The sky appears to be the limit for wireless local area network (WLAN) technologies. WLANs provide a good fit for the usage model aimed at high-bandwidth consumers, with its low infrastructure cost compared to other wireless data technologies such as third-generation (3G) cellular or point-to-multipoint distribution systems. They can be easily adapted for business or residential use and for low-mobility environments such as airports, coffee shops, hotels and other locations where a need for broadband Internet access exists.

In June 1997, the IEEE approved the 802.11 standard for WLANs, and in July 1997, IEEE 802.11 was adopted as a worldwide International Standards Organization (ISO) standard. The standard consists of three possible physical (PHY) layer implementations and a single common medium access control (MAC) layer supporting data rates of 1 Mb/s or 2 Mb/s. The alternatives for PHY layer in the original standard include a frequency hopping spread spectrum (FHSS) system using 2 or 4 Gaussian frequency-shift keying (GFSK) modulation and a direct sequence spread spectrum (DSSS) system using differential binary phase-shift keying (DBPSK) or differential quadrature phase-shift keying (DQPSK) baseband modulation.

Testing of WLAN products offers unique challenges to the test engineer. When performing system-level tests, the worst-case environment must be simulated in the lab to ensure that the operation of the installed system meets performance specifications and marketing claims. When evaluating the overall quality and utility of a particular vendor's system, customers will need to review the products not only in terms of the physical layer operation in their own environments, but also in terms of network management software, security and site survey utilities.

Testing the physical layer to the standard requires that specific measurements be made for both versions of the standard.

The physical layer basics

Knowing what some of the physical layer terminology means is essential to understanding the intricacies of 802.11.

• GFSK is a modulation scheme in which the data are first filtered by a Gaussian filter in the baseband, and then modulated with a simple frequency modulation. “2” and “4” represent the number of frequency offsets used to represent data symbols of one and two bits, respectively.

• DBPSK is phase modulation using two distinct carrier phases for data signaling providing one bit per symbol.

• DQPSK is a type of phase modulation using two pairs of distinct carrier phases, in quadrature, to signal two bits per symbol. The differential characteristic of the modulation schemes indicates the use of the difference in phase from the last change or symbol to determine the current symbol's value, rather than any absolute measurements of the phase change.

Both the FHSS and DSSS modes are specified for operation in the 2.4 GHz industrial, scientific and medical (ISM) band, which has sometimes been jokingly referred to as the “interference suppression is mandatory” band because it is heavily used by various electronic products. The third physical layer alternative is an infrared system using near-visible light in the 850 nm to 950 nm range as the transmission medium.

At the forefront of the new WLAN options that will enable much higher data rates are two supplements to the IEEE 802.11 standard: 802.11b and 802.11a, as well as a European Telecommunications Standards Institute (ETSI) standard, High Performance LAN (HIPERLAN/II). Both 802.11 and HIPERLAN/II have similar physical layer characteristics operating in the 5 GHz band and use the modulation scheme orthogonal frequency division multiplexing (OFDM), but the MAC layers are considerably different. The focus here, however, will be to compare the physical layer characteristics of 802.11a and 802.11b. With HIPERLAN/II sharing several of the same physical properties as 802.11a, many of the same issues will apply.

Another standard that warrants mention in this context is IEEE 802.11g. With a ruling from the Federal Communications Commission that will now allow OFDM digital transmission technology to operate in the ISM band, and the promise of interoperability with a large installed base of 802.11b products, the 802.11g extension to the standard begins to garner the attention of WLAN equipment providers. Although not detailed here, it will offer data rates equal to or exceeding 22 Mb/s with products available late in 2002.

What to know about 802.11b

802.11b, which was approved by the IEEE in 1999, is an extension of the 802.11 DSSS system previously mentioned and supports higher 5.5 and 11 Mb/s payload data rates in addition to the original 1 and 2 Mb/s rates. Products are now widely available, and the installed base of systems is growing rapidly. 802.11b also operates in the highly populated 2.4 GHz ISM band (2.40 to 2.4835 GHz), which provides only 83 MHz of spectrum to accommodate a variety of other radiating products, including cordless phones, microwave ovens, other WLANs, and personal area networks (PANS). This makes susceptibility to interference a primary concern. The occupied bandwidth of the spread-spectrum channel is 22 MHz, so the ISM band accommodates only three non-overlapping channels spaced 25 MHz apart. To help mitigate interference effects, 802.11b designates an optional frequency agile or hopping mode using the three non-overlapping channels or six overlapping channels spaced at 10 MHz.

802.11b uses eight-chip complementary code keying (CCK) as the modulation scheme to achieve the higher data rates. Instead of the Barker codes used to encode and spread the data for the lower rates, CCK uses a nearly orthogonal complex code set called complementary sequences. The chip rate remains consistent with the original DSSS system at 11 Mchip/s, while the data rate varies to match channel conditions by changing the spreading factor and/or the modulation scheme.

To achieve data rates of 5.5 and 11 Mb/s, the spreading length is first reduced from 11 to eight chips. This increases the symbol rate from 1 Msym/s to 1.375 Msym/s. For the 5.5-Mb/s bit rate with a 1.375 MHz symbol rate, it is necessary to transmit 4 bits/symbol (5.5 Mb/s/1.375 Msym/s) and for 11 Mb/s, an 8 bits/symbol. The CCK approach taken in 802.11b, which keeps the QPSK spread-spectrum signal and still provides the required number of bits/symbol, uses all but two of the bits to select from a set of spreading sequences and the remaining two bits to rotate the sequence. The selection of the sequence, coupled with the rotation, represents the symbol conveying the four or eight bits of data. For all 802.11b payload data rates, the preamble and header are sent at the 1 Mb/s rate.

802.11b testing issues

The 20 MHz-wide bandwidth of WLAN signals makes power envelope measurements difficult because most spectrum analyzers have resolution bandwidth filters that are limited to 10 MHz or less. Therefore, the signal is considerably attenuated by the time the power is measured within the instrument. Vector signal analyzers are available with information bandwidths that are considerably greater than 20 MHz, making WLAN signal analysis more accurate.

The 802.11b standard uses error vector magnitude (EVM) as a measure of modulation quality. This measurement has become common for most wireless applications. The underlying philosophy of EVM is that any signal deteriorated by a noisy channel can be represented as the sum of an ideal signal and an error signal. The test instrument determines the error signal by reconstructing the ideal signal based on detected signal information and subtracting it from the actual signal at each sample point.

What to know about 802.11a

While 802.11a was approved in September 1999, new product development has proceeded much more slowly than 802.11b. This is due to the cost and complexity of implementation. This standard uses 300 MHz of bandwidth in the 5 GHz unlicensed national information infrastructure (UNII) band. The spectrum is divided into three “domains,” each having restrictions imposed on the maximum allowed output power (see Figure 1). The first 100 MHz in the lower frequency portion is restricted to a maximum power output of 50 mW. The second 100 MHz has a higher 250 mW maximum, while the third 100 MHz, which is mainly intended for outdoor applications, has a maximum of 1.0 W power output.

OFDM operates by dividing the transmitted data into multiple parallel bit streams, each with lower relative bit rates and modulating separate narrowband carriers, referred to as sub-carriers. The sub-carriers are orthogonal, so each can be received without interference from another. 802.11a specifies eight non-overlapping 20 MHz channels in the lower two bands; each of these are divided into 52 sub-carriers (four of which carry pilot data) of 300-kHz bandwidth each. Four non-overlapping 20 MHz channels are specified in the upper band. The receiver processes the 52 individual bit streams, reconstructing the original high-rate data stream. Four complex modulation methods are employed, depending on the data rate that can be supported by channel conditions between the transmitter and receiver. These include BPSK, QPSK, 16-QAM, and 64-QAM.

Quadrature amplitude modulation is a complex modulation method where data are carried in symbols represented by the phase and amplitude of the modulated carrier. 16-QAM has 16 symbols. Each represents four data bits. 64-QAM has 16 symbols with each representing four data bits.

BPSK modulation is always used on the four pilot sub-carriers. Although it adds a degree of complication to the baseband processing, 802.11a includes forward error correction (FEC) as part of the specification. FEC, which does not exist within 802.11b, enables the receiver to identify and correct errors made during transmission by sending additional data along with the primary transmission. This nearly eliminates the need for retransmissions when packet errors are detected. The data rates available in 802.11a are noted in Table 2, together with the type of modulation and the coding rate.

802.11a products are expected to begin arriving in the first half of 2002. Some of the companies developing chipset solutions for 802.11a are touting the availability of operational modes that exceed the 54 Mb/s stated in the specification. Of course, because faster data rates are out of the specification's scope, they require the use of equipment from a single source throughout the entire network.

Considering the composite waveform resulting from the combination of 52 sub-carriers, the format requires more linearity in the amplifiers because of the higher peak-to-average power ratio of the transmitted OFDM signal. In addition, better phase noise performance is required because of the closely spaced, overlapping carriers. These issues add to the implementation cost of 802.11a products. Application-specific measurement tools aid in the design and troubleshooting of OFDM signals and systems.

802.11a testing issues

Design of devices using 802.11a with OFDM signals and operating at 5 GHz will bring new challenges in testing, particularly because the data rate will be increasing by a factor of five and using the same bandwidth (20 MHz) to do it. The high peak-to-average power ratio representative of multicarrier OFDM signals dictates the need for highly linear and efficient amplifiers, as well as a method to characterize them.

Transmitted signals such as OFDM, which do not have a constant power envelope, are not well-characterized by peak-to-average power ratio. This metric is not useful, as the true peak power may not occur often. It is usually more meaningful for OFDM signals to associate a percentage probability with a power level.

A more meaningful method for viewing OFDM signal power characteristics uses the complementary cumulative distribution function (CCDF). This metric links a percentage probability to a power level. In this measurement, an instrument with time-gating capability is used to select only the active portion of the burst (see Figure 2 lower trace).

If time gating were not used, the periods when the burst is off would reduce the average power calculation. The CCDF, which is simply the more common cumulative distribution function (CDF) subtracted from 1.0, shows the number of decibels above the average power on the horizontal axis, and percent probability on vertical axis (see Figure 2 upper trace). A CCDF measurement would be made over several bursts to improve the accuracy of the measurement.

Pros and cons of 802.11b, 802.11a

A drawback of the 5 GHz band, which has received considerable attention, is its shorter wavelength. Higher-frequency signals will have more trouble propagating through physical obstructions encountered in an office (walls, floors, and furniture) than those at 2.4 GHz. An advantage of 802.11a is its intrinsic ability to handle delay spread or multipath reflection effects. The slower symbol rate and placement of significant guard time around each symbol, using a technique called cyclical extension, reduces the inter-symbol interference (ISI) caused by multipath interference. (The last one-quarter of the symbol pulse is copied and attached to the beginning of the burst. Due to the periodic nature of the signal, the junction at the start of the original burst will always be continuous.) To contrast, 802.11b networks are generally range-limited by multipath interference rather than the loss of signal strength over distance.

When it comes to deployment of a wireless LAN, operational characteristics have been compared to those of cellular systems, where frequency planning of overlapping cells minimizes mutual interference support mobility and seamless channel handoff. The three non-overlapping frequency channels available for IEEE 802.11b are at a disadvantage compared to the greater number of channels available to 802.11a. The additional channels allow more overlapping access points within a given area while avoiding additional mutual interference.

Both 802.11b and 802.11a use dynamic rate shifting where the system will automatically adjust the data rate based on the condition of the radio channel. If the channel is clear, then the modes with the highest data rates are used. But as interference is introduced into the channel, the radio will fall back to a slower, albeit more robust, transmission scheme.

Network considerations in 802.11

Network planning is critical to the development of an optimized system. Each network must be customized to satisfy the planned applications and the physical environment. Requirements must be researched and well-documented, including anticipated roaming and data rates needed for applications to be used at specific locations. A site survey must be thorough and realistic to adequately characterize the RF environment of the proposed wireless network in terms of range, channel interference and delay spread.

It would be unrealistic to expect to realize the full data rate capability (54 Mb/s) of 802.11a if the access points of an existing 802.11b network — optimized to operate at full speed (11 Mb/s) — were simply replaced. But as has been shown, 802.11a is faster than 802.11b at any range. Cost vs. performance requirements need thorough analysis during the network planning stage to arrive at the appropriate implementation decision.

Testing is critical to any product development process. WLAN products require that special attention be given to design verification and characterization because standardized operation across multivendor products may be required. To provide an efficient development environment, test tools are available to quickly diagnose problems and isolate them throughout all design segments. These tools can be used within the manufacturing process to generate and analyze production metrics for process and product improvement.

Even during these lean economic times, when there is a reduced demand for technology products, the new, but already robust WLAN market is projected to grow by an order of magnitude over the next five years. These wireless networks will require increasing data rates to provide the simultaneous distribution of Internet data, high-quality video and audio in the office or at home. In addition to higher data rates, it is almost a foregone conclusion that end-users will be demanding continuous improvements in functionality, ease-of-use and reliability.

About the author

John Hansen is currently a product manager for Agilent Technologies' Electronic Product and Solutions Group. He has more than 20 years of extensive and progressive experience in systems engineering and new product development within the wireless, microelectronics and defense industries. At Agilent, he is responsible for the development of seminars, technical articles and application notes regarding the detailed aspects of various wireless technologies. Prior to joining Agilent, Hansen worked at Hughes Network Systems, where he participated in the development of terrestrial cellular and satellite terminal products as an engineering test manager. Hansen received his B.S.E. from UCLA, an M.S.E.E. (with a communications systems emphasis) from USC, and his MBA in marketing from San Diego State. He is also a registered professional engineer in the state of California.