Machine Design

Servosensors Eliminate the Controller

Embedding a controller into sensor electronics results in a faster, simpler closed-loop system

Jerry Johnson
Paw-Taw-John Services
Rathdrum, Idaho

Jesse Russell
MTS Systems
Cary, N.C.

Most industrial closed-loop control systems have a fairly standard look. They usually consist of a controller, actuator, servovalve or motor, and a feedback sensor. A newly developed servosensor departs from this norm by eliminating the external controller. Instead the design closes the loop with a servocontroller inside the sensor. The result is a more compact package at a lower cost, with response time about half that of other closed-loop systems.

Developed through an agreement between Paw-Taw-John Services, an engineering firm specializing in closed-loop motion control, and MTS Sensors Div., the platform for this unique design is an MTS III magnetostrictive linear-position sensor. It offers fast displacement measurements, high resolution, a rugged housing, and it is routinely installed in servoactuators. With the recent availability of smaller, more powerful microprocessors and advances in surface- mount technology, for the first time it is possible to fit the servocontroller, drive module, and sensor electronics in the sensor head.

Servosensors hold a number of advantages over conventional closed-loop control systems and tops on the list is performance. One key to fast closed-loop response is the MTS III containing a proprietary microprocessor with a 4-GHz counter that permits high resolution with just a single interrogation of the sensor. In the past, slower clock speeds required interrogating the sensor several times to obtain high-resolution position information. That, in turn, would slow servoloop updates. The MTS III provides resolution to 0.0005 in., in stroke lengths from 1 to 65 in. Position update time depends on the sensor length but can be as fast as 7,500 Hz.

The servosensor also integrates proprietary MTS Temposonics sensing algorithms directly into the servocontroller. Therefore, it handles position and error information quite efficiently, which speeds servocontrol outputs.

Physical layout is a factor as well. For example, data is processed in the sensor head and this contributes to fast response. Also, the servosensor typically mounts inside an actuator, so just a short driver cable — usually less than 1-ft long — runs from sensor to valve.

In a conventional control system, on the other hand, the sensor sends signals to a controller that is typically mounted in an enclosure some distance away. The controller reads the position signal, accepts target information, closes the loop, and sends a control signal to the servovalve.

The bottom line is that servosensors close the loop much quicker, with servoloop times generally 1 msec or less. More conventional systems have typical response times of 2 msec. Servosensors have run successfully with servoloop times of 500 μsec at resolutions of 0.0005 in.

As a result, servosensors are well suited for many applications that require high accuracy and fast response, and they can also improve the response of systems with lower-performance valves. The units can directly control servovalves and some proportional valves. Maximum drive current is ±50 mA, but valves that require higher current generally have poor response and are not used to control highly dynamic systems. Servosensors can also drive proportional valves that have on-board electronics. The devices are presently used in fluid-power applications but are not limited to fluid-power control.

Servosensors are simpler, smaller, and cost less than conventional systems because they eliminate an external controller, reduce control-cabinet panel space and mounting hardware, and run on a single 24-V power supply. Most applications require only one cable from the servosensor (located in the actuator) to the host computer and power supply. Other systems typically require two per axis. While cost varies with the length of the sensor, the price for a typical system compares with that of a single-axis controller, in the range of $2,000. Add in the expense of external sensors, cabinet, power supply, and so on and the total cost for a conventional system can be significantly higher.

Servosensors store PID and operational software and are programmed with PCs, PLCs, or “smart” handheld devices. The user defines servoloop dynamics that include gains and zero adjustments, as well as operating parameters such as velocity, acceleration, dwell time, travel limits, sensor resolution, and emergency stops.

Output from the controller to servovalve is ±0 to 50 mA or ±0 to 10 Vdc. Currently, an RS-485 interface is standard and converter boxes for RS-422 and RS-232 communication networks are available. DeviceNet and other protocols, as well as analog interfaces, are in development.

In present applications, the servosensors operate as slaved units with no stored set points or machine-operation software. However, for simple systems, the controller can store parameters such as set points and dwell times and operate independently.

Servosensors are suitable for any system that demands high-speed, closed-loop motion control from a package that is smaller and less expensive than more-traditional control systems. They are currently being used in wood-product applications such as line-bars. Other applications currently under consideration include aerospace, food products, machine tool, and steel and plastic manufacturing.

© 2010 Penton Media, Inc.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.