Embedded Wireless and AntennasThursday, May 31, 2007
Yesterday I received my new Panasonic notebook (CF-W5) with embedded Verizon Wireless EV-DO Rev A (Sierra Wireless). I have been using several EV-DO Rev A cards for some months now, but this is my first experience with an embedded solution. The Verizon Wireless broadband manager, once installed, recognizes not only the Verizon Wireless network, but any Wi-Fi networks, making it an ideal client for running wireless from a notebook. (Sprint and AT&T also offer this type of solution). Having a single client that recognizes both wide-area and Wi-Fi networks is the right way to provide a wireless connectivity wireless user interface and it makes life a lot simpler.
The first thing I did was to make sure the system was working properly out of the box (it was), then I set about changing my Verizon Wireless account from a PC Card to the embedded module. The entire process, including loading the software, took only a few minutes, and in a few minutes more I had the system up and operating on the Verizon Wireless network and my own Wi-Fi 802.11g network.
It is difficult to take typical measurements as close to a cell site as I am, but at first blush I have to say that the embedded card, and embedded antenna system, works as well if not better than the PC Card with an external antenna. This is no fluke since Panasonic has its own engineering lab in Japan that runs tests on its internal antenna systems. Once I have a chance to run some tests in other parts of the country, I will be able to provide a better comparison between the PC Card and embedded approach, at least for this system.
One thing I noticed right away is that the Panasonic 802.11g system works better than either my Sony VAIO or the Toshiba Satellite I use for Wi-Fi and wide-area testing. My own Wi-Fi access point is located at the point of entry of my DSL connection in a different building about 400 feet away. Using the other two systems, I have a marginal signal in my office. But using the Panasonic laptop in exactly the same location, the signal is being received stronger and the data throughput is better (1.2 Mbps as opposed to 784 Mbps for the Sony and 650+ for the Toshiba). It appears to me that Panasonic engineers have done a great job with this notebook.
Unfortunately, many companies don’t understand the importance of antennas when it comes to wireless. In many devices, the weak point in a system and how well it works is the antenna or its placement. As an example, in our work testing muni-Wi-Fi systems, we have two (and sometimes three) laptops we use as well as a device provided by the Wi-Fi vendor to extend the range of the system indoors. At each location, we take measurements with each of the laptops and with the external device.
Our tests include how many access points (SSIDs) we can see from each location, how strong the signal is from the SSID we want to test, the signal-to-noise ratio and then a set of uplink and downlink speed tests. We have noticed over the past year that there is a big difference in performance between the laptops. The best of the bunch is the IBM ThinkPad, next is the Sony VAIO and in last place is the Toshiba Satellite. The interesting thing about this is that all three have the same Intel Wi-Fi module inside them. The difference between how well they work comes down to the antennas in each system.
Some notebook companies design their notebooks for the best computing experience, insert Wi-Fi and wide-area modules and almost literally throw in an antenna wherever it fits. They don’t understand that the most critical element of any wireless system is the antennas—both on the device and on the network. One claim being made by the Wi-Fi Alliance is that 802.11n on the 2.4-GHz band will offer about 120% more range than either 802.11b or g, in spite of the fact that n has a much higher data rate. Part of the reason for this increase in coverage is that many of the n devices are employing what are called MIMO (multiple-in, multiple out) antenna systems. These systems are carefully designed to enhance both transmission and reception.
Antennas are a critical part of any wireless system and one way to boost performance, which in turn provides higher data rates to customers, is to design antenna systems that can actually add a measure of performance to the system. Designing antennas is a multipart process―the first is to design an antenna that is efficient (and, many times, efficient on different portions of the spectrum), transfers RF signals well and, if possible, increase the overall system gain. Once that is accomplished, measurements need to be taken to show the pattern of the antenna (omni-directional, which provides 360-degree coverage, is the goal). The next issue is the placement of the antenna(s), especially in a notebook computer. Placement is not simply dependent on where it will fit, but how close it is to the CPU, screen and other noise-generating devices inside the computer. When embedded into a notebook or other device, the characteristics of the antenna can change dramatically and some of the performance gains designed into an antenna can be lost if it is not positioned properly inside the device.
So far, I have not seen any of the companies that test laptops do a really complete evaluation of coverage for either Wi-Fi or wide-area wireless. One notebook vendor does study each of its competitors’ products and its own, but the results are not published anywhere. How much does it really matter? In our testing of one muni Wi-Fi system, it made a difference in whether we could even connect to the network and it made a big difference in the data rates we experienced once we were connected. The connect speed between the “best” and “worst” laptop was sometimes as much as 50% from the same location.
My point is that building a product to work on a wireless network takes a different type of expertise than building one to work using a wired Ethernet connection. Many of the vendors of these products do make an effort to work with RF design experts, but as more devices become wireless-capable, the differences in performance will become greater. And, of course, the result of a poor device is that the customer blames the network, not the device.