AN2372

Experimental implementation using ST7FMC

7/13

2.1

Speed and absolute position estimation

The control scheme requires instantaneous position of back emf A to generate sinusoidal

PWM pulses for motor control. However, with only a Hall sensor feedback, that generates

only two edges within a 360° electrical cycle (corresponding to say 0° and 180°),

instantaneous position information is not available. To obtain the intermediate values, an

estimation of rotor speed is required, so that an integration of rotor speed at the pwm update

period gives the rotor position at these instants.

The basic estimation scheme is as follows:

θ is the estimated angle

θ

ω is the rotor speed

At the instant of a Hall edge,

θ=θ

H

For intermediate positions, for use in pwm update routines,

θ ⇐ ω.T

Figure 5.

Hall edge detection and period measurement using ST7FMC

However, there are two different methods to estimate the rotor speed:

1.

division method

2.

PLL method

There are some commonalities between these methods. Figure 5 shows a simplified

diagram of Hall feedback connection to ST7FMC. IS1 and IS0 bits are used to connect one

of the Hall inputs to the comparator that will next change its output state after the current

one. When the comparator detects a state change, the contents of a free running timer

(MTIMH:MTIML) are captured into (MZREG:MZPRV) and the timer is reset to zero, and an

interrupt (C) is generated. Inside the C ISR, the Hall sensor states are read,

θ

and

θ=θ

In the division method, rotor speed is calculated on the basis of time difference between two

consecutive edges. With 1 or 3 Hall sensors feedback, consecutive edges correspond to

180° or 60° respectively. The existing motor has 3 sensors and hence captured period

corresponds to 60 electrical degrees.

MTIM : MTIML

MZREG : MZPRV

+

-

MUX

IS0

IS1

Hall A

Hall B

Hall C