Enhanced Communication Reliability and Performance for Motor Control Encoder Applications

Rotary encoders are widely used in industrial automation systems. A typical use of this type of encoder is for electric machines where the encoder is connected to the rotating shaft, and thereby provides feedback for the control system. While the primary purpose of the encoder is angular position and speed measurement, additional features, such as system diagnostics and parameter configuration, are common as well. Figure 1 shows a motor control signal chain using RS-485 transceivers and microprocessor to interface between the absolute encoder (ABS encoder) slave and industrial servo drive master for closed-loop control of an ac motor.

The RS-485 communication link between the servo drive and ABS encoder typically requires high data rates up to 16 MHz and low propagation delay timing specifications. The RS-485 cabling typically extends to a maximum of 50 meters, but in some cases can be as long as 150 meters. Motor control encoder applications are challenging environments for data communications, because electrical noise and long cable lengths affect the integrity of RS-485 signaling. This article focuses on key benefits for motor control applications using Analog Devices 50 Mbps (25 MHz) ADM3065E RS-485 transceiver and the ADSP-CM40x mixed-signal control processor.

Figure 1. Using RS-485 to interface between the absolute encoder slave to servo drive master for closed-loop control of an ac motor.

The ADM3065E RS-485 transceiver is designed for reliable operation in harsh environments such as motor control encoders, with added noise immunity and (IEC) 61000-4-2 electrostatic discharge (ESD) robustness.

Noise Immunity

RS-485 signaling is balanced, differential, and inherently noise immune. System noise couples equally to each wire in an RS-485 twisted pair cable. One signal emits the opposite of the other signal, and electromagnetic fields coupled onto the RS-485 bus cancel each other out. This reduces the electromagnetic interference (EMI) of the system. In addition, the enhanced ADM3065E 2.1 V drive strength allows greater signal-to-noise ratio (SNR) in communications. Adding signal isolation to the ADM3065E can be easily implemented using the ADuM141D. The ADuM141D is a quad-channel, digital isolator based on Analog Devices iCoupler® technology. The ADuM141D can operate at a data rate of up to 150 Mbps, making it suitable for operation with the 50 Mbps ADM3065E RS-485 transceiver (see Figure 2). Direct power injection (DPI) measures the ability of a device to reject noise that is injected into the power supply or input pins. The isolation technology used in the ADuM141D has been tested to the DPI IEC 62132-4 standard. The ADuM141D noise immunity performance exceeds that of similar products. The ADuM141D maintains excellent performance over frequency, but other isolation products exhibit bit errors in the 200 MHz to 700 MHz frequency band.

Figure 2. Signal isolated, 50 Mbps RS-485 solution (simplified diagram—all connections not shown).

IEC 61000-4-2 ESD Performance

ESD on the exposed RS-485 connectors and cabling for the encoder to motor drive is a common system hazard. The system-level IEC 61800-3 standard relating to EMC immunity requirements for adjustable speed electrical power drive systems requires a minimum ±4 kV contact/±8 kV air IEC 61000-4-2 ESD protection. The ADM3065E exceeds this requirement with ±12 kV contact/±12 kV air IEC 61000-4-2 ESD protection. Figure 3 shows the 8 kV contact discharge current waveform from the IEC 61000-4-2 standard compared to the human body model (HBM) ESD 8 kV waveform. Figure 4 shows that the two standards specify a different waveform shape and peak current from one another. The peak current associated with an IEC 61000-4-2 8 kV pulse is 30 A, while the corresponding peak current for the HBM ESD is more than 5× less, at 5.33 A. The other difference is the rise time of the initial voltage spike, with IEC 61000-4-2 ESD having a much faster rise time of 1 ns, compared to the 10 ns associated with the HBM ESD waveform. The amount of power associated with an IEC ESD waveform is much greater than that of an HBM ESD waveform. The HBM ESD standard requires the equipment under test (EUT) to be subjected to three positive and there negative discharges—while in comparison, the IEC ESD standard requires 10 positive and 10 negative discharge tests. The ADM3065E with the IEC 61000-4-2 ESD ratings is better suited for operation in harsh environments compared to other RS-485 transceivers that state varying levels of HBM ESD protection.

Figure 3. IEC 61000-4-2 ESD waveform at 8 kV compared to the HBM ESD waveform at 8 kV.

EnDat Communication Protocol

A number of communication protocols are used for encoders; for example EnDat, BiSS, HIPERFACE, and Tamagawa. Despite their differences, the encoder communication protocols have similarities in regard to implementation. The interfaces of these protocols are serial bidirectional pipes that comply with either the RS-422 or RS-485 electrical specifications. While there are commonalties in the hardware layer, the software required to run each of the protocol is unique. Both the communication stack and the required application code are specific to the protocol. This article focuses on hardware and software implementation of the master side of an EnDat 2.2 interface.

Impact From Delays

Delays fall into two categories: first, there is the transport delay of the cable, and second, there is the propagation delay of the transceivers. The speed of light and the dielectric constant of the cable determines cable delay with typical numbers of 6 ns/m to 10 ns/m. When the total delay exceeds half a clock period, the communication between the master and the slave breaks down. At this point, the designer has the following options:

  • Lower the data rate
  • Bring down the propagation
  • Introduce delay compensation on the master side

Option 3 compensates for both cable delay and transceiver delay and therefore is an effective way to ensure that the system can run with high clock rates on long cables. The disadvantage is that the delay compensation increases the system complexity. In systems where delay compensations are either not possible, or in systems with short cables, the value of using transceivers with a short propagation delay is evident. A low propagation delay enables a higher clock rate without having to introduce delay compensation in the system.

Master Implementation

A master implementation consists of a serial port and a communication stack. Because the encoder protocols do not comply with standard ports, such as a UART, the peripherals found on most general-purpose microcontrollers cannot be used. Instead, the programmable logic of an FPGA enables implementation of dedicated communication ports in hardware and support of advanced features such as delay compensation. While an FPGA approach is flexible and can be tailored to the application, it also comes with disadvantages. When compared to a processor, a FPGA is costly, power hungry, and has significant time-to-market.

The implementation of the EnDat interface discussed in this article is done on the ADSP-CM40x from Analog Devices, which is a processor targeting motor control drives. Besides peripherals for motor control, such as pulse width modulator (PWM) timers, analog-to-digital converters (ADCs), and sinc filters, the ADSP-CM40x has highly flexible serial ports (SPORTs).

These SPORTs are capable of emulating a number of protocols, including encoder protocols such as EnDat and BiSS. Because of the rich peripheral set of the ADSP-CM40x, it is possible to perform advanced motor control, as well as interfacing to an encoder with the same device. In other words, the need for an FPGA is eliminated.

Test Setup

The EnDat 2.2 test setup is shown in Figure 4. The EnDat slave is a standard servo motor from Kollmorgen (AKM22) with an EnDat encoder (ENC1113) mounted to the shaft. Three pairs of wires (data, clock, and power lines) connect the encoder to the transceiver board. There are two transceivers and power supply for the encoder on the EnDat PHY. One of the transceivers is used for the clock and the other transceiver is used for the data line. The EnDat master is realized with ADSP-CM40x using a mix of standard peripherals and software. Both the transmit port and receive port are implemented with flexible SPORTs.

Figure 4. Experimental setup.

The EnDat protocol consists of a number of different frames of varying length. However, these frames are all based on the same sequence, as seen in Figure 5. First, the master issues a command to the slave, then the slave processes the command and performs the necessary calculations. Finally, the slave sends the result back to the master.

Figure 5. EnDat transmit/receive sequence.

The transmit clock (Tx CLK) is generated by the processor ADSP-CM40x. Because of delays in the system, the data from the encoder will be out of phase with the transmit clock before they get back to the processor. To compensate for transport delay, tDELAY, the processor also issues a receive clock (Rx CLK), which is delayed by tDELAY compared to the transmit clock. Bringing the receive clock in phase with the data received from the slave is an effective way to compensate for the transport delay.

The clock signals from the processor are continuous, while the EnDat protocol specifies the clock must only be applied to the encoder during communication. At all other times the clock line must be held high. To handle this, the processor generates a clock enable signal, CLK EN, which is fed to the ADM3065E’s data enable pin. After exactly two clock periods (2T) the master starts clocking out the command on Tx DATA.

The command is 6 bits long and is followed by two 0 bit. To control the data direction through the transceiver, the processor sets Bit Tx/Rx EN high while transmitting.

While the slave prepares a response, the system enters a wait state where the master continues to apply clocks, but the data line is inactive. When the slave is ready to respond, the data line receive data is pulled high and the response is sent immediately after. After receiving the n bits response, the master stops the clock by setting CLK EN signal low. At the same time, the ENC CLK signal goes high. The data flow is half duplex and the traffic on the combined data line is shown as ENC data.

Figure 6. EnDat data exchange.

作者

Generic_Author_image

Jens Sorensen

Jens Sorensen 是ADI公司应用工程师,负责工业应用的电机控制解决方案。他的主要兴趣在于控制算法、电源电子和控制处理器。Sorensen目前专注工业应用,而在早期职业生涯中致力于开发家用电器和汽车应用中的电机控制和电源电子元件。

Richard Anslow

Richard Anslow

Richard Anslow毕业于爱尔兰利默里克大学,获工程学士和硕士学位。他从事新产品定义工作,并且担任面向客户的角色,处理有关ADI隔离式通信产品系列的问题。