Designing FMC-test products requires wireless-LAN and cellular expertise

Agilent Technologies and Azimuth Systems have created a suite of FMC-test products that combine Agilent's cellular-testing know-how with Azimuth's wireless-LAN-test expertise.

Dan Strassberg, Contributing Technical Editor -- Test & Measurement World, 9/9/2008 8:20:00 AM

You can now get cellular phones that allow you to initiate or take calls via wireless LANs at home, at work, or at wireless hot-spots and then seamlessly continue the conversation via the cellular network as you move out of range of the WiFi LAN. Less common but also becoming available are cellular-modem cards for laptop PCs that let you do much the same thing while you access the Internet at speeds of 300 kbps and more.

Moreover, the processes work equally well if you begin the conversation or online session via a cellular network (today, usually GSM, the second-generation cellular technology on which most FMC-fixed-mobile convergence-activity has focused) and then move within range of a WiFi LAN. For mobile-phone users and their employers, these FMC capabilities are more than just convenient; they can save huge amounts of money, because, at least in the US, you can access WiFi LANs, unlike most cellular services, without incurring connect-time charges.

But arranging this marriage of WiFi and cellular networks is far from simple, and demonstrating the trouble-free performance of equipment, such as dual-mode (WiFi/GSM) mobile phones, requires test hardware and software designed by people intimately familiar with the eccentricities and challenges of both cellular and LAN technologies. In fact, Agilent Technologies and Azimuth Systems maintain that, to create their suite of FMC-test products, they had to call upon engineers from both companies. The Agilent people brought the cellular-testing know-how; the Azimuth folks provided the wireless-LAN-test expertise.

One of the most difficult problems in FMC systems is the handoff that occurs when you leave the wireless-LAN domain and enter the cellular domain or vice-versa. Sandy Fraser, product manager at Agilent's Mobile Broadband Division, says that although cellular-service providers and cellphone manufacturers have been working on the handoff problem in purely cellular systems for more than 15 years, such systems still occasionally drop calls when you leave one cell and enter another, even though you never leave the cellular domain. Fraser says he expects that cellular-service providers and manufacturers of wireless-LAN APs (access points) and dual-mode FMC phones will be working on the FMC-handoff problem for years to come-but that doesn't mean acceptable QoS (quality of service) won't exist until the last kink is worked out. In fact, he says, acceptable FMC QoS already exists.

A problem unique to FMC handoffs is that WiFi wasn't originally designed to handle roaming in the cellphone sense. WiFi uses a break-before-make approach to handle devices that move among the coverage areas of different APs. That is, WiFi systems don't locate an AP that offers better reception of a device's signal until communication problems arise between the device and its current AP. Still, for a technology that wasn't designed for it, WiFi handles roaming surprisingly well.

In contrast, cellular networks, which are, of course, designed for roaming, use a make-before-break approach, which establishes contact between a mobile phone and a new cell before communications with the current cell become seriously impaired.

Reliably demonstrating reliability

For equipment designers and manufacturers and network operators, demonstrating that FMC products and systems reliably met network requirements and provided a satisfactory customer experience was, until recently, a frustrating exercise. There was no alternative to over-the-air testing, which is fraught with such problems as signal impairments, for example, fading, RF shadows, and multipath propagation, and the presence of transitory and unpredictable interference (the most popular WiFi frequencies are still in the unlicensed 2.4-GHz band, which is home to such powerful emitters as microwave ovens). As a practical matter, the impairments and interference essentially precluded reproducible results, although the use of large, expensive, shielded anechoic chambers could partially ameliorate some of the difficulties.

Agilent Technologies and Azimuth Systems have addressed these problems by eliminating over-the-air testing and substituting for the large test chambers small shielded enclosures that are only slightly larger than the UUT (unit under test). Software and hardware enable the use of repeatable test signals that simulate signal impairments and interfering signals. Because the test conditions are now known and understood, the results are reproducible. Moreover, standards can prescribe meaningful conditions for qualification testing.

Both companies supply portions of the test-system hardware (figure) and software. The Agilent components include the 8960 (E5515C) wireless communications test set, the E6701F GSM/GPRS (general packet-radio system) lab application and the E6962A, which simulates a GANC (generic-access network controller). The GANC is an important part of a cellular network; it controls traffic on the ultra-high-speed fiber-optic links that connect the carrier's core network to multiple cell sites.

The US price for the Agilent portion of the system is $75,000. Figure 1 also shows the Azimuth hardware, except for the PC, although the PC's monitor and keyboard appear in the photo. US pricing for the Azimuth equipment and software starts at approximately $50,000 to $60,000. For this price you can test the effect of varying signal strength on the WiFi LAN. Signal impairments, such as multipath propagation are extra-cost options. For pure-cellular implementations, the Agilent 8960 (E5515C) hardware supports W-CDMA as well as GSM. So when FMC encompasses W-CDMA, as it almost surely will, the cost of extending the test system to support that implementation should be modest.