the head gasket or the engine head itself. Regardless, the head will still have to be pulled to assess the damage so either one is a moot point as far as the diagnosis is concerned.

Even with the reasoning on this problem being very solid at this point, I would not be satisfied until proven with a test if I were the one working on this vehicle. The idea of the test is to prove this theory. If the test outcome is not as expected, something was missed.

Selecting the correct test is equally as important. Testing the coolant system would be a waste of time in this case, coolant will leak out due to its higher than atmospheric operating pressure. Safe to say that we have no external leak here. What we suspect is a break in the head gasket at this point, so a compression test or leak-down test would be appropriate. Sure, pressurizing the coolant system may show coolant leaking into a cylinder, but this test could be difficult to determine where the leak is. A small amount may leak into the cylinder or its even possible not at all – the compression cycle has a much higher pressure so pressurizing the coolant system may not show a small leak.


This one occurs on my own car, a 04 accent with the usual 1.6L under the hood. Symptoms were no start in cold or especially damp conditions. The engine would turn over with the starter just fine, but would not catch. It was as if no fuel or spark was present. However, this problem was very intermittent. Very hard to catch and reproduce and was very dependent on the environmental conditions and even how long it sat. that’s a big clue right there…interestingly the car never let me down. Crank it for a couple minutes and it would fire right up and run fine from there on in.

My code scanner revealed code P0562 ‘system voltage low’. That indicates that the ECU is not getting the power it needs at the instant that code was thrown. I pulled the related troubleshooting info from the hmaservice web site (as you should always do when possible!) only to find that it was very brief and everything checked out OK.

Normally during a no start condition like this if the engine is turning, I check for Fuel, Air, Spark or FAS for short. Unfortunately what made this such a ******* of a problem was that it was too brief to check for these! By the time I was able to get the hood open and a spark detector on the wire the problem would be gone. Same goes for putting a stethoscope on one of the injectors.

But that right there is a clue…one interesting difference between mechanical and electrical parts is that when a mechanical part breaks, it’s done. It may try to function but it’s not going to start working again. Electronics can sometimes…and usually the failure is either outright complete failure or it works. Rarely is it in between as happens with the mechanical parts. This leaves me to believe that it’s an electrical problem.

Now I admit, I got lucky with this one. Without further info, this is very difficult to diagnose. Once I got the extra clue you’ll see why I picked this scenario:

The golden clue was both oxygen sensor heaters were dead according to the code scanner. This is a case of the ECU reporting something in error….sort of. What do you suppose the chances of both o2 sensors going south at exactly the same time? Almost nothing – especially since I’d replaced one about a month prior. No coincidences here. I also had the low system voltage error present as well.

So not believing in coincidences here, I asked myself: “what could cause failure of power to both o2 sensors AND the ECU?”

Once I pulled up the schematics it became quite obvious: there is a single relay supplying both power to the ECU and those two o2 sensor heaters.

Now I don’t believe that this is the exact wiring diagram for my particular car, but it is very similar. Please ignore the pink circle…wiring schematics are getting difficult to find now that hamservice site has gone to a pay format….we are looking at some of the components within that circled area though. You’ll have to excuse me if this is not 100% accurate, but its close.

Note how power initially comes in through the engine compartment relay box via the switch labeled C22. without that, no power gets to the ECU (labeled PCM at the lower right). Further down the line, no power gets to the oxygen sensors, fuel tank diagnosis module or the purge control valve. Well the computer doesn’t care about either until you’re running and getting started was my issue.

I did attempt to test this part. Seemed to function fine with power applied to the correct terminals and some test leads…no abnormal resistance across the terminals, etc. tried to mist it with water, even put it in the freezer with water on it. Wouldn’t fail! That doesn’t prove its good though. There are times when a leap of faith is necessary, but only as a last resort. This was one of those times. $10 later and my car was fixed.


Once again, this is my own vehicle. 2004 Accent, 1.6L, manual trans. At first there were no obvious symptoms, other than my check engine light. Code scanner revealed P0441 evap system incorrect purge flow. At first I didn’t care, it passed emissions testing with flying colours. Then slowly it began to develop very small misfires while at idle accompanied by random misfire trouble code.

I have to admit, I couldn’t really feel the misfires. However, computer was detecting it. I didn’t care much either, this was only at idle. Then one day the emissions laws where I live changed. No more are they doing proper tailpipe testing, instead its just an ECU scan. I won’t harp on what I think of that nonsense. The result was that I was now forced to fix this long standing issue.

I started by troubleshooting the EVAP system. The usual suspect is the purge valve so I started there, tested fine. Put some low pressure air on one side of it to check for leaking, there was none. Then, tested operation of the valve by applying voltage to the terminals. Again, worked fine.

After that, applied low pressure air to the EVAP system to check for leaks as described earlier on. There was no leaks.

The usual suspects were checked for the misfires. Plugs, wires, intake, etc. no joy.

Here is where my mistakes began. Mistakes that I learned from, so I’m not too disappointed J first, I found the valve would flutter open and closed while at idle. I incorrectly assumed that this was causing the fault. I assumed that fuel vapour from the tank was going into the intake when it shouldn’t be and was causing my misfires. I then replaced the valve. This was mostly because the valve was cheap and I wasn’t sure where to go from there after my testing. Of course, this was a mistake. New valve functioned just the same. I now suspected the ECU was faulty, but since I was having no other issues and due to the high replacement cost (even from a wreckers) I was reluctant to replace that on a guess.

Good thing I backed off there. Not sure what to do next, this was the one time with my car that I went to the dealership for a diagnosis. Having a bit more experience with this particular issue, they knew exactly what to check next: my intake system.

The mistake here was I assumed that the faulty EVAP system was causing the misfires. Turns out it was the other way around. I should have been troubleshooting for the misfires first. it was the misfiring that was causing the EVAP system to function incorrectly. This can also be seen as a case where you can’t take the error codes at face value. The computer only knows so much, sometimes the fault it reports isn’t the problem at all – but it’s still a clue. You have to ask yourself ‘what else can cause this fault?’

The confirmation was in that they noticed my idle was slightly rough. Since it came on gradually, I never noticed. The idle was not severely rough either so it was an easy thing to miss. If I had found that, this story would have been quite different indeed.

To check for an intake leak, they used argon gas. By spraying argon gas around the intake gaskets, they discovered the intake manifold gasket was leaking. The argon gas temporarily sealed the leak and the idle smoothed out. This gasket is a known issue with this particular engine and the part has since been revised. A total of $32 for my new gasket and I had this fixed in less than an hour.

Now, look into this a little deeper. Misfires are caused when the fuel/air mixture is too lean or the mixture is just not being ignited. What are the causes of these scenarios?

This is an easy one. Either plugs, wires or the coils. I did check all of these items but one piece of information rules this out: misfires only occurred at idle. This leads to a mixture problem as the spark system should fail to ignite at any RPM range and not just idle. In fact, it would be more likely in the higher RPM range.

Air/fuel mixture:
Too lean being the condition, there are only a few possibilities. The air flow sensors being incorrect so that the ECU doesn’t put enough fuel in, low fuel pressure/volume into cylinders, or a leak that is bypassing the sensors entirely. Sensors can be ruled out here because, again this is only at idle. A bad sensor would always be bad, not just at idle. Same goes for low fuel into cylinders. The pump supplies a constant volume/pressure to the fuel rail and injectors, so it stands to reason that misfires would occur in the higher RPM range or under heavy engine load when more fuel is required.

That leaves only an intake leak that is bypassing the sensors. The ECU simply had an inaccurate value for the air going into the engine at any given time. This was a small leak which explains why only at idle. It also explains why the ECU was not able to detect a lean mixture condition. The leak was just too small.

Also consider volumes involved. That leak has a specific volume of air that will be able to flow through it. That leak size isn’t changing very fast. At idle, the engine pulls in very little air so the ratio of air coming in past the sensors and what isn’t through the leak is very large compared to wide open throttle. More air is coming through the intake, but the leak through the gasket has not increased very much at all if any. At that point the mixture is close enough and burns just fine.

This finally concludes my beginner’s guides for car repair. Hope this was help to some of you. This guide was meant to be more of a primer to get you started. The intent is to start you thinking in different ways about repair and begin to start diagnosing instead of random part replacement.

It is a long process to learn enough to effectively diagnose difficult problems. Hopefully this will get some novices on the right path J

1,788 Posts
Discussion Starter · #2 ·
this is an update to case study #2 :

found the schematics i was looking for. as you can see, there are a few things fed by the engine control relay (first pic, extends to second). note the two oxygen sensors (second pic) and the ECM. those explain all received codes and even though purge control valve and tank leakage diagnostics module are also powered off this relay, the ECU can't do much with those until after the car is running.

this would be a good example of how the trouble codes are not always as they seem. sometimes a little bit of 'think outside the box' is required. as you can see, this is almost identical to the original schematic i had posted.
1 - 2 of 2 Posts