Finding the right sensor for linear displacement

Les Schaevitz
Design-for-Application Product
Manager Macro Sensors Div.
Howard A. Schaevitz Technologies Inc.
Pennsauken, N.J.

Engineers continue to increase their use of linear-position and displacement sensors for providing real-time information to feedback-hungry PLCs and microcontrollers, which are taking over more control of linear-position/actuator-based systems. For example, they are now used for control surfaces on aircraft, vane position in gas turbines, valve position in steam turbines, and to measure dimensions for quality assurance. But fitting the right sensor to the application requires that design engineers have a working knowledge of the operating characteristics behind today's position sensors.

These devices, known as linear-position and displacement sensors, are proportionaloutput devices that continuously indicate position. They are not threshold-type devices, like proximity sensors, which only determine if an object is present and whether it has moved to or passed a specified location. These sensors are also classified as either absolute or incremental (relative) sensors. Absolute sensors generate the same output after a power outage as before. Conversely, incremental sensors, such as encoders, cannot make meaningful measurements after power outages until they are rezeroed or returned to a “home" position.

Although most of the sensing technologies discussed here can be used in rotary and angular-position sensing, we will stick to linear applications, specifically those used in most industrial and commercial applications of 100min. to 100 in. (2.5mm to 2.5 m). Choosing a sensor is a function of three factors:

• Basic economic issues of price and nonrecurring expenses.
• Technical performance.
• Physical attributes such as packaging, connectivity, and environmental considerations.

OEMs buying sensors are heavily driven by economics. System integrators, on the other hand, focus on the physical attributes as they relate to installation. And researchers probably choose sensors based on technical performance.

The most frequently used electromechanical linear position and displacement sensors can be divided into six basic technologies: resistive, capacitive, inductive, magnetic, time-of-flight, and pulse encoding.

Resistive sensors or “pots" (short for potentiometer) are the best-known and most frequently used resistive sensors. They use a moving contact sliding against a fixed resistive element to generate changes in resistance. Hooked to a dc source as a voltage divider, they produce a proportional voltage output when used with highimpedance loads. Pots are easy to use, relatively economical, and require almost no support electronics. However, because they are contact devices, they have poor repeatability, large hysteresis, and output tends to deteriorate over time due to wear, particularly when exposed to vibrations. This makes them unacceptable for applications needing long-term reliability.

Magnetoresistive sensors, a contact-free variation on the pot, use moving magnets, thus eliminating wear problems. However they typically suffer from relatively large temperature coefficients, which are not acceptable in many applications. Both magnetoresistive and pots are available with full-scale measuring ranges from about 0.1 to 20 in. (2.5 to 500 mm).

Capacitive sensors rely on the changing capacitance between two plates, one fixed, the other free to move, to detect changes in linear position of up to 0.4 in. (10 mm). They require somewhat complex support electronics and are susceptible to humidity changes and temperature extremes. They are used primarily for short-range dimensional gaging of mechanical parts, and in these applications, they have high resolution and good accuracy.

Inductive sensors measure inductance variations caused by movement of a flux-concentrating element. They are probably the most versatile of all position sensors, with a wide range of operating characteristics. Inductive sensors are contact-free, inherently robust, and have infinite resolution with high repeatability. They are often used where long-term reliability is important, particularly in harsh and hostile environments. Being contact free, or frictionless, is important when mechanical resistance cannot be tolerated.

There are three basic types of inductive sensors: LVDTs (linear-variable-differential transformers), LVRTs (linearvariablereluctance transducers), and LVITs (linear-variable-inductance transducers). And all of them need ac signal conditioning or support electronics that may either be inside the sensor or placed in an environment more amenable to electronics than where the sensor is.

LVDTs have high outputs with reasonable temperature performance. They are best for measuring movements between 0.01 and 10 in. (0.25 to 250 mm).

LVIT's are simple to build and are the most economic of the inductive sensors. However, they operate in the 2-MHz frequency range, making them sources of and susceptible to electromagnetic interference. They can be found in ranges up to 40 in. (1 m).

LVRTs are usually configured as half bridges and are more popular in Europe where they are used as spoolposition feedback sensors in servovalves and in short-range dimensionalgaging probes. They are thought of as less expensive than LVDTs but have lower outputs for a given size and, to be used properly, require support electronics. LVRTs typically have ranges under 0.4 in. (10 mm).

Magnetic sensors or Hall-Effect sensors, produce output voltages proportional to the strength of a nearby magnetic field generated by a moving magnet. They have relatively poor temperature performance, but can be effectively used for short-range position sensing where cost is most important and temperature is not an issue. Hall sensors work best when movements are less than an inch (25 mm).

Time-of-flight sensors send out some type of wave (usually sound or light) toward a moving target and measure the time it takes to receive a reflection off the target. One widely used timeof-flight sensor, called MLDT ( magnetostrictive linear displacement transducers) uses a moving magnet to create a torsional pulse in a metallic waveguide. The pulse moves down the waveguide at a known speed to a wave detector. By measuring the transit time, MLDTs produce an output proportional to the magnet's position relative to the detector. They are somewhat expensive for shorter ranges and have difficulty handling large shocks, but magnetostrictive sensors offer high resolution, high repeatability, and exhibit good temperature stability over a limited temperature range. MLDTs are particularly useful for relatively long ranges, typically from 6 to 120 in. (150 mm to 3 m) or more.

Ultrasonic and laser time-of-flight sensors-depend, of course, on timing the reflection of sound and light, respectively. Both are contact-free, but depend heavily on the target and its orientation. For ultrasonics, air temperature, humidity, and wind also can introduce errors into readings. Still, ultrasonic sensors are generally economical and reliable enough for range finding and liquid-level sensing where precision is not a key concern. Depending on the operating frequency, ultrasonic sensors measure from a few inches up to 33 ft (0.1 to 10 m).

Laser sensors offer high precision and excellent repeatability, but are relatively expensive and generally require targets with proper surface geometry and reflectivityfor optimal results. Lasers also pose a hazard to human eyesight that must be considered. Typical ranges are from fractions of an inch to hundreds of inches (a few millimeters to dozens of meters).

Pulse-encoding sensors are usually incremental sensors, although there are some absolute encoders. They optically read linear graduations on a glass scale or detect magnetic poles uniformly spaced on ferromagnetic material. They perform with high precision generating quasi-digital output, and are used primarily in motion-control systems, robots, machine tools, and other applications where their relative fragility and high cost can be accommodated. Encoders' limited frequency response, however, makes them best suited for slower dynamic applications. Measurement distances range from 6 to 60 in. (0.15 to 1.5 m).