Select sensor for motor feedback
In many motion control applications, it is necessary to know the position, speed and even acceleration of the motor rotor or its load. Depending on the application and design details, the motor controller may need to know these parameters accurately, roughly or not at all. By knowing the motor condition and rotor state, the motor controller has a closed loop condition, as shown in Figure 1.
In many motor management and control applications, the real-time details of the rotor position and/or speed provided by the sensor assembly are critical to effective closed loop feedback and therefore accurate performance on the target. Source: Bill Schweber
Figure 1: In many motor management and control applications, the real-time details of the rotor position and/or speed provided by the sensor assembly are critical to effective closed-loop feedback and therefore to the accurate performance of the target. Source: Bill Schweber
Of course, the speed, position and acceleration of the motor are closely connected. Because speed is the derivative of the position (time rate of change), acceleration is the derivative of speed, so even if you only know one of the factors, you can determine all three factors (also note that the speed is the integral of the acceleration), the position is the speed integral).
However, in practice, this method of determining related parameters is usually (but not always) inadequate due to resolution and noise. For example, knowing that the rotor completes another rotation will tell you all three variables, but the resolution is very low and usually unacceptable. Depending on the application, the resolution and accuracy required can range from coarse to medium to precise. CNC machine tools require precise rotor information, automotive power window controllers can accept approximate data, and washers or dryers require only rough information.
To detect rotor position or motion, the most common options are resolvers, optical or capacitive encoders, and Hall-effect devices, roughly in descending order of accuracy, resolution, and cost. The physical design, implementation and electrical interface of these sensors are very different, so the user must understand what is needed, the best choice for a given application, and how they connect the sensor to the controller's circuitry.
Incremental encoders (used only when relative positions are required, or cost is a problem) are often used with AC induction motors. In contrast, absolute encoders (which provide different binary outputs at each location, thus absolutely determining the axis position) are typically paired with permanent magnet brushless motors in servo applications. Of course, the application is the main factor in determining whether incremental or absolute information is needed.
While most motor control is now done through a digital control loop, the sensor signal itself is either fully analog, digital, or digital, but voltage and other properties make it incompatible with standard digital circuits. While some feedback sensors provide "raw" output that can be customized as needed, many feedback sensors also have conditionalized, standard-ready, and interface-compatible outputs that are compatible with standard I/O ports, formats and protocols.
While more resolution seems like a good idea, it may not be the case in practice. Too much is obviously a good thing - resolution - you can slow down the system by asking for extra processing of unwanted or useful information, so it is a good idea to limit the resolution to the minimum required.
"I am here to solve..."
The resolver is a very precise, rugged, absolute position sensor. They are based on the basic principle of a transformer, a primary winding plus two secondary windings, orthogonal to each other (90°), Figure 2. The effective turns ratio and polarity between the primary and secondary windings depend on the axis. angle. The primary coil is excited with a constant frequency reference AC waveform with a frequency range of 50/60 Hz to several hundred kHz, and the output of the secondary winding is out of phase due to its physical location. The secondary peak voltage will vary with the rotation of the shaft and will be proportional to the angle of the shaft. Demodulate these outputs by using the main signal as a reference,
The resolver uses the primary winding and a pair of orthogonal secondary windings to evaluate the angle; it requires AC excitation and demodulation, but is accurate, robust, and provides absolute position information at power-up. (Source: Analog Devices, Inc.)
Figure 2: The resolver uses the primary winding and a pair of orthogonal secondary windings to evaluate the angle; it requires AC excitation and demodulation, but is accurate, robust, and provides absolute position information at power-up. (Source: Analog Devices, Inc.)
The resolver is not only accurate, but also rugged. There is no physical contact between the primary side and the secondary side, there is no separate brush or bearing other than the motor itself, there are no friction points that can cause component wear, and there is no chance of contamination (such as oil) interfering with operation. Due to their mechanical robustness and performance, rotary transformers are widely used in challenging situations, such as angle measurements in military guns.
However, rotary transformers tend to be larger and relatively expensive than alternatives and require a relatively large amount of power, which is generally unacceptable in low power applications. They also require relatively complex circuitry to generate and demodulate AC waveforms, although this is much less of a hindrance to modern ICs. They provide an absolute position indication when "powered up" and do not require any motion to index or determine the initial angle. Of course, this feature is "must have" in some cases and "don't care" in other cases.
Coding location, not data
Optical encoder (herein the term "encoder" is independent of the encoding of digital data) in incremental position readings, which use a light source (LED), two orthogonal photosensors and a glass or plastic disk between them, 3. The disc has a fine etched line radiating from its center, and as it rotates, the sensor sees a pattern of light and dark.
The number of lines on the disk and some other techniques determine the resolution, typically 1,024, 2048 or even up to 4,096 counts per revolution. Unlike transformer-like resolvers, optical encoders are not mass-market devices until long-life LEDs and high-efficiency photoelectric sensors are developed.
The optical encoder has a light source, an orthogonal light sensor and a wired insertion disk; it is small in size, low in power consumption, very easy to connect to the circuit, and provides excellent performance. (Photo: National Technology Intensive Learning Program (NPTEL), a project funded by the Government of India)
Figure 3: The optical encoder has a light source, an orthogonal light sensor and a wired insertion disk; it is small in size, low in power consumption, very easy to connect to the circuit, and provides excellent performance. (Photo: National Technology Intensive Learning Program (NPTEL), a project funded by the Government of India)
The physical arrangement of the sensors allows the encoder to determine the direction of rotation. The basic circuit converts a sequence of pulses from two sensors (called A/B outputs) into a pair of bitstreams indicating motion and direction, as shown in Figure 4.
The optical encoder's A/B quadrature and index outputs are compatible with many interface and motion control processor I/O ports. Source: Bill Schweber
Figure 4: The optical encoder's A/B quadrature and index outputs are compatible with many interface and motion control processor I/O ports. Source: Bill Schweber
However, the encoder is an incremental rather than an absolute motion indicator. To determine the absolute position, most encoders add a third track and photosensor as the indicator "zero reference track"; the axis must be rotated enough to pass the zero reference position to signal. There are ways to add true relative position readings to the optical encoder, but this adds to the complexity of the unit.
Optical encoders provide very good resolution, but they are not as rugged as a resolver. Dirt can interfere with the light path and the encoder disk can get dirty. However, their performance is sufficient for many applications, and they are small in size, light in weight, low in power, easy to interface, and low in cost.
A typical optical encoder for motor and spin applications is a similar HEDS-9000 and HEDS-9100 dual channel module from Avago Technologies (Broadcom). These high-performance, low-cost modules include a lensed LED source and a detector IC packaged in a small C-shaped plastic package, as well as drive and interface electronics, Figure 5. They have a highly collimated source and a special source for photodetection. The physical arrangement of the devices, so they are very tolerant of misalignment.
Figure 5: The Avago HEDS-9000 and HEDS-9100 dual-channel modules offer small size and mounting flexibility; the inserted discs are ordered separately and have the required count resolution per revolution. (Source: Avago Technologies / Broadcom)
Note that the disk called the code wheel is purchased separately, the HEDS-9000 has a resolution of 500 CPR and 1,000 CPR, and the HEDS-9100 has a resolution of 96 CPR and 512 CPR. These modules provide two TTL-compatible A and B digital output channels that require a 5 V supply, as shown in Figure 6.
The Avago optical encoder includes an alignment of physical components and internal electronic components. (Source: Avago Technologies / Broadcom)
Figure 6: The Avago HEDS-9000 and HEDS-9100 dual-channel modules offer small size and mounting flexibility; the inserted discs are ordered separately and have the required count resolution per revolution. (Source: Avago Technologies / Broadcom)
There is an alternative to optical encoders based on the principle of capacitance rather than optical principles, as shown in the CUI AMT10 series shown in Figure 7. These encoders offer a range of rugged, high-precision modular units available in both incremental and absolute versions, allowing users to select up to 12 (4,096 count) resolutions from 16 values with a four-position DIP switch. The CMOS-compatible A/B quadrature outputs of these units are reported via the standard SPI interface.
The CUI AMT10 capacitive encoder may look like an external optical encoder, but the basic working principle is very different. (Source: CUI, Inc.)
Figure 7: The CUI AMT10 capacitive encoder may look like an external optical encoder, but the basic working principle is very different. (Source: CUI, Inc.)
Unlike optical encoders, CUI AMT devices use conductors of repeated etched patterns on the moving and non-moving portions of the encoder. As the encoder rotates, the relative capacitance between the two portions increases and decreases, and this change in capacitance is sensed somewhat like the output of a phototransistor in an optical encoder. Dirt and other contaminants have almost no adverse effects here.
Keep in mind that a resolver or encoder is also a mechanical device with installation precautions and electrical compatibility requirements. To minimize inventory and inventory issues, CUI offers a variety of sleeves, covers and mounting bases for the AMT10 series, as shown in Figure 8, so the same basic encoder can be used for a variety of shaft diameters and installations.
In practice, the encoder must address a variety of axes and installations; CUI offers a full range of color-coded bushings and other accessories, so a single encoder can serve many applications. (Source: CUI, Inc.)
Figure 8: In practice, the encoder must address a variety of axes and installations; CUI offers a full range of color-coded bushings and other accessories, so a single encoder can serve many applications. (Source: CUI, Inc.)
The resolver and encoder can produce basic readings with resolutions up to 1/100 degrees (0.6 arc minutes) or better, but with different resolutions (again, some applications focus on one rather than the other). Whether the design uses a resolver or an encoder, the source of the error is due to temperature, tracking speed of change, unwanted phase shifts, and other factors. However, vendors of these units have devised ways to eliminate, cancel or compensate for many of these shortcomings, typically by using IC-based circuitry between the raw sensor output and the conditional output entering the system controller.
Hall effect devices become powerful
Yet another type of coding or sensor device is also based on the time principle, but requires modern semiconductor electronics and packaging to become widely available, available, and efficient. In addition, critical interface circuits can now be used on-chip, which can simplify the use of the technology by utilizing tiny voltages and easily interfacing them with the system. Hall effect devices can be used to sense current through a conductor that is part of the sensor, or if there is a nearby magnetic field.
The Hall effect we know is discovered by Edwin Hall in 1879: a potential difference - Hall voltage - produced on an electrical conductor, at right angles to the current in the conductor, perpendicular to the current of the current, Figure 9.
The principle of Hall effect devices involves currents, voltages and magnetic fields that are orthogonal to one another. (Photo: National Technology Promotion Learning Program (NPTEL), a project funded by the Indian government)
Figure 9: The principle of a Hall effect device involves current, voltage and magnetic fields that are orthogonal to each other. (Photo: National Technology Promotion Learning Program (NPTEL), a project funded by the Indian government)
Some Hall effect sensors far more than just the sensor elements themselves. The Melexis MLX90367 Triaxis Position Sensor is a monolithic absolute sensor IC that is sensitive to magnetic flux density that is orthogonal and parallel to the IC surface. It is sensitive to the three components of flux density, allowing the MLX90367 (with the correct magnetic circuit) to decode the absolute position of any moving magnet (for example, a rotational position of 0 to 360°).
Internally, this 12-bit resolution device includes on-chip signal processing with a microcontroller and DSP, Figure 10, so it can perform the required calculations and corrections to inherent nonlinearities, as shown in Figure 11. It also supports a wide range of user-selectable features and features, as well as a variety of output formats, including an advanced format with built-in error correction, called SENT (SAE J2716-2010), which is widely used in automotive applications.
The Melexis MLX90367 is more than just a Hall effect sensor; it includes amplifiers, digitizers, processors, firmware and I/O. (Source: Melexis NV)
Figure 10: The Melexis MLX90367 is more than just a Hall effect sensor; it includes amplifiers, digitizers, processors, firmware and I/O. (Source: Melexis NV)
The processing power of the MLX90367 enables it to significantly improve performance by correcting some of the avoidance errors in the linearity of the basic Hall effect sensor. (Source: Melexis NV)
Figure 11: The processing power of the MLX90367 enables it to significantly improve performance by correcting some of the avoidance errors in the linearity of the basic Hall effect sensor. (Source: Melexis NV)
Most Hall effect magnetic encoders use wheels that are connected to the motor shaft, and the wheel has a set of magnetized north and south poles at its periphery; it is the magnetic analog of the optical encoder slotted wheel. The wheels are usually made of injection molded ferrite embedded with a pole array. A typical wheel is magnetized by 32 poles (16 north poles and 16 south poles), so the resolution is much lower than that of an optical encoder or resolver, but in many cases it is usually sufficient. Typically mounted with three Hall effect sensors, the electrical spacing is 120° to sense the commutation of the wheels.
Designers who must detect motor position, speed or acceleration can choose from a variety of options, covering many key parameters and performance attributes. Rotary transformers, optical and capacitive encoders, and Hall-effect devices all have long-term, reliable records and are supported by application know-how.
The choice may be driven by one of the most important factors - such as ruggedness or low power - or through traditional and customary use in certain situations. Once the basic technologies to be used are determined, each still has many viable suppliers and components, so decisions on specific devices may require some research to better understand the trade-offs.