1 April 2003
Wireless ain't wireless!
By Graham Moss
What's what in wireless automation
Just as users have found that "oils ain't oils" in engine care, industrial wireless users have found a big difference in their experience. While some engineers have had good success with wireless instrumentation and automation and are already onto their third or fourth wireless project, other engineers have suffered "burnt fingers" and don't want anything more to do with the technology.
Here are some differences in wireless technology, where it should be applied, and where it shouldn't.
At the heart of all wireless devices is a radio, and an understanding of radio physics explains a lot of the differences in performance.
Most industrial wireless uses license-free bands because obtaining a license for a dedicated radio frequency (RF) channel in most areas is anywhere from difficult to nearly impossible. And the license-free bands do give high-reliability performance if used within their performance parameters.
The two common license-free bands are 900 megahertz (MHz) and 2.4 gigahertz (GHz), and both require spread spectrum for use. Spread spectrum is a catchall phrase covering several different techniques. Basically, it means that radio messages are not transmitted on a single fixed radio channel but spread over multiple channels. The group of radio channels is the radio band.
Spread spectrum is good news—it allows a large diversity of operation and allows many users to share a band reliably. The technology was originally developed for the U.S. Navy in World War II to overcome enemy interference on fixed radio channels, so it can actually give a higher degree of security than licensed channels.
Both of these radio bands fall into the line-of-sight RF spectrum, but neither requires line of sight. Confusing? Line of sight, in radio parlance, means that over long distance, radio waves will travel in a straight line rather than follow the earth's curvature or bounce off the ionosphere, as low-frequency radio waves can. But over short distances, as in plants and factories, the radio signal will penetrate walls and buildings. It'll penetrate through and around vessels and steelwork. These bands fall into the same spectrum area as cell phones. And how often do you use a cell phone when you can actually see a cell tower?
There are differences between 900-MHz and 2.4-GHz performance. The 2.4-GHz band is much wider; its channels have wider bandwidth, allowing transmission of much higher data rates. Generally, higher frequency means higher data rate, but higher frequency also means diminished radio propagation. Higher-frequency RF becomes more absorbent, with higher attenuation through air and obstacles.
In instrumentation and automation applications, it's the ability of a radio message to penetrate through a plant that's important, not how far it will travel line of sight. It's also where most users experience problems.
There's a significant difference between the penetrating performance of 900 MHz and 2.4 GHz, and to confuse matters further, different products generate different amounts of radio power. Common 2.4-GHz technologies are the 802.11 Wi-fi wireless LANs and Bluetooth. The win factor of 2.4 GHz is its high data throughput. And 2.4 GHz is license free in the main markets of the world, whereas 900 MHz is unique to North America.
This is a significant influence on wireless manufacturers. Both 802.11 and Bluetooth were born with commercial applications in mind over short-distance uncluttered radio paths. They will work in plants and factories, but they won't work far when the path becomes obstructed. In most cases, 900 MHz works where 2.4 GHz doesn't. Does the lower data rate of 900 MHz matter? No, not for instrumentation and automation, where you are transmitting a handful of tag values.
The Federal Communications Commission (FCC) allows devices on these license-free bands to generate 1 watt (W) of RF power. The FCC also allows high-gain antennas to focus the radio signal into an effective 4 W. But high-gain antennas are directional—the way a lens is used to focus a torch beam—and not particularly useful in plants and factories, where much of the RF signal is being reflected and bounced in all directions.
A 1-W, 900-MHz device will reach half a mile to 2 miles in an industrial environment. There are very few applications up to half a mile that won't work because of the high amount of reflected RF signal bouncing around the immediate area.
Some wireless device designs call for shorter distances and generate less than 100 milliwatts (mW) of radio power. Just because a product is 900 MHz doesn't mean it will reach half a mile through steel work. You need to check the power rating. The same applies to 2.4 GHz. The regulatory limit for 2.4 GHz in Europe is 100 mW, and many 2.4-GHz products will generate only this power.
The propagation difference between 100 mW and 1 W is approximately 30%. That is, if a 1-W device will reach 2,000 feet over a particular path, expect a 100-mW device to fade out at 600 feet using the same antennas.
|One- and two-way systems|
One-way vs. two-way devices
Many wireless I/O products are one-way devices. A radio transmitter device has input connections, and a separate radio receiver device has output connections. These products are designed for one- or two-instrument signals. A common complaint from users is they later want to add signals or a control signal in the reverse direction, and they have to install additional transmitters and receivers.
Just as with cell phones, once users experience the benefits of wireless instrumentation, there's no going back. But also just like with cell phones, cars, or any other product, you have to take your future requirements into account.
Two-way products use transceivers. They are generally multi-I/O devices that have the advantage of transmitting in both directions. The big advantage, however, is not as obvious. Two-way devices can talk to one another. Think about a telephone or walkie-talkie radio; think about the difference between one-way and two-way communications.
The big advantage of two-way systems is acknowledgments—the receiving unit can transmit back an acknowledgment that it received the radio message. Equally important is the transmitting unit can retransmit the message if the first transmission is unsuccessful. Two-way devices give a significant step up in operating reliability.
In industrial environments, the probability that a single radio message will not be successful is 5% to 30%, depending on the installation. If the transmitting device knows the transmission was not successful and can transmit a second message on an alternate radio channel, the probability is much better. If you continue this to five attempts (if you try different radio channels to transmit the message), the probability of failure drops to about 10–8. You can make five attempts in 1–2 seconds and eventually signal failure as an immediate alarm.
The other advantage of transceiver devices is repeater functionality—using an intermediate unit to repeat a radio message onto the destination. In industrial environments, black spots will always exist where radio signals are weak, just like with cellular networks. You can overcome this by relocating the radio antenna. However, you'll need to install antenna cable, and the driving benefit of using wireless is not installing cable.
Being able to bounce a radio message via other units in a system is a big advantage. An extension of this is a multiple repeater path—bouncing the radio message through a preplanned route. Extending further, systems become autorouting. The system works out for itself how to get a message from A to B and can automatically reroute if necessary. This is the same principle used in cell phone systems when a cell phone roams through a cell network.
A looming problem for wireless is the growth of dedicated wireless loops. A plant installs a gas analyzer with embedded wireless, then a wireless level gauge, then a radio modem link to a plant programmable logic controller.
Suddenly, there are three or four antennas on the control room roof. This is called the porcupine principle. How many antennas will you need to cover 100 wireless applications?
Several degrees of system expandability are becoming important. First, station expandability means additional I/O or data bus connectivity to existing wireless units.
The second level, network expandability, means you can add new units or nodes to the network without significantly reworking the network.
An enhancement of this is peer-to-peer communications. Peer to peer allows any unit in a network to communicate with any other unit, as opposed to communicating only with a network master. The advantages are not being dependent on a single master unit as well as simultaneous communications links within the same network.
The third degree is global expandability: the ability to interconnect to external systems such as distributed control system, Internet systems, or public wireless systems. One of the inherent benefits of wireless is being able to share information.
Of course, the ability to share information is a double-edged sword. The openness of wireless is a growing concern to plant managers, particularly in recent times.
Radio transmissions do not stop at plant or factory boundaries. So can someone steal wireless information and use it as industrial espionage? Or worse, can someone inject wireless commands into a plant to cause malicious damage?
Fortunately, the information technology industry is leading the way with several security techniques applicable to wireless communications. Spread spectrum gives a good initial level of protection. Spread-spectrum techniques use a form of frequency encryption that a hacker would need to unravel before he could do any damage.
If you add onto this simple data encryption techniques developed for the e-business industry, then wireless can be very secure, certainly to similar or better levels than traditional wired loops and systems.
There is no doubt wireless is a growth area. Several wireless vendors experienced 100%+ growth last year—in a market where anything with positive growth was doing well.
Market studies predict industrial wireless to grow at a compound rate of 40%+ per year in the next decade. This means the average plant and factory will have hundreds of wireless loops.
This sort of growth will need vast improvements in product technology, particularly in communications. In the wired LAN domain, improvements in hardware bandwidth have continuously resulted in higher information throughput. Communication protocols have not seen the same degree of development, simply because there hasn't been a need.
But the wireless domain has a limited bandwidth and will require large improvements in communications efficiency to limit the radio density. Low-density, event reporting protocols instead of high-density polling or timed updates will become very important. P
Behind the byline
Graham Moss is general manager for remote monitoring and control products at Elpro Technologies in Stafford, Queensland, Australia.