Brushless induction-less control is widely used in practical applications, and people have been studying it for a long time. There are two main control difficulties: first, starting the motor; second, rotor position detection. For the high-voltage induction-free scheme, besides the difficulties in software, the hardware design can not be ignored. For example, if the hardware design is slightly improper, it will cause great interference to the whole control board, thus increasing the difficulty of the success of the whole scheme.
Below we mainly discuss the low-voltage induction-free scheme, for the low-voltage induction-free scheme, the hardware design on the market are very similar, the way to detect the rotor position is almost all using back-EMF detection method.
1. Why is it so difficult to start a sensorless motor?
For brushless motors, the operation of the motor is controlled by electronic switches commutation, then want the normal and efficient operation of the motor, we must know the rotor position, then the normal commutation, the problem is, the motor has no sensors, also did not rotate, so the rotor position is unknown, so no inductive Start to start automatically, let the motor rotate at a certain speed, in the process of motor automation, we know the rotor position by detecting the back-EMF, so as to get the correct phase commutation.
The self-starting of the motor is easy to say and difficult to do. In the process of debugging many sensorless schemes, I summarize the following experience for reference:
(1) First of all, rotation, rotation must let the motor run smoothly, can not shake, but also can not cause large current. This is a very critical step in the success of startup. How to achieve this effect, you must adjust the PWM duty cycle and commutation time in the process of debugging.
(2) The number of starting steps can not be too small, nor too many, generally more than ten steps is enough, and so on motor operation after ten steps began to detect the back-EMF, when the detection of the correct back-EMF after this time the normal operation of the motor.
2, how to detect back EMF.
There are two methods to detect the back-EMF. The first is to compare the back-EMF signals sampled by the AD in the microcontroller, and the second is to compare them directly by the comparator. The two methods are the same, but according to personal experience, the comparator scheme is more reliable, better performance, especially when the motor speed requirements are very high, AD sampling method is almost infeasible.
Although comparator schemes have more advantages, why are AD sampling methods common in the market? This is mainly due to the cost of the product. To do so with the comparator scheme, or not to add a comparator IC outside, not only increase the cost, but also increase the space of PCB cloth board, the second is to find a microcontroller with AD inside, and this microcontroller is relatively expensive.
The following is a circuit reference diagram for detecting the back EMF.
20130807151543937.jpg
This is a general circuit diagram of AD detection scheme and comparator comparison scheme. If a single chip microcomputer has a comparator, then it is connected to the port of the comparator. If not, it must be connected to the AD port.
Taking the comparator scheme as an example, when the motor spins, the comparator interrupts (the comparison object is the voltage value of the neutral point voltage and the suspension phase), commutates the phase immediately when the comparison interrupt arrives, and sets the comparator's comparison object after commutation, that is, the voltage value of the neutral point voltage and the current suspension phase, to wait for the next comparison interrupt. The coming.
The above control mode is a control mode without delay of 30 degrees. In general control system, this control mode is feasible, especially for the scheme of large torque without inductance and large current, the control mode without delay will be more stable and the carrying capacity will be stronger. Of course, there is also a drawback to doing this, that is, efficiency is not as high as 30 degrees of delay. The specific way to do it depends on the actual product.