The presence or position of an object can be determined using a variety of sensing techniques. These generally fit in two categories: single point sensing and continuous sensing.
Single Point Sensing
Single point sensing provides a yes/no indication of an object’s position. The simplest type is a switch activated by an object’s movement. A common example is a microswitch with a lever arm. A slight movement of the arm closes (or opens) the switch. Microswitches are often used as limit switches, to prevent overtravel of a mechanism. Physical contact with the object being sensed is required, which makes this method unacceptable for many applications.
A simple non-contact position detector uses a reed switch. This switch activates in the presence of a magnetic field. A common application is to detect an open door or window in a home security system. The photo shows a reed switch (with three screw terminals) and the activating magnet. The switch is mounted to the frame and the magnet is attached to the door. Another popular use for reed switches is in float-type level detectors. A reed switch mounted in the tube is activated as the float, with magnet inside, rises to the preset level. This photo shows a level switch with two floats for high/low level detection.
In many applications, attaching a magnet to the moving object is not possible. For these situations, inductive, capacitive and optical techniques are available. All have electronic circuitry and require external power. An inductive proximity sensor (left) uses a coil to detect the eddy current from a conductive metal object. Typical detection range is 1 to 5 mm. The voltage output is either pull-up (to the +supply voltage), or pull-down (to the -supply voltage). While microswitches and reed switches can carry currents of 1A or more, inductive sensors are usually limited to <100mA. Their maximum operating rate is typically >500Hz, which is considerable higher than mechanical switches. Capacitive sensors use an electric field to detect metal objects. Their housings are similar to the inductive sensors. However, they are more sensitive to dirt and are less popular as single-point sensors in industrial applications.
For applications requiring a greater detection range, optical (photoelectric) sensors are a popular choice. These typically use a light-emitting diode (LED) or laser as the source and a phototransistor as the detector. The beam may be visible (typically red) or infra-red. The detector can be in the same housing as the source or in a separate housing. This photo shows a combined housing (right). There are two operating modes with this style. Depending on the object’s shape, color and surface, it may reflect light back to the detector. In another arrangement, the object blocks the light returned to the detector from a reflector located behind it. If the source and detector are in separate housings, the detector can be placed behind the object, in place of the reflector. A separate detector can also be positioned to capture a beam reflected off the target object at an angle. Optical sensors have detection ranges from <10cm to >10m and operating rates up to 500Hz. Some units use a modulated beam to reduce sensitivity to ambient light or a polarized beam to reduce false triggers from shiny objects. Optical and other externally powered sensors may include additional features, such as time delay, sensitivity adjustment and activation indicator.
Continuous Sensing
Continuous sensors measure the position or displacement of an object. Rather than an on/off output, they provide a continuous (typically analog) output signal. The linear potentiometer (top) is a low-cost transducer. It produces an analog electrical signal directly proportional to the position of the 
wiper as it slides along a resistive element. A common application is in a light dimmer, where a person’s hand is the moving mechanism. Resolution and linearity depend on the resistor composition. Wiper contact wear limits operating life.
Non-contact sensing techniques provide extremely long life. A linear variable differential transformer (LVDT) detects the position of a ferromagnetic core as it slides inside a primary and two secondary coils. A connecting rod attaches the core to the moving object. Units with 1mm to 1m travel are available. This high precision sensor has a wide operating temperature range and can detect minute changes in position. However, the LVDT requires specialized ac signal conditioning circuitry, which adds complexity and can be difficult to implement in an industrial setting. A linear variable differential transducer (DC LVDT) has oscillator, demodulator and amplifier circuits packaged inside the sensor body. It operates off DC power (typically 10-28V) and produces a scaled DC volt or mA output. This versatile sensor can be also used to measure other mechanical parameters. For example, with a spring-loaded rod in contact with a moving object, it will measure flatness or runout.
Similar sensing techniques can be used to measure rotation. When connected to a DC voltage source, a circular potentiometer easily converts shaft position into a DC voltage level. Resolution and accuracy are determined by the pot specifications. A large diameter pot, such as a Helipot, has a large resistive element the gives it high linearity and high resolution. Most pots have stops that limit rotation to <300°. Some without end stops allow continuous rotation. With the addition of a suitable mechanism, these pots can also sense linear motion. A common example is gasoline level in a car’s fuel tank. A float and lever assembly rotates the pot as the fuel level changes.
The rotary variable differential transducer (RVDT) uses a non-contact method to detect rotation. This sensor (middle) has a ferromagnetic core that rotates between primary and secondary coils. DC-powered RVDTs are limited to a sensing range less than ±75° but can accurately measure <0.001°. RVDT applications are diverse. One model of electronic viscometer uses an RVDT as the sensing element.
Rotary encoders (right) track shaft position. Many have a pulse output that can be read by a counter or rate meter. Some have a digital output that directly interfaces with a controller or computer. An absolute encoder reports the shaft position in relation to a fixed reference point. An incremental encoder tracks the position vs a relative (variable) starting point. Either type can be used on a shaft that rotates only a few degrees or turns at thousands of RPM.
A mechanism is sometimes added to one of these basic sensing techniques to create another type of sensor. A string potentiometer (aka cable-actuated position sensor) can measure large movements along one axis. This drawing shows the working principle of a Tyco string pot. It has a flexible measuring cable wrapped around a spring-loaded spool. The free end of the cable is attached to the moving 
object. A tension spring keeps the cable taut as it is pulled out and retracted. A rotary sensor attached to the spool tracks the position of the cable. Internal electronics convert this into a DCV, mA or digital output signal. Measurement ranges less than 1 to greater than 30 meters are available.
This article described only the more popular types of position sensors used in industrial applications. Even in these categories, a wide variety of products are available. Most have outputs that can be easily read by analog, digital and bargraph meters. Weschler lists electronic sensors in the Position Sensor and Level Sensor categories. Tachometer Measurements has additional information on optical, inductive and Hall sensing techniques for rotation detection.