Fuel Delivery Inspection and Verification
When a vehicle presents itself with a crank no start situation, a few items must first be determined. One of which is, “Does it have fuel?” While this might seem like a simple question of whether or not the gas tank is empty, what we really need to know is whether or not there is adequate fuel delivery. This question is answered with a standard fuel pressure and volume test. If the answer is no, its time to figure out why.
First, let’s talk about that initial fuel pressure and volume test. Modern “Port Fuel Injected” engines have a test port with a Schrader fitting to allow for convenient measurement. On most BMWs this is located on the fuel rail. However, some chassis have the test port built into the fuel filter and pressure regulator assembly. Pressure specifications vary, so be sure to look them up. For a while all BMWs had a standard spec of 3.5 Bar. Over the last decade that has changed. We now see pressures up to 6 Bar, and on some vehicles a variable pressure. As far as volume goes, a good rule of thumb is one liter in 30 seconds.
A fuel pressure gauge can quickly answer the question of whether or not there is adequate fuel supply.
No or low pressure or volume can be caused by an empty tank, bad pump, pump control circuit, or a leak in the system. The primary difference being that while no fuel delivery will certainly cause a no start condition, low or inadequate delivery may allow the engine to run. In this latter scenario the complaint will be a lack of power, misfires and check engine light. In all cases a complete vehicle scan should always be performed. Fuel trim faults indicating a lean mixture in both the “Additive” and “Multiplicative” ranges may indicate a failure to deliver enough fuel volume. Newer vehicles may also have a fuel pump plausibility fault.
Fuel delivery starts at the tank, so let’s look at it first. BMW uses what is referred to as a “saddle tank”. This refers to the shape. In order to achieve optimal weight distribution, the fuel tank is located beneath the rear seat. To allow room for the exhaust and drivetrain, the tank is shaped like a saddle. This results in two storage areas with a common connection on top. There is only one fuel pump, located on the right side, so fuel must be transferred from the left side of tank in order to feed the pump. These design elements are where we can encounter the first possible issues when dealing with no fuel delivery.
Is there gas in the tank? Because of the design, there are two fuel level senders. One on each side. Depending on chassis, these are direct inputs to the instrument cluster, JBE or FEM. The control module evaluates both signals to determine the total fuel level to display to the operator. Obviously a problem with one or both sensors can result in a false display. Another problem can be the failure to transfer fuel from the left side of tank to the right side. In both of these scenarios, the vehicle will effectively run out of gas while the gauge does not show to be empty.
Using an enhanced scan tool, with the ability to communicate with the instrument cluster, you can view live data of both fuel level sensors. You can also go into “Instrument Cluster Test Mode” and call up fuel level data directly out of the vehicle without a scan tool via test number 6. Test 6 may also displays the raw data, or actual resistance, of each sender unit. This is displayed in Ohms. The first thing to note is the comparison of both sides. There should never be more fuel on the left side than the right side. If so, there may be a transfer problem or a failure with a level sensor. Due to the possibility of a sensor problem, physical fuel level must be confirmed visually. Removal of the access covers for both sides of tank is relatively straight forward on most models, with the notable exception of “Z” chassis.
Using test 6 within instrument cluster is an easy way to check status of fuel level sensors. Some vehicles even allow the resistance values to be displayed.
If, after visual inspection, it is determined that the level readings are incorrect, continue with electrical diagnosis of the level sensor circuit. Sensors are potentiometers, so a voltage drop test while moving the float arm is the easiest test. When the replacement of a fuel lever sender is mandated, either as a result of failure or as part of a fuel pump assembly, there is an important step that is often overlooked. Because this is an analog signal being generated by a mechanical component, wear results in signal degradation. This problem is addressed using adaption values stored in the instrument cluster or appropriate processing control module. Now if the sensor was bad, the software has been applying a lot of inaccurate adaptations to that signal. As a result, when the sensor is replaced, the signal will be modified to a point that it too is inaccurate. This can, and often times does, lead to the same symptom the vehicle first presented itself with until adaptations shift back to zero, which can take several full tank refills. To avoid this problem and a dissatisfied customer these adaptation values must be reset. This resetting is accomplished via test number twenty-one in the instrument cluster test mode. Remember, the cluster must first be unlocked. When asked to enter the unlock code, simply add the last five digits of the VIN together. For more information on cluster test mode, visit the BMWTIS website and read the “Driver Information and Displays” section of the body electric training manuals.
Because the fuel level sender is often times incorporated into the pump assembly, remember to clear the adaptation values for sensor circuit in the instrument cluster when replacing.
If visual inspection reveals that the left side level is in fact higher than the right side, there is a problem with the syphon jet pump circuit inside the fuel tank. This system relies on return flow and pressure after the fuel pressure regulator. A restriction in the system, like a clogged filter could cause low enough flow. A bad fuel pump can also result in not enough flow. Another possibility is an internal leak. A hose in this syphon jet circuit can break or come loose at a service connection fitting. While the access covers are still off, activate the in tank pump. In the case of a broken or disconnected line, there will be fuel spraying out from this leak. A disconnection due to improper installation may be fixable. However, a broken line will likely require the replacement of the entire tank. The plastic corrugated lines used inside of BMW fuel tanks are starting to fail. This is probably a result of ethanol content in gasoline.
When this hose was crimped onto fitting, it was not completely inserted. The result was an internal leak.
Now, if initial testing did not reveal any problems with fuel levels, the fuel pump circuit must be examined. On early vehicles the pump control circuit is pretty straight forward, consisting of little more than a fuse protected relay. Now all BMWs utilize a control module to operate the in tank pump. At this point we will be doing some electrical circuit diagnosis. Before starting, be sure the vehicle battery and charging system are in good order. If the vehicle doesn’t start, obviously the charging system can’t be tested. However, the battery is likely low on charge due to the several attempts to get it started. Electrical system test results are only accurate with appropriate system voltage. It is recommended to have a vehicle power supply connected to maintain consistency throughout the diagnostic process.
To test the fuel pump circuit, you will first need to activate it. Depending on the model and year, this will be done one of two ways. On earlier BMWs the pump circuit is controlled using a conventional relay. This relay may be located in many different locations from the engine compartment to the trunk. Be sure to look up the component location using BMWTIS or any other service information source like All Data or a Bentley book. Once the relay is located simply jump it via a purpose built tool or a small fused wire with spade connectors on each end. You will need to connect terminal 30 to terminal 87 at the relay socket.
Starting in the early 2000’s BMW started varying the fuel pump speed to match load conditions. This first appeared on the E39 and E46 Motorsport vehicles and was later introduced on all chassis. The reason for this was to only deliver the necessary volume to meet the current load demands and a little extra to ensure the return system could still transfer fuel via the syphon pump system in tank. This concept saves energy, reduces wear on fuel pump, reduces the overall temperature of the electrical motor and therefore prevents unnecessary fuel vaporization in the tank. All in all, a great feature. In order to do this, the conventional relay is replaced by an ECU. On those early “M” cars, this ECU was little more than a smart relay. The engine control module, or DME, out puts a pulse width modulated (PWM) square wave pattern to indicate the desired delivery. On these models you can jump the circuit the same way as you would on earlier relay controlled circuits.
Now all BMWs have a fuel pump control module that is on a bus network. These ECUs are called EKP modules. The first chassis to use this was the E65/66. Although on this vehicle, it is a unique setup. Instead of a dedicated ECU for fuel pump regulation, the electronics were incorporated into a preexisting control unit. The module chosen for this task was one of the safety system satellites. Specifically, the right side B pillar satellite, or SBSR. Now, the SBSR is on the Byteflight network. This means that the command path for fuel pump control goes from the DME over PT-CAN to the central gateway module, or ZGM. From there the activation signal is sent over Byteflight to the SIM and passed on to the SBSR. Suddenly the importance of a complete vehicle scan is becoming clear. On all other BMWs with a dedicated EKP module, the module is on the PT-CAN. So, the signal path is much less complicated. As a side note, the EKP module contains one of the two terminating resistors for PT-CAN.
On the E65/E66 chassis, an airbag system control module, the SBSR, is used for EKP regulation. Because the SBSR incorporates accelerometers used for crash detection, do not manipulate it while the key is on. This SBSR is a common point for fuel pump circuit failure.
Because the EKP modules are on a vehicle network, they can be accessed via a scan tool. Using your scan tool live data like pump speed and current can be viewed, fault codes can be stored, and the pump can be commanded to run. The speed is controlled via fixed frequency variable duty cycle, also known as PWM. To ensure accurate circuit analysis, you will have to test using a lab scope. If using a DVOM, the low sample rate will display an average. For example, 12 volts at a 50% duty cycle will be displayed as 6 volts. If you were unaware of this, you would find yourself chasing a power supply problem that doesn’t exist. To avoid this issue, either use a high sample rate digital oscilloscope, or when commanding the pump to run with scanner, select the option of running pump at 100%.
Now that the circuit is complete, or commanded on via scan tool, perform a voltage drop test across the fuel pump connector under the rear seat. You should be measuring within half a volt of system voltage (some will be lost through various connections). If measurement is within spec, and you can not hear the pump running, clearly the problem is in the tank. Remove the electrical connecter and visually confirm the contacts are not burnt before proceeding with a pump replacement.
A DVOM and a solid understanding of Voltage Drop theory is all you need to fully analyze the fuel pump circuit.
However, if the measured voltage is more than half a volt lower than system voltage, there is a problem with the circuit. First determine if the problem is on the power or ground side. This is done by doing a voltage drop test of the complete circuits on either side of pump to the battery. So, connect the ground lead of DVOM to the negative post of battery and the positive lead to the ground pin on the pump connector. This is a complete ground circuit voltage drop test. Anything over half a volt is a problem. To check the supply side of circuit, place the positive lead on the positive post of battery and the negative lead to the power pin of pump connector. This is a voltage drop test of the entire positive circuit. This test can determine if there are any bad connections or open circuits. Don’t forget, a blown fuse is an open circuit.
Once you have determined which side of the pump control circuit the problem is on, it is time to further isolate the issue. Using a wiring diagram, locate all the connections in the circuit. Continue to perform the described drop test at additional locations to divide the circuit. Potential issues can be anything from a burnt connector, loose ground, worn relay contacts, or even cold solder joint failures on circuit board of EKP control module. I will note that the SBSR is a common failure item with regards to solder joint failures.
Everything discussed here can also be applied to the low pressure side of fuel supply system in vehicles equipped with direct injection engines. The primary difference is that fuel pressure is controlled by EKP regulation using feedback from a fuel pressure sensor located in the fuel line. This is opposed to a mechanical fuel pressure regulator. These systems do not have a pressure test port. Measurements are done via live data stream in scan tool, or by using an adapter to “T” into the fuel line.
In closing, I will leave you with some service notes. When replacing a faulty fuel pump, you should always replace the fuel filter too. If the repair requires replacement of the EKP module, it will likely need coding or programming. If the SBSR is being replaced, after coding and programming, don’t forget to restart the system time for the passive safety system. And lastly, always take great care when working with fuel.
Always replace the fuel filter when replacing a faulty pump. Even this in tank filter is heavily contaminated after less than 120,000 miles.]
Thanks for sharing this. In regards to your 1 Liter per 30 sec statement, this is a dynamic test run on a return system at system pressure? I’ve tried to tell guys that they need to do a dynamic flow test, meaning measuring flow under system pressure. What I’ve seen on healthy higher HP vehicles is .5 gal min @4 Bar. If my math is correct, that’s 0.95 liters in 30 seconds. So your spec is good. What I’ve seen is some test by dumping the hose into a measureable bucket with no load. Since a lot of vehicles now are non return systems, how can you test these under load? We have a fuel flow meter (CODA) but it doesn’t have a bleed off for measuring flow. Do you recommend a tool for this purpose that you like to use?
Great point! On a non return system that still uses a mechanical fuel pressure regulator located in the tank, I use fuel pressure gauge with a bleed off line while the engine is running. However, in the case of vehicles using an ECU to control the speed of the pump things get different. Now flow is only going to be whats required for current load. If you bleed any off to measure, this will result in a drop of pressure. This drop will be immediately recognized by Engine Control Module resounding by increasing the fuel pump speed. As a result, a fault code will be set. Now, you could begin by commanding the pump to run at full speed. In this case a pressure relief valve in the tank will function like a mechanical pressure regulator and you can run test as you would on an earlier system. All while the engine is running of course.
That being said, I have found that the logic used to deliver the appropriate amount of fuel based on load in these newer systems is so precise that a failure resulting in inadequate flow will be caught by ECU as a result of pump being commanded to run faster than nominal. In fact, this logic has proven to be so accurate, BMW has now removed the low side fuel pressure sensor and relies exclusively on a software algorithm! If a failure occurs, it will be caught on the high side of the system now.
I've never really thought about the effect of ethanol on the plastic lines in the tank. It's crazy that the lines aren't replaceable. I anticipate this becoming an issue in the future. Is there anybody you know that makes replacement lines?