When a normal diagnostic procedure becomes a massive headache

Dec. 1, 2018
Conflicting information can cause what would be a normal diagnostic procedure to become a massive headache, costing time, money and in some cases, sanity.

The vehicle that caused this is a 2011 BMW X3 2.8i (F25 Chassis) with a 3.0L (N52T) turbocharged engine. It was towed into the shop for stalling while driving and would not restart. I did not physically look at this vehicle on its original diagnosis but did offer the tech assigned to it some direction to find the problem. However, I was asked to get more involved and assist with the diagnosis when a problem still existed after the original repair.

2011 BMW X3 (F25) with an inline 6 cylinder 3.0L (N52) engine

The vehicle was initially dropped off with the concern that while driving, the engine sputtered and stalled. The engine cranked over fine but would not restart. Once at the shop the technician put the vehicle on a set of GoJaks Car Dollies to get it inside. Normally a couple of people would have just pushed the vehicle inside so they don’t have to work out in the 115 degree Las Vegas summer heat, but this vehicle, like many other modern BMWs, is not allowed to be shifted into neutral if the engine is not running making it very difficult to move.

No fuel or no fuel pressure?

Once inside the shop the technician did his basic checks and decided that fuel was the cause of the crank no start. He stated that he had no fuel pressure with the engine cranking. I asked, “Is there fuel in the tank?” to which he informed me that there had to be since the fuel gauge was reading at half a tank. Based on my experience, and probably anyone who has dropped a gas tank to replace a fuel pump only to find the tank completely empty, I recommended that he add a couple of gallons of gasoline before condemning the fuel pump just to be on the safe side. He did and the engine started within a couple seconds of cranking. 

Using the test plan in the factory scan tool gave me information on how the fuel level sensor circuit functioned. Most importantly, it contradicted what the dealer had stated and verified the vehicle does in fact have two fuel level sensors.

So the fuel pump was working but the fuel level was not being read correctly. The fuel gauge still read at half a tank, which cannot be correct for only having a couple of gallons of fuel in the tank. He came back to me later after some further diagnostic work and said that he had ordered the fuel pump assembly that came with the fuel level sensor. Well, OK running the pump with no fuel didn’t do it any favors so I agreed with his conclusion. On a side note, I later found out that the fuel level sender is available separately without purchasing an entire fuel pump. The dealer did tell the tech that the fuel level sensor was part of the pump assembly when ordered.

New part, same problem

The new fuel pump assembly was installed and the technician noticed that the fuel gauge still read a little under half a tank. He then added more fuel to the tank and test drove the vehicle. The needle still remained at almost the halfway point. A call to the dealer parts department was made and the parts person stated that this model only has one fuel level sensor on the right side and no left side fuel level sender was listed. He stated he also checked for TSBs and any matches in Identifix but came up empty handed.

The code pointed to the right fuel level sensor being shorted to battery positive, however that sensor was already replaced with an OE unit.

At this point the tech had hit a wall since he replaced the only part responsible to indicate the fuel level in his mind, but the same problem remained. Frustrated and out of ideas he asked to have me assist with the diagnosis.

Noting that the vehicle has a saddle tank, I find it odd that it does not have a sensor on both sides. The rear seat is still removed and it is easy to see that there is an access hole on the passenger side for the fuel pump, but there is no cut out on the body on the left side for an additional fuel pump sender or jet pump. Next I look up the fuel level sensor theory of operation. I do find that (depending on the series) one or two fuel level sensors are installed in the vehicle. So at this point it looks like the dealer parts person very well could be correct. As I read on, I find that voltage is supplied by either the Junction Box Electronics module (JBE) or the Rear Electronics Module (REM) or Body Domain Controller (BDC) or the Hybrid pressure refueling electronic control unit (TFE). Well we can rule out the last one since this is not a hybrid vehicle, but reading that there are three possible modules that supply voltage doesn’t make this diagnosis any easier. 

The test plan instructs me to test the resistance of the right side fuel level sensor in the installed state, which agrees with the nearly empty fuel tank level.

One of these control modules then measures the voltage drop across the potentiometer(s). From this a resistance value is added and it is transmitted to the instrument cluster (KOMBI) in a PT-CAN message.

Where is the information?

My first step is hook up BMW ISTA, which is the factory scan tool for the vehicle. While the system is gathering its preliminary information, I also print out a wiring diagram to get an understanding how the circuit is designed. 

Something interesting occurs during both. First, I have a code in the Junction Box Module (JBE) and the Instrument Cluster (KOMBI) for the Right fuel level sensor: Short circuit to B+. Several times codes may not be stored in the powertrain module but other modules may have codes that give clues to the fault being experienced. It is always a good practice to check for codes in all modules. This is something that is commonly done on vehicles, especially ones with Guided Fault Finding (GFF). Since all modules communicate over multiple Controller Area Networks (CAN), they all need to be included when performing a scan to help see problems in all system circuits. Second, according to the wiring diagram and ISTA’s circuit description, the fuel level sensor signals go directly to the Junction Box Module (JBE). This knocks out the other ones listed in the circuit description I had referenced earlier. Most important, the circuit diagram shows two fuel level sensors, not just a right-side level sensor like the dealer showed on their system. This is later verified by the same information in ISTA. 

The test plan instructs me to remove the fuel level sensor and use an ohmmeter to verify the resistance at full and empty levels while physically moving the float arm.

With that information I proceed with the GFF in ISTA. I am instructed to disconnect the 6-pin connector to the right-side pump/fuel level assembly and measure the resistance of the right-side fuel level sensor. I am looking for a range between 50 and 990 ohms and it passes. Next, I am instructed to remove the right-side pump and sensor assembly to test the range of the fuel level sensor while I physically move the float arm. Note that the lower limit of 50 ohms is actually corresponds to the high fuel level of the tank and the high limit of 990 ohms would indicate the tank is almost empty. The range indicated by the ohmmeter as I was moving the float from empty to full shows that the right-side fuel level sensor is working correctly. When asked by the GFF section of ISTA if the measured resistance values within the listed specifications I click on “Yes.” The response I received was disheartening, “The right fuel level sensor is OK. No fault can be currently found in the component group tested.” 

After performing the test with satisfactory results, ISTA informs me that no fault can be found with the right side fuel level circuit. However, the same code returns shortly after it is cleared.

What's next?

Next, I clear the code and see if any changes occur in the fuel level on the gauge, thinking that if there is a code in the system, it may substitute a default value until the fault is rectified. There is no change on the fuel gauge (still stuck at a little below half full) and the same code for a shorted right-side fuel level sensor signal returns on the next key cycle. Knowing that the GFF in ISTA already did some circuit tests in the background during the beginning of the test plan, I’m fairly confident that the Junction Box Electronics (JBE) module is working correctly and the Instrument Cluster (KOMBI) is simply displaying the information it received from the PT-CAN. 

So now what? Run the test plan again? Replace the JBE? Replace the KOMBI?Since all test procedures are contained within the factory scan tool and there are no other test procedures that I can find, I created a fault with both fuel level sensors by unplugging them and causing the module to code for the left side fuel sensor as well. Just a reminder that the left side fuel sensor is not accessible from under the back seat like the right side, there is no access to it even when the tank is removed from the vehicle. 

After inducing a fault code for both sensors by disconnecting the connector at the fuel tank I am given an option to choose the test plan for the left fuel level sensor.

I now have an option to pick the test plan for either the right or left fuel level sensor. Something interesting that I noticed when given that option was a note in ISTA, “If no fault cause can be determined during the check on one of the two fuel level sensors, you must start the test module again and select the checking of the other fuel level sensor.” 

The left fuel level sensor can only be tested in its installed state since there is no way to remove the sensor for testing like we did on for the right side fuel level sensor.

The test plan of the left side fuel level sensor is similar to the right for the first part, but I cannot physically remove the sensor to move the float through its full travel while observing the ohmmeter. When measuring the resistance of the left fuel level sensor I notice a strange value of 577 ohms. Strange because the fuel level sensor has a range of 50-990 ohms corresponding from full to empty and this fuel tank has less than two gallons of fuel in it. From that observation, I selected “No” when the test plan asked if the measured resistance in the specified range.

This connector is all that is visible of the left side fuel level sensor. Due to the saddle style fuel tank, the float for the left side cannot be seen even while looking into the opening of the right side level sensor/fuel pump assembly.

Based on that response, the next screen in ISTA stated there was a fault in the left-side fuel level sensor. The repair was to replace the fuel tank. Using a mirror, I looked into the opening for the right-side fuel pump/level sensor assembly hoping to see the float on the left side of the tank so that I might be able to move the arm with a long screwdriver to see if the resistance on the sensor changes but only found the transfer tube and wiring disappearing into the darkness. 

Pulling the trigger on the tank

The cost of a new fuel tank (which includes the left side fuel level sensor) is over $1000, not to mention the 5.7 hours of labor it calls for to replace it. I’m sure all of us have felt that doubt when asked, “Are you absolutely sure that will fix it?”

With that large amount of parts and labor, I wanted to be sure that the rest of the circuit and components were working as designed for myself, but how? There was no way to test the sweep of the left side sensor. Even if we filled the fuel tank, I could not see the left side float to tell if it was actually moving. So, I decided to simulate the fuel level sensors with a pair of resistance decade boxes. 

A resistance decade box is a device that will create a specific resistance using a combination of switches. With it I can create any resistance from 1 ohm to over 11 megaohms (11 million ohms) in 1-ohm increments.

After determining the left fuel level sensor is not reading correctly, I am instructed to replace the entire fuel tank assembly which is the only way to replace the defective level sensor.

I replace each fuel level sensor with a decade box attached at the 6-pin connector at the right-side fuel pump assembly. Each box was wired to the specific pins of its substituted fuel level sensor. Setting each box to 50 ohms and then changing each to 990 ohms, I was able to see the corresponding change on the fuel gauge from full all the way to empty. In fact, changing to various resistances, I could place the fuel gauge anywhere I wanted to.

Since the fuel tank and labor required to replace it was very expensive, I installed a resistance decade box in place of each fuel level sensor and verified that the fuel gauge responded correctly to changing the resistances throughout the ranges specified in the test plan.

This proved that the Junction Box Module (JBE), the Instrument Cluster (KOMBI) and all wiring were all OK and I could confidently make the call to have the fuel tank and labor authorized as this would fix the concern.

I do not know why the right-side fuel level sensor having a short to battery + was indicated when the left side sensor was at fault. Looking at the wiring diagram, each sensor’s value was determined by individual circuits to the Junction Box Module (JBE) and they were then added together. The sensor’s circuits appear to be completely isolated from each other, so the left sensor should have been the one that coded and I also don’t understand why it was a short to battery positive fault. Perhaps it is just a software error, but it definitely caused a lot of frustration and confusion to something that if coded correctly to identify the fault component would have made the diagnosis fairly straight forward.

Sponsored Recommendations

Best Body Shop and the 360-Degree-Concept

Spanesi ‘360-Degree-Concept’ Enables Kansas Body Shop to Complete High-Quality Repairs

ADAS Applications: What They Are & What They Do

Learn how ADAS utilizes sensors such as radar, sonar, lidar and cameras to perceive the world around the vehicle, and either provide critical information to the driver or take...

Banking on Bigger Profits with a Heavy-Duty Truck Paint Booth

The addition of a heavy-duty paint booth for oversized trucks & vehicles can open the door to new or expanded service opportunities.

Boosting Your Shop's Bottom Line with an Extended Height Paint Booths

Discover how the investment in an extended-height paint booth is a game-changer for most collision shops with this Free Guide.