Code readers not that good?

[John Wilson]

It's the 21st century. Gone are the days of having to know about your engine. Just plug in your computer and the answer comes out. Only, judging by this forum, it never seems that easy. There is currently running a thread with poor starting issues, and, dispite the use of a code reader, discussion is still going on about what could be the fault. Although this makes it more interesting for those of us who were working on cars when such things were only on Star Trek, it leaves me wondering why this is the case. It seems to be mainly with the various sensors (I just love them- how did we ever manage without them). Why is it at so often they are changed and the problem still is there. Surely the code reader can give a definitive answer. My only experience of them is with taking the Z to our local garage with old fashioned engineers who actually know in depth about cars from years of traditional diagnostic methods, trying to scroll through dozens of menus on their code reader to discover a fault before giving up and fixing it by looking at the cars workings themselves.

[Robert T]

The problem at the moment isn't that the code readers are not helpful, but rather we don't know how to interpret what they are telling us. For example, if a misfire is detected, then BMW will know that there are a set number of things that cause this, which need to be checked, whereas we have to make educated guesses. Also, without knowing how a particular error code is detected, then it is very difficult to attach meaning to it, foe example how does the ECU know that there is a misfire - what sensors does it use to detect this? Most of us are fairly new to code readers, and I expect that as we go on we should be able to build up our own knowledgebase of what to check when a particular code (or combination of codes) is recorded.

To add to that, it is my intention to try and build up such a bank of knowledge which is specific to our cars (not generic as per most of the other sites I have looked at). If you put that together with our already very good knowledge of how our cars work, we should be in a good position to help people.

Finally, think of the error codes not as a replacement for mechanical knowhow, but as a supplement to it.

Cheers R.

[OldskoolRS]

I get the feeling that code readers don't give enough information so there is still an element of skill required by the mechanic. I use the INPA software on an old laptop and it seems (so far) quite specific about what was causing the problem as it clearly stated that I needed an exhaust camshaft sensor, which I replaced, cleared the error codes and the car ran much better afterwards.

My old RS2000 can be serviced with a penknife and a hammer (I joke, but you get the idea): It is easier to fix if it isn't working, but having said that the performance tails off more quickly than with a more modern car like the Z3 between services. In fact with the RS I would tend to service it more often (or at the least regap the plugs and adjust the points) just to keep it running better for more of the time. Newer cars with catalytic converters need the various sensors to control the fuel and ignition to keep running correctly and of course to keep the emissions down, so we're stuck with them. However, when you see the horsepower of some modern engines compared to the 110PS of my 2 litre RS you can see that they are more efficient with all the fuel injection, sensors, etc. My 3.0 Z3 couldn't run the camshafts it does without having the variable timing (and associated control mechanisms) so at least I can get the best of both worlds and not have to suffer a lumpy tickover and poor low rev performance.

[Del]

I think a lot of the problems that find their way onto a forum are the trickier, non-specific ones where the codes just refer to general fuel trim or misfiring codes. I also suspect that describing the symptoms fully on a forum is sometimes difficult. Also when people find the cause I suspect they are sometimes so overjoyed they forget to post-up what they found the cure to be.

Having said that, I don’t find all professionals necessarily better. I had a non-specific, sometimes intermittent (cold start) misfire issue and took the car to an independent garage specialising in German cars where the proprietor is said to be an ex-BMW master grade technician. Very helpful and pleasant, plugged in the expensive professional code reader, but ended up only guessing at items like the MAF, fuel filter and changing the rubber MAF bellows as a precaution i.e. clearly going down the route of speculative replacement of items.

I think Robert T’s suggestion of supplying some form of “likely cause” list (based on experiences) for Z3s when particular non-specific codes pop up, is a very good idea.

[John Wilson]

I understand there could be ambiguity in some instances, but the current thread 'car wont start all the time' a camshaft sensor was replaced after using a code reader. The problem was still there. Surely the computer would know if a sensor was working or not.

[Del]

Hi John - I think the challenge with that particular forum member's car is that he says he is getting no codes at all and there is the added issue of a major engine swap in the past. I agree that normally if your camshaft position sensor was bad you would expect to get a straightforward fault code telling you so regardless of whether you were using OBD II or specific BMW fault finding kit.

[John Wilson]

I think that's entirely my point. What you are saying is that the code reader is saying there is no fault with the cam sensor, but because code readers, or the system that tells the code reader, is not very good, you might as well fork out for a new one anyway.

[Robert T]

I would not say that at all. The systems are actually pretty sophisticated and can detect a wide variety of failures - they are particularly good at detecting failed sensors which start to give erratic readings - however not everything has a sensor, or sensors can fail in some non-predicatable way, and then the codes take more understanding to interpret. I think also, where problems either log no codes, or multiple codes, then it is quite difficult to see the problem - if just one code is recorded, it tends to be much more obvious.

About the best example I can think of is that you have sensors on your bath that tell you whether the water is too hot or cold and whether it has reached the desired level, but the sensors will all read situation normal and carry on filling the bath even if the bottom of the bath splits open and water is going everywhere.

Cheers R.

[OldskoolRS]

Simple, you just need a leak detector sensor below the bath.

[John Wilson]

Robert
I think I'm understanding it a bit better. A sensor might not be completely broken but just be giving occasionally eratic readings, and in that case the fault might or might not pick it up. I was just looking at another current post where a MAF sensor was successfully picked up by a code reader so it was obviously of value there. I was looking at it too much as black and white. The sensor (or fault)was either there or not. If a fault is intermittant and is not picked up by a reader, is the opposite ever true where a fault is shown on a sensor that turns out to be perfectly OK? If so, why would this happen?
By the way i am realising that I've done this post at the same time as the post about people getting together with code readers to solve owners problems. I wasn't trying to pour cold water on this idea, which I think is going to be an invaluable resource.

[Robert T]

Yes, in theory it is possible for errors to be flagged up erroneously, but I think this is quite rare due to the way problems are detected.

Let's take the example of the bath above a bit further. Imagine that the temperature sensor on the tap fails in such a way that the temperature reading is always zero. In this scenario, the tap would be made to fill the bath with just water from the hot tap, potentially resulting in a burnt botty. However the designers were aware of this possibility, and they set limits on the values the temperature sensor is allowed to return - if the reading were 0C, then we would have ice, which clearly wouldn't be flowing out of the tap at any rates of knots, so we must have an error. Similarly if the water temperature were above 100C we would have steam not water and something is very wrong with you hot water system. So we can safely say that the temperature sensor has failed if the value is outside of the "acceptable range" of 0-100C. Of course this doesn't take into account the failure mode whereby the temperature sensor sticks as a value of 25C, but that one would assume is much less likely.

Where this system would break down is if it were a problem with the circuit in the control unit that reads the sensor. If this fails, then we can't tell whether it is the sensor that has failed, or the control unit itself. However as the sensor is the bit exposed to the elements, the likelihood is that it will be the sensor that has failed. It is also unlikely that if the control unit fails, only one sensor would be affected - more likely that several subsystems will appear to have failed at the same time. Also bear in mind that some failures have knock on effects - particularly where fuelling the engine is concerned, and it is not unusual for emissions problems to be flagged when the problem is actually with the air/fuel mix going into the engine.

Hope that helps. I'm in IT, so I am quite happy with cars having computers in them, but I agree that when things do go wrong, it can make it harder to diagnose the problem - but then modern engines are very much more complex than that such as is in my other car - the air/fuel ratio on that is set by a carburettor which really has no other moving parts then the throttle butterflies! To me that's just voodoo.

Cheers R.

[BladeRunner919]

I understand there could be ambiguity in some instances, but the current thread 'car wont start all the time' a camshaft sensor was replaced after using a code reader. The problem was still there. Surely the computer would know if a sensor was working or not.

One of the big problems, and I refer specifically to the above mentioned thread, is that there are two types of code readers for most manufacturers. OBDII is a worldwide standard designed originally to cover off emission testing in the US. There are a limited number of things that OBDII is designed to capture and report on, but they are very generic because the codes have to be used by many different manufacturers.

When you have a problem with a car you really need to pick up manufacturer-specific codes to have a better chance of a successful diagnosis because, for example, BMW may have a dozen different codes that equate to one generic OBDII code.

Unfortunately people are buying generic code readers and then getting frustrated that they're not really getting any closer. That, coupled with the misconception that a code reader will output a shopping list of faulty parts to replace, is the root of many of the very frustrated threads.