How Encoders Make Automated Motion Safe

At its core, automation is defined by control loops, but automated machine safety is more aptly described as an arrow. And that arrow’s sharp point is often an encoder that makes it possible for the control system to know where it is and how fast it’s moving. With that knowledge, the system can not only avoid trouble, but act appropriately when circumstances bring trouble to its door.

By C.G. Masi, Control Engineering August 1, 2008

At its core, automation is defined by control loops, but automated machine safety is more aptly described as an arrow. And that arrow’s sharp point is often an encoder that makes it possible for the control system to know where it is and how fast it’s moving. With that knowledge, the system can not only avoid trouble, but act appropriately when circumstances bring trouble to its door.

Incremental encoders need only two tracks to provide displacement, speed, and direction information. To get true position, however, they maintain a running sum of displacements. Absolute encoders use much more space, but report true position at all times.

Trouble, of course, can take many forms. It is generally any event that can cause harm to the machine itself or to people or equipment around it. For moving-machinery applications, such events include intrusions into the machine’s work space, failures of machine components, or power upsets.

“The biggest machine safety issue is unannounced power outages,” says Jim Marshall, consulting design engineer for Sick Stegmann. “On startup, you need to know where axes are without physically moving them. For machine tools and tool changers that’s critical, as it is for automated warehouses.”

Encoders decoded

“What we’re concerned about is machine axis control from a position point of view as well as velocity,” says Tom Wyatt, national sales and product manager for the automation group at Heidenhain.

Of all safety sensors (e.g., light curtains, proximity switches, and E-stops), only encoders can provide forewarning of trouble ahead. All types of encoders can report both position and velocity.

Encoder systems generally have three main components:

  • A source, such an LED or permanent magnet, which mounts on the static part of the machine and provides a steady excitation.

  • The encoder mounts on the machine’s moving part and modulates the steady excitation.

  • A detector receives the modulated signal and decodes it to produce an electronic output indicating position and velocity.

The simplest example is an optical rotary encoder. “In a contained system with a bearing, you have a glass disk aligned to the bearing center, and a stator portion that holds the electronics and scans the disk as it turns, says Wyatt.

A light source, typically an LED, provides a narrow light beam aimed at the detector, which might be a photodiode. Both are rigidly mounted to the stator-side of the rotary joint’s bearing. The encoder is an opaque disk with transparent slots or windows in a circular pattern mounted on the bearing’s rotor.

As the bearing rotates, the encoder alternately allows the beam through (when a window is in position), or blocks it. The photodiode output thereby alternately goes high and low to signal position changes. The photodiode’s output then goes to an electronic circuit that decodes it into position and speed readings.

Originally, stator-side electronics were built of separate components. Sales volumes and improvements in manufacturing techniques have allowed some commercial vendors to integrate most or all of the stator-side electronics into a single application-specific integrated circuit (ASIC). “The ASIC makes the encoder more reliable,” points out John Pindroh, business development manager, BEI Duncan Electronics. “It allows for higher MTBF because of the low part count, fewer electrical connections, and fewer points of failure.”

Generally, a rotary encoder comes as a complete unit enclosed in a stator housing with a rotating shaft. The housing mounts to the bearing stator side, while the shaft couples to the bearing rotor. Commercial units typically have multiple “tracks” of modulating elements on the disk, and multiple detectors for those tracks.

“We offer encoders with complementary outputs,” says Greg DePue, product engineer, BEI Duncan Electronics. “There will be A channel, and B channel, and the complements of those. Some engineers use the complements to double check each original signal.”


There are two non-exclusive ways to classify encoders. They can be absolute or incremental, and they can be rotary or linear.

Absolute encoders provide a unique output for every position within the encoder’s range. Various coding schemes are available, such as Gray code, and pseudorandom codes. Absolute encoders’ big advantage is that it is impossible for the encoder to get “lost.” Catastrophic power loss, for example, doesn’t hamper the function of absolute encoders. As soon as power returns, they instantly know exactly where they are. Their big disadvantage is complexity.

Incremental encoders tick off pulses every time the position changes by one resolution unit. It is then up to the control electronics to count pulses to keep track of the actual position. Incremental encoders are very simple. Two bands providing regular waveforms that are identical except for a 90° phase shift are all that is needed to provide position, speed, and direction information. Incremental encoders’ big disadvantage is that anything that interferes with their keeping track of the count—such as signal interruptions during motion, power failures, glitches, noise, or intermittent electronic faults—can destroy the system’s position knowledge.

Rotary encoders, as described above, include a rotating encoder disk with the modulating elements applied in a circular pattern. The disk then rotates between the excitation source and detector fixed to the encoder housing.

Linear encoders unwrap the circular pattern and lay it out on a long tape. Because linear encoder measurement ranges may be indefinitely long, they may not be enclosed in a housing. Rather, the tape may actually be mounted to the machine’s static part, while the source and detector electronics slide back and forth along it, mounted to the machine’s moving part.

Encoders and safety

Just as a spear tip isn’t much of a weapon by itself, an encoder by itself does not make a system safe. Safety is a property of the overall machine as a system. As such, planning for machine safety is part of the machine integration process. Planning for machine safety starts with a formal risk assessment, which can be performed at any stage of the machine’s life cycle—even as a retrofit to an existing machine. Preferably, risk assessment starts in the design phase.

Risk assessments

“The risk assessment will help determine whether a simple safety system, a system requiring some kind of speed monitoring, or some other type of system would be required,” says Bill Sinner, portfolio manager for the standardized business at Rockwell Automation. “The risk assessment would determine if the information coming out of a single encoder was sufficient for protecting the operator.”

Kevin Zomchek, global product manager for safety components at Rockwell Automation, says, “In an application where speed is not monitored, such as with a safety PLC, it’s going to remove power from the machine. Encoders give us the ability to monitor if there is still motion in the machine after it has been commanded to stop and act accordingly, such as by limiting access by personnel to that area until it either detects a standstill, or below a safe speed limit as determined by the risk assessment. Being able to monitor speed allows the machine builder to increase its productivity while providing a safer application.”

Once risk assessment has identified the potential failure modes, the next step is to provide a mitigation strategy that will allow the machine to recover. For example, if the machine is an automated multi-spindle drill press, the obvious response to any safety event would be to hold the work fixture steady and retract all spindles to their mechanical stops. At that “home” position, a system using incremental linear encoders to track spindle position can recover its position memory.

If the machine is a multi-axis robot, however, or a conveyor system, making any movement without first knowing its starting point could be a recipe for disaster. In such cases, an absolute encoder on each axis is a safety must.

Another important safety principle is redundancy. Each component of a safe machine system needs a backup. That goes for encoders, too.

“A safety encoder should have two independent scanning methods inside the encoder,” says Wyatt. “You’re scanning the same disk, but you create two positions or two velocities. The controller compares them and, if there’s a difference, there’s a problem.”

A structured graduated disk rotates relative to a scanning point while photovoltaic cells convert light into electric signals. The graduated disk determines absolute position.

Do you need both channels?

In some cases, however, there may be no need for both channels to report absolute position. The system might be able to reach the typically required safety integrity level (SIL) of SIL-3 with a relative encoder backed up with an absolute encoder.

In general, there is no universal formula for how to use encoders to achieve machine safety. Proper safety system design in general, and proper use of encoders for machine safety in particular, depends on the results of a thorough risk assessment that takes into account the entire machine as well as its operating environment.

ONLINE extra

Below is a transcript of an online exchange between Control Engineering senior editor C.G. Masi and Dr. Simon Stein, Manager R&D for Safety and Connectivity at SICK Stegmann GmbH in Donaueschingen, Germany regarding the use of encoders in machine safety systems.

CGM: What aspects of machine safety involve encoders?

Dr. Stein: Encoders usually support position or speed-related safety functions of machines. A main field is safety-related motor-feedback, where encoders are used to sense the position and speed of servo drives used in safety applications.

According to the relevant standard IEC

Safe-position output of an encoder can usually be achieved by two means:

The latter approach can be more cost effective if supported by the encoders. It is useful for synchronous servo drives where safety functions like safe operating stop can be monitored both by comparing encoder position value with drive commutation information.

Other drives (asynchronous, DC) or general machinery usually must rely on the redundant sensing approach. Here, the second channel of position can not be derived from other sources and the encoder alone has to fulfill redundancy requirements.

CGM: What encoder technologies are used in machine control applications?

Dr. Stein: Apart from safety technology (number of channels, diagnostic coverage), the basic sensor technology of an encoder can be versatile. As long as the technology allows for necessary diagnostic coverage there is no obstacle from the safety point of view.

At our company, encoders with optical, magnetic, and capacitive sensing technologies are used for safety applications. This requires that all mentioned sensor technologies be able to process analog, independent sine and cosine signals related to the position value of the encoder. These two signals must follow the trigonometric identity that the squares of the sine and cosine of any angle must sum to unity, resulting in a high diagnostic coverage.

CGM: How do these encoder technologies impact machine safety strategies?

Dr. Stein: As mentioned above the sensing technology has limited impact on overall machine safety strategies. The safety architecture of the encoder however influences the safety strategy.

Where a safety application would need two standard encoders and potentially elaborate diagnostics within the PLC, a single safety encoder can suffice to fulfill the same role. Both cost and ease of implementation are affected by this choice.

CGM: What machine safety issues surround incremental encoders?

Dr. Stein: Incremental encoders can only support safety functions related to speed (including stop functions). In contrast, absolute encoders can enable position-related safety functions. All other considerations apply for both variants of encoders.

Author Information
C.G. Masi is a senior editor at Control Engineering. Contact him at