New products are emerging that could drive his costs for wireless networking to less than $5 per node. As well, this new technology promises robust connections in a self-configuring, self-healing mesh that minimizes power requirements.
Wireless networks employ radio frequency (RF) links in one of three main configurations. In a point-to-point network, the simplest configuration, a dedicated connection is established between two radios that are in direct communication with each other. The second configuration, point-to-multipoint, uses a shared link between a base station and multiple client radios, or nodes. Each point must be within range and have a clear line of sight back to the base station.
Both point-to-point and point-to-multipoint rely on detailed site planning and antenna positioning. Says Tyco’s McCoy, “Traditional wireless networks using point-to-point or point-to-multipoint are dependent on a site survey to establish lines of sight, align antennae and build repeater stations for obstructed areas. After those networks are installed and commissioned, they have a relatively short life before some change in the plant destroys your scheme. For instance, if a new tank is built, you have to go out and re-commission the system.”
Network, heal thyself
The third, and newest wireless configuration is the multipoint-to-multipoint, or mesh network. The wireless mesh topology is similar to the wired Internet in its ability to connect multiple nodes and provide multiple redundant communication paths. Nodes do not have to communicate back to a single base station, or master. Rather, nodes in the mesh can send and receive messages, and act as routers to relay packets of data to other nodes. Each node, then, becomes part of the network infrastructure by sending data over multiple hops. If one node should fail, network traffic is rerouted over an alternate path.
This is extremely important in industrial environments, where networks are subject to electrical noise and radio interference. “Mesh networking is inherently more reliable than other wireless topologies,” says Andy Wheeler, founder and vice president of engineering at Ember Corp. “In manufacturing environments, things like motors turning on and off cause the RF link to vary substantially from one minute to the next. The algorithms we’ve developed allow nodes to reconfigure and reroute data, resulting in a self-organizing, self-healing network.” (See Mesh Network Topology diagram, p. 43.)
Ember, based in Boston, is one of several suppliers that have developed embedded wireless mesh networks for sensing and control. Current products include radio modules that give processors the ability to communicate in a wireless spread-spectrum network as both endpoints and embedded routers. An evaluation kit includes the EmberNet software stack, sample hardware and development tools to support multiple radios and microcontrollers.
Ultra-small size and ultra-low power characterize i-Bean mesh networks from Cambridge, Mass.-based Millennial Net. Slightly bigger than a dime, the i-Bean is claimed to be the world’s smallest wireless sensor networking device. It combines analog and digital interfaces, a radio transceiver, and a microcontroller for sensor signal processing, control and networking. Millennial Net uses radio technologies that include low-power, narrowband solutions as well as IEEE 802.15.4 wireless personal area network (WPAN) components. Like Ember, Millennial Net’s target markets include industrial automation, building automation, supply chain management, metering, security and environmental monitoring.
Millennial Net’s topology uses i-Bean endpoints to interface to sensors and actuators, routers to extend network coverage areas, and gateways to interface to a host, local area network or the Internet. “Mesh networks provide additional eyes and ears into the process,” says Millennial Net president Todd Riedel. “i-Beans are used to extend the network into areas it couldn’t go before.”
Mesh Benefits
Wireless mesh networks present several benefits to industrial users. Adrian Tuck, Ember executive vice president, summarizes three. “Where wiring is a significant expense, wireless mesh will help drive costs out of the installation. The second benefit of mesh is its ability to assist manufacturers in complying with government regulations. Mesh networks address the need to measure and monitor more things, such as environmental variables and security factors.”
The third benefit of mesh networks is in the diagnostic monitoring of intelligent devices. This monitoring network can reside outside of the existing control network and is used to notify the control system if a device failure occurs. Explains Tuck, “As manufacturers try to squeeze more years out of existing plants, they can retrofit with intelligent maintenance systems that monitor and improve overall performance.”
It was the drive to reduce wiring installation costs that first led Tyco’s McCoy to consider mesh networking. Tyco manufactures heaters for industrial and commercial applications that are used primarily to keep pipes and tanks from freezing, or to hold their contents within a specific temperature range. The electrical heaters may run several hundred feet in length and use 120 V to 600 V power.
To monitor and maintain the specified temperature, Tyco Thermal Controls, based in Menlo Park, Calif., provides a thermostat control system for the Tyco heaters. The system measures temperature, compares it to a set point and switches the heater on and off as necessary. Resistance temperature detectors (RTDs) are mounted on the pipes and tanks for temperature measurement. Because distances often exceed several hundred feet, Tyco uses a three-wire RTD, in which two wires measure change in resistance while the third wire corrects for the resistance of the lead wires.
“This is where Ember’s mesh network fits in,” says McCoy. Ember wireless nodes connect the RTD circuitry to the controller. The RTD probe, which is surface-mounted on the pipe or tank, is connected by a short lead—typically 12 in. or less—to the mesh network node. (See photo of node cross-section below.) An antenna in the node box transmits data from the RTD through the mesh network to the control room. “We’ve eliminated the cost of running those three-wire leaders from the control room all the way out to each RTD circuit. In a large plant there may be thousands of RTDs, each requiring 20 to 200 feet of wire. It’s not the wire that’s so costly, it’s the installation.” McCoy estimates the cost to install wiring to be $7 to $25 per foot. In a large facility—for instance, an oil recovery application—9,000 individual RTD circuits may be needed. Savings to eliminate wiring can easily range from $1 million to $10 million.
McCoy has been working with Ember development tools for about 18 months to develop a field-ready product. One of the key issues is power supply and management. In initial field trials, which began in June and July, the Tyco wireless mesh system is using ground power. In the future, McCoy is looking to install long-life rechargeable batteries at the nodes. “The mesh networking concept has been very attractive to our application. We have a lot of points with good potential for a dense mesh. It looks solid for the stuff we do.”
The next step
Ember’s Tuck looks beyond reduced wiring costs to new applications. “Manufacturers started to use wireless networks because they wanted to eliminate wiring and reduce costs. Now we’re finding that some of our customers are starting to re-architect their systems to take advantage of the inherent benefits of mesh networking.”
Continues Tuck, “For example, many control systems today employ distributed controllers talking to sensors on the network. With a wireless mesh network, sensors can talk directly to each other.” And in the future, Tuck believes manufacturers may even be able to eliminate some of the controllers completely and have sensors communicating through a gateway bridge back to an Ethernet or other plant network.
See sidebar to this article: Wireless standards move markets
