what you know about rotary encoders

[URL="http://en.wikipedia.org/wiki/Image:Rotatory_EnDat_Encoder.jpg"][/URL] [URL="http://en.wikipedia.org/wiki/Image:Rotatory_EnDat_Encoder.jpg"][/URL]

[LEFT]The absolute digital type produces a unique digital code for each distinct angle of the shaft.
A ****l sheet cut into a complex pattern is affixed to an insulating disc, which is rigidly fixed to the shaft. A row of sliding contacts is fixed to a stationary ****** so that each contact wipes against the ****l sheet at a different distance from the shaft. As the disc rotates with the shaft, some of the contacts touch ****l, while others fall in the gaps where the ****l has been cut out. The ****l sheet is connected to a source of electric current[/URL], and each contact is connected to a separate electrical sensor. The ****l pattern is designed so that each possible position of the axle creates a uniquebinary code[/URL] in which some of the contacts are connected to the current source (i.e. switched on) and others are not (i.e. switched off).
This code can be read by a controlling device, such as a microprocessor[/URL], to determine the angle of the shaft.
The absolute analog type produces a unique dual analog code that can be translated into an absolute angle of the shaft (by using a special algorithm)..

[URL="http://en.wikipedia.org/wiki/Image:Encoder_disc_%283-Bit_binary%29.svg"][/URL][URL="http://en.wikipedia.org/wiki/Image:Encoder_disc_%283-Bit_binary%29.svg"][/URL]
Rotary encoder for angle-measuring devices marked in 3-bit binary. The inner ring corresponds to Contact 1 in the table. Black sectors are "on". Zero degrees is on the right-hand side, with angle increasing anticlockwise.


An example of a binary code, in an extremely simplified encoder with only three contacts, is shown below.
Standard Binary EncodingSectorContact 1Contact 2Contact 3Angle1offoffoff0° to 45°2offoffon45° to 90°3offonoff90° to 135°4offonon135° to 180°5onoffoff180° to 225°6onoffon225° to 270°7ononoff270° to 315°8ononon315° to 360°
In general, where there are n contacts, the number of distinct positions of the shaft is 2n. In this example, n is 3, so there are 2³ or 8 positions.
In the above example, the contacts produce a standard binary count as the disc rotates. However, this has the drawback that if the disc stops between two adjacent sectors, or the contacts are not perfectly aligned, it can be impossible to determine the angle of the shaft. To illustrate this problem, consider what happens when the shaft angle changes from 179.9° to 180.1° (from sector 4 to sector 5). At some instant, according to the above table, the contact pattern will change from off-on-on to on-off-off. However, this is not what happens in reality. In a practical device, the contacts are never perfectly aligned, and so each one will switch at a different moment. If contact 1 switches first, followed by contact 3 and then contact 2, for example, the actual sequence of codes will be
off-on-on (starting position)
on-on-on (first, contact 1 switches on)
on-on-off (next, contact 3 switches off)
on-off-off (finally, contact 2 switches off)
Now look at the sectors corresponding to these codes in the table. In order, they are 4, 8, 7 and then 5. So, from the sequence of codes produced, the shaft appears to have jumped from sector 4 to sector 8, then gone backwards to sector 7, then backwards again to sector 5, which is where we expected to find it. In many situations, this behaviour is undesirable and could cause the system to fail. For example, if the encoder were used in a robot arm, the controller would think that the arm was in the wrong position, and try to correct the error by turning it through 180°, perhaps causing damage to the arm.

[URL="http://en.wikipedia.org/wiki/Image:Encoder_Disc_%283-Bit%29.svg"][/URL][URL="http://en.wikipedia.org/wiki/Image:Encoder_Disc_%283-Bit%29.svg"][/URL]
Rotary encoder for angle-measuring devices marked in 3-bit binary-reflected Gray code (BRGC). The inner ring corresponds to Contact 1 in the table. Black sectors are "on". Zero degrees is on the right-hand side, with angle increasing anticlockwise.


To avoid the above problem, [URL="http://en.wikipedia.org/wiki/Gray_code"]Gray encoding[/URL] is used. This is a system of binary counting in which two adjacent codes differ in only one position. For the three-contact example given above, the Gray-coded version would be as follows.
Gray CodingSectorContact 1Contact 2Contact 3Angle1offoffoff0° to 45°2offoffon45° to 90°3offonon90° to 135°4offonoff135° to 180°5ononoff180° to 225°6ononon225° to 270°7onoffon270° to 315°8onoffoff315° to 360°
In this example, the transition from sector 4 to sector 5, like all other transitions, involves only one of the contacts changing its state from on to off or vice versa. This means that the sequence of incorrect codes shown in the previous illustration cannot happen here.

In commercial absolute encoders there are several formats for transmission of absolute encoder data, including parallel binary, SSI, ISI, Profibus, CAN DeviceNet, CANopen, Endat and Hiperface, depending on the manufacturer of the device

An incremental rotary encoder, also known as a quadrature encoder or a relative rotary encoder, has two outputs called quadrature outputs. They can be either mechanical or optical. In the optical type there are two gray coded tracks, while the mechanical type has two contacts that are actuated by cams on the rotating shaft. The mechanical types requires debouncing and are typically used as digital potentiometers on equipment including consumer devices. Most modern home and car stereos use mechanical rotary encoders for volume. Due to the fact the mechanical switches require debouncing, the mechanical type are limited in the rotational speeds they can handle. The incremental rotary encoder is the most widely used of all rotary encoders due to its low cost: only two sensors are required.
The fact that incremental encoders use only two sensors does not compromise their accuracy. One can find in the market incremental encoders with up to 10,000 counts per revolution, or more.
There can be an optional third output: reference, which happens once every turn. This is used when there is the need of an absolute reference, such as positioning systems.
The optical type is used when higher RPM's are encountered or a higher degree of precision is required.
Incremental encoders are used to track motion and can be used to determine position and velocity. This can be either linear or rotary motion. Because the direction can be determined, very accurate measurements can be made.
They employ two outputs called A & B which are called quadrature outputs as they are 90 degrees out of phase.
The state diagram:
Gray Coding for Clockwise RotationPhaseAB100201311410
Gray Coding for Counter Clockwise RotationPhaseAB110211301400


[URL="http://en.wikipedia.org/wiki/Image:Quadrature_Diagram.svg"][/URL][URL="http://en.wikipedia.org/wiki/Image:Quadrature_Diagram.svg"][/URL]
Two square waves in quadrature.


The two output wave forms are 90 degrees out of phase, which is all that the quadrature term means. These signals are decoded to produce a count up pulse or a count down pulse. For decoding in software, the A & B outputs are read by software, either via an interrupt on any edge or polling, and the above table is used to decode the direction. For example if the last value was 00 and the current value is 01, the device has moved one half step in the clockwise direction. The mechanical types would be debounced first by requiring that the same (valid) value be read a certain number of times before recognizing a state change.
If the encoder is turning too fast, an invalid transition may occur, such as 00->11. There is no way to know which way the encoder turned; if it was 00->01->11, or 00->10->11.
If the encoder is turning even faster, a backward count may occur. Example: consider the 00->01->11->10 transition (3 steps forward). If the encoder is turning too fast, the system might read only the 00 and then the 10, which yields a 00->10 transition (1 step backward).
Rotary sensors with a single output are not encoders and cannot sense direction, but can sense RPM. They are thus called [URL="http://en.wikipedia.org/wiki/Tachometer"]tachometer[/URL] sensors.
This same principle is used in old ball mice to track whether the mouse is moving to the right/left or forward/backward.
A variation on the Incremental encoder is the Sinewave Encoder. Instead of produce two quadrature square waves, the outputs are quadrature sine waves (a Sine and a Cosine). By performing an Atan function, arbitrary levels of resolution can be achieved.

If the manufacturer moves a contact to a different angular position (but at the same distance from the center shaft), then the corresponding "ring pattern" needs to be rotated the same angle to give the same output. If the most significant bit (the inner ring in Figure 1) is rotated enough, it exactly matches the next ring out. Since both rings are then identical, the inner ring can be omitted, and the sensor for that ring moved to the remaining, identical ring (but offset at that angle from the other sensor on that ring). Those two sensors on a single ring make a quadrature encoder.
For many years, [URL="http://www.mathematik.uni-bielefeld.de/~sillke/PROBLEMS/gray"]Torsten Sillke[/URL] and other mathematicians believed that it was impossible to encode position on a single track so that consecutive positions differed at only a single sensor, except for the two-sensor, one-track quadrature encoder. However, in 1996 Hiltgen, Paterson and Brandestini published a paper showing it was possible, with several examples. See [URL="http://en.wikipedia.org/wiki/Gray_code"]Gray code[/URL] for details.

[URL="http://en.wikipedia.org/wiki/Image:Basic-Rotary.jpg"][/URL] [URL="http://en.wikipedia.org/wiki/Image:Basic-Rotary.jpg"][/URL]
Optical tachometer (no quadrature output)


[URL="http://en.wikipedia.org/wiki/Image:Odometergearcropped.jpg"][/URL] [URL="http://en.wikipedia.org/wiki/Image:Odometergearcropped.jpg"][/URL]
Hall-effect quadrature encoder, sensing gear teeth on the driveshaft of a robot vehicle.


Encoders may be implemented using a variety of technologies:

The quadrature variant is most prevalent in industrial use even though more sophisticated and tougher absolute transducers have been on the market for some time. Most applications are satisfied with an initial homing function on power up to achieve the desired absolute positioning. The simple wiring associated with quadrature encoders along with its relative ruggedness are likely the primary reasons for its success. And as such, it has become notably cheaper than all other precision options. The only serious contender I've noticed is the resolver and its linear variants. This will be due to the resolver being capable of withstanding very harsh environments like operating in liquids.
Another trend that may be happening is modern transducers are designed to output quadrature signalling while internally they are not actually quadrature encoders at all - The power of mass production and digitally reconfigurable encoding.
During the 1980's and 1990's an assembly known as the mouse[/URL] with two rotary quadrature encoders inside was massively popular as a partner to the rising desktop[/URL] phenomenon. Starting out targeted as a workstation[/URL] but saw much bigger acceptance as a home computing[/URL] and gaming device then later absorbed by the clone PC[/URL] market. The rotary encoder saw a rapid decline in this role as the far more sophisticated "optical" mouse arrived on the scene in the early 2000's(?). As a side note, these "opticals" also produces the quadrature signalling, even though the massive PC market has always used the serial command port for gathering the deltas.
With the rise of digital in control systems a final popular use of the quadrature variant as a replacement for the humble potentiometer on the control panel has flourished. Throw out the absolute markings and have a display showing the level and there is no need for knowing the absolute position. At low resolution, it's super cheap, as rugged as desired and completely immune to adding noise.