Reinventing absolute rotary encoders
A new magnetic absolute rotary encoder design achieves the high resolution and accuracy of optical absolute encoders in a more compact, robust, and durable product. The encoder’s strength stems from a nonperiodic pattern and a patented method of signal evaluation to generate Gray code and high-resolution absolute positions from one circular track.
[Note: Gray code, reflected binary code, is where two successive values differ by one bit; named after Frank Gray.]
Optical rotary encoders
Conventional incremental encoders use a disk with either one or two concentric tracks. Each track is a pattern of equally spaced increments that are either transparent or opaque. A beam of light passing through the disk is detected by an optical sensor, which outputs equally spaced pulses as the disk rotates. In a dual-track arrangement, shown in Figure 1, each track is fitted with a light emitter-detector pair; the two sensors generate two sinusoidal analog signals with a phase difference. In a one-track arrangement, two signals are similarly produced by two sensors positioned at a quarter-period offset. These analog signals are then converted into digital signals (square waves) in which one channel leads the other by 90 electrical degrees. By monitoring the phase difference of the two output channels, the direction of rotation can be determined.
Optical encoders typically have an incremental pattern containing a number (n) of periodic sections, ranging from 250 to several thousand, spaced equally around the disk. The signal digitizing process produces 4xn counts for every turn. The analog values of the two signals are used to calculate the exact position within one period of the incremental pattern, thus achieving an incremental position with a high resolution. However, to obtain an absolute position within one turn, the starting position must be known.
To determine absolute positions, the disk of a rotary encoder includes an additional absolute track with a binary code pattern; an example is shown in Figure 2. The absolute track is composed of segments of different length, according to the pattern design. Each segment is a multiple of the divisional step, which is defined by the resolution.
The resolution of the absolute track must be as high as the resolution of the incremental track. The length of each divisional step of the absolute pattern is therefore equal to the length of the periodic section (graduation period) of the incremental pattern.
To read the absolute track, several sensors read consecutive steps of the pattern. The distance between adjacent sensors must be equal to the divisional step of the absolute track or the graduation period of the incremental track. Due to the consequent small size of such sensors, custom-designed miniature optical sensors are typically encapsulated as a sensor array on an application-specific semiconductor chip. Figure 3 shows an example of a sensor arrangement of an absolute encoder.
The digitized output signals from the sensor array provide a Gray code, a binary code in which two successive values differ in only one bit. Because successive incremental position codes differ by just one binary digit, the Gray code prevents the introduction of erroneous code at transitions between positions.
Magnetic rotary encoders
Magnetic encoders, which are based on the same principle as incremental encoders, have the advantage of being more robust than optical encoders, since they are less sensitive to shock, vibration, and contamination. They also are more durable because there is no degradation of light-emitting diodes.
However, since the magnetic field rapidly decreases as the distance from the magnet surface increases, the number of periodic sections of the pattern is no more than a few dozen. Moreover, if the graduation period is very small, the magnetic sensor must be very close to the magnet surface to sense a distinct transition.
Magnetic encoders commonly compensate for a small number of periods by performing analog processing at a higher resolution. This results in a higher sensitivity to electrical noise. In addition, the signal in one period is less precise, rendering the overall accuracy of magnetic encoders inferior to that of comparable optical encoders.
Optical, magnetic disadvantages
Both optical and magnetic encoders have several disadvantages. They require at least two tracks and an array of sensors to determine the absolute position of the rotating disk. In particular for magnetic encoders, arranging two concentric magnetic patterns on an encoder disk is difficult. In optical and magnetic absolute encoders the reliability of position detection depends largely on the accuracy of the code projection onto the sensor array.
Because of the low dimensional tolerances of the code tracks, absolute encoder disks must be manufactured with extreme precision and their size can only be as small as is feasible. This explains why absolute encoders commonly have 256 divisions, while incremental encoders of the same size usually have 1,024 increments.
New design advantages
A new magnetic absolute rotary encoder design overcomes many of the disadvantages of conventional absolute encoders, as shown in Figure 4. A number of permanent magnets of different sizes are positioned along the outer edge of the encoder disk in one circular track, forming a magnetic code track whose pattern is nonperiodic.
Magnetic (Hall) sensors are fixed to a static part of the encoder, spaced equidistantly from each other, arranged concentrically and in close proximity to the magnetic code track.
A patented algorithm generates a Gray code with a maximum number of positions for a given number of sensors from a nonperiodic pattern of one magnetic code track.
In addition, the sensors’ analog output signals directly provide a high-resolution absolute position, without needing additional incremental readings. The sensors produce electrical signals proportional to the strength of the magnetic field generated by the facing magnet. These analog signals are first digitized by comparing them to a threshold value, thus generating a Gray code that describes an absolute position at a low resolution. A configuration of seven sensors and seven magnets, for example, creates a Gray code that identifies 98 positions.
To achieve a higher absolute resolution, an additional patented method for signal evaluation is applied. Two analog signals are associated with each Gray code according to a predefined signal table. The absolute position of the disk corresponds to the associated position value in a prerecorded position table for the analog signal whose value is closest to the threshold.
Such a design can deliver resolution of 20 bits using a 12-bit analog-to-digital converter and a configuration for seven sensors.
Rugged design is particularly reliable for motor-feedback applications exposed to severe shock, for example due to emergency braking, or high vibration, as in the mining, steel, cement, and paper industries.
– Dr. Markus Erlich is vice president marketing, Servotronix Motion Control Ltd. Edited by Mark T. Hoske, content manager, CFE Media, Control Engineering, email@example.com.
- New design puts the high resolution and accuracy advantages of optical encoders in a compact, robust magnetic absolute rotary encoder.
- Magnetic encoders are less sensitive to shock, vibration, and contamination than optical encoders.
- Rugged design suits motor-feedback applications exposed to severe shock, such as emergency braking or high vibration, for mining, steel, cement, and paper industries.
New designs may afford higher reliability for a rotary encoder application.
See additional information about the products described in an online-only story linked below.