EGR and Turbo Boost Control

In trying to work out what the ECU is detecting that triggers the error code I started trying to understand how the EGR and turbo is controlled. These are my latest thoughts

EGR Control

Why have an EGR valve? In practice, combustion using the exact ratio of fuel to air has some significant problems. The formation of soot, nitrous oxides and diesel knock (explosive detonation). To counter these problems the air is mixed with cooled, inert, exhaust gas to dilute the oxygen content.
When does the EGR operate? The EGR in most modern engines is controlled by the ECU. In order to decide when to operate the valve the ECU must have to know two things. The mass flow of gas through the engine and the flow of air entering from the atmosphere. Since the exhaust gas is being recirculated the two are not equal. However, the volume of gas being 'pumped' by the engine can be calculated from the engine revolutions, engine size, air pressure and temperature. The flow entering from the atmosphere can be  measured by the MAF. So the ECU should be able to calculate and control how much exhaust gas is being recirculated by the EGR valve.
So when does the EGR operate. Unlike a petrol engine a diesel does not have a throttle. A diesel simply alters the amount of fuel to control the power.Less fuel less power, more fuel, more power, up to the point where the quantity of oxygen in the air will burn no more.
As this point is reached the combustion process begins to get problematic as detailed above. The combustion process is at its maximum temperature and air fuel mixing becomes difficult. Therefore a solution is to recirculate the now oxygen depleted exhaust gas back into the engine as this point is reached.
So when the accelerator is depressed it effectively is requesting the ECU to increase the fuel flow. From the engine revs and MAF the ECU knows when the critical maximum fuel to air ratio is being reached and opens the EGR valve.

Turbo Control

Why a turbo? A turbo, like a supercharger, is a way of forcing more air and therefore oxygen through an engine. This means more fuel can be burnt and more power produced. This also results in higher efficiency by utilising some of the wasted energy contained in the exhaust gas.
How does it operate? Most modern turbos use variable vanes on the exhaust side. When closed they create the maximum pressure differential across the turbo and generate the maximum power.
However, under normal driving conditions the amount of oxygen entering the engine will exceed the power and fuel required. Therefore, the ECU will open the vanes and maintain the flow at atmospheric pressure.
However, as more power is requested and the maximum fuel:air ratio that can be achieved at normal pressure is reached, the vanes are closed and the turbo begins to raise the pressure and consequently air flow. Therefore, more fuel can be burnt and more power produced.

Oh dear. Not entirely fixed.

It's puzzling that the MIL light still comes on periodically to indicate the same fault. But the time period between these occurrences is now usually very long and so not as concerning. Sometimes it clears itself and other times I clear it out of curiosity to see how long it will take to reoccur. For example I drove 200+ miles of mixed urban and motorway driving before it relit. It's like I've turned the clock back to when the problem first started.

So to me this indicates the cause has been an accumulation of problems with one still left to be identified. In retrospect I think the major culprit has been the EGR valve not the MAF. Most probably cleaning with specialist MAF cleaner would have been sufficient.

Last time the MIL illuminated was under heavy acceleration. Is this a piece of information that could indicate the cause?

Dashboard MIL still off fitting new EGR valve.

After clearing the error codes on my BMW320D with Torque app the dashboard MIL is still off. What a relief after all the work and investigation that I've carried out. However, the experience does raise many questions.

One of the most important question is the accuracy, or perhaps more accurately, the interpretation of the error codes. It didn't help that the P3263 code generated by the 'P'ower train ECU was very difficult to find documentation for as it appeared to be manufacturer specific. That is, a code created by BMW. Eventually, it was found to be indicating a MAF (mass air flow) metering problem. So the MAF meter was replaced at great expense without fixing the problem! Fortunately, a simple improvised BlueTack test on the EGR valve indicated it was this unit all along.

Much of the work and expense would have been avoided if more information was available on what the ECU is measuring and then calculating to raise an error. It undoubtedly is an air flow discrepancy but not a faulty MAF meter.It appears that it is an error generated by measuring an unexpected airflow value caused by the faulty EGR valve. So, lesson learnt. Don't take the codes so literally.

My overall feeling though is that the manufacturers may have adopted the open standards but only reluctantly.

EGR Valve fitted

The BlueTack remedy on the EGR valve resulted in a significant improvement. In fact so much so that the dashboard engine management light turned itself off for a short period. However, under hard acceleration the car could still generate considerable smoke (to the annoyance of cyclists!) and a replacement EGR was necessary.

An EGR valve has now been obtained from H.R. Owen, North Acton, London via eBay (and who provided excellent service to resolve problems with the delivery). Although it's fitted I've only been able to take the car a short test run as yet. What I'm eager to do is collect the engine data with Torque and see if the air flow/ manifold pressure etc now generate a more orderly pattern of results.

New EGR valve due anytime.

I found that the EGR valve on my BMW 320D is no longer stocked by many of the after market parts dealers so I made a quick decision to buy when I saw one being advertised on eBay.

So the new EGR valve has been ordered and is due to be delivered very soon. I'll be so disappointed and confused if it doesn't resolve the problems.

And just to recap the original diagnosis indicated by the ECU and backed by the dealers diagnostic equipment was the MAF meter. This was replaced but made no difference.

Getting closer to a diagnosis! - Got to be the EGR valve.

I found a video on Youtube from someone demonstrating an easy way of checking for a faulty EGR valve. His valve had a faulty vacuum diaphragm and he demonstrated a way of testing by forcing the valve open and then closing the inlet pipe using his finger. When released to let the valve close then if the diaphragm was good the valve would stay open and, if not, it would close.

Although there was certainly evidence of my EGR working many of the symptoms still pointed to this rather than the MAF meter,as diagnosed by the garage diagnostic system, and subsequently replaced with no effect. I therefore tried sucking the inlet pipe and noted a flow of air.

Next, I tried the above YouTube test and a partial vacuum was held but the valve did close. I then decided to block the small vent hole in the diaphragm body to see if this made a difference when driven. The theory being that this would stop the flow of air whilst still allowing partial operation of the valve. (I used bluetack to cove the hole).

The road test resulted in a marked improvement. The old car, the one I'd almost forgotten, was back. Smooth acceleration. No hesitation. Reduced knock. More power. However, as expected some symptoms were still noticeable.

With this drastic improvement it has to be the EGR valve. Let's hope that the £100+ for a new valve will be the solution.

Long time no post!

A house move meant that work to understand and diagnose the problem with my BMW 320D diesel was put on hold despite the symptoms getting worse! Just to recap it has been an erratic fault resulting in the dashboard light being illuminated and an error code of P3263.

The symptoms are intermittent loss of power. Noticeable but not total and some smoke at times under acceleration. The car will respond after a small delay by applying more accelerator but obviously this results in worse fuel consumption. The other noticeable side effect is diesel 'clatter' during periods when the power drops.

Long distance road test results are in.

After the disappointment of installing a new MAF meter I was looking forward to evaluating the performance of the car over a long motorway journey. Unfortunately, there were no pleasant surprises such as a miraculous improvement.

Overall there are so many confusing observations to make. To begin with I filled up at the beginning of the journey with the higher grade diesel as had been recommended by the BMW dealership. They advised that this was a good idea every fourth tank or so. (Presumably it contains less impurities and perhaps some cleaning additives). But as I've noticed before the performance problems can seem more acute after a fill up?!! However, I'm reluctant to get side tracked down fuel related routes.

So after a journey returning 50+mpg at 70mph I once again connected my ELM interface and collected data over an urban route with my smartphone. This was after the hesitation and faltering seemed to be getting particularly annoying.

I then disconnected the vacuum pipe to the EGR valve to see if it was in any way related to this device. It did reduce the symptoms but not completely and performance seemed flatter but was preferable to before. For urban driving the mpg was certainly in the 40's so nothing dire was happening to the fuel consumption. It seems that the EGR valve is more of an emissions related device and so I decided to keep it disconnected for the return leg up the motorway and collect data once again.

Having looked at the graphed data it is radically different to that collected previously under no load! There is very little in the way of a correlation between MAF, boost etc vs engine revs as there was before. Whether that is to be expected or not is unknown at this point.

One thing I did think that must have a correlation if boost pressure vs air flow rate and this is shown below with the EGR valve disconnected i.e. permanently closed. I'm intrigued that above a certain pressure the flow rate becomes erratic. An air leak maybe?

This is with the EGR connected during the urban driving. There are obviously a number of differences such as considerably lower flow rate for a given boost pressure which would be expected given that the EGR valve is letting in exhaust gas. But there is also no sign of the valve shutting and the flow increasing either.

Another exhaustive search of the internet revealed many more BMW 320D owners with similar problems and spending a fortune with specialist but no magic bullet!

Next step check for air leaks after the turbo.

New MAF now installed!

Apologies for the delay in posting. I'm moving house which is proving very time consuming!

After spending £130 + new air cleaner with MAF Shop in Berlin the new MAF is now installed and the result is ........... only a slight improvement! :-(

Possibly that might be an understatement because I've not had to chance to take the car a long trip yet. (That will be this weekend). Perhaps the fuel consumption does look better but the 'misfire' is still there and the MIL dashboard light is still lit.

My first reaction is annoyance with BMW who, to be honest, I'm sure are no different to any other manufacturer. Why don't they disclose the real meaning of the codes! And I mean the REAL meaning of the error code. Would it undermine their business to disclose what analysis the control unit is performing to generate the exception. It seems that they have been forced to comply with international regulations to provide a standard method of reporting but they have only reluctantly complied!

So the investigation continues. It's interesting that

• After resetting the light with Torque it illuminates almost immediately after the engine is started. There is some initial diagnostic taking place that is checking various parameters and finding them out of limits.
• The misfire/rough running occurs around the time the turbo is beginning to work and is generating positive inlet manifold pressure. 

This is now making me suspicious of the EGR operation despite the fact it has been cleaned and now seems to operate freely instead of being carboned up. What I have to find out next is how and when the valve operates. I know it's vacuum operated and I know where the electric solenoid valve is. But is it operating correctly? My suspicion is fuelled by the fact that the misfire seems to occur at 1800rpm which, as I understand it, is about the time the EGR valve is closing due to the inlet pressure being greater than the exhaust gas pressure.

All that said I'm still suspicious of the volumetric efficiency curve. This is calculated by measuring the theoretical air flow through the engine using inlet pressure, inlet temperature and rpm and comparing it against the measured air flow. The ratio of the two is the volumetric efficiency. 

I'm surprised that the maximum recorded value when the turbo is working is still less than 100%. (I'll try logging when the car is being driven rather than sitting on the drive). The second 'anomaly' is why the dip around 2500rpm?!

Vital difference between petrol and diesel engine operation

As I struggle to find simple ways of interpreting the data obtained from my car there has been one thing that has puzzled me. (In reality there are many more than one!) The question is, where is the throttle on my diesel? All the other engines I have experience of (all petrol) have a butterfly valve controlling the airflow into the engine. Where is it on my diesel? Well the answer provided by Denbury Diesels is that there isn't one!

The explanation is that petrol and diesel engines operate in two different ways. Whereas a petrol engine maintains a constant air to fuel ratio the diesel just alters the amount of fuel. Using a fire as a crude analogy the petrol engine controls the air vent to vary the heat output where the diesel is equivalent to getting more heat by adding another log. This basic information answers many questions raised by articles on volumetric efficiency etc that assumed the reader knew the engine was petrol.

So looking back at the volumetric efficiency graphs it could be expected that the efficiency is flat until the turbo kicks in. The really interesting work will be to see how the graphs looks when the new MAF is installed.

How the fault is affecting performance

First thing to say is that cleaning the MAF meter improved performance for a couple of days but then deteriorated! Back to the drawing board. And for information it might be of interest to describe how the fault is affecting performance.

Way back when the fault first started the first component to be suspected and checked was the EGR valve. This, as I have learnt,  lets exhaust gas back into the intake which reduces the amount of oxygen available for combustion and helps, amongst other things, reduce 'diesel knock'. One of the symptoms at this time was excessive smoke from the exhaust under acceleration and logic might suggest that this could be too much recycled exhaust causing incomplete combustion. However, this was the condition of the EGR.

About as bad as it gets I would think and looks like no recycling was occurring! However, 'diesel knock' was and still is more pronounced than it should be.

After much cleaning the EGR was put back and an immediate improvement noticed in fuel consumption and smoke emission. On a long run and being very economy conscious fuel consumption improved from 50 to 60 mpg! However, fuel consumption variability continues to be a characteristic of the fault despite cleaning the EGR.

Sometimes the car performs as though there is nothing wrong and then like a switch being thrown the deterioration kicks in. Fuel consumption increases and a small misfire and hesitation can be noticed around 1000 - 1500 rpm under low to moderate acceleration or load. But floor the peddle and the performance is still there.

The BMW dealer suggested two possible causes. One was the MAF and the other carbon build up in the cylinder head. Having taken off the inlet manifold to inspect the latter the condition was as follows. This is the manifold port. Not good, so the manifold and head was carefully cleaned  as carefully as possible.

After this, unfortunately, no improvement in performance was detected  but it's worth also noting that two vacuum pipes that actuate the EGR and swirl flaps were in poor condition and were replaced.

Also what I think is the MAP sensor at the back of the manifold was found to be almost clogged.





This is the MAP sensor.

Overall, the engine in a modern car is so crowded and these items are not easy to inspect without disassembly.

As an aside modern diesels (perhaps petrol as well) seem to have variable inlet flow paths. On the BMW the path is controlled by swirl flaps and these can be seen in the rectangular ports of the inlet manifold below. Once again they are controlled by the ECU and the purpose of these is to modify performance. There is much written on  internet forums about removing them as there also is with regards the EGR valve. However, it must surely cause havoc with the ability of the ECU to manage the engine efficiently.

In fact testing the effect of disabling the EGR resulted in excessive exhaust smoke (but perhaps this was caused by the suspected MAF fault).









Finally, having run out of further ideas the decision has been made to buy a new MAF meter!

Can volumetric efficiency help confirm my suspected faulty MAF meter?

I hadn't heard of volumetric efficiency until I found the excellent articles on the MOTOR magazine web site. These contain a wealth of practical information on engine operation and faults and have been the inspiration for my on going work.

Basically, volumetric efficiency concerns how efficiently air flows through the engine and is the ratio of measured flow to theoretical flow. On a normally aspirated engine this is not going to be 100% because of several factors. The more obvious of these are the resistance to air flow offered by the air filter, valves and ducting coupled with the back pressure etc from the exhaust system.

The torque app calculates volumetric efficiency and so, just like the MAF data, I plotted this against engine revolutions.

This was the data before cleaning the MAF meter. (Note: at approximately 2500 rpm the turbo begins to provide boost. At approximately 1800rpm, I think, the EGR valve should be closed and therefore the MAF meter is reading all of the air flow into the engine.)



And this after cleaning


It clearly demonstrates the effect of cleaning the MAF! Why is the improvement so much more noticeable on this data than the simple MAF figures? I'm not quite sure but air flow temperature and manifold pressure are used in the calculations and these might now be measuring more accurately as well.

One thing that does surprise me is that a 100% wasn't reached. My surprise is because the engine has a turbo charger and so the air is being forced through the engine which will tend to negate the air flow resistance mentioned earlier. So could this mean the MAF meter is reading low? That would be consistent with information I have read on the most common characteristic of a faulty MAF meter. Perhaps some readings taken whilst driving may be useful. (The above and earlier graphs are based on data obtained with the car stationery).

To buy a new MAF or not to buy?

Although there does appear to be some noticeable improvement does this indicate the MAF is the key cause of the problems? Could it just be a contributing factor and perhaps another component like the EGR valve?

Without a doubt a modern internal combustion engine is a complex system. Before starting this work I had very little knowledge of turbo's, exhaust gas recycling (EGR), mass air flow (MAF) meters etc. but with the help of my smartphone, Torque Android app and device to connect to the car management interface my knowledge is increasing rapidly. I never thought a computer would become as important for maintaining my car as my spanners, allen keys etc!

On my diesel engined BMW I now know that the air passes first through the filter and then the flow rate is measured by the MAF. Next the air enters the turbo where the pressure is boosted and sent through the intercooler before it enters the EGR valve. The EGR valve then mixes the clean air with exhaust gases (for combustion reasons to reduce diesel knock and emmissions) before it enters the inlet manifold and two possible routes. One is a longer route and controlled by a butterfly valve in each inlet manifold port. The purpose of this route, as I understand it, is to get better driveability e.g. smoothing pulling away/ torque at low revs.

From the above description it can be appreciated that there is great margin for error and great necessity for monitoring and control which, of course, is the job of the Engine Control Unit/Engine Management Computer/Power Train module or whatever you want to call it. And as I've found there are a multitude of sensors measuring things such as pressure and temperature and actuators and controls to make things happen.

On my car the actuators tend to be vacuum powered. For example, the EGR valve is opened and closed by an electrical valve which controls a vacuum which acts on a diaphragm and opens a hole to allow the exhaust gas into the stream of clean air. But the decisions the ECU has to make are complex.

For example, the ECU has to know that the exhaust gas pressure is higher than the inlet pressure. So it has to close when the turbo is beginning to develop pressure (which implies the EGR valve only operates at low revs). Then there is the question of how much exhaust gas to admit. This is deduced by measuring the air flow by the MAF and knowing how fast the engine is turning. Overall, there are many things that can begin to go wrong as they get old, dirty and worn.

On my diesel, one of the first things I looked at was the EGR because comments I had read in other forums and blogs seemed to point to this being a potential cause of some of the symptoms. I took it off, simply cleaned it, and got 10mpg better fuel consumption on a long run. That was an amazing return for simple cleaning. However, unfortunately the necessity for cleaning, I'm reasonably confident to state, has been  caused by the problem I suspect is the MAF meter. The coking up of the EGR valve really does look like too much unburnt fuel and I hope this is not affecting things like the turbo or exhaust system. But I want to be confident it is the MAF before spending a lot of money on a replacement.

One of my biggest problems is that I don't know what a correctly operating engine should look like. So are there any other parameters I can measure to help? Well I've come across something called volumetric efficiency. This might give me further indications. It's a measure of how much 'gas' is actually passing through the engine compared to the theoretical amount. More about this in the next post I think.

Can the ECU provide more information to diagnose the problem

With a genuine replacement MAF meter costing upwards of £170 (dealer price £312.02!!) is there any further diagnostic work that can help determine the cause of the error code. And as well as the error code is deteriorating performance and fuel economy. In attempt to diagnose the problem with more confidence the Android Torque app in conjunction with the ELM 327 interface was used to obtain some basic engine operating data.

This is the air flow rate entering the engine plotted against engine revs.


Next the MAF was cleaned and the air flow rate measured again.

Interestingly the correlation looks cleaner i.e. not so many spurious plots and for reference this is the cleaner used.

BMW Recommendations

Having reached a dead end resolving the meaning of the standardised OBD fault code P3263 I decided that, since my car was with the dealer, I would take advantage of the opportunity and pay for an official BMW diagnostic. This returned their code of 3FF0.

To my satisfaction this also pointed to an air flow problem and the advice was to look at two possible reasons. The simplest was the air mass air flow meter and the second was coking up of the cylinder head. However, the latter being rare on my car but more common when the engine is fitted in an X3. A strange occurrence but a relief since a decoke would be a very expensive fault to remedy. However, determining if a decoke was necessary would mean I would have to remove the inlet manifold for a closer look at the air ways.

BMW dealer does not understand standardised OBD fault codes!?

My car is now due for its oil service and so I thought this might be a good time to seek some professional input on the error code P3263. I use a BMW dealership because they've treated me well in the past and my hope is that they know as much, if not more, than anyone else. Unfortunately to my surprise they informed me that the OBD fault codes obtained from my ELM device are meaningless to them! The 'official' BMW diagnostic tools access another set of codes! So I have two choices. To put my faith in the dealership to efficiently solve the problem or continue investigation. Having read several horror stories of people spending hundreds of pounds at garages and the same fault re occurring soon afterwards I favour some further investigation.

To this end I've been studying more about MAF meters, EGR valves, turbo boost control and how that relates to volumetric efficiencies and fuel metering! The tool I've been using is my bluetooth ELM327 interface and Torque app on my Samsung phone. It has allowed me to log these parameters to a spreadsheet  in order to try and analyse what may or may not be happening. The results are interesting but it's still difficult to draw a conclusion without the benefit of a set of results for a 'healthy' engine. One thing it is easy to see is that the EGR valve is working following it's de-coke. Disconnecting the vacuum pipe that opens and closes the EGR causes the results to change dramatically. It also causes black smoke from the exhaust when revved. A common sight from a diesel but something I guess has more than one reason.

More later as I look more closely at the figures or take the plunge and buy a new MAF meter.

MIL On

One reason for this blog was the periodic illumination of the MIL (Malfuncion Indicator Light) on my car. Having read so many horror stories of the cost of professional diagnosis I wondered if it was within my capabilities to diagnose the cause.

Using my OBD adapter and the free Torque app for my mobile phone it was easy enough to obtain the information code being reported by the engine management computer. It was (still is) P3263. The P means the code has been generated by the Powertrain module i.e. engine control unit (ECU) rather than air bag unit etc. More about these OBD codes can be found here. However, whereas the meaning of these codes are often common to all cars and publicly available mine starts with a 3 meaning manufacturer specific! Unfortunately, it seems manufacturers still guard sets of these codes closely and so create or rather hold on to the car owners dependency on the dealer or specialist. (Although they may say the codes should only be interpreted or acted upon by personnel with specialist training).

After much, much searching the best interpretation was relating to the measurement of air flow going into the engine. More accurately a problem with the calculated amount of air compared to the measured amount and this has started a whole new chapter in learning about how the ECU uses the mass airflow meter (MAF) meter, turbo and exhaust gas recirculation valve (EGR). A very steep learning curve but I'm getting there!

To finish this post on a positive note the first part of my investigation and remedial work was to clean the EGR valve. The result was about 8 miles per gallon better fuel consumption! More on EGR's etc later.