By Andreas Schroter
Edited by Miles Budimir
The inductive rotary encoder sits between resolvers and optical rotary encoders. It is an absolute position device. By virtue of integrated electronics, it can generate an electronic ID label and handle diagnostic functions as well. Inductive rotary encoders are also good for applications where accuracy requirements are relatively low, generally in the range of 10 arc minutes or higher. The only real limitations are susceptibility to extreme temperatures and shock, which makes resolvers a better choice in harsh environments.
INSIDE THE INDUCTIVE ENCODER
There are three basic components to an inductive encoder; an excitation coil, a series of receiving coils, and movable shielding disks. A signal with an 800-kHz carrier frequency is applied to the excitation coil, generating a magnetic flux. The receiving coils sit within the excitation coil. A current is induced through the existing alternating field into the receiving coils.
The excitation and receiving coil are on the same PCB. The conductive shielding disk is located on a separate PCB that is attached to the shaft so that signals do not have to be transmitted to any moving parts. As the shaft rotates, the shielding disk passes over one of the two receiving coils. This induces a current in that coil, weakening its magnetic field. The induced voltage around the remaining surface of the receiving coil predominates, modulating a signal onto the 800-kHz carrier. As the shielding disk moves, the amplitude of the signal changes depending on the proximity of the shielding fields relative to the receiving coils.
This modulated carrier signal then feeds into a mixedsignal ASIC where it's demodulated to yield a sine wave. There are two tracks on the shielding disk; one track has a resolution of 32 signal periods and the other has one signal period. The absolute value is determined by evaluating the signals from both tracks in the ASIC. The absolute position value is then transmitted via the EnDat interface (see box). The transfer time for the position value in the multiturn version (29 bits), at a clock frequency of 2 MHz, is only 23 sec.
The encoder also has the raw 32 signal periods of the first track available in a 1-V peak-to-peak, sinusoidal output. This can be interpolated within the subsequent electronics to determine either position or speed. This feature is for applications that require absolute data only at power up at which point they go back to incremental data for either speed control or position. Using 12-bit interpolation with the 32 1-V peak-to-peak signals yields a resolution of 131,072 counts/rev.
There are two forms of absolute feedback with rotary encoders: single turn and multiturn. Applications requiring angle measurement work within one rotation. These are considered single-turn applications because there is no need to determine the total number of revolutions.
Applications involving motor position require multiturn capability. The controller needs to know the position within a revolution as well as the number of revolutions. This particular encoder is 29 bit; 17 bits/turn with 12 bits or 4,096 distinguishable turns. This gives 536,870,912 positions within one complete cycle.
The multiturn part operates with the same principle used in optical multiturn encoders; namely, gear reduction and permanent-magnet graduated disks. In contrast to electronic revolution counters, the mechanical gears do not need servicing and are not susceptible to interference signals and battery-buffer failure.
In contrast to optical encoders, the inductive version does not have an integral bearing. The encoder shaft is fixed to the motor shaft, and the encoder flange is secured directly to the motor flange. The air gap between the scanning PCBs is adjusted with simple tools during setup. Changes in the size of the air gap (e.g., from thermal expansion of the motor) influence the signal amplitudes. So developers took into consideration the tolerance for axial motion of the motor shaft; electronics corrects for any errors.
The encoder is also fairly robust. Typical motor applications need to survive 115°C. The encoder can run at up to 12,000 rpm, and take 100 g's of shock and 10 g's of vibration.
An EEPROM stores automatic commissioning data, with 1.4 kbytes of available memory. The motor-encoder unit receives an electronic ID label. The evaluation electronics of the encoder are also equipped with diagnostic functions which recognize and report malfunctions.
The EnDat (Encoder Data) interface is a bidirectional, pure binary serial interface. It's used to transmit position information from the encoder to the controller. Motor data such as voltage ratings, current ratings, and other details can store in the encoder memory. So on power up, the encoder dumps its memory to the controller. It also provides warnings and alarms for scheduled maintenance.
EnDat contrasts with fieldbuses and standard point-to-point communications. A fieldbus, for instance, typically has slower transmission rates than point-to-point systems. A fieldbus connection with several sensors and actuators could take as long as several milliseconds for transmission, which could result in unacceptable dead times in the control loop. One solution is to use a faster fieldbus. But fieldbus-equipped encoders are more expensive than those without.
Encoders with built-in fieldbus are best for applications with low-to-medium control dynamics requirements such as positioning accuracy, resolution, and transmission speeds, or where a fieldbus is already in place.
Typical point-to-point configurations also have drawbacks such as low data transmission rates and low transmission reliability. There is also no support for automatic parameter input or safety monitoring and diagnostics.
EnDat permits cable lengths of up to 150 m with maximum clock frequencies of 2 MHz. Redundancy techniques and simultaneous transmission of absolute and incremental position values makes encoders with an EnDat interface suitable for use in machines such as presses with critical safety requirements.