Technical MB General

Technical non-specific MB

Mercedes ML W164 V6 Diesel engine tapping under load

Investigating a problem with a good friends Mercedes ML MY2006 320CDI who had noticed a heavy tapping noise from the engine under under anything above moderate load. It sounded expensive

Just like a top end rattle similar to a noise which would be if there if there was an issue with a piston, little-end or ring/gudgeon. Nothing could be heard at idle, which was odd in itself. Lifting the engine revs slowly, whilst stationary, would result in a smooth increase in RPM without the accompanying tapping noise. He reported that performance was down a little too, feeling that top end maximum speed would be unattainable, he was right.

I noticed the service indicator was illuminated on the dash and the mileage rested at some 3500 miles over when the oil should have been routinely dropped. Still, dirty oil alone wouldn’t usually cause a sound like this, but we needed a baseline to work with. A full engine service was carried out, including the twin air filters which frankly looked like leafy undergrowth! The diesel filter was replaced, filling the new one to the brim with a branded diesel system cleaner. The owner had already added ‘in-tank’ cleaners, thinking it could be an injector related noise but without a positive result. Once the neat diesel cleaner had been forced through the direct injector system the car was taken out for a test drive. The noise was still apparent, whenever you pushed the pedal in determination, there would be a tapping soundtrack running parallel to the normal induction noise of the engine.

I started thinking about the noise. Although the tapping was following the RPM rotations of the engine it was not every rev. It was audibly about half the frequency of the indicated RPM on the tacho. As a 4 stroke engine (Diesel or Petrol) does 2 full revolutions per combustion cycle, it was easy to determine it was either injector or valve related. It was highly unlikely that a con rod little end would only knock once every two revolutions of the engine, especially as the forces on the piston would be either dragging it up or down the bore, changing linear direction twice per revolution. I know that sometimes noises only occur on the power stroke when forces are a maximum, but this noise was far lighter than you would expect from a bearing surface that was failing or wearing oval.

The above video illustration explains nicely what I mean. It was most probably an issue with something valve or cam related or as the owner had wondered, an injector.

I dug a little deeper into the engine, removing the plenum and air filter housings to physically inspect the injectors with a good torch. I was looking for the infamous ‘Black Death’ (see video linked below) that often plagued more commonly the earlier 4 and 5 cylinder diesel types as you would see in the T1N Sprinter and many MB passenger cars of the time. The Mercedes OM642 V6 power plant has been known to suffer similar issues with combustion gasses leaking past the injector seals, in some cases making a clicking/tapping sound under load and causing a hard, often shiny black carbon build up around the mating surfaces of where the injector body enters the head (injector pocket)- nothing found!

After running the ML for a few more miles there would be occasions where the tapping-under-load noise would mysteriously disappear. That was great news. Noise from the mechanical failure of a con rod, piston or camshaft would not usually come and go, so that narrowed the field splendidly.

More ‘Italian tune-up’ accelerations and driving style where conditions allowed, over the following week in the hands of ML’s owner, saw the noise diminish and vanish completely over a period of about 10 days.

So, what was it!

I suppose we will never know for sure, but here is my view:

Considering all of the above, noise and lack of sparkling performance, it could have been either – A sticking hydraulic tappet that was suffering from the lack of lube from dirty or plain old worn-out oil, or the lifter just not filling to full pressure due to sludge etc?

Or was it a dirty Piezo diesel injector that was simply misfiring and not performing correctly under load?

This tale goes to illustrate that you should always give a little time to your diagnosis before jumping in with both feet and undertaking what could well turn out to be costly, fruitless unnecessary work.

Everyone was happy with the outcome on this occasion. My advice: Look, listen, make some changes, WAIT a while and observe!

Mercedes ML W164 will not go above 4mph and all wheels brake.

My Mercedes ML320CDI (2007) had a mind to play up on the motorway and put me in some hot water with regard to the vehicle being totally disabled.

Looking back, I had noticed a few symptoms over a month or two that I had not connected to the real underlying issue that would eventually completely put me out of action on the M6 only ten miles from home!

The month running up to event I had noticed an occasional hard change down at slow speed when coasting to a halt, I simply put this down to the foibles of the 7G-tronic gearbox and ignored it. There had been a very slight groaning noise on particular right hand bend on my daily journey to work and again, unless I was listening for it it just got ignored. Then one particularly wet day the ABS, ESP and Tyre pressure monitoring inoperable warning appeared on the dash – everything seemed fine, I found a safe place to turn off and restart the car, everything was fine.

I had a very early morning trip up to Liverpool and it was an awful day with heavy rain and large amounts of surface water on the road, I got to Liverpool without drama and pulled up to a set of traffic lights as I slowed there was a really loud groan, almost like a passing jet under power take off – I actually dismissed this as just that, thinking it was a low flyer out of John Lennon Airport. I parked up and went about my business.

About three o’clock I set off from the carpark, through the town and joined the M62 back home. Later joining the M6 southbound is where the fun would really start!

Outside lane of one our marvellous SMART motorways saw the dashboard light up like a Christmas tree, ABS, ESP and Tyre Pressure Monitoring not working. I eased off the throttle and it became obvious the car was holding in a single gear. As I could not quickly get over to lane one (not as it would have done me any good as there was no hard shoulder now thanks to the SMART motorway improvements!). It took me a few miles to get myself to a place of less burden. Each time I pressed the throttle the ESP triangle would flash stopping doing so when I lifted off the power. I made the decision I was going to try a rolling restart in an attempt to reset the problem – big mistake. After selecting N and turning the engine off and on again I had no drive at all and the gear selector stayed in N no matter what I did with the lever. Attempt 2: off and on, this time I had drive in what seemed like second gear, useful rpm’s at 50mph! Enough was enough and given the circumstances I decided to hang in there and sweat it out, locked in a single gear – it was certainly far better for me than the ‘no drive at all’ option I had adopted earlier.

So I got off the motorway and pulled over in the nearest layaway I could find – what a relief! A restart and off we go, the ABS and ESP warnings still lit, as soon as I hit 4mph the brakes activated on what seemed like all wheels and bought me to a halt, it was impossible to move the vehicle above a slow walking pace. Many restarts later it was obvious this was not going to rectify itself and I called the RAC….. That will be four hours wait sir! I called up a friend as I was only ten miles from home and he took a detour past my house and collected the Autel code reader en route. He arrived some minutes later and we plugged in. Wow! anything speed related had flagged a fault, including gearbox, steering, ETC/ESP Instrument cluster, the lot. I cleared them off one by one which eventually allowed me to clear every warning on the dash. I set off again and exactly the same happened, 4mph and all wheel ABS braking arrested the vehicle to a halt with no intervention my foot apart from pressing the ‘go-go’ pedal like mad, which I will add had no affect whatsoever.

So sit it out for the RAC man… I had already described the fault over the telephone as ‘no drive’ and hoped they would send the appropriate vehicle to recover me. They did, three and a half hours later this recovery vehicle arrived, my jaw dropped…

The rest is history, the ML was delivered home in style, I was impressed by the way the set up worked and to be honest was fascinated by the mechanical wizardry of it all. My car was limped into its parking place, the gates closed and left overnight. Still I might add with not one fault showing on either the dash or code reader.

The following day I had convinced myself that it was a traction control problem and coupled up the Auto in live data mode. A short run down the road at crawling pace revealed the rear left wheel speed sensor to be counting at about half the impulse counts of the other three wheels. Sure enough as the speed rose above 4mph the brakes applied and halted the vehicle.

Off came the wheel and the sensor removed, there looked little wrong with it, no rust debris stuck to it as you often find. I used a torch down into the sensor pocket and it looked clean and dry. I have seen similar sensors on Jaguar hubs that live in a debris filled pocket that builds up and wears into the seal of the hub bearing. Once water gets to play in here the magnetic poles embedded into the bearing seal start to delaminate. Once this happens there is no hope in changing the sensor. The complete hub bearing has to be replaced to renew the magnet pole ring within the bearing oil seal. Fortunately the ML’s seal looked in fine condition as far as I could see down into the sensor pocket on the hub carrier.

The later wheel speed sensors we see on these and later stable Mercedes are ‘Active’ sensors. These have some electronics built into them and are basically a ‘Hall Effect’ device that reacts to magnetic fields passing through their target area. The ring of ‘miniature magnets’ that are moulded into the bearing seal are ordered North, South, North, South and so on making up a full 360 degree angular device. Because the sensor is electronic it can determine wheel speed far more accurately than older ‘Passive’ inductive types (the toothed wheel of yester-year) because of the modern arrangement the sensor can determine rotational speed from a standstill and even detect direction of rotation!

I think the problem with my sensor was it was only detecting either North or South poles from the sensor ring and not both, this would have halved the impulses measured as I saw on my live data. There was no fault present since I cleared them because the ESP traction control was controlling exactly as it should given the wheel speed information. Let me explain. One wheel gives speed information slower than the other three, the electronic controller thinks three wheels are spinning and applies forced braking to those faster moving wheels to control the vehicle. Of course this is not actually the case but in terms of what the controller sees it is reacting as it was designed.

Replace the sensor. Very simple one small bolt on the hub carrier and the sensor withdraws. There is a single connector about 250mm away on the upper suspension arm from which the sensor simply unplugs. Mercedes Benz want an eye watering £99.00 plus vat for a genuine ATE wheel speed sensor – err don’t think so guv! Our local motor factors had a pattern part on the shelf £24.00 No brainer?

So replacement sensor fitted, a final clear down of any fault codes and we were away. Perfect fix. A couple of other things became apparent, the slow speed gear change harshness had gone, the slight groaning no longer present in the particular corner on the way to work and how about this… – The Hill Start Assist actually worked far better than it had ever done, only previously working as well in reverse! Transformational and a fantastic low cost fix to what appears on the face of it to have been a ‘big ticket’ job.

Sprinter >2006 Cutting out and None Start – A Few Notes – Causes and Things to Check


Sprinters up to model year 2006 share some similarity with the later model when it comes to ‘mysterious’ cutting-out and repeated intermittent non-start issues, this article covers specifically the earlier 4 cylinder Sprinter >2006, although some of the items covered are also common to the later model. The points covered here are some common failures and do not represent an exhaustive list of issues that may be causing problems with your vehicle. You may find at least some clues that help towards your own fault finding from the information given below.

To start a Mercedes CDI diesel engine, there must be a set of conditions met that the ECU requires to initiate running the engine.


The fuel rail must provide an absolute minimum cranking pressure of between 250 and 300 Bar, any less and the ECU will not allow the engine to start. Causes for this pressure not to be maintained could be due to internally leaking injectors, faulty fuel pressure regulator or its O-ring seal (see covering article here), faulty rail pressure sensor, High or Low pressure fuel delivery pumps. Equally any air resident in the fuel delivery system will hamper starting. Usually the seals used to terminate the clear plastic pipes from the low pressure to high pressure pump give rise to air bubbles visible in these pipes, which is not a good sign and will need to be further investigated. It is worth noting that unlike any circulating diesel feed system, the fuel exiting the high pressure pump is not self bleeding and any air held within the rail will take some time to clear should you open the system to atmosphere (for instance if you do a poor job of pre-filling a new fuel filter at service).

Sprinter >2006 Cutting out and None Start 1

It is worthy of note that the diesel pressures within the system of a running engine could be in excess of 1,600 Bar that is very capable of doing you some damage! – always use caution when inspecting and working on running CDI fuel systems.

Should at any point the measured pressure drop within the fuel rail, say if there were a serious ‘high-side’ leak or if you were to open a feed union to an injector the ECU would instantly cut the fuel delivery and shut down the engine as protection. Equally there is overpressure protection for the fuel system, where if it ever rises beyond a predetermined ‘safe limit’ the ECU operates the shut off valve on one of the three piston element heads of the HP diesel pump thus dramatically reducing delivered pressure to a safe level. This shut off device is the large tower like solenoid that extends from one element (there are three) of the high pressure pump, it has a two pin electrical connection and loom plug at its end.

Sprinter >2006 Cutting out and None Start 2

‘Live data’ is the best way of determining fuel system pressures and the operation of the engines attached control devices, but obviously in a running engine. You will have to inspect the fuel system of a dead non-start engine using the functional details above as a guide. Unless you have a major fuel leak or the delivery pumps have catastrophically failed, fuel related problems are not often sudden and tend to show themselves as intermittent transient problems for a period of time before they get to the stage where the vehicle will no longer start. More often non-start problems arise from the electrical system.

A really useful link on the Sprinter fuel injection system here


For the most part the electrical system will be the major cause of any none or poor starting issue. There are two main timing sensors fitted to the engine: Crank Position Sensor and Camshaft Position Sensor.  These relay engine positional information to the ECU so that it can determine when to accurately fire the injectors so that the engine can run.  Any synchronisation errors that occur between these two primary angle measuring devices will cause an engine to slip into limp home if running, and if stopped, the engine will not restart until it gets valid signals from both.  The cam position sensor is mounted on the top rear of the valve cover slightly to the left of cylinder 4 (rearmost looking in) and gets its positional signal from a lobe on the cam. Due to its location directly above the exhaust manifold, it is subjected to regular extremes of heat and daily cycles of hot and cold.  It is not uncommon for this sensor to exhibit thermal problems, where a cold engine will start and run fine, then once warm, would falter or not restart until cool. Replacement of this sensor is quite cheap – sub £30, and even if it does not cure any problems you may have, it is best to ‘eliminate it from enquiries’ if there are any synchronisation errors recorded.

The crank sensor lives on the block seam of the engine and gearbox bell housing, just above the starter motor, it picks up its pulses from teeth on the flywheel. Generally it is quite reliable – if you have problems starting the vehicle it is often a good quick check to watch the tacho needle flick slightly as you crank the engine. If it flicks, then this is normally a sign of an impulse from the sensor reaching the ECU. If however the tacho needle does not flick on spinning the engine over, then suspect crank position sensor problems.  It is possible this sensor can suffer from thermal problems but not as frequently seen as with the cam sensor.

The connecors and cabling to the injectors are important things to check, if any connector to any single injector becomes disconnected, then the engine will not run, equally if one injector solenoid becomes disconnected when the engine is running the engine will die.  It is a known problem for the injector wiring loom that lies under the black plastic cable tray beneath the fuel rail to have issues whereby the injector cables rub and abrade against the metal surface of the alloy rocker cover over time. Once the cable insulation weakens and any of the conductors short to ground, it has a similar effect to disconnecting an injector – the engine will stop. You can cut strips of plastic bottle and slide them under the cable tray to insulate the cable from the alloy cover to test if you suspect this is an issue.

Sprinter >2006 Cutting out and None Start 3

The injector solenoid wiring can be easily checked for continuity and insulation to ground by removing the injector connectors and metering out the wires directly to the ECU under the dash.  The injectors connections are the large gauge wires on the push in connector to the far right looking at the ECU under the dash.  Release the ECU retaining springs to drop it down into the passenger footwell to make it easier to work on.

If you detect anything untoward with the wiring during your continuity checks, the first obviously place to look for broken cables and damage is to cut open the outer insulation on the section of loom that drops from the engine, over the left engine mount on its way to the rear of the battery shelf area. Here you will find the most mechanically vulnerable selection of sensor and engine management cables – be sure to check here first. Note: although not related directly to non starting, as previously mentioned in other posts, the turbo control, vacuum solenoid valve loom, hooks round the front of the radiator from behind the passenger headlamp and travels across the front crossmember. It passes the horn onward to the vac valve situated under the air box. Commonly this loom is found to be damaged or broken in the area where it snakes past the headlamp tinwork on the front panel. See here for more related fault finding detail.

If all fuses prove good, and the engine cranks but will not fire, then take a look at the ECU connectors. Remove each of the plugs and check that water, usually from a poor windscreen seal or rust-rotten screen surround, finds its way onto the ECU.  Over time the fine connectors and pins start to corrode and all manner or electrical issues ensue.  To remove the ECU plugs with lever tabs, push in the small locking pip just behind the upright grey lever, on the black part of the shell. Then bring the grey lever down over it and continue to move the lever to the horizontal in an arc, this will eject the connector from the multiplug.

Sprinter >2006 Cutting out and None Start 4

I have had instances where intermittent cutting out has been due to dry joints on the PCB inside the ECU alloy box.  To check this out – once all the cables have been removed undo the six torx head screws holding the cover plate onto the pressed aluminium ECU box. Once removed, lift off the plate and remove the PCB assembly from the pressed case.  You can now inspect each pin solder joint with a magnifying glass to look for a broken or poorly soldered/corroded joints.  Often careful re-soldering of the suspect joint will provide a suitable repair, obviously if its corroded rather badly, the chances are that the multilayer board is damaged and at worst a replacement ECU could be needed (very rare).

Sprinter >2006 Cutting out and None Start 5

If you ever get any code reader messages relating to ‘capacitor voltage low’ (Autel or none-Star diagnostic tools) I have always found these to be a good indication of faults connected to the CPS, and after replacement or repair to its connecting wiring, it has always fixed the problem which was originally accompanied by an engine that will just cease running for no apparent reason. My first advice would be when the only recovered code is one relating to ‘capacitor low’ then more often than not – its related to the CPS and well worth a try first!

Sprinter >2006 Cutting out and None Start 6

Notice the relay on the right has a ‘pushed-in’ terminal

If the starter does not crank at all, then it is good to understand that the start command is issued by the ECU and not the key!  This is because the ECU monitors the key security module and will disallow starting unless a valid key (with correct internal chip) is being used in the ignition switch.  Any problem that displays ‘KEY ERROR’ in the dash LCD module is connected to using an incorrect key, a broken or missing key code chip, or a problem with the key reader detection coil (around the ignition switch) or the key security module itself (SKREEM).  If the key signal is valid the start signal will be actioned by the ECU. In auto versions it will also look to see if the transmission is in neutral or park.  Once an ECU start output is active, the start relay is commanded.  It lives beneath the driver seat, accessed through the flip off side panel, once caused to pull in – this puts voltage onto the coil of the starter motor, via a black with yellow tracer wire in the loom.

Sprinter >2006 Cutting out and None Start 7

more clearly – the pushed in relay contact

I have seen odd instances where because of the nature of the size of the relay spades, they ‘give-way’ when pushed into the fuse board receptacle, pushing them up inside the relay case – misaligning the internal switch contact faces. Often a root cause of intermittent starting/cranking with just a ‘click’ – should obviously the battery and starter cabling prove to be in good order.  The dead giveaway here is if the headlamps stay bright and hardly dip when an attempt to crank is made, if it is not the relay and good voltage is reaching the stud terminal on the starter solenoid, then the chances are the solenoid itself is faulty and the starter would need replacement.

Sprinter >2006 Cutting out and None Start 8

Inside the relay – the silver plated contacts are pushed out of alignment making for poor or high resistance switching

You can of course quickly test the starter by ensuring the vehicle is out of gear and then shorting the large 13mm power stud/nut on the starter to the smaller 10mm nut on the solenoid with a robust old spanner or jump lead clip. If the starter is healthy, then the motor should crank. If not, then you are most probably looking at a faulty starter motor.  If it does crank, and will not turn over from the key, then you will have to investigate why following the information in the above text.

Sprinter >2006 Cutting out and None Start 9

I hope the above gives you a little information to help you fault find a few little known anomalies with the Sprinter and hopefully it leads to identifying  your starting issues.





Mercedes S500 (W220) and others – ‘Fly-by-wire’ throttle pedal fault


There had been occasion when I would decelerate to an intersection or junction and the engine would not settle to an even idle and remain rather ‘lumpy’. By tapping the pedal only slightly to raise the RPM’s by only the smallest amount resulted in the engine dropping back to its normal smooth idle.  It was also noticed that when pulling away from a junction the accelerator pedal occasion required more travel to get the car moving than it normally would, and when it did begin to pick up speed it moved perhaps more swiftly than intended.  There were no reported engine codes and also nothing illuminated on the dash to indicate a problem.

Asking around, a faulty throttle pedal position sensor was the confirmed as the most likely cause and the consensus of opinion was that they rarely ever caused a fault code to be registered, in fact one of my best contacts in the trade said that this condition directly pointed to the pedal issue, doubly confirmed by the fact that no code was reported.

Mercedes Electronic Throttle Pedal Problem 1

So I purchased a used part from a local breaker for £35 and fitted it in less than quarter of an hour from start to finish.

All that is involved is to remove the floor mats if fitted to gain unhindered access to the accelerator pedal.  In the rear moulding there is a plastic circular bung about half way up the assembly, using a small screwdriver or pick lever this out.  Beneath lies a recessed 10mm nut and washer, undo this with a small socket and extension.  Now grasp the pedal and lift it upward from the bottom, levering it outwards slightly to clear the fixing stud, once clear, pull down to disconnect the pedal assembly from its top locator.  The pedal will now be free, retained only by two wiring/loom connectors.

Disconnect the upper multiway plug to the position sensor model and remove the smaller two pin connector from the ‘kick-down’ switch from behind the pedal.  Note how the kick-down wire is threaded along the side of the pedal assembly and remove it.  The pedal can now be replaced with the new unit, threading the thinner loom back in its correct position and then connecting both wiring connectors back onto their respective plugs.

Feed in the pedal assembly, upper section first. Once located correctly, swing down the pedal to mate with the fixed mounting stud and replace the fixing.  Job done, all you then need to do is replace your mats and test the vehicle.

Mercedes Electronic Throttle Pedal Problem 2

On the S500 (W220) there were two types of pedal assembly used, one with a part number A220 300 01 04 for the post face-lift version circa 2004 onward and  part number A220 300 00 40 for the earlier model ranges to this point. Its a good idea to remove your own pedal, as it is such a simple task, to check the part number of the item fitted before searching for a spare.  A new non Mercedes replacement should cost in the order of £150.00 and a genuine part closer to double that.  A good used item should be in the range £35-£70 – the choice is yours!

Throttle pedal fault Mercedes Benz 1

Interestingly there is a large number of people who still believe there is a potentiometer within the pedal assembly, this is of course not the case and the often read advice to spray switch cleaner on or around the sensor is largely futile.  I have taken (broken) this unit apart so that you can see inside and confirm that in this case and most of the Mercedes models from about year 2000 on, this pedal assembly is all solid state and non serviceable.

Throttle pedal fault Mercedes Benz 2 Throttle pedal fault Mercedes Benz 3

What you can see within the housing is a pair of hall effect semiconductors mounted to a PCB. Each sensor is fixed within a static soft iron slotted core and as the pedal is rotated a small annular magnet rotates about this core.  Minute changes in magnetic field are detected by each of these two ‘Hall Effect’ sensors (compared electrical outputs from both sensors are checked for integrity, before a plausible position signal is accepted by the ECU as a failsafe, much in the same way as there is mechanical failsafe protection built into the pedal assembly by having two individual springs controlling the pedal return, just in case one breaks, allowing the pedal to return to a safe ‘throttle closed’ position rather than simply falling to the floor in a ‘full on’ scenario – Which would be a grand ‘brown trouser’ moment.)

Throttle pedal fault Mercedes Benz 4

Each degree of throttle pedal rotation at the rotating hinge point, relates to a predetermined magnetic field strength and this is interpreted as positional data referenced back to the ECU.

Throttle pedal fault Mercedes Benz 6

So, not a carbon track of a potentiometer to break or wear out, just electronics.  Obviously failures occur but in general this is a far more reliable, if slightly more complex solution to throttle pedal position sensing.  I would imagine the most common problem with failure of these pedal assemblies is the ageing of the semiconductor devices, once one device gives a slightly different output at any stage of its sweep, this is interpreted by the ECU as a mismatched signal and it will be ignored – hence dead pedal!  Once the signals are again in unison the throttle position is relayed and off you go…

Mercedes Sprinter 2007-on and VW Crafter Coolant Level Warning Problem



The later Mercedes Sprinter and Volkswagen Crafter (and probably many others) uses a round plastic header/expansion tank in the coolant system.  This nylon spherical tank houses a rather clever float level switch that not only indicates a ‘low fluid level’ but also will illuminate the warning lamp in the dash binnacle if the ‘specific gravity’ of the coolant is less than specified – indicating to the driver that all is not too well with the coolant circuit.

Mercedes Sprinter Coolant Level Sensor Problem 4

On one of the fleet Crafter vans the coolant level warning lamp had come on and the driver had reported the fact.  When I investigated the issue there was no warning light on the dash, water level was fine but it could be clearly seen that the float inside the header tank had broken up, the plastic becoming somehow brittle and breaking away from its vertical retaining slide/guide. The internal float section is not available as a separate part so a ‘used’ second hand bottle was purchased.  After disconnecting the electrical connector at the base of the sphere, it is possible using a couple of sets of brake hose clamps to remove all the rubber hoses from the bottle in situ.  Once the single torx screw has been removed from the bottle mounting foot, orientating the bottle directly down into the void beside the engine, the connected hose-work can be removed quite easily and the bottle withdrawn upward without losing almost any coolant.  What coolant remains in the expansion tank can then be decanted back into the new bottle once fitted and piped.

Mercedes Sprinter Coolant Level Sensor Problem 3

On closer inspection once the bottle was removed, it could be seen that although the float inside the expansion tank was in this case the obvious problem (the actuating magnet it transpired, had been removed from the loose float by the driver, to turn off the lamp – why do they do that!)  Further dismantling of the float switch and level sensor revealed exactly how it functioned and also highlighted a further weaknesses that could in some cases cause issue with level measurement and warning lamp activation.

Mercedes Sprinter Coolant Level Sensor Problem 5

In the base of the tank is a moulded central vertical column that forms an upright guide that the tubular float moves on. The float rides up and down this central guide tube. Beneath and fitting into this tube is the sensor device. Although moulded in a plastic holder, this sensor is actually houses a small glass tube type reed switch that is triggered by the passing float (and magnet). The contacts of the reed switch close as the float magnet descends past a predetermined point and the warning lamp is activated. As the float rises above the reed switch, the magnet no longer causes the contacts to close and they spring open, resulting in extinguishing the coolant warning lamp.  There is a small resistor across the contact terminals of the sensor (around 800 ohms – the one I measured) This I am certain enables the ECU to detect in addition to reporting the coolant level status, a disconnected plug to the sensor, as it would have four possible measured electrical ‘states’ present between the sensor terminals as measured from the ECU. (1): Reed switch contact open, 800ohms = Correct fluid level. (2): Reed switch contact closed, short = Low fluid level or weak antifreeze water mix, Warning lamp lit. (3): Open circuit = Connecting cable disconnected or cut, warning lamp lit. (4) Cable shorted together = Connecting cable linked or fused together, warning lamp lit.

Mercedes Sprinter Coolant Level Sensor Problem 1

It can be seen from further ‘distructive’ disassembly the level sensor that it is susceptible to corrosion around the resistor and wire junctions, obviously due to the service environment it which it lives.  This discovery is important as if the float inside the water tank looks in good order, it may just be that the level sensor is faulty.  Depending on how keen you are with your fault finding, you could just slot a 1k resistor across the sensor plug and see if the warning lamp extinguished.  If it does, the chances are the sensor is faulty and can simply be rotated through 90 degrees and removed from the bottle base without any draining down or tank removal. Slotting in a working sensor, turning and reconnecting its plug to fit the replacement part.

Hopefully by exploring a little deeper how the float level switch works gives you a better understanding should you ever needed to trouble shoot the circuit.  The coolant level warning lamp is extinguished on the dash as you can see below, however you may notice the engine management lamp illuminated just behind the speedo needle hub – this is due to a faulty glow-plug and is next on the list for repair!

Mercedes Sprinter Coolant Level Sensor Problem 7