Incremental Encoders

Interesting facts about these rotary sensors

Guide for Incremental Encoders

Incremental Encoders

Optical and magnetic encoders with incremental output

Incremental encoders are used wherever angles, rotational speeds or angular velocities have to be measured with high precision. Incremental encoders provide output signals in the form of pulses that are counted by an external evaluation unit. The sensors of the high-quality products used by MEGATRON are based on non-contact measuring principles such as optoelectronic and magnetic (Hall effect) sensor technology.

Optical incremental encoders are insensitive to external interference fields and offer the highest precision for positioning or adjustment processes. Magnetic encoders are extremely durable and very robust against vibrations. Thanks to the wide variety of designs and output options available, there is an incremental encoder best suited for almost any application.

However, special applications often require technical adaptation, which we at MEGATRON implement even for relatively small quantities. It is our claim to offer each customer individually the best functional and economical product for the application. We support you as a reliable partner from the enquiry to the start of series production up to the end of the product life cycle with high delivery reliability and quality assurance.

Guide for Incremental Encoders

What is an incremental encoder?

Incremental encoders are rotary encoders that output their output signal in the form of pulses. One pulse corresponds to one period, the increment, after which this type of encoder is named. Incremental encoders are also called rotary pulse encoders because of their characteristics as encoders for rotary motion and their signal form. The use of pulses for measurement is a fundamentally different principle than, for example, potentiometers and absolute encoders.

The most important property for determining the angular accuracy of an incremental encoder is the number of pulses generated per full revolution of the shaft at the output (pulses per revolution, ppr.). This value can be found in every data sheet of an incremental encoder.

To evaluate the signals of incremental encoders, an external evaluation unit such as a counter is always required.

  • If the angle is to be recorded, the number of pulses must be counted by an evaluation unit. If the angle is to be recorded, the information from the increments must be evaluated. If the incremental encoder delivers e.g. 360 ppr., then 1° corresponds exactly to one pulse.
  • For a measurement of angular velocity (angle change per time unit) the number of pulses per time unit is calculated

In general, there are some things to consider when evaluating the signals, see Signal Evaluation of Incremental Signals.

Operating Principles

There are various sensor principles that can be used to implement technically incremental encoders. The most widespread technology is probably optoelectronic detection, which is used in optical encoders. Another possibility are magnetic measuring principles. "Hall encoders" are also offered with incremental outputs. MEGATRON exclusively uses modern gradient based Hall sensors.

Optical incremental encoders

The figure shows the imaging measurement principle of an optical encoder in a simplified form. The two detectors A and B are illuminated with a spatial offset during the rotation of the coding wheel (black), thus generating pulses.

Optical sensor technology has several advantages that make optical encoders the most important incremental encoders. First, the fact that the measuring method of the integrated sensor elements itself already generates increments and therefore the use in incremental encoders is obvious. The overview of optical encoders can be found here.

The optical system of a modern optical incremental encoder consists of at least the following components:

  • A light emitting diode (LED), which generates light
  • A collimator, which directs the light of the LED in parallel
  • A coding wheel having alternating permeable and non-permeable (or reflective and absorptive) regions
  • The photodetector that detects the incident light of the LED and converts it into an electrical signal

Two processes have established themselves on the market: The transmissive (imaging) and the reflective (interferential) method. With the transmissive method, the coding wheel is transilluminated, whereas with the reflective method, the light beam is reflected by the surface of the coding wheel and interference effects are used.

Brief explanation of the transmissive method:
The light is collimated (parallelized) and passes through the coding wheel. The wheel ensures that light and dark areas alternate periodically on the detectors. The signal from both photodetectors is usually 90° out of phase. As a result, the direction of rotation can be determined via the sequence of the signals or their spacings in the output signal.
The structure varies depending on the requirement. Additional elements in the design of the sensor, for example, generate a reference pulse that only generates a signal on a third channel once per revolution. With this reference, the absolute angle can be calculated. I.e. the number of pulses is counted from the reference. If the counter value is lost due to an interruption in the power supply, the reference angle can be used to restore the information about the absolute angle.

Coding wheels for optical encoders

The coding wheels are made of different materials, usually metal, glass or plastic. With low-cost encoders, plastic is mainly used. Metal coding wheels are very robust. If metal is compared with glass or plastic, it is not possible to achieve such high optical resolutions in the transmissive process with identical diameters made of metal. With the reflective method, the incremental structure is printed on the coding wheel and it is possible to realise finer structures.

Hall effect incremental encoders

Hall effect encoders are also available with incremental outputs. As with optical encoders, the measuring technology is contactless and therefore hardly affected by wear (apart from the bearing). The advantages of Hall-effect incremental encoders are primarily the practically unlimited lifespan of the sensor technology (no ageing of LEDs) and the excellent shock resistance. A disadvantage can be the sensitivity to external interference fields and the fact that the signals are transmitted with a slight time delay (update rate). For an explanation of the measuring principle of Hall-effect encoders, see the Absolute Encoders guide. For a more detailed analysis of the advantages and disadvantages of various encoder technologies, see Rotary Encoder Guide.

Signal evaluation of incremental signals

Channels, resolution and direction of rotation

Incremental encoders usually have several signal outputs. If an incremental encoder outputs several signal packages, the term channel is used in this context. For example "Channel A" and "Channel B". In the literature, the term "track" is also used instead of "channel".

If the data sheet of an incremental encoder specifies the value 360 ppr and the encoder has the electrical signal outputs "A" and "B" ("Channel A" and "Channel B"), then 360 pulses per one revolution of the shaft (per 360°) are output at output "A" and another 360 ppr 90° ahead or behind the pulses of channel A are output at output "B". In total, the encoder generates 720 ppr per full shaft revolution (360°) for both channels A and B.   

The number of pulses per revolution (ppr) is also referred to as the resolution, the higher the value  ppr), the higher the angular resolution of the encoder.

The square wave signals of "Channel B" are either 90° ahead or 90° behind the signals of "Channel A". Whether the signal of "Channel A" is 90° ahead or 90° behind that of "Channel B" depends on the product and is specified in the data sheet. In most cases there is an illustration of the signal output function in connection with the indication of the direction of rotation, in which the signal sequence of the channels is shown.

In the illustration on the right, CW (clock wise) is defined as the direction of rotation. If the encoder is viewed from the front (the shaft end of the angle encoder is facing the viewer) and the shaft of the encoder is rotated clockwise, the signal output of the signal of "channel B" is delayed by 90° to the signal output of "channel A". However, if the shaft is rotated counterclockwise, the signal from "Channel B" is 90° ahead of the signal from "Channel A".

This relationship can be used in an evaluation unit to detect the direction of rotation. The number of pulses, the pulse length and the period duration of track A and track B are identical. When exchanging an encoder for another model, these characteristics are decisive, since the programming of the evaluation unit does not have to be changed if the signal sequences of the products to be exchanged are identical.

Z-track / Index Signal

Often an additional track can be selected as an option, the so-called index track or "Z-track". At signal output for track Z, an index signal in the form of a single square pulse is output for each full shaft revolution (360°).

The index signal essentially has two functions:

  • As zero point reference: After a voltage-free period, a defined zero point can be approached with the help of the index pulse.
  • As a reference pulse: Especially for encoders, which are operated at very high actuation speeds, the reference pulse has a control function as a separate counting pulse for a full actuation / rotation.

Case study:
A check is made whether between two consecutive index pulses the number of "normal" counted pulses matches the expected ones. If, for example, an angle encoder with the specification 16000 pulses/revolution is used and less than 16000 pulses are counted by the evaluation unit per full revolution, then an error has occurred.

Edge evaluation / quadrature signal

The 90° signal offset of the square wave signals of channels A and B has an advantage. For each track and signal period, a square-wave signal has one rising and one falling signal edge.
The edge sequence for tracks A and B of a signal period are as follows:

Track A rising edge (1) → after a ¼ period Track B rising edge (2) → after a ½ period Track A falling edge (3) → after a ¾ period Track B falling edge (4)

If an evaluation unit evaluates not only the rising edge of a track, but also the rising and falling edges of both tracks A and B, then the number of pulses can be quadrupled using this method. This corresponds to an increase in accuracy by a factor of four, without making any structural changes to the encoder.

If a resolution of 1024 ppr. is specified in the data sheet of the incremental encoder, then this would be four times as high with an edge evaluation, which corresponds to 4096 signals per revolution per channel. The edge evaluation just described is also called "quadrature signal with directional information". An edge evaluation can, for example, be based on the integrated circuit LS7083 offered by MEGATRON.

Maximum speed and cut-off frequency

Incremental encoders cannot be operated at arbitrarily high speeds. There are mechanical and/or electronic limitations.

The mechanical limitations can be determined from the data sheet and have the following causes:

  • Max. Speed of the shaft bearing (only applies to encoders with their own shaft bearing, see Shaft Encoder). Here the maximum permissible actuating speed is often below 10000 rpm.
  • The eccentricity (imbalance) of the mechanics. With optical encoders this is caused in particular by the imbalance of the coding wheel. However, the maximum actuating speed here can be as high as 60,000 rpm. Actuation. With magnetic kit encoders, however, this limitation does not usually exist.

The electronic limitation can be calculated: here the result of the calculation is called the "theoretical maximum possible actuating speed".

  • The reason for this lies in the cut-off frequency of the electronics. The electronics cannot process a higher frequency than the cut-off frequency. The higher the cut-off frequency and the lower the resolution of the encoder, the higher is the theoretically possible actuating speed.

The following formula can be used to calculate the theoretical maximum actuating speed from the cut-off frequency:

\(max. rpm =\frac{\text {cut-off frequency} \frac {1} {s} * 60 }{ \text {number of pulses}}\)

Two examples of how to calculate the theoretical maximum actuating speed are given below.

Example 1:
A resolution of 512 ppr is required. The cut-off frequency in the data sheet of the encoder is specified as 100 kHz. You get

\({100000 \cdot 1/s\cdot 60 \text{ s} \over 512} = 11718 \text { rpm} \)
Result: the theoretical maximum permissible actuating speed is 11718 rpm.

Example 2:
The desired resolution is 10000 ppr. The cut-off frequency is specified in the data sheet of the encoder as 100 kHz. Result: The theoretical maximum actuating speed of the actuator is 600 rpm.

\({100000 \cdot 1/s \cdot 60 \text{ s} \over 10000} = 600 \text { rpm} \)

A comparison between the maximum theoretical and mechanically permissible actuation speed shows which one counts for the application: the lower of the two values is relevant!

Tolerances and deviations of optical incremental encoders

No incremental encoder delivers perfect signals. For optical incremental encoders, the following describes the uncertainties or tolerances that must be observed for the signals from these encoders. The optical system includes the encoder wheel itself and the encoder module or assembly containing the LED and the photodetector. All elements interact to produce a certain deviation from the ideal, rectangular signal shape and the ideal position of the edges. These tolerance relationships are often described in the data sheet of an optical incremental encoder and help the user to make a more precise analysis of the measurement data.
In most cases, the signals of channels A, B and, if necessary, Z are represented as an image. With the aid of the adjacent figure, the relationsships are then explained using examples.

The symbols have the following meaning:
C corresponds to a period
P stands for a ½ signal period
S for ¼ Signal period
Ф is the phase reference between channels A and B

In the ideal case holds C = 2 * P = 4 * S = S1 + S2 + S3 + S4.

Example for description of the tolerance field of a quarter period duration

One increment and thus one period duration ideally consists of four equidistant signal components (C/4). Since in practice a signal period is not divided into four equal parts, the possible ratio and thus the tolerance band of the four parts of a signal period (T) to each other is described. The following term describes that a quarter of the signal period can vary by one twelfth of the signal period:

\(S1,S2,S3,S4 = \frac {C} {4} \pm \frac {C} {12}\)

Example for description of the tolerance field of a half period duration

One increment and thus one signal period ideally consist of two equidistant signal components (C/2). Since a signal period does not always consist exactly of two wave pairs of equal length, the possible ratio of both wave pairs of a signal period (T) to each other is described. The following term describes that half a period, or half a signal period, or half a wavelength can vary by plus-minus one twelfth of the ideal.

\(P = \frac {C} {2} \pm \frac {C} {12}\)

Description of the possible phase shift between channel A and B

Ideally, the phase shift between channel A and B is exactly 90° (ninety degrees). The 90° are shown in the relationship C/4. So a quarter of a signal period corresponds to 90°. The error in this case can be ± C/24, i.e. plus-minus one twenty-fourth. One twenty-fourth corresponds to 360°/24, which corresponds to a possible phase error of plus-minus 15°. Thus, the relationship of the increments between channels A and B can be in a range of 90° ±15° and the phase reference between channels A and B can be in a range of 75°...105°.

\(Ф = \frac {C} {4} \pm \frac {C} {24}\)

Description of the tolerance band of the index pulse length (channel Z)

The index pulse is output once every 360° if the shaft is continuously actuated in one direction. One period duration corresponds to C. The representation C/4 means that the index pulse ideally ¼ corresponds to the length of one signal period. The pulse width of the index pulse can deviate from the ideal, i.e. the length of a quarter signal period (=C/4), by plus-minus one twelfth of a signal period.

This means that the pulse width of the index pulse can vary between 1/3 (=C/3) and 1/6 (=C/6) of a signal period.

\(Po = \frac {C} {4} \pm \frac {C} {12}\)

Sine/Cosine Interpolation

The more pulses per revolution that are realized with an optical encoder, the smaller the line width of the increments on the coding wheel. However, the optical system of an angle encoder is only able to detect increments up to a certain bar width. For example, a coding wheel with a diameter of 10 mm cannot have 10,000 lines due to its small size. If incremental encoders with a small housing diameter and high resolution are to be implemented, this is often done on the basis of sine/cosine interpolation.

In this method, the optical system of the encoder is not used as in a conventional optical incremental encoder, so that there are abrupt changes of state between transmission and transmission interruption, respectively reflection, reflection interruption. Instead, the transition between no and maximum transmission or reflection is as seamless as possible. The stepless transition leads to a sinusoidal function of the signal. To produce a second channel, which generates a cosine signal, another LED and a phototransistor are required. The sine and cosine signals are then digitized. Usually a continuous sampling rate is used here.

When using an encoding wheel from which 8 sine periods are obtained, this corresponds to a resolution of 3 bits. If, however, this sinusoidal signal is sampled with 10 bits, this results in a (digitization) resolution of 213 bits, which corresponds to a resolution of 8192 ppr. The advantage of the principle is therefore obvious.

Also, available are optical and magnetic encoders with analogue output, which provide sine and cosine analogue signals. With the help of such an encoder, subsequent interpolation is possible.

Output interfaces

Incremental signals are particularly well suited to work with digital circuits due to their characteristics (high-low, on-off, Boolean logic). Many incremental encoder series therefore offer interfaces that allow easy integration into such circuit networks:

  • OC (Open Collector)
  • Standard voltage output or TTL (transistor transistor logic)
  • PP (Push Pull)

Open Collector (OC) Output

The open collector circuit is an obvious standard for output circuits for incremental signals, and as a major advantage it allows the output to be connected to a different voltage level defined by the application. This is possible because no pull-up resistor is integrated in the encoder and the collector is led out of the housing (open collector). The transistor thus functions as a switch.

The following example is applicable for a bipolar Si-NPN transistor:

High level at signal output:

  • At low level (<0.7 V) at the base of the transistor, the transistor blocks and the supply voltage (VSUP) is applied to the collector.

Low level at the signal output:

  • If the base of the transistor is high (>0.7 V), the voltage at the collector (VSUP) is pulled to ground.

With the open collector circuit it is usually necessary to place a pull-up resistor between the supply voltage and the signal outputs A, B and Z of the encoder (collector). This ensures that the levels can be detected by the evaluation unit as low and high levels. A typical value for a pull-up resistor can be 4.7 kOhm. The maximum collector voltage depends on the transistor used and is usually specified in the data sheet of the encoder. Since it exceeds 50 V, incremental signals with very high signal levels can be transmitted over long distances. Due to the variability of the collector voltage level level conversion is also possible.

TTL output

The TTL output is often simply referred to as voltage output. The difference to the open collector output is that the required pull-up resistors are already integrated in the encoder housing, and the levels are thus fixed. A variable level conversion as with the open collector circuit is therefore not possible.

These levels are for standard TTL logic:
< 0.4 V for the low level
> 2.4 V for the high level

Push / Pull output

The push / pull output circuit is based on a complementary transistor pair (n-channel and p-channel). It alternately blocks one of the two transistors. 

During the high level of the output signal it is at the level of VSUP and in the low state it is approximately at ground. The advantage of a push/pull circuit is that no additional pull-up or pull-down resistors are required. If no level conversion is required, encoders with push-pull output circuitry can be used as a universal replacement for open collector and TTL/voltage outputs.

Low level at the transistor inputs: NPN disables and PNP opens
High level is VSUP

High level at the transistor inputs: NPN opens and PNP disables
Level Low approximately to ground

33 results found
Ergebnisse anzeigen
My products