April 2009

Factory Automation

No Wild West wireless

Automotive manufacturers want wireless standard that meets their special requirements, not those of a process world


  • Automakers still looking for standardization for wireless manufacturing.
  • ISA wireless standard emphasizes process, not discrete needs.
  • Factory automation group rallies for standard changes to meet discrete needs.
By Ellen Fussell Policastro

Automotive manufacturers still struggle in the midst of financial turmoil, yet they forge ahead in technical innovation and continue to sound the horn of standardization needs, especially in the field of wireless for discrete manufacturing.

In a presentation at ISA EXPO last October, Mike Read, senior technical specialist in IT at Ford Motor Company in Dearborn, Mich., voiced the struggles of automotive manufacturers in obtaining next-generation wireless requirements.


The specific requirements for discrete parts manufacturing screams for standards writers to take another look at the ISA100.11a standard, which so far is process heavy. And since the automotive industry does not yet have a wireless standard, they are putting together a set of requirements with the intent for these requirements to become part of the ISA100.11a standard.

"When we conceived ISA100 would be a family of standards, we envisioned immediately that manufacturing automation would need tighter constraints on the time issue," said Wayne Manges, the ISA100 chair and program manager at Oak Ridge National Labs. "We're glad to see that group becoming active in the ISA100 family. In fact, VW is now moving to Chatanooga, Tenn. And I'm trying to get them to join. They have plans to build a multi-billion dollar plant. And other companies are building nearby to support that plant."

In his talk, Read pointed out the major differences in today's ISA100.11a standard between process and discrete manufacturing requirements and explained how the standard could   better meet discrete manufacturers' needs.

Robots need no tangles

There are places where wireless is a requirement because the wired equivalent is not practical, such as on robots, where the end effector is continuously flexing. "On a robot, you have six axes with joints moving," Read said. "So you have the base axis 1 turning clockwise or counterclockwise. Axis 2 makes the robot bow down. Then you work your way up to the sixth axis, at the wrist. All those axes have to get wired to the end effector, and it has to go past all those axes. So every axis that flexes is flexing that wire. If you continuously flex a coat hanger it'll break. The same thing happens with copper wires."

That is the classic example, but similar applications, such as festooned cables (as with a shower curtain) mean a tool goes on a rail forward and backward. It is connected on a base, and "rather than a rail, we use a festoon, which is equivalent to a hanger on a shower curtain that droops down," Read said. "So as it pulls away from its base, the festoon cable stretches out, like a shower curtain unfolds. It's the same problem with a high-speed machine; most cables will sway and slap into each other, with a lot of flexing of cables."

Teaching, tight-spaces, quick stops

Most manufacturers want wireless I/O for equipment where wiring is either high maintenance, not physically possible, or cost prohibitive. But wireless is not intended to be the replacement for all currently wired networks.

Some typical high-speed I/O examples include robot end effector, where wiring is rerouted through the robot and is high maintenance due to continuous cable flex and tight radius joints. Other high-speed I/O examples include places where hard wiring is not an option and bus communications are high maintenance (such as carriers on electrified monorail systems or track-mounted and rotary equipment).

Wireless would also be effective for emergency stops and for wireless teach pendants. "There are cases where a red button on a machine is the emergency stop, and it's normally hard-wired to the machine-on an operator console or on a cord attached to the machine," Read said. "It would be good to have a wireless emergency stop, something the operator could take with them that's wireless and meets all the requirements for safety."

A teach pendant is a portable robot device operators use to teach the robot, "to jog each access to the robot to where you want it to go," Read said.

Wireless standard revamp?

While the ISA100 committee has not yet decided on whether or not the ISA100.11a standard will be revamped to accommodate the requirements for factory automation (including automotive and other discrete and hybrid industries), an ISA100.16 Factory Automation working group is preparing to document use cases and requirements specific to this market space, said Cliff Whitehead, manager of business development at Rockwell Automation in Mayfield Heights, Ohio, and co-chair of the ISA100.16 Factory Automation working group. "Those requirements may result in modifications to ISA100.11a in the future or may be manifest in some other form," he said.

At the end of the day, however, ISA100.11a might not be the only vehicle automotive manufacturers use to define the discrete and hybrid market wireless solution. The IEEE standard, 802.11x (and especially 802.11n as it emerges on the market) and the existing industrial Ethernet protocols (such as EtherNet/IP) are attractive to the discrete and hybrid market for quite a few use cases. "As a standards group focused on the broad range of solutions for our constituents, we are obligated to look at all available solutions on the market, as well as documenting requirements for technology solutions that may not exist today," Whitehead said.

The most obvious discrete requirement, which would mean a change in a future revision of ISA100.11a, is speed. "As we all know, discrete processes demand high-speed communications in the sub-10 millisecond range.

The current scope of ISA100.11a, as documented in the Draft 2 version of the standard, says the standard's application focus addresses performance needs for "non-critical monitoring and control applications where latencies on the order of 100 milliseconds can be tolerated, with optional behavior for shorter latency."

"The implication is that the 'optional behavior for shorter latency' is going to have to be explored and implemented for the standard to be more valuable to discrete and hybrid users," Whitehead said. While important aspects do exist in the standard today, such as coexistence and security, "mesh may not be as important in some discrete applications, and our group is exploring that angle as well," he said.


What automakers want

Read presented a matrix or chart of ISA100.11a priorities versus what automotive manufacturers are looking for. The chart explains the whole reason for a factory automation working group, pointing out the need disparity between discrete and process manufacturers.

Take a look at the "Optimized for battery life" row. For ISA 100.11a, that is high priority "because they want a wireless transmitter that'll sit on a remote reactor or tank," Read said. "It doesn't update very fast, and so an efficient battery-operated device is effective for reporting tank levels. Their rate of change is not very fast, not fast enough for what we need for discrete in most applications. We happen to have power at the discrete device we want to monitor very quickly, so battery life isn't as important."

It is the same thing in the gateway to plant network and mesh network. The priorities for these things are high for process industry, which has a low priority for discrete I/O and fast update rate. But it is the opposite for automotive requirements.

Rating performance requirements

Read's pick for the most important requirements to the automotive industry, which are different from where ISA100.11a is now, includes:

1. Performance issue: Automakers need 10 milliseconds

In the automotive industry, requirements are high for discrete I/O and fast update rate. "We need a report and input for writing to an output every 10 milliseconds. In the ISA100.11a standard, it is 10 times slower," Read said. "It's perfectly acceptable for an ISA100.11a device to update every 100 milliseconds. But ISA100.11a compliant devices today are not fast enough. We need them to update every 10 milliseconds, and they don't do that."

2. Emphasize discrete I/O

The automotive industry needs a bigger emphasis on discrete I/O because in the process industry most things are analog measurements, (of temperature, flow, and pressure), and in the discrete parts world, they are discrete on/off devices.

"In discrete manufacturing you're looking at sensors that indicate a part is present or a clamp is closed," Read said. "In our outputs, when we tell a machine to do something, the actuators we talk to in discrete are on and off. So you tell a solenoid to turn on, or you start a motor, or you turn on a light. It's more simple. In the process industry, you're reading a temperature (with an analog value); it's not 'on' or 'off,' it's 'how hot is it?' That goes to your control device where you're doing a PID loop that's now giving an analog signal back to a valve positioner to control steam or chilled water."

3. Density of devices

The automotive industry deals with a large number of wireless devices in a very small area. "So in process you've got a number of wireless devices probably at a fairly long distance away from the central control. By contrast, in discrete manufacturing, you have a local controller in a cell talking to wireless devices all within 10 meters-very close," he said.

This makes it easier for wireless because of the short range. It adds another level of complexity though because operators have multiple wireless conversations going on in some areas, so there are multiple chances for interference. "In the 801.11a area, there are ways to put 802.11 access points that are having independent conversation into close proximity, by using different channels," he said. "So we're asking in the design of wireless devices or controls that we have the ability to have many parallel conversations in the same space without interfering with each other."

Who benefits most from changes?

If the standard is in fact revised to address factory automation requirements, "the objective would be to benefit the market as a whole," Whitehead said. "End users would benefit most because they would be the recipients of common investment, development, design, and delivery by a multitude of vendors who would be serving process, discrete, and hybrid markets.

Avoid Wild West, lay down laws

Read said he believes the industry is "spending a lot of money and downtime on hard-wired devices where a simple wireless replacement would be desirable. The robot end effectors is a good example," he said. "It costs a lot of money to run wires through a robot and a lot of downtime to repair."

The direction in which the industry is headed concerns Read and the rest of the automotive industry, he said. "Today I can go out and buy discrete I/O wireless devices, but from multiple suppliers. The problem with that is they're not interoperable. Because there's not a wireless standard for the discrete industry, as the industry evolves, we don't want to see more of this diverse 'nonstandardization' so to speak.

"Right now it's like the Wild West. We want to get in with a standards body like ISA as early as possible and lay down the requirements for the discrete industry. Then, once those specifications are instituted, we can have the manufacturers develop products that meet those standards to give us the performance and interoperability we're looking for."


Ellen Fussell Policastro is the associate editor of InTech. Her e-mail is efussellpolicastro@isa.org.

Automotive research group

The U.S. Council for Automotive Research is the umbrella organization for collaborative research among Chrysler LLC, Ford Motor Company, and General Motors Corp. Founded in 1992, the goal is to further strengthen the technology base of the domestic auto industry through cooperative research and development.

Some performance requirements include: 10 milliseconds throughput time (PLC to device or device to PLC); 1.5 seconds or less power-up time; 0 to 60°C operating range; deterministic response time; high density; high EMF/RFI tolerance; and no interference with existing 802.11 b/g/a/n infrastructure.