R_Lefebvre
Expedition Leader
Edit: just wanted to make sure that the most important detail here isn't lost in the detail here. The MAF sensor for 99-02 P38 and 03-04 D2 (and probably the 99-02 but I don't have details yet) is cheaply and easily available from any VW parts stockist. Skip down to the bottom for the good stuff.
Just thought I'd open up a thread that would be easy to find for people who have trouble with this commonly failed part. I recently went in depth on this thing since mine failed. I did some research in Bosch technical literature, as well as cross referencing info from some other vehicles. The VW, Porsche and BMW guys were pretty advanced in their knowledge.
Started out with a CEL, after which I decided to buy a ScanGauge II which I'll detail another time. I read the codes and got P1884 and P0102. More research and work showed that P1884 is more of an issue with the ScanGauge and the Land Rover computer not getting along perfectly, so ignore it. I never had any drivability issues at all, I did fit a snorkel last winter, I've never had a K&N filter which is reported by many to cause a MAF sensor failure. K&N vigorously deny this claim, with some pretty interesting data. I wonder if it's possible that K&N is being falsely blamed for MAF failures that would have happened anyway. Anyway, that's not the point of this topic.
So, from the Rave CD we have:
Now, inside the MAF sensor is a Wheatstone bridge which is a fairly simple circuit, and I had hopes of opening it up and delving further. Dead end here as the circuit board is quite small and cemented inside a chamber in the sensor. Anyway, basically what happens is that as the velocity of air flowing across the MAF sensor element increases, the voltage signal from the MAF to the ECU increases from 0 to 5V. The ECU then infers this velocity reading in the tube into a mass airflow, and generally speaking it tries to match up this mass airflow with fuel from the injectors at a 14.7:1 ratio.
When the MAF sensor fails, the computer reverts back to a back-up system which looks at the throttle position to guesstimate the airflow. There's a fixed table programmed by the engineers at Solihul. The effectiveness of this scenario is related to how close your atmospheric conditions are to those in Solihul. :Wow1: All is not lost however, as the guestimate is reinforced by the O2 sensor readings which the computer uses to trim the fuel at lower loads. Basically, my truck ran fine with the MAF failed, or even with the MAF completely unplugged. YMMV, and I strongly caution against running at high loads, because O2 sensor fuel trim correction is disabled.
I was leery to just go ahead and replace the MAF for $265US (probably $500 by the time I got it in Canada) because some people reported that they were getting the P0102 code when it wasn't the MAF at all. Buying a used one was ruled out because the failure rate on these things is quite high, chances are a used unit might fail in short order anyway. I wanted to confirm it was the MAF sensor before buying a new one.
It was suggested by some to try cleaning the sensor, but this is hopeless on this sensor because the element is buried deep inside the sensor housing. The metal you can see is the backside of the circuit board, not the element itself.
It's easy to check the voltage supplied to the sensor from the ECU by checking the chassis side of the connector. Simply disconnect the sensor, and probe the pins. Pin 3 supplies a ground, Pin 4 supplies a 5V signal and Pin 2 supplies 12V power, all when the key is turned on, engine running or not. This is the first step to make sure you don't have a gross chassis harness failure.
Now comes the tricky part, trying to read back the signal from the MAF to the computer. There's two ways to do this, both involve needle probes, or the poor man's substitute, needles from your wife's sewing kit. I used the latter. Some people push the needles through the wire insulation into the wire, but I hate compromising the insulation like this. The approach I took was to disassemble the back of the connector, and push the pins in through the o-rings sealing the backside wire entry.
DO NOT let the needles touch or let anything short them out. But if you don't probe the 12V supply, this should be much safer.
First, I confirmed a solid 5V with the engine running. Then I checked the signal coming from MAF, which showed 1.009V at idle. I shut the engine off, key on, and ... 1.009V. Not good. With the engine running again, I revved the engine and confirmed the signal was pegged at 1.009V. The Porsche boys stated that their MAF should read 1.0V or close to it with the engine off, zero flow, and 1.4V with the engine idling. I would appear that one leg of my Wheatstone bridge is fried, the MAF is toast.
Now, why am I talking about Porsche guys? This is where it gets good:
http://www.6speedonline.com/forums/...-cures-they-dont-want-you-know-about-diy.html
These Bosch MAFs are pretty common, basically all the European manufacturers use them, or a similar part. In fact, the exact BOSCH part number on my 04 D2 MAF sensor housing is 0 280 218 010, and is also used on BMW and Mercedes cars. The M5 in particular. Yes, that's right, the 03-04 D2 and 99-02 P38 is the exact same part as a BMW M5.
I then discovered that the sensor insert which screwed into the housing is the EXACT same part as countless other Bosch MAF's, only the housings are different. It's possible a cheap repair could be done if one could find a cheaper Bosch part for... say a VW, and just remove the sensor insert and install it into our housing.
The Bosch part number on the sensor insert is F 00C 2G2 029. It turns out that some of the BMW M5 guys are using the sensor insert from a VW, F 00C 2G2 032, it's cheaper, and the calibration difference is reportedly 4%. The computer can easily compensate for that. The only other difference is the 029 has pins to secure the connector, the 032 is missing the pins. The connector fits, but is only somewhat secure. Some are running just with the friction from the pins, some are tie strapping it down. I found a cross-reference somewhere, and there are no cheap MAFs which use the 029. The 029 is used in all BMW's, Porsches and Rovers.
The 032 part comes in a MAF for a VW Golf, and in fact is quite cheap due to a class action lawsuit. Apparently they actually sued because the sensors fail so often, and they (Bosch, VW, Dealers) were making so much money on it, they dropped the price.
I picked up a sensor from a local independent VW specialist. It's from a 2000 Golf GL 2.0L Gas. VW PN: 06A 906 461A, Bosch PN 0 280 218 002, it has the Bosch OOC 2G2 032 sensor insert which can replace the 029 insert in our sensor. I got it for $150CDN, that's a great price. Turns out this shop could have ordered the real LR MAF sensor and housing for a little more money but... oh well, I was pressed for time. Most importantly, if you have a problem with your MAF sensor in BFN, you can stop in your nearest VW dealer and buy this sensor, they should stock it on the shelves by the case.
I removed my sensor insert from the LR housing and tossed the insert in the bin. Removed the insert from the VW housing, and tossed the housing in the bin. Insert the VW insert into the LR housing and screw it down. The hardest part is the tamper proof screws they use. 5 point Torx-like tips with the pin in the middle. Luckily, you can turn them just using pliers on the heads of the bolts! The part is an EXACT match. Even the plug fits well. It doesn't have the positive locking tabs, but it's held secure. I could tie strap it but...
Fire the truck up, code cleared and has not come back after a month of use. I've also monitored the long term fuel trims in the ECU with no signs of trouble. I'm showing -7% and -6% on the left and right banks. Perfectly in range.
I'd be curious to look into the situation on the 99-02 D2's as they appear to also be Bosch parts, but a different part number. Somebody would have to give me the Bosch pn off the housing. As for the GEMS motors, sorry, sucks to be you. That engine mismanagement system seems to be a complete POS. I'd suggest just gutting it and running an aftermarket system, which you could do for the price of one MAF sensor.
Just thought I'd open up a thread that would be easy to find for people who have trouble with this commonly failed part. I recently went in depth on this thing since mine failed. I did some research in Bosch technical literature, as well as cross referencing info from some other vehicles. The VW, Porsche and BMW guys were pretty advanced in their knowledge.
Started out with a CEL, after which I decided to buy a ScanGauge II which I'll detail another time. I read the codes and got P1884 and P0102. More research and work showed that P1884 is more of an issue with the ScanGauge and the Land Rover computer not getting along perfectly, so ignore it. I never had any drivability issues at all, I did fit a snorkel last winter, I've never had a K&N filter which is reported by many to cause a MAF sensor failure. K&N vigorously deny this claim, with some pretty interesting data. I wonder if it's possible that K&N is being falsely blamed for MAF failures that would have happened anyway. Anyway, that's not the point of this topic.
So, from the Rave CD we have:
Now, inside the MAF sensor is a Wheatstone bridge which is a fairly simple circuit, and I had hopes of opening it up and delving further. Dead end here as the circuit board is quite small and cemented inside a chamber in the sensor. Anyway, basically what happens is that as the velocity of air flowing across the MAF sensor element increases, the voltage signal from the MAF to the ECU increases from 0 to 5V. The ECU then infers this velocity reading in the tube into a mass airflow, and generally speaking it tries to match up this mass airflow with fuel from the injectors at a 14.7:1 ratio.
When the MAF sensor fails, the computer reverts back to a back-up system which looks at the throttle position to guesstimate the airflow. There's a fixed table programmed by the engineers at Solihul. The effectiveness of this scenario is related to how close your atmospheric conditions are to those in Solihul. :Wow1: All is not lost however, as the guestimate is reinforced by the O2 sensor readings which the computer uses to trim the fuel at lower loads. Basically, my truck ran fine with the MAF failed, or even with the MAF completely unplugged. YMMV, and I strongly caution against running at high loads, because O2 sensor fuel trim correction is disabled.
I was leery to just go ahead and replace the MAF for $265US (probably $500 by the time I got it in Canada) because some people reported that they were getting the P0102 code when it wasn't the MAF at all. Buying a used one was ruled out because the failure rate on these things is quite high, chances are a used unit might fail in short order anyway. I wanted to confirm it was the MAF sensor before buying a new one.
It was suggested by some to try cleaning the sensor, but this is hopeless on this sensor because the element is buried deep inside the sensor housing. The metal you can see is the backside of the circuit board, not the element itself.
It's easy to check the voltage supplied to the sensor from the ECU by checking the chassis side of the connector. Simply disconnect the sensor, and probe the pins. Pin 3 supplies a ground, Pin 4 supplies a 5V signal and Pin 2 supplies 12V power, all when the key is turned on, engine running or not. This is the first step to make sure you don't have a gross chassis harness failure.
Now comes the tricky part, trying to read back the signal from the MAF to the computer. There's two ways to do this, both involve needle probes, or the poor man's substitute, needles from your wife's sewing kit. I used the latter. Some people push the needles through the wire insulation into the wire, but I hate compromising the insulation like this. The approach I took was to disassemble the back of the connector, and push the pins in through the o-rings sealing the backside wire entry.
DO NOT let the needles touch or let anything short them out. But if you don't probe the 12V supply, this should be much safer.
First, I confirmed a solid 5V with the engine running. Then I checked the signal coming from MAF, which showed 1.009V at idle. I shut the engine off, key on, and ... 1.009V. Not good. With the engine running again, I revved the engine and confirmed the signal was pegged at 1.009V. The Porsche boys stated that their MAF should read 1.0V or close to it with the engine off, zero flow, and 1.4V with the engine idling. I would appear that one leg of my Wheatstone bridge is fried, the MAF is toast.
Now, why am I talking about Porsche guys? This is where it gets good:
http://www.6speedonline.com/forums/...-cures-they-dont-want-you-know-about-diy.html
These Bosch MAFs are pretty common, basically all the European manufacturers use them, or a similar part. In fact, the exact BOSCH part number on my 04 D2 MAF sensor housing is 0 280 218 010, and is also used on BMW and Mercedes cars. The M5 in particular. Yes, that's right, the 03-04 D2 and 99-02 P38 is the exact same part as a BMW M5.
I then discovered that the sensor insert which screwed into the housing is the EXACT same part as countless other Bosch MAF's, only the housings are different. It's possible a cheap repair could be done if one could find a cheaper Bosch part for... say a VW, and just remove the sensor insert and install it into our housing.
The Bosch part number on the sensor insert is F 00C 2G2 029. It turns out that some of the BMW M5 guys are using the sensor insert from a VW, F 00C 2G2 032, it's cheaper, and the calibration difference is reportedly 4%. The computer can easily compensate for that. The only other difference is the 029 has pins to secure the connector, the 032 is missing the pins. The connector fits, but is only somewhat secure. Some are running just with the friction from the pins, some are tie strapping it down. I found a cross-reference somewhere, and there are no cheap MAFs which use the 029. The 029 is used in all BMW's, Porsches and Rovers.
The 032 part comes in a MAF for a VW Golf, and in fact is quite cheap due to a class action lawsuit. Apparently they actually sued because the sensors fail so often, and they (Bosch, VW, Dealers) were making so much money on it, they dropped the price.
I picked up a sensor from a local independent VW specialist. It's from a 2000 Golf GL 2.0L Gas. VW PN: 06A 906 461A, Bosch PN 0 280 218 002, it has the Bosch OOC 2G2 032 sensor insert which can replace the 029 insert in our sensor. I got it for $150CDN, that's a great price. Turns out this shop could have ordered the real LR MAF sensor and housing for a little more money but... oh well, I was pressed for time. Most importantly, if you have a problem with your MAF sensor in BFN, you can stop in your nearest VW dealer and buy this sensor, they should stock it on the shelves by the case.
I removed my sensor insert from the LR housing and tossed the insert in the bin. Removed the insert from the VW housing, and tossed the housing in the bin. Insert the VW insert into the LR housing and screw it down. The hardest part is the tamper proof screws they use. 5 point Torx-like tips with the pin in the middle. Luckily, you can turn them just using pliers on the heads of the bolts! The part is an EXACT match. Even the plug fits well. It doesn't have the positive locking tabs, but it's held secure. I could tie strap it but...
Fire the truck up, code cleared and has not come back after a month of use. I've also monitored the long term fuel trims in the ECU with no signs of trouble. I'm showing -7% and -6% on the left and right banks. Perfectly in range.
I'd be curious to look into the situation on the 99-02 D2's as they appear to also be Bosch parts, but a different part number. Somebody would have to give me the Bosch pn off the housing. As for the GEMS motors, sorry, sucks to be you. That engine mismanagement system seems to be a complete POS. I'd suggest just gutting it and running an aftermarket system, which you could do for the price of one MAF sensor.
Attachments
Last edited: