High-Speed Interface for Absolute Position Measuring Systems Position encoders are frequently integrated in control systems based on programmable logic controllers (PLC). These systems are usually designed with point-to-point connections. While such designs do ensure fast transmission of the positional value to the PLC, they are also saddled with the disadvantage of requiring complex cabling configurations. Figure 1: Common communication configurations for rotary encoders The present trend is to reduce cabling complexity by using field bus systems (such as CAN, Interbus-S, and Profibus-DP) between the PLC and the actuators and sensors. A field bus takes more time than the common point-to-point configurations to transmit the positional value. A field bus with several actuators and sensors typically takes one or more milliseconds for transmission, resulting in dead times in the control loop that are unacceptable for many applications. Also, a position encoder with field bus interface is relatively expensive. Position encoders with field bus interface are therefore best suited for applications with low-to-medium requirements on control dynamics, and for systems that already require a field bus.   Point-to-point communication Bus structure Common methods EnDat Data transmission Unidirectional Bidirectional Bidirectional Transmission rate Medium High Low Data security Low High High Price for position encoder Low Low High Cabling costs High High Low Table 1: Characteristics of various interface types A hybrid solution can be found in the application of decentralized input/output components or decentralized PLCs that stay close to the sensors and actuators in order to handle tasks requiring quick response. The position encoders are connected in a point-to-point structure with extremely short cable lengths to the decentralized I/Os or PLCs, while less time-critical data is sent over field bus to the higher-level PLC. Such a solution exploits the advantages of both designs: short transmission times over the point-to-point connections as well as reduced cabling for the field bus — particularly over long distances. For applications where time is a critical factor, point-to-point communication from the encoder to the control electronics remains essential. However, most point-to-point configurations used today for synchronous-serial position transmission have the following drawbacks: Low transmission reliability, Low data transmission rate, No support for automatic parameter input, No support for safety monitoring, Little support for monitoring and diagnostic functions, Suitable only for certain types of absolute position measurement. With the development of its new EnDat (for Encoder Data) interface, HEIDENHAIN has succeeded in creating an interface for point-to-point communication between position encoders and subsequent electronics that is standardized for all types of absolute encoders and avoids all these disadvantages. Standardized Interface for Point-to-Point Communications Position encoders with EnDat interface transmit bidirectionally in synchronism with a clock signal sent by the subsequent electronics. Not only position values, but also parameters can be transmitted quickly and reliably over only four lines. The subsequent electronics send a mode command to the encoder to determine whether it is to transmit position values or parameters. Position value transmission (Fig. 2) begins with a start bit followed by an alarm bit. This is followed by the position value in pure binary code and five CRC bits for security. The alarm bit is a collective message for all monitored functions and can be used for fault monitoring. The exact cause for activation of the alarm bit is stored in the encoder memory and can be read by the subsequent electronics. Figure 2: Position value transmission begins with a start bit and an alarm bit The encoder has various memory areas (Fig. 3) that can be read from, and some written to, by the subsequent electronics. The memory area for operating parameters contains the datum shift, which can be adapted to the motor or machine either by the OEM or the end user. The memory area for parameters of the OEM is freely definable and can be write protected. The memory area for parameters of the encoder manufacturer is write protected and contains all data specific to the encoder. Figure 3: The memory areas can be read, and some written to, by the subsequent electronics As an option, additional sinusoidal incremental signals with 1 VPP levels can be transmitted. Cable lengths up to 150 meters are permissible. The maximum clock frequency for serial data transmission is presently 2 MHz. Automatic Parameter Setup Encoder parameters comprise information specific to the encoder model, such as the encoder type (absolute singleturn, absolute multiturn, absolute linear), measuring steps per revolution or number of resolvable revolutions. At present, the widespread practice is to have someone enter these parameters manually at the control unit panel or at a computer. This method, of course, is time consuming and highly susceptible to error. Encoders with EnDat interface therefore have all encoder parameters stored by the encoder manufacturer in a separate memory area. This makes it possible to have the parameters read and entered automatically. To support maintenance and service personnel, the encoder can store two separate words, each with 16 bits, describing causes of encoder fault conditions or the violation of tolerance limits critical to the long-term function of the encoder. These words are distinguished as either an alarm or a warning. An alarm is set when a malfunction of the encoder can result in incorrect position values. Alarms are released, for example, if the power supply voltage is too low, the light source fails, or the amplitude of the scanning signal is too low. Warnings simply show that tolerance limits for certain internal values of the encoder have been exceeded. These include the maximum electronically permissible speed, the permissible operating temperature, and the expiration of control response from the light-source. Warnings do not necessarily indicate that incorrect position values have been transmitted. Rather they make it possible to perform preventive maintenance when it becomes necessary, thereby reducing the idle times of expensive systems. Support of Safety Monitoring In order to reduce cost, manufacturers of machines and systems would like to be able to do without limit switches. For systems in which safety is particularly critical, two position encoders are often used to provide redundancy. Some manufacturers wish to use only one position encoder, which provides a "safe" position value both at a standstill and at the maximum permissible speed. A multiturn absolute encoder can provide an example of how this can be achieved. In its basic design it consists of several graduated disks, some of which are connected together mechanically in gear ranges. Scanning the individual tracks of the coded disks immediately provides digital information at the input of a gate array where it can be logically combined to form the position value. Logical gating eliminates the effect of the unavoidable play between the gear ranges. The position value is transmitted over a line driver to the subsequent electronics. Since a malfunction can occur at any of the components in this chain, a function monitoring system must be able to monitor all of them. The functions to be monitored can be divided into three groups: Light source, Schmitt triggers at the inputs of the gate array, shift registers at the output of the gate array and line driver Gate array. The light source is monitored by means of a closed-loop control that releases an error message if it fails to remain within the permissible control range (Fig. 4). To monitor the Schmitt triggers at the inputs of the gate array, the shift register at the output of the gate array, and the line driver, test voltages that result in defined changes in the states of the output signals are applied to the Schmitt triggers while the encoder is stationary. If the source signal from the photocell lies near a switch threshold this results in impermissible status changes in the output signals. The system recognizes these changes and generates an error message. Figure 4: Security methods for absolute rotary encoders Since absolute rotary encoders permit shafts speeds up to 12,000 rpm, the dynamic monitoring of position value acquisition must be very fast. Under such time conditions it is presently not possible to complete the same routine as described above for a stationary encoder. It is possible, however, to monitor transmission at high shaft speeds as well by using redundant transmission of absolute position values over the serial interface, separate incremental signals in quadrature, and a dynamic code-track congruity check. The dynamic code sequence check results in an error message if the positions of the scanning signals relative to each other exceed certain tolerances. These tolerances are influenced both by play in the gears between the multiturn ranges and by the phase angle error that results from the unequal effects of low-pass filtering in the photocell signal amplifiers for the various tracks. With these deviations, an accuracy of ±1 LSB (least significant bit) in the absolute position value can no longer be realized. In order to receive the required absolute position information in spite of the high shaft speeds, the encoder transmits to the subsequent electronics two 90° phase-shifted incremental signals in addition to the absolute position values. The absolute position value received must agree with the incremental position value within an "accuracy frame" defined for the respective shaft speed (for example ±1 LSB up to 1500 rpm and ±50 LSB up to the maximum permissible speed). If these "accuracy frames" are exceeded, this can be determined in the subsequent electronics (Fig. 5). Figure 5: Redundancy checking techniques ensure high data security Table 2 lists the main characteristics of various point-to-point interfaces for position encoders. The EnDat bidirectional interface is characterized above all by short transmission times and high data transmission security. Unlike the unidirectional data transmission widely used today, it supports installation, monitoring, fault diagnosis, and safety techniques.   Synchronous Asynchronous Common methods EnDat RS-485 Data transmission Unidirectional Bidirectional Bidirectional Clock frequency 100 kHz to 1 MHz 100 kHz to 2 MHz 0.6 kHz to 38.4 kHz (600 to 38,400 baud) Maximum cable length 100 m 150 m 100 m Number of lines1) 4 4 2 Transmission time2) 3) /clock frequency minimum: 25 µs/1 MHz 50 µs/1 MHz 4) typical: 50 µs/500 kHz 100 µs/500 kHz 4) minimum: 20 µs/2 MHz typical: 40 µs/1 MHz minimum: Approx. 3 ms at 38,400 baud typical: Approx. 11 ms at 9600 Bd Additional incremental signals Optional Data checking methods None, double transmission, parity bit CRC Parity bit and CRC Reliability of check None or low High Low 1) Only for serial transmission of the absolute position value 2) Including data checking 3) At least 8192 measuring steps per revolution and 4096 distinguishable revolutions and transmission of the complete position value 4) With double interrogation for data security Table 2: Characteristics of synchronous and asynchronous point-to-point communication Position Encoders for Differing Application Needs The above described characteristics make the EnDat interface ideal for the transmission of absolute positional values in many fields of automation. On machine tools, textile machines and printing machines, for example, digital control loops with extremely good positional performance can be realized in combination with virtually delay-free sinusoidal incremental signals. Thanks to redundancy techniques and the simultaneous transmission of absolute and incremental position values, encoders with EnDat interface can be also be used in machines with critical safety requirements, such as presses. The interface is also suited for automation tasks with relatively low time requirements for transmission of the actual value, for example on foil cutting machines. Regardless of the application, the EnDat interface always uses the same input hardware in the subsequent electronics. Specific application needs can be met through a wide variety of encoders, all of which possess the same interface (Fig. 6). Figure 6: Singleturn absolute encoders, multiturn absolute encoders, absolute angle encoders, and absolute linear encoders with the standardized EnDat interface Absolute rotary encoders of the ROC 400 and ROQ 400 series have the same mounting dimensions as incremental encoders of the ROD 400 series and are intended for mounting by the system manufacturer or end user. The standard models provide IP 64 and IP 67 protection and are available in singleturn (ROC 413) and multiturn (ROQ 425) versions. In addition to the absolute position values they provide sinusoidal incremental signals with levels of 1 VPP. They are therefore suitable for the safety techniques described above, and for applications with high demands on dynamics and resolution. The ROC 413 singleturn absolute encoder has the same outside dimensions as an incremental rotary encoder. Its multiturn counterpart, with its special scanning configuration and highly integrated components, is only 7 mm longer in the version with radial flange socket. Singleturn and multiturn absolute encoders with these dimensions are also available for field bus systems. The encoders for Profibus-DP function in accordance with the recommendations of the Profibus user organization Profibus-Nutzer-Organisation PNO for rotary encoders and support the entire Class 2 range of functions. The absolute rotary encoders ECN 1313 and EQN 1325 were developed particularly for installation in drives with digital speed control and natural cooling. They have a stator coupling that compensates axial and radial offset between the motor shaft and the encoder shaft. The rigid coupling of the rotary encoder via taper shaft to the motor shaft permit control loops with large bandwidths. A permissible operating temperature up to 115 °C (239 °F) permits smaller motor dimensions for a given rated torque. Since these encoders provide 2048 or 512 signal periods per revolution, subdividing the incremental signal in the subsequent electronics by a factor of 1024, for example, results in 2,000,000 or 500,000 measuring steps per revolution, respectively. The motor manufacturer installs the encoders in the motor and programs them through the EnDat interface for the motor e.m.f. This eliminates time-consuming adjustments. Together with a special design of the graduation and other design measures in the scanning process, subdivision of the sinusoidal incremental signals in the subsequent electronics results in positional deviation within one signal period of less than ±1% of the signal period. This means that for 2048 signal periods per revolution, the positional deviation (error resulting from subdivision) within one signal period is less than ±7". The same signal quality with 512 signal periods per revolution results in positional deviation up to ±26". The specified positional deviation attained through subdivision of the sinusoidal signals is maintained also at relatively high operating temperatures, which makes it possible for drives with digital speed control to achieve low speed ripple and high control bandwidths within the typical range of operating temperatures. The RCN 220 and RCN 723 sealed absolute angle encoders attain an accuracy of ±2.5 arc seconds and 2 arc seconds, respectively, and are equipped with an integrated stator coupling. The IP 64 protection rating and the permissible vibration of 100 m/s² (50 Hz to 2,000 Hz) make these absolute angle encoders particularly attractive for harsh environments. The outputs are 20 bit and 23 bit resolution (8,388,608 measuring positions per revolution), respectively. Both outputs have absolute values and additional 16,384 and 32,768 incremental signal periods per revolution, respectively, permitting measuring steps as fine as 0.0003° and 0.000 04°. The RCN 220 and the RCN 723 are mechanically compatible with sealed incremental angle encoders of the RON 200 series and RON 706/RON 786 series, respectively.. The LC 181 absolute linear encoder was conceived primarily for machines on which, for technical reasons or due to the axis kinematics, a reference run to ascertain the absolute position value after power interruption is either impossible or very difficult and would therefore entail long idle times. It is also suited for synchronous linear motors that need an absolute position value immediately after switch-on for commutation. It is intended for measuring steps of 0.1 µm to 1 µm and is offered in measuring lengths up to 3 meters. Conclusion The EnDat bidirectional synchronous-serial interface now makes it possible to meet the demands of the future on the transmission of absolute position values from rotary and linear encoders to their subsequent electronics. In comparison with the synchronous-serial data transmission commonly used today it enables faster transmission times, greater transmission security, and provides a monitoring and diagnostic capability. Moreover, it is independent of the type of absolute position value acquisition, permits automatic parameterization of the subsequent electronics with the characteristics of the encoder, and supports servicing and preventative maintenance. With its capability of transmitting over cables as long as 150 meters between position encoder and subsequent electronics, and in combination with additional sinusoidal incremental signals with levels of 1 VPP, the EnDat interface can play an essential contributive role in the design of high-quality control loops.   About the author Dr. Rainer Hagl studied mechanical engineering at the University of Stuttgart, where he also completed his graduate work. Since 1992 he has been with the Dr. Johannes Heidenhain company in Traunreut, Germany, where he heads the technical department for rotary and angular encoders.