BMW Forums : BimmerForums banner

1 - 15 of 15 Posts

·
Registered
Joined
·
190 Posts
Hello all,

Lots more to come on this, but as I've got a wonky issue and a flatspot with my engine I guessed this might be useful for people. This is a guide to testing most of the sensors on the M52 engine, I've got a E36 323i so that's what this guide is based on but I'm sure a lot of it will cross reference with other similar vehicles; besides most of this is stuff is generic concepts which you'll find across most cars in the car-parc.

Disclaimer: I am a hobbiest at this stuff, I am not an electronics engineer or qualified in any way connected to these items laid out in this guide. It's provided 'as-is' and for entertainment purposes only. Please take advice of a suitably qualified individual before carrying out any tests contained in this guide. No responsibility for any damage that will occur will be taken on my behalf.

These diagrams were drawn to stave insomnia so please point out any errors and I'll happily redraw them! They are provided to illustrate the basic concept only, they are NOT manufacturer schematics so please don't attempt to use them as such!

Glossary
ECU - I'll use the almost generic term 'ECU' but accept this as any computer connected to the engine management, DME, PCM etc are other common phrases.


MAF Sensor
The M.A.F. (Mass Air Flow) sensor is rather an ingenious little piece of kit, it allows the engine to calculate how much air is being drawn in and could be considered one of the cardinal sensors of engine management. This is being slowly superseded by the MAP (Manifold Absolute Pressure) sensor which is a mechanical switch in most modern engine management systems but the MAF is still accurate and reliable generally in use.

The MAF acts as a potential divider, and relies on one of the simplest rules of electronics, when a wire is heated it becomes more resistive, when it is cooled it becomes less resistive. The MAF has a heated coil of wire inside it (heated by a current from the ECU), this 'hot-wire' is constantly being heated and as air passes over the wire (the air being sucked in to the engine) it cools it back down again. A basic division circuit uses this to vary a voltage being returned to the ECU. A very simplified diagram:

(Excuse that these are hand drafted, I can't get the hang of doing these on the PC!)

That's the theory taken car of, now let's test the damn thing, you'll need a multimeter and some test leads.
Pin NumberColour TracePurpose
1Red/WhiteBattery Voltage
2YellowOutput to ECU
3OrangeDivider Circuit (Will do inverse to 2)
4BlackGround


Firstly, located the MAF. Mine's held on with these natty cable-ties, but your one will be properly clipped on!
Turn the ignition to 'RUN' position but DO NOT start the engine.



Release the connector on the side by unscrewing the collar


Using the multimeter in DC volts mode check for a voltage very close to the battery one from Pin-1 to ground, and Pin-1 -> Pin -4


Providing this checks out okay replace the multiplug on to the MAF then slide away the protection boot on the rear of the multiplug


With the plug replaced start the engine, then with the negative lead of the multimeter connected to an earth point probe in to the back of the Red/White wire connection (Pin-1), there should be a strong voltage here with the engine running;


With the engine idling, check between Pin-2 and earth, the tolerance here is for around 0.6-1VDC with the engine idling.



Now increase the engine RPM, the voltage should steadily climb - my engine is about 3,500 RPM at this point. The increase should be smooth, any dropping and rising should be considered suspect at a constant RPM.



REPLACE THE BOOT!

Troubleshooting.

IssueResolution
No voltage on Pin-1 (Red/White Wire)Double check meter settings and connections - Points towards fuse failure or serious ECU/electrical issue. Requires further investigation.
Voltage output doesn't change with RPM, or changes are not smooth, go up&down etcCheck meter settings and connections - points towards a failed MAF. Remove the MAF and with some alcohol very carefully clean the hotwire elements with a cotton bud and retry. Providing the battery voltage test appears okay suspect failed MAF. Check very carefully for air-leaks around the MAF!
What voltages should I expectThe data I have here suggests 0.6-1VDC idling and around [email protected] The figures aren't absolute, a nice smooth increase and decrease are more important but if you've strayed far from these values clean the MAF, check for air-leaks then retest.

That's all the in-car testing, there are bench tests I can cover but they're getting beyond the scope of this simple guide.
 

·
Registered
Joined
·
190 Posts
Discussion Starter #2 (Edited)
Throttle Position Sensors

These are potentiometers (voltage dividers) then are used as a voltage input by the ECU to calculate how far open the throttle plate is.


Is the basics of the circuit, imagine here 5v is being sent from the ECU, when the throttle is 50% open 50% of the incomming voltage is diverted back to the ECU (so it sees 2.5V), 50% is pushed out back to earth. This is a bit like a see-saw, at small throttle openings the voltage is mostly sent one way, as the throttle plate is opened it's swung back the other way. There can never be gain in this circuit so it can never exceed 5V nor can it drop to 0V.

That's dandy 'n' all, but how do I test it?

This is pretty easy but remember if you have ASC+ you'll have TWO throttle position sensors, one for the main engine throttle and one for the 'backwards' throttle for the ASC system, the smaller on the left being the 'normal' one and the larger one on the right the ASC+ body one. Running issues are mostly likely to the concerned with the 'normal' one, and this is where you'll be looking the majority of the time, it's the same test procedure for both though.

The TPS use a 5V voltage, not battery voltage, so expect 5VDC as the baseline reading.

Locate the TPS, and unclip the multiplug. Keys out the ignition here, it's a passive test!


Measure the resistance between the most left/right hand pins, I got 4.7kΩ which is the entire resistance across the whole assembly, open and close the throttle (you can reach underneath the throttle body and rotate the quadrant), this should remain totally unchanged throughout.


Measure between the centre and righthand pin, the resistance with the throttle resting shut should be around 1kΩ-2kΩ, slowly open the throttle and watch the meter;



Now, this is the critical bit, this sensor is perfectly linear and the reading should rise smoothly, any 'flat-spots', readings jumping around or open circuit position points squarely at a defective TPS. These 'wonky bits' will be felt when driving and can create awful flat-spots when accelerating. Mine is toast as it wanders about quite a bit then shorts-out at about 75% of pedal travel.

Replace the multiplug!
 

·
Registered
Joined
·
437 Posts
nice write up, for voltage checks the meter will be fine but for checking the output of the maf you should really use a scope as there can be a bad waveform that the meter just won't see.
i bought a 4 channel scope from amazon for just over £100.00, very useful for trouble shooting digital and analouge outputs.
chris.
 

·
Registered
Joined
·
190 Posts
Discussion Starter #5
nice write up, for voltage checks the meter will be fine but for checking the output of the maf you should really use a scope as there can be a bad waveform that the meter just won't see.
i bought a 4 channel scope from amazon for just over £100.00, very useful for trouble shooting digital and analouge outputs.
chris.
The MAF is a voltage divider - a good, fast-switching meter will give you a good approximation within £5 worth of kit, although I do totally agree you'll get a FAR greater resolution with a 'scope. I have a 'scope and I'm willing to add on the use of one - BUT - my reservation being that I could add loads to this regarding using an oscilloscope potentially pointlessly as if you happen to have an oscilloscope sitting in your garage you're likely to have a fair idea how to use it anyway! :p

A good compromise would be to use an analogue meter, but hands up who has got one of those lovely bits of kit knocking about in the garage!
 

·
Registered
Joined
·
190 Posts
Discussion Starter #6 (Edited)
Sorry, rain has stopped play but more to come!

Honourable Mentions

Knock Sensor
The knock sensor is a glorified microphone (yes, I'm really oversimplifying that!) - Frankly it'll work or it won't, the sensor is tuned to physically listen for 'knock' - should the frequency the microphone is tuned for occur in the engine the sensor will generate a tiny pulse of current. There is a lot of conjecture about how effective knock sensors are, if a knock sensor issue is expected short of replacing it, it's always worth checking the torque of the tightening bolts (they're finely tuned, a physical force holding it to the engine will affect its tuning. The spec' I have for the M52 engine is exactly 15Nm. If you get error codes that knock is constantly being detected seek further advice as a) The sensor could be faulty b) Your engine could be a big knocky mess! Unless you've found some 25 year old petrol somewhere and tipped that in the engine!

Oxygen Sensor (aka. Lambda probe, O2 probe, HEGO etc)
This is a biggie, but I'll side with caution here. If you understand how to test one then you won't need this thread anyway.

Oxygen sensors are incredibly fragile electronically, I work with oxygen sensing instruments all day (Nernst cells) and can be a bit of a bore on how they work, I'll stick to the 7" Radio edit here, if anybody would like the 12" Extended Remix of testing or theory please shout. One thing I must stress is DO NOT, EVER attempt to measure the resistance across the output of the sensor, to do this requires a current to be driven in to the sensor..... Like I said they're very fragile.

Oxygen sensors used generally in cars work by comparing the oxygen present in the exhaust gas to atmospheric oxygen. If the exhaust is laced with free oxygen the engine is running lean (not enough fuel) the sensor will deliver a poor output (nominally 0.1V) conversely if there is little oxygen in the exhaust gas then the rise rapidly indicating a rich mixture (nominally 0.9V). When the engine is running it's looking for a mid-ling value indicating the engine is running in the 'Goldilocks' zone, as we're dealing with a feedback loop the engine will drift in one direction, the injection system will take over and try and drive the engine in the opposite direction, this switching should swing around 0.45V when cruising but should be seen to switch fairly quickly a bit rich/a bit lean/a bit rich etc... Allowance has to be made for sudden acceleration where the fuelling will be allowed to become richer so the nominal values apply to partial throttle cruising. The switching periods are variable, and beyond the scope of this guide. Oxygen sensor readings are ignored until the engine has reached operating temperature due to chemistry that's not very interesting.

Suspect oxygen sensors can be tested, most usefully would be the use of a 'live-data' type trace of the output or with an oscilloscope in a workshop BUT if you don't have access to this kit you'll be a bit stuck. One 'home-brew' test that does work is to monitor the output of the sensor in a vice, then using a blow-lamp heat the living piss out of the end of the sensor with a very hot flame - the output should hit around 0.9v rapidly, remove the lamp and it should fall dead-low as it's comparing the air around the ends of the sensor which should be similar.

If it's sluggish or not swinging the output then it's knackered, BUT this test is forcing the sensor to work at the extremes so will potentially only indicate a REALLY knackered sensor.

ONE VERY IMPORTANT BIT OF ADVICE
The car parts market is littered with shit oxygen sensors. I mean, utterly useless ones. If they look cheap for no good reason discount them, I'm all up for thrifty car maintenance but you will 'buy twice' with a Chinese £15 O2 sensor. You don't have to use a BMW one but speak to your local parts shop and ask them which ones they recommend and what kind of warranty they'll offer. If they pull a face and say well I can sell you an X- for £20 with 30 Days or a Y- for £50 with two years, take the wise choice.

Oxygen sensors are fairly generic (the vast majority of cars use only a handful of sensor types) but get one with the correct multiplug, there were some spectacularly nasty ones about a few years ago that had flying leads and required salvaging the OE plug and soldering it on to the new sensor.... Not worth saving £5 over!

Oxygen sensors are usually refered to as by the number of wires, for example "three wire type", "four wire type", 1-5 wire types exist;
WiresConnections
1Old-skool narrowband sensor, just about obsolete now. No heater or reference, just a signal and grounded through the case.
21x Signal wire, 1x Ground. As the '1 Wire' this is just about unseen these days.
31x Heater, 1x Signal, 1x Ground
42x Heater connections, 1x Signal, 1x Ground USED ON THIS ENGINE.
52x Heater connections, 1x Signal, 1x Reference, 1x Ground (Latest type)
Yeah, if you think it's failed and with nothing to prove it's not then consider biting the bullet and change it with a decent one.
 

·
Registered
Joined
·
10 Posts
Just wanted to say thanks for posting this information, I;ll be trying this out tomorrow, I'm getting some maf codes and think might be a bad maf.
 

·
Super Moderator
Joined
·
12,946 Posts
Ya can see what all these sensors and things are doing and enable stuff remotely along with live data with INPA diags ya know...
 
1 - 15 of 15 Posts
Top