Wireless II: Wireless in Action

As wireless technology becomes more sophisticated, it is increasingly popping up in industrial applications. Although it’s widely used in the process industry, its uptake in discrete processing and the motion control sector has been slower. In part, the issue has been signal quality; as we detailed in Part I, current wireless systems are not deterministic enough and do not operate with low enough jitter for highly coordinated motion applications such as packaging or web processing.

For less demanding applications, however - for example, passing start/stop commands, speed set points, downloading parameters, or interrogating sensors - wireless can be a useful solution. 

“Today we see wireless being used regularly in applications like automated guided vehicles, overhead cranes, any device where you have a controller somewhere stable and you have other controllers or devices you need to talk to that are moving,” says Jeremy Bryant, industrial communications manager at Siemens Energy & Automation Inc. (Norcross, Georgia).  “The reliable update limit of wireless today is 16 ms.  Start/stop signals, run commands, all these things can be handled within that limitation.”

There are wireless motion applications outside of the industrial sphere, as well.  At Preston Cinema System (Santa Monica, California), a servo motor-driven lens adjustment system allows camera assistants to fine tune focus, iris, and zoom with a flick of a slider or a joystick on a handheld unit.  The systems use wireless to send not only start/stop commands but real-time control data. Yes, there are challenges to wireless motion control, says company president and founder Howard Preston, but it’s quite feasible with the proper design.

“Most people are using off-the-shelf wireless systems and they’re not really designed for real-time control,” he says.  “We can't afford to lose data during a transmission, so maintaining a robust wireless connection is extremely critical.  We actually developed our own wireless transmission system and network protocol so that we could achieve high reliability.”

The Preston system operates at 2.4 GHz, a band that is already crowded by Bluetooth and WiFi devices.  To avoid having their signals swamped by mobile phone and laptop communications, the Preston team chose a direct-sequence spread-spectrum approach. “It’s inherently robust against reflections, it had a reasonably high process gain, and supports real-time control applications,” Preston adds.

The result has been an effective, high-performance system that dramatically changes how film cameras are used.  In 2007, the Academy of Motion Picture Arts & Sciences recognized the work of Preston and his colleague, Mirko Kovacevic, with a Science and Technology Oscar for contributions to the filmmaking industry.

Making sense
Another application for wireless in motion systems is data communications.  Wireless systems can interrogate sensors and components for everything from fault history to the output of sensors for temperature, pressure, and vibration.

“It’s a natural evolution,” says Bryant. “First, systems started leveraging Ethernet so you could plug in anywhere on the network to get data and information.  Now, customers want to do that without even plugging in, so they’re putting wireless on a machine for diagnostics and troubleshooting.  It's considered a non-critical but very valuable maintenance tool.”

At Honeywell Sensing & Control, engineers have used wireless sensor communications to detect torque in test stand applications, for example.  “There are big benefits for any sensing activity where the sensor is moving or rotating and it’s just not practical to communicate with it using wires,” says Frank Turnbull, chief engineer at Honeywell Sensing & Control.  Torque sensing is one example, monitoring bearings with embedded sensors is another. 

Go too far with it, though, and the issue of speed raises its ugly head again.  “The problem is that when you start to bus these things into networks, the response time drops because you have to be able to interrogate multiple sensors with one system at the same time,” says Turnbull.  “In normal terms, you really have a fairly slow rate of response.  To be able to use wireless for control and rotating equipment, you’d need to go point-to-point, and that would just be a wire replacement.”

Connecting a large number of sensors over point-to-point wireless connections is certainly more expensive and labor intensive than connecting via a wireless network, but even then the cost and maintenance benefits of replacing cabling with wireless still make it worthwhile for some applications.  “I’ve done a multiple-axis wireless system that involved gathering data and integrating with [the customer’s] data management system,” says Melvin Foo, principal engineer at system integrator Kazio Networks (Leesport, Pennsylvania).  “It had a controller and smaller subnodes that created a centralized coordinator. It was more for I/O; the main coordinator actually controlled the subnodes at a lower cost per node than a wireline system.”

Wireless in motion
Although the speed and jitter issues remain, in certain applications, wireless can perform sufficiently well to hand down motion commands.  Consider the life sciences market, for example, which tends to have tabletop machines for applications like DNA testing and drug discovery.  Such machines feature 10 to 20 axes of motion, each of which requires motors and drives, as well as the associated cabling.

“There’s a need to reduce cabling costs and the pain that goes along with these cabling issues,” says John Guite, business unit manager, Parker Hannifin Corp., Electromechanical Automation (Rohnert Park, California).  At the same time, the motion requirements of such biotech machines are typically modest, which Guite says makes them a good target for wireless solutions.  “What we propose is not to eliminate the PC-to-control connection but to go the next step and eliminate the wires between the controller and multiple axes of drives.”

Instead of running cabling from controller to 16 different drives, such a life sciences system would pass the commands wirelessly.  “By going to a wireless motion control solution, a customer could realize savings on the order of 15 to 25% of total machine cost of hardware and reduce their footprint by similar percentages,” says Guite.  “On top of that you have savings in labor costs.”

With the ability to control their own trajectories and feature their own homes and limits, today’s smart drives provide a semi-distributed control architecture that would help ensure machine performance even in the case of a network outage.  A regular status check of the system to check availability could flag failures, with users determining frequency, fault conditions, and response - for example, whether to shut down the drives, raise alarms, or just classify it as a soft error.  All of these technologies can help provide sufficient reliability - for the right application.

“This would not be a competitive motion network to the likes of Powerlink or SERCOS,” says Guite.  “If you have a web process or a following application, you have multiple axes of motion that have to be synchronized within 500 ns.  You really cannot tolerate any error. This is not the network that we would put into those types of applications; we would not put it into a packaging machine, for example.”
 
The life sciences market tends to have much less demanding requirements.  Considering a sampling application, in which a sample in a well of a DNA plate needs to be aspirated and then dispensed into another location.  “The motion requirements just simply aren't that high,” says Guite.  “They can afford to make a move on an X trajectory and a Y trajectory.  They don't need to be moving it in an arc or a 3-D trajectory or anything like that.”

So there is a niche for wireless motion in scientific applications, but is it ever likely to take over the plant floor?  “It's a case of development of the technology to the point where you have 100% confidence in it,” says Turnbull.  “In my experience of wireless so far, those who’ve got more to win than lose will adopt early and they'll convince others that maybe would hang on a bit longer.”

The adoption of wireless motion is likely to follow that of industrial Ethernet.  Yes, there are engineering problems, but they are not fundamental physical barriers.  Judging by the speed of technical advances, it seems reasonable to think that in five to ten years, wireless will surmount its current obstacles and find broad use throughout motion applications.

“We may not be there yet,” says Foo, “but it’s coming real fast.”