In 1901, Guglielmo Marconi used Morse code to communicate the letter “s”. This transmission was an analog equivalent of a “digital” signal. By 1914, the Marconi Company was engaged in experimental voice transmissions that laid the groundwork for broadcast radio. With the advent of short wave radio, costs came down and reliability went up, ensuring commercial viability. At first, wireless transmission was unregulated, but as broadcasting began to develop, it became obvious to lawmakers that some type of regulation was needed to provide for orderly use of the airwaves. To fill this need, the Federal Communications Commission was formed on June 19, 1934, when President Franklin Roosevelt signed the Communications Act of 1934.

The earliest form of wireless communication on the factory floor was the use of walkie-talkies, where manufacturing engineers communicated with the maintenance engineers to troubleshoot production or machine problems. It was a crude wireless network, but a wireless network nonetheless. Problems with these networks began to surface as other companies in the area started using the same walkie-talkies. Messages would be garbled when radios keyed at the same time from multiple locations collided with each other, resulting in misinterpreted messages and forced re-transmission of information.

Cirronet's radio modems

The first wireless data communications had its share of other problems as well. It used licensed narrowband technology, which operated over a narrow range of spectrum and was susceptible to multipath fading and interference or jamming. Multipath fading occurs when multiple copies of the same signal arrive at the receiver with slight differences in timing or phase. These differences in timing and phase occur when the radio frequency (RF) signal reflects or bounces off objects, thus taking a longer path than the signal that goes directly from the transmitter to the receiver. When the signals arrive out of phase and slightly later in time, they tend to reduce the signal strength of the direct path signal, causing “fading.” Interference occurs when another device, such as a microwave oven, generates RF frequency noise at the same frequency as the radio.

The FCC helped alleviate another problem faced by wireless communication— the licensing of radios, and the limited availability of frequency licenses. To ensure that as many people as possible had a reasonable chance to obtain a license, the FCC limited the amount of bandwidth and transmission power for each license. To facilitate deployment of these local licenses, frequency coordinators were designated to assign licensed frequencies to individual facilities. While this limited interference, it also limited the number of radios a facility could have and the narrower bandwidth resulted in lower data rates.

As technology improved, so did the narrowband radios—at least in terms of interference from radios far removed in frequency. But the issues of licenses and coordination remained; as did the effects of in-band interference and multipath fading. As a result, licensed narrowband radios were used primarily in outdoor applications, such as SCADA. With outdoor use, multipath fading was reduced. And since SCADA systems sent small amounts of data, the low data rates were tolerable. But this still left a need for factory floor
wireless communications.

The Introduction of Spread Spectrum Technology

Frequency Hopping Spread Spectrum (FHSS) technology was developed in large part by the military for secure voice communications. These techniques were classified until some time after World War II. This form of spread spectrum uses a narrowband radio signal, but the frequency of the RF signal varies rapidly. So, at one moment the radio could be transmitting at 915MHz, and a few milliseconds later it could be transmitting at 927MHz. Anyone trying to detect and listen to a radio transmission would have to change frequency at the exact same time and to the exact same frequency as the
transmitting radio.

Another type of spread spectrum technology called Direct Sequence Spread Spectrum (DSSS) came about after World War II as a result of working with coding and sampling techniques. In Direct Sequence, a high strength narrow band signal is spread using coding and sampling techniques to become a low level wide band signal. This makes the signal hard to detect. The receiving radio reverses the technique to recover the high strength narrow band signal.

In addition to making radio signals hard to detect and intercept, spread spectrum offered another advantage for factory applications. By operating over relatively large chunks of spectrum, it was less likely that an interfering signal would block the entire band and that multipath fading would affect the entire band equally. In the case of frequency hopping radios, if the radios are hopping fast enough over a large enough number of channels, other radios could be used in the same area since the radios would only interfere with each other when they were using the same channel at the same time. This might reduce data throughput but would allow communications. Thus the FCC set a minimum number of channels covering a minimum amount of spectrum and a minimum amount of time in which each channel must be used at least once.

For direct sequence radios, if the signal is spread over a wide enough range, the signal intensity will be low enough that its signal will not interfere with other radios. Thus the FCC set a minimum amount of spreading to be legal. For both methods of spreading, the FCC set power limits as a further guarantee that multiple users could operate in the same band. Thus spread spectrum held promise for factory floor wireless data applications.

Because of their lower data rates, frequency hopping radios provide better range than direct sequence radios. The ability to co-locate more radio networks using frequency hopping technology provides more flexibility. And without the need for computing-intensive protocol, frequency hopping radios can be easily connected to devices with limited intelligence. Many frequency hopping radios have been designed to work with several industrial communication busses, such as MODBUS.

Factory Floor Communications

In 1985, the FCC recognized the potential benefit of spread spectrum technology and the need to use radio transmission for commercial in-building communications systems. The FCC allocated three separate bands for low powered systems that did not require licensing. The Industrial, Scientific, and Medical (ISM) band was broken down into 900 MHz, 2.4 GHz and 5.8 GHz. Ham radio operators were given priority in these bands and non-Ham operators were required to employ spread spectrum technology.

As you might imagine, use of the ISM band turned into a free-for-all as a large number of proprietary systems were developed with no interoperability. Radios were being designed to maximize performance in specific applications. The 900MHz radios were the first technology on the market. The 900MHz radio, by virtue of its lower frequency, had better propagation characteristics than 2.4 GHz and 5.8 GHz—a fancy way for saying it goes farther. However, the 900MHz band is not as wide as the 2.4GHz or 5.8GHz bands. As a result, you cannot get as many channels or as high a data rate at 900MHz. A second limitation is that paging systems operate at a frequency close to the 900MHz band. Paging transmitters are allowed to transmit at up to 3,000 watts. By comparison, 900MHz systems are limited to 1 watt. Thus the potential for interference from paging systems is clearly present—even from systems not close by. Also, 900MHz is not an unlicensed frequency in most of the world, so 900MHz products can only be sold in North America, parts of South America and Australia. The emergence of 900MHz cordless phones that operate in the same frequency band has also caused concern due to the large number of these phones in use.

Meanwhile, 10Mbps Ethernet became the standard in office environments. Thus wireless Ethernet devices needed to provide similar speed connections to the network. Given the FCC rules for the ISM bands, it was difficult to achieve this level of performance in the 900MHz band. Fortunately, RF technology had advanced sufficiently so that radios operating in the 2.4GHz band could be produced at a reasonable cost. The advantage of the 2.4GHz band was that it was wide enough to allow radio systems to be built with 1Mbps + data rates. In addition, 2.4GHz is an unlicensed frequency in most of the world. Some disadvantages of 2.4GHz are that it has slightly worse propagation characteristics than 900 MHz. Also, microwave ovens operate in the 2.4GHz band. Nevertheless, a large number of radio manufacturers produced
2.4GHz radios.

The 802.11 Standard

When 1-Mbps and 2-Mbps wireless connections failed to gain wide acceptance due to the lack of a standard, the IEEE created 802.11, a standard for wireless LAN products. (Note: The initial standard had no letter after it.) The 802.11 standard was established as an “after the fact” standard encompassing existing products. As a result, it was decided to start over and develop a true standard to which new products could be built rather than trying to fit a standard to existing products. So work began on the 802.11b standard.

The 802.11b standard has turned out to be an excellent standard for short range, wireless LANs. However, 802.11b products were tried in many applications where their success was very limited. The factory floor was one of those less than successful application areas.

To understand the shortcomings of 802.11b for factory applications, it is helpful to understand the benefits of wireless on the factory floor. First and foremost is removal of the need for wires. In most instances, the biggest cost of an industrial control system is the cost of running the wire. Often, to run wires or conduit, production lines must be stopped, resulting in lost production. Another benefit of wireless is the flexibility it provides. If lines are changed or moved, no wires have to be moved.

To provide these benefits, a wireless solution must be able to cover the distances commonly found in a factory—typically hundreds of feet. When 802.11b radios were deployed in factories with the expectation that they could cover the range while providing the maximum data throughput, they failed miserably. The solution was to deploy many access points around the factory, typically in the ceiling. But access points need to be wired to the network, so one of the main benefits of wireless was lost. And when many access points are needed, 802.11b has only three non-overlapping channels. Thus only three access points could be deployed in a factory without fear of interference.

In the meantime, applications specifically designed for use on the factory floor were by and large proprietary solutions. While frequency hopping radios could not boast the high potential data rates promised by 802.11b radios, it turned out that the high data rates were not needed. What were needed were radios that could stand up to the high RF noise interference of a factory floor environment, be deployed easily, and work reliably. Radios employing the FHSS technology meet this need today.

802.11b has some areas where it performs well on the factory floor. The most common is in stock room inventory applications. The ranges are not as long as those on the factory floor and they are further removed from
equipment such as lighting systems, welding machines and motor starters that may cause interference. With the proliferation of notebook computers and even handheld devices that have PC card slots, 802.11b radios can be used as a short range connection to production machinery for diagnostic or configuration purposes. Like the narrowband licensed radios that found their niche in outdoor SCADA applications, 802.11b radios have exhibited their usefulness on the factory floor.

What’s Next?

Already there are two new radio technologies being discussed as the next great thing—ZigBee and Ultra Wide Band (UWB). ZigBee is designed to be a low power, low data rate mesh radio technology. A mesh network is one in which there are multiple paths between points. ZigBee holds promise for industrial sensing applications where a large number of sensors need to be read at a fairly slow rate. Data throughput will range up to 250Kbps and latencies will vary by network size, but will typically be on the order of 100 milliseconds. ZigBee products are expected to hit the market in volume in the second half of 2005.

UWB is a technology that uses a very wide chunk of spectrum but at a low RF power to obtain theoretically extremely high data rates. Reports have circulated recently reporting data rates in the hundreds of Megabits per second. The tradeoff will be a very short transmission range. The UWB products currently being tested are focused on location-finding devices such as wall stud finders. It remains to be seen where UWB will find its niche in industrial applications.

Conclusion

While it has taken some time and caused frustration, robust and reliable wireless products are now available for factory floor applications. There is not one size that fits all applications, but by having a clear understanding of the application and with the help of a well-trained provider, the benefits of wireless communications can now be realized on the factory floor.

By Tim Cutler
Cirronet

Originally Published: Dec. 1, 2005