Recently, there have been some inquiries about the operation and diagnosis of electronically commutated motor (EC) fuel pumps. I think this is a good topic, so I share it with you here.
We must learn and master the working principle of EC. In the future, EC can use three-phase motors to power hybrid vehicles or pure electric vehicles.
To make a long story short, here is a 3.0-liter V6 gasoline car Audi SQ5, the engine code is CWGD, there is a problem of insufficient fuel pressure in the fuel pump (G6).
Please note that the fuel pump G6 is integrated in the fuel delivery device assembly, combined into a fuel supply unit (GX1), and then installed in the fuel tank. The fuel pump is controlled by the external fuel pump control unit (J538), in which the conversion from direct current to three-phase alternating current is realized .
The thing to remember is that under any load situation, we require the fuel pump to deliver enough fuel from the fuel tank to the engine .
In Figure 1, we connected the Pico oscilloscope 4823 to capture the three-phase voltage and current signals on the failed fuel pump.
Figure 1 Voltage and current of each phase
So, why use such a complicated control system on the fuel pump?
Performance, limited control, reliability and durability are all answers to this question. Except for supporting bearings, this type of motor hardly wears. Since there are no brushes, there is no contact between the moving motor parts. This eliminates harmful friction and arcing. (This kind of motor is called BLDC, brushless DC motor)
Brush motors usually suffer from wear and arcing (sparks), as shown in Figure 2 and Figure 3.
Figure 2 Wear of brushed motor
Figure 3 Sparks appear in the brush motor
In addition to some of the EC motor principles I introduced above, you can also watch the following explanation video:
Brushless DC motor working principle
To make the rotor in the motor rotate, we need to generate a rotating magnetic field around the stator, and the rotor will rotate with this magnetic field. If you connect the rotor to the pump, you can convert the rotational motion into physical pressure.
This working principle is applicable to all applications, whether it is connecting the EC motor to the gearbox, wheel or output shaft.
In Figure 4, we zoomed in on the waveform to analyze the voltage and current changes during pump/motor operation. Please pay attention to how the voltages of channels A, B and C are cut off at 0 V, but the current is reversed at this time!
Figure 4 Correspondence between voltage and current
The DC voltage signal we captured does not explain all the problems, because we are measuring the voltage to ground. In fact, the voltage signal in the figure is reversed, which is to reverse the winding currents of each phase captured by channels D, E and F.
If you want to capture negative voltage, you need to use a differential probe to measure the phase you want to know. Especially when testing high-voltage systems, you will need to use differential probes, and make sure you are properly trained and equipped with relevant protective equipment.
In summary, by measuring the "current", it is possible to reveal the working conditions of the entire motor in a non-intrusive manner and provide some data as evidence. The measured current can show:
- Other operating characteristics of the motor
- Whether the magnetic field/coil winding is intact
- Motor/pump action
- Whether the control circuit is normal
- Motor frequency/speed
- Motor load condition
Here I mention that magnetic fields do have an effect on voltage and current . The best example is that there is a turning point when measuring the injector current. In Figure 5, we captured the initial movement of the needle valve (the injector is opened), then the magnetic field around the coil winding changes ( thus causing the current signal to turn ), and when the needle valve returns to the seat ( (Injector closed) again induced voltage (back electromotive force).
Figure 5 Injector voltage and current
So what does this have to do with our BLDC motor?
The waveform in Figure 6 shows that there is a period of zero current between the positive peak current and the negative peak current, and the rotor pole and the stator pole are "N pole" and "S pole" respectively.
Figure 6 Determine the rotor position
Near each start and end point of the A channel signal in Figure 6, the voltage signal in the gray rectangular box is quite special (each voltage signal has the same characteristics). The phase voltage signal is the induced voltage generated in the winding during the current from power-on to power-off. The fuel pump controller determines the position of the rotor based on this voltage, without the need to add a resolver or Hall-effect position sensor. The position of the rotor can be determined.
Knowing the position of the rotor is essential for determining the energization sequence of the stator windings and generating a rotating magnetic field (EC).
Please note that due to the above reasons, we cannot see the reverse induced voltage at the end of each voltage phase (please refer to the introduction in the paragraph below Figure 4). In other words, we can see a gap at the end, but in this gap, the negative voltage appears for a short time and disappears instantly .
One more point, before we do mathematical operations, we must pay attention to the relationship between the power supply frequency and the rotor/pump rotation frequency.
Figure 7 The relationship between power supply frequency and rotor rotation frequency
The relationship between the power frequency and the rotation frequency of the motor (rotor/pump) is related to the number of magnetic pole pairs, the number of magnetic pole pairs = the number of rotor poles/2.
Suppose our pump contains a 4-pole rotor (1 pair of N pole and 1 pair of S pole), so the number of rotor pole pairs is 4/2=2. In other words, the power frequency of a 4-pole rotor divided by 2 is the rotation frequency of the rotor. In other words, for a 4-pole rotor, it takes 2 power cycles to make the rotor make one revolution.
If you don't know the number of poles of the rotor, you can use an optical sensor to capture the motor rotation frequency signal (when conditions permit), and also use an oscilloscope to capture the current of one of the three phases.
Then in one rotation cycle of the motor, calculate the number of cycles of a certain phase current signal, and then multiply it by 2 to get the number of rotor poles. Figure 8 calculates the number of rotor poles of the three-phase motor is 30 through the above method.
Figure 8 Calculate the number of rotor poles
Please note that due to the reduction gear and other reasons, the rotor may not be directly connected to the optical sensor, resulting in inaccurate capture of the rotor rotation frequency, which will definitely cause an error in the calculation of the number of rotor poles.
Now that we return to the faulty fuel pump, what information can we get from the raw data captured in Figure 1?
Use the math channel LowPass((abs(D)+abs(E)+abs(F))*0.333,50) to determine the average current consumed by the fuel pump (including currents of all three phases).
LowPass can smooth the AC ripple, which is low-pass filtering;
(Abs(D)+ abs(E)+ abs(F))* 0.333 is the average current value of three-phase rectification;
50 refers to the frequency of low-pass filtering (50Hz).
In order to calculate the rotor/pump speed, use the math channel 60*2*freq(D)/4 (60*2*power frequency/rotor pole number).
60 is to convert Hz to RPM; because there are positive and negative alternating currents, it needs to be multiplied by 2; divided by 4 because our rotor has 4 poles.
Note: The rotor/pump speed depends on the power frequency and the number of rotor poles.
- Increasing the power frequency will increase the speed, but will reduce the torque.
- Increasing the number of rotor poles will reduce the speed, but will increase the torque.
Figure 9 Faulty fuel pump
In Figure 9 above, it can be seen that the fuel pump runs at a fixed speed of 10,000 rpm and consumes an average current of 7.6A.
Now, compare it with the waveform captured by the new normal fuel pump in Figure 10.
Figure 10 Normal fuel pump
There must be a difference. Check the speed and current consumption of the fuel pump in Figure 10. The speed is about 3200rpm and the current is 5.4A at no load . It should also be noted that the frequency of the current between the time scales (channel D) is reduced to 109.1 Hz, which causes the pump speed to decrease. When the fuel pump works under maximum load conditions, the speed is about 7787rpm and the current consumption is 10.4A.
To sum up, the current of the new fuel pump is maintained at 5.4 A at 3200rpm. This is to maintain sufficient fuel pressure under no-load conditions (the current is smaller and the speed is lower to obtain sufficient fuel pressure). Due to insufficient fuel pressure, the old fuel pump has a current of 7.6A and runs at 10000rpm.
To be sure, measuring current can reveal the working condition of the fuel pump, which is very obvious in the waveform of the fuel pump under load in Figure 10. So, what happened to the old fuel pump?
Remember that the direction of pressure is opposite to the direction of fuel flow. The fuel pump signal captured in Figure 9 shows that the pump delivers fuel at 10,000 rpm, but where did the fuel go? Let's look at the diaphragm of the fuel pressure regulator integrated in the fuel supply unit (GX1).
Figure 11 Diaphragm rupture
The diaphragm in the fuel pressure regulator ruptured, causing most of the fuel to flow back to the fuel tank instead of being transported along the fuel pipe to the engine compartment. This is the root cause of the failure of the old fuel pump.