For about the past two years, every time I fill up my R1150RT, the fuel gauge only reaches 8 bars out of 10. I've read in the past that this is due to the use of a common ground for the fuel pump, low-tank warning and fuel level circuits. Today I cleaned the connector and the gauge now reads 10 bars. The common ground wire is indeed the root of the problem but because I wanted to know for sure that the contact of the ground wire was the problem, and that after cleaning the connection was better, I made several before and after measurements that shed light on what's going on.
The first thing I tried was turning the bike on and not starting the engine. The RID came alive, the fuel pump ran for 2 seconds and shut off, and the RID showed 8 bars. I then waited about 15 seconds and the RID showed 9 bars and then waited another 15 seconds and the RID displayed 10 bars. The waiting time is because the RID averages readings before changing so that the display doesn't jump around. The behavior of the bars going from 8 to 10 is consistent with the Common Ground theory. Once the pump stops, its effect on the gauge is eliminated.Before Measurements:
Fuel Pump Current, measured at the fuel pump relay-- 5.25 ampsFuel Level Sensor Voltage with pump running-- 285 millivolts
Fuel Level Sensor Voltage with pump off-- 85 millivolts
Resistance of Fuel Level Sensor, full tank-- 4 ohmsBefore CalculationsDrop Across Ground Lead at Fuel Tank Connector: 200 millivolts (285-85 millivolts)
Resistance of Connector Pins in Tank Connector: 40 milliohms (200mV/5.25A)
Current in Fuel Sensor Circuit: 20 milliamps (85mV/4ohms)
Full Scale Voltage for Fuel Sensor Circuit 1.2V (20mA*60ohms, I found the 60 ohms in a white paper on the RID)
Interpreting the above data shows that the voltage at the RID varies from about 0V with a full tank to about 1.2V with an empty tank. The 0.2V drop across the ground pins in the connector from the Fuel Pump current adds 0.2V to voltage the RID is measuring, leading to a measurement of about 80% full, just what I was seeing.
Although 40 milliohms resistance in the ground lead is a very low, it is too high in the application where the BMW engineers ran two sensors and the fuel pump through one ground wire. And who can blame them for saving about a buck with with this engineering shortcut?
The solution was to clean the ground connector pins and get the resistance of those pins below 10 milliohms. I didn't have contact cleaner so used isopropyl alcohol and a clean cotton rag. The pins were mated and unmated several times to get further scrubbing action and then the connectors were plugged together. After Measurements:Fuel Level Sensor Voltage with pump running-- 120 millivolts
Fuel Level Sensor Voltage with pump off-- 80 mVAfter CalculationsDrop Across Ground Lead at Fuel Tank Connector: 40 millivolts
Resistance of Connector Pins in Tank Connector: 7 milliohms (40mV/5.25A)
Now that the contact resistance is under 10 milliohms and the voltage drop only 40 mV, the circuit is stable and the excess ground drop insignificant, and the RID fuel display reads a full 10 bars. The moral of the story is that the Fuel Tank electrical connector has to be very clean for the gauge to read correctly give BMW's cost-cutting.