This invention relates to the measurement of the signal offset of a Rotor Position Transducer (RPT) in an electrical machine.

An RPT is used to measure the angular rotation of the rotor of an electrical machine. For example, Figure 1 is a schematic representation of RPT signal generation on an electrical machine 2. A magnetic ring 6 mounted on the rotor shaft 4 excites a Hall-effect magnet sensor 8 mounted on the electric circuit board (not shown) to generate rotor position information for motor control software. The North to South / South to North switching points are shown at 10, and high voltage and low voltage are indicated at H and L respectively.

Due to assembly tolerances in the magnet ring 6, rotor shaft 4, and the magnet sensor 8, the RPT signal may not be perfectly aligned to the appropriate reference point of rotor-stator geometry, leading to errors in the measurement of the RPT signal. To achieve the desired response and efficiency, such errors should be compensated for when applying control laws to energise the stator phases.

Previously, a method of assessing the signal offset of an RPT from a motor involved the following steps:

a) dissemble the electrical machine and mark a reference point on the circumference of the rotor shaft or impeller; b) remove all control electronics, use a DC power supply directly to energise stator phase A;

c) when the rotor is locked to phase-A, mark the position on the machine housing using the reference point, (thereby identifying the "rotor locked-to-phase-A position");

d) rotate and repeat step c) for each rotor pole, until the locked-to-phase- A position has been identified for all four rotor poles;

e) re-assemble the electronics;

f) using an oscilloscope or multi-meter to monitor the RPT signal;

g) manually rotate the rotor to identify the position where the RPT signal falling edge occurs, mark the position on the machine housing using the reference point, (thereby identifying the "RPT-falling-edge position");

h) repeat step g) for each rotor pole, until the RPT-falling-edge position has been identified for all four rotor poles;

i) Measure the distance between the locked-to-phase-A positions and the RPT-falling-edge positions;

j) Using the known machine circumference and diameter, and the distances identified in step i), establish the angle of RPT signal offset for each of the four rotor poles;

An aim of the present invention is to provide apparatus for and a method of measuring RPT signal offset in an electrical machine, which enables a simpler, faster and more accurate measurement of RPT signal offset than currently employed methods. Accordingly the present invention provides a method for measuring the RPT signal offset of an electrical machine as claimed in claim 1 of the appended claims.

The encoder means may comprise a rotary encoder or a software encoder algorithm.

An embodiment of the present invention will now be described by way of example and with reference to the accompanying drawings in which:

Figure 2 is a schematic representation of a rotor and stator geometry of an electrical machine;

Figure 3 is a Pulse Width Modulation (PWM) profile of the A, B and C phases of the electrical machine of Figure 2;

Figure 4 is a schematic representation of a production end-of-line test rig setup for undertaking a method in accordance with the present invention comprising a rotary encoder;

Figure 5 is a table of the relationship between "PWM Walk" and rotor position in accordance with a second embodiment of the present invention; and

Figure 6 is a schematic representation of a production end-of-line test rig setup for undertaking a method in accordance with the present invention comprising a software encoder.

RPT Signal Offset Learning with Rotary Encoder

A first embodiment of the invention comprises a method of learning the RPT signal offset from an electrical machine 12 using a controller, and encoder means comprising a rotary encoder mounted on a rotor shaft of the electrical machine.

Referring to Figure 2, the electrical machine comprises six stator poles, A1 , B1 , C1 , A2, B2, and C2, and four rotor poles, α1 , β1 , a2, and β2. The coils on the stator poles are connected into 3 phases; phase-A comprising A1 and A2, phase-B comprising B1 and B2, and phase-C comprising C1 and C2; i.e. the coils on A1 and A2 are always energised simultaneously, (or B1 and B2, or C1 and C2).

The control system drives the motor in regulated current mode using a controlled voltage, wherein the voltage is controlled by PWM. A voltage walk, in this case comprising a "PWM walk" is used, wherein the PWM controlled voltage being applied to a particular phase being stepped up, or stepped down, by a predetermined value, and consequently the current in the coils of the electrical machine is also controlled.

At any time, a PWM controlled voltage is applied to one phase and also to a second, adjacent phase. Initially, the PWM controlled voltage applied to the first phase is set at a predetermined maximum duty cycle value, and the PWM controlled voltage applied to the second phase is set at a predetermined minimum duty cycle value (which could be zero).

The minimum duty cycle applied to the second phase is subsequently increased in a "voltage walk", in this case a "PWM walk", i.e. the PWM controlled voltage which is being applied to the second phase is gradually increased in pre-defined steps, up to the maximum duty cycle PWM controlled voltage value. When the PWM controlled voltage being applied to the second phase reaches the maximum duty cycle value, the PWM controlled voltage being applied to the first phase is gradually decreased in another "PWM walk", i.e. the PWM controlled voltage which is being applied to the first phase is gradually decreased in pre-defined steps, down to the minimum duty cycle PWM value, and subsequently (in the case where the minimum duty cycle PWM value is not zero), to zero.

For example, initially a PWM controlled voltage is applied to phase-A at a maximum duty cycle value, and a PWM controlled voltage is applied to phase-C at a minimum duty cycle value. The PWM controlled voltage applied to phase-C is "walked up", i.e. gradually increased, up to the maximum duty cycle value. When the PWM controlled voltage being applied to phase-C reaches the maximum duty cycle value, the "voltage walk up", in this case a "PWM walk up" on phase-C stops, and the duty cycle on phase-A starts a "voltage walk down", in this case a "PWM walk down", i.e. the PWM controlled voltage being applied to phase-A is gradually decreased in predefined steps, until the minimum duty cycle PWM controlled value is reached, at which point the "PWM walk down" on phase-A stops.

As a result of the "PWM walk up" of phase-C and the subsequent "PWM walk down" of phase-A, the rotor is thereby moved from lock-to-phase- A to lock-to-phase-C in a controlled manner. The steps moving from phase- A to phase-C, especially at the proximity around the locked phase, can be controlled to arbitrarily small values by applying multiple "PMW walk" steps.

In detail, and with reference to Figure 3, the following steps are followed to achieve the PWM Walk:

1 ) Energise phase-A (A1 and A2) with a PWM controlled voltage, where the duty cycle is PWM_HI. Stator poles a1 and o2 are attracted to A1 and A2, respectively from their original position and stabilised at what is so-called "locked position", i.e. where the rotor is locked to the energised phase of the stator;

2) Phase A is still energised with PWM controlled voltage, duty cycle is PWM_HI. Phase-C is energised with PWM controlled voltage, duty cycle starts from PWM_LO, and is then stepped up until it reached PWM_HI; 3) Phase-C is energised with duty cycle PWM_HI voltage, phase-A is energised with a gradually reducing PWM controlled voltage from PWM_HI to PWM_LO;

4) The above procedure is repeated for phase-C and phase-B;

5) The above procedure is repeated for phase-B and phase-A;

6) The process is repeated from step 1 ) four times to complete a 360° rotation.

With reference to Figure 3, from step 1 to step 5, the rotor will be locked from phase-A to phase-C, then to phase-B, and finally again to phase-A. There will be a 30° stepping clock-wise when it is locked from one stator pole to the next. After 90°, it can be seen that an electrical cycle completes, β2 is locked to stator pole A1 .

The rotary encoder is used in conjunction with the PWM walk above in order to ascertain the RPT signal offset from the electrical machine. The complete method of learning the RPT signal offset therefore involves the following steps:

1 ) Locking the rotor to phase-A, initialising the encoder driver, Encoder- Angle^ ⁰ ;

2) Stepping the motor using the "PWM walk" as described above;

3) Over one revolution, recording encoder readings when RPT signal H-L transition occurs. Four angular readings will be obtained: Θ1 , Θ2, Θ3, Θ4. 4) Compare Θ1 , Θ2, Θ3, Θ4 with the reference geometry points to calculate the RPT signal offset for each rotor pole.

An example of a production end-of-line test rig for undertaking the above method using a rotary encoder is as illustrated schematically in Figure 4.

RPT Signal Offset Learning with Software Encoder Algorithm

A second embodiment of the invention comprises a method of learning the RPT signal offset from an electrical machine, using a controller, and encoder means comprising a software encoder algorithm.

When two phases are energised with different controlled voltages (in this case PWM controlled voltages), the rotor will be stabilised at a position in between the lock-to-phase positions, for example, as shown in Figure 2. The position of the rotor is a function of the PWM duty cycle on phase-A and phase-C. At the proximity around the position of lock-to-phase-A, an empirical set of data which describes the relationship between "voltage walk", in this case "PWM Walk", and the step angles can be obtained. The PWM duty cycle count in the table of Figure 5 is based on a particular type of electrical machine. A positive rotor position means the rotor pole is after lock-to-phase-A position, a negative rotor position means the rotor pole is before lock-to-phase-A position, and lock-to-phase-A position is 0[deg], as shown in Figure 2. It can be seen that when the phase-B voltage is "walking down", in this case in a "PWM walk down", the rotor pole is moving towards lock-to-phase-A position. After it is locked to phase-A, with the phase-C voltage "walking up", in this case in a "PWM walk-up", the rotor pole is leaving phase-A, as shown in Fig. 3.

The above relationship may vary from machine to machine. The variations are mainly correlated to coil resistance which deviates current in phase coils and hence torque, which consequently varies the relationship between rotor position and the PWM duty cycles.

Test data have shown that on the same batch of production, such variation is well within the tolerance requirement on RPT signal offset accuracy.

The RPT signal offset learning procedure using the software encoder, as illustrated schematically in Figure 6, comprises the following steps:

1 ) Locking the rotor to Phase-A;

2) Stepping the motor using the "PWM walk" as described above;

3) Over one revolution, record phase-A, phase-B and phase-C PWM duty cycle counts when RPT signal H-L transition happens. Four sets of readings corresponding to four RPT signal transitions will be obtained: C1 , C2, C3, C4. 4) Referring to the table of Figure 5, C1 , C2, C3 and C4 values will provide corresponding RPT signal offset values for each rotor pole.

An example of a production end-of-line test rig for undertaking the above method using a software encoder is as illustrated schematically in Figure 6.

In alternative embodiments, other means of motor control could be used, for example direct current limiting. Although the method discussed above is based on a 3-phase 4-6 rotor-stator machine, the application of the technology is not limited to this type of machine.

In the above embodiments, pulse width modulation (PWM) is used as a means for controlling voltage. In other embodiments, any other suitable means for controlling voltage in a multiphase machine, in which the current flowing in the coils can be modulated, may be used. Other suitable means include current chopping and torque control.