Motor Encoder-less Position Detection and Control Technology
Can you imagine a motor that is controlled accurately without an encoder?
Nidec has long been focusing on replacing stepping motors and other existing motors with brushless DC motors, which consume significantly less energy, and are lighter, thinner, shorter and smaller than their counterparts. Here is a question: What component is necessary to adjust a motor exactly to its designated position, or eliminate its rotation fluctuation? The answer to the question is the encoder. An encoder is essential in rotating a motor accurately; however, miniaturizing it is very difficult, and it accounts for approximately 20 to 40% of the motor’s cost. This circumstance has led Nidec to develop a motor “that does not require an encoder to control its rotation accurately.” For those of you wondering how we did it, here is the explanation on our newly developed “Technology for Encoder-less Detection and Control of Motor Rotor Position.”
Improving accuracy by learning the hall sensor’s “dispersion pattern”
Nowadays, almost all brushless DC motors are equipped with hall sensors, a component that is far more inexpensive and can be made more compact than the encoder. Nidec’s past brushless DC motors were equipped with hall sensors and an encoder whose position error of mechanical angle was 0.9°.
In general, a brushless DC motor uses hall sensors to detect the rotational position of a rotor (magnet); however, accurate detection of it is difficult, with a margin of error generated by “dispersions” in magnet magnetization, hall sensor installation, sensitivity and more. Nidec has succeeded in developing an algorithm to correct the “dispersions” to improve the position control precision by using hall sensors only. This new technology can eliminate the need for an encoder itself.
Conventionally, motors were controlled on an error-free, ideal motor model, which would create “gaps” between the model and the real motor's behavior, due to differences among individual motors caused by, for example, the hall sensor’s “dispersion,” though the motors would look as if they were controlled accurately. This is why conventional brushless DC motors would exhibit inconsistent rotations and a decline in precision.
To counter this problem, Nidec built a motor model that assumes such differences in individual motors. This new model’s algorithm that Nidec created takes advantage of those differences, which usually work adversely in motor control, and runs software to learn signals in the brushless DC motor’s hall sensors, to detect the motor axis’s angular position.
In detail, the software statistically processes and corrects the hall sensor signal’s “dispersion.” Any correction algorithm’s principle is based on the hypothesis that, however widely a detected hall sensor signal may vary, the very same signal waveform can be detected every time the motor rotates 360°. In its detection process, the software first extracts a hall sensor signal for a rotation or a period. Repeating this rotation several times enables the software to understand (learn) the extent of the “dispersion” accurately. As shown in the illustrations, this process connects the “line segments” created by connecting the intersections of the waveforms that multiple hall sensor signals draw, to draw a single polygonal line, and align the signal waveforms’ heights and centerlines (normalization). Thus, correcting “line segments” enables one to make the aforementioned polygonal line almost straight. Processing based on this correction algorithm “line segment connecting method (CLH)” enabled us to reduce differences among individual motors, and realize a high-precision control (positional error: mechanical angle of ±0.25°, and environmental temperature: 0 to 60℃). Not only did we confirm this precision in a simulation, but also an actual product-based evaluation verified a significant precision improvement in positioning of 3.6 times compared with conventional encoder-equipped brushless DC motors. Our activities will continue for even further precision improvement.
Further development of the motor control technology and its future forecast
In recent years, robots, automotive, and other fields are planning to promote a trend towards modularization of motors, which integrates drive/power circuit, and a control circuit including a motor control unit. The motor control technology based on the Nidec-developed correction algorithm, i.e., the line segment connecting method (CLH), and the motor control technology based on the CLH have reduced the cost and improved the precision of encoder-less motor control modules, enabling us to sufficiently meet the needs for smaller and high-precision applications. We expect the products that utilize this technology to be applied to industrial robots and mobility systems for China, where competition for lower costs will intensify. While realizing the encoder-less position detection and control technology to win the cost competition, Nidec will research and accomplish better product performance to revolutionize the robot and mobility markets.