Wireless data network technology, mainly 802.11, is essential for home network installations, primarily for use by laptops, portable touch screens, keypads, and so on. The purpose of this TIP is to help you understand some of the limitations of 802.11 technology, how to measure it’s performance in the field using my favorite WiFi gadget, and to pass along some tips on improve the reliability and bandwidth of your installations.

Somfy Automated Shading Solutions

Currently, the fastest standardized wireless network technology available is 802.11g. If you’re installing wireless as part of home networks, this is what you should be using. If the customer has older 802.11b equipment you should offer to replace it for free—a good customer relations move that will also make your job easier.

I’m sure you’ve heard a lot about 802.11n, the newer WiFi technology. Unfortunately, 802.11n, in my opinion, is not ready to be professionally installed as a wireless infrastructure for your customer. It’s major problems right now are:

* It’s still in draft form and will be until mid-2009, the major reason that no commercial WAP utilizes 802.11n. Interoperability is still a problem.
* It has so many mode variations—576 to be exact—it’s just not possible to predict performance between any two devices.
* in the presence of competing ‘g’ networks (even in a neighbors home) it will “fall back” to g speeds.
* It is subject to many of the interference problems of ‘g’ .
* It can seriously screw-up existing ‘g’ networks.
* Something that most people forget is that to utilize it’s potential higher bandwidths (over about 60 Mbps) requires it connect to a 1000Base-T wired network. If your customer wants to experiment with 802.11n, fine, but make sure everyone under-stands it’s an experiment and performance may actually be worse than 802.11g.

WAP Performance and Bandwidth Limitations

The quoted maximum bandwidth of 802.11g (54 Mbps) is just theoretical. It is the maximum rate date is transferred while in the middle of a frame. If you placed a laptop computer right next to a wireless access point, inside an RF shielded room to prevent any external interference, and both devices were using 802.11g and streamed a large file to the laptop from an Ethernet wired connection, then assuming the only bottleneck to transmission speed was the wireless connection, the actual average data transfer rate might be something close to 45 Mbps. You can assume this is the highest rate achievable by the technology. Any other condition will degrade performance.

But exactly what factors do distance, house construction, interference, WAP specifications, and “client” device specifications have on bandwidth and how can you tell what you can count on? Until recently, equipment to accurately measure what’s going on was cost prohibitive so most of us used the old “plug-and-pray” technique – hoping the wireless install would work OK.

WiFi Transmitter/Receiver performance and Signal Characteristics

First, you need to understand typical WiFi receiver/transmitter performance. Table 1 lists the 802.11g performance of an actual good quality Wireless Access Point (WAP). This illustrates the data-rate vs. received signal strength given in dBm’s. 0 dBm = 1 mW (milli-Watt). Transmitter power is typically measured into the antenna but specs might refer to the effective radiated power (ERP) taking into account the gain of the antenna. Anything below 0 dBm is written with a minus such as -40 dBm (10,000 times less power than 1 mW).

Transmitter specs. (output power settable on the transmitter)

Receiver Specs. (received signal strength vs. max. data rate)

200 mW (23 dBm) ERP assuming a 3 dBi gain antenna

63 mW (18 dBm)

30 mW (15 dBm)

20 mW (13 dBm)

10 mW (10 dBm)

5 mW (7 dBm)

1 mW (0 dBm)

-66 dBm 54 Mbps*

-71 dBm 48 Mbps

-76 dBm 35 Mbps

-80 dBm 24 Mbps

-83 dBm 18 Mbps

-84 dBm 12 Mbps

-85 dBm 11 Mbps

-86 dBm 9 Mbps

-87 dBm 6 Mbps

-88 dBm 5.5 Mbps

-89 dBm 2 Mbps

-92 dBm 1 Mbps

* or maximum Ethernet data rate

Figure 1 plots the data rate vs. received signal strength of two good quality WAPs. The data assumes ideal conditions (no competing WiFi signals, no interference). Notice the pronounced “cliff” effect where the useful data rate drops rapidly. This is due to the nature of the 802.11g data modulation technique of OFDM (orthogonal frequency division multiplexing). It’s reminiscent of the signal capture effect of FM radio.

Figure 1. Data rate vs. received signal strength for two popular WAPs. Both exhibit a common “cliff” effect where data rate drops off quickly.

This means that to achieve anything close to the bandwidth needed to support the streaming of a SD DVD movie (about 10-12 Mbps) and nothing else, the received signal needs to be not less than -80 dBm across the channel. And that’s the receiver spec. of a good WAP. You can assume that the receiver performance of a laptop or wireless client device is going to be worse. Good luck getting any specs from the manufacturer. I use a “rule of thumb” level of NLT -72 dBm across the channel at a client device (more on this shortly) to insure best performance. Since you have little control over the customer’s laptops or other wireless devices they purchase, the best you can do is use the best WAP you can and place it accurately.

The graph also implies that as the receiver moves away from the location of the installed WAP the bandwidth will remain good up to a point then drop to an almost unusable level quickly. When in the “cliff effect” zone, the receiver is highly susceptible to interference and data rates will fluctuate rapidly. If streaming video or even audio, there will be dropped frames. The only way to know what performance to expect on a given installation is to measure it.

Measuring Signal Strength and Bandwidth

To accurately measure signal strength and bandwidth you’ll need two tools: a software tool to measure bandwidth, and a hardware/software tool to measure signal strength and interference.

To measure bandwidth you can use a utility such as NetMeter for PCs or NetMonitor for Mac’s. Both have the ability to monitor and graphically display Ethernet receiver/transmit speeds as well as record bandwidth performance over time. These utilities require the monitoring computer to be the source or destination of data transfers. For more sophisticated monitoring capability between any two devices on a network check out Qcheck. Ixiacom specializes in software and hardware based monitoring products. These utilities can also be used to measure the bandwidth requirements of media servers and clients as well as accurately check ISP performance.


For measuring signal strength and interference, Wi-Spy from Metageek (www.metageek.net) is a MUST HAVE tool. Their Wi-Spy 2.4 will turn your PC or Mac into a 2.4 GHz spectrum analyzer capable of accurately measuring signal strength (in dBm) and signal characteristics of anything in the 2.4 GHz band throughout the home. And it does this for under $400. The product consists of a small USB receiver with a dipole antenna (Figure 2) that attaches to a PC or Mac. It’s supplied with software for the PC (Chanalyzer). Software for the Mac (EaKiu) is available from Cookware Inc.

Figure 2. The Wi-Spy USB 2.4 GHz spectrum analyzer receiver with dipole antenna. The device ships with spectrum display software (Chanalyzer).

It has a serious dynamic range of -6.5 to -110 dBm which means it can measure almost anything including interference from baby monitors, Bluetooth, phones, WiMax, neighboring networks, neighboring 2.4 GHz devices, and so on. It is the most useful network tool for the money I’ve ever used. Just get it.

Figure 3. A typical Wi-Spy spectrum display using the Mac EaKiu software showing the transmission to/from a WAP and laptop on channel 11. The display is annotated with the channel 11 limits and horizontal dashed line showing the minimum signal level necessary to achieve the best bandwidth.

Figure 3 show a typical spectrum display using the Mac EaKiu software. This shows the entire 2.4 GHz band display with frequency (channel number) along the bottom and amplitude in dBm vertically (color coded). This screen capture shows the transmission between a WAP and laptop on channel 11. This is a classic spread spectrum display showing a major “hump” centered on the band and progressively smaller humps spread throughout the band. This illustrates that a strong nearby signal in channel 11 can easily interfere with a weak signal in channel 4 for example (and you thought a WiFi signal was isolated to a single channel). The display is annotated to show the channel 11 band limits and the minimum signal level necessary to achieve the best bandwidth performance for 802.11g—about -72 dBm, the edge of the “cliff” (for this particular device). To achieve maximum performance, all of the signal throughout the channel should be above this lower limit. Anything less and bandwidth drops. WiFi signals down to about -90 dBm are usable IF there are no interfering signals. This display was captured in a fairly quiet RF environ-ment with few other signals. The yellow “grass” at the bottom of the display is background RF noise.

Both software product’s have a very useful “waterfall” display (Figure 4). Here the vertical axis is time (how much is configurable) as the display continually scrolls up to show signal history. Signal amplitude is indicated by color corresponding to the signal strength color on the right hand vertical scale. Display data can be recorded to capture the RF environment over a long period of time—perfect for finding intermittent interfering sources.

The Figure 4 display shows a streaming video file on channel 11 with a microwave oven being used occasionally. The microwave oven noise “signature” is easy to spot once you’ve seen it. This caused bandwidth on channel 11 to drop about 20% generating a few drop outs and freezes. Continuous interfering signals show up as vertical lines (you can see a few weak ones in the display). Many more examples of interference “signatures” can be seen on the Metageek web site.

Figure 4. A “waterfall” display showing a “bursty” streaming video signal on channel 11 with intermittent microwave oven interference. The display shows about 60 seconds of signal capture with time increasing upward.

Figure 5. Metageek Chanalyzer display showing three views of the same signals.

Figure 5 shows a screen capture from Chanalzyer 2.1 software showing a data transmission on channel 1 and portable phone interference across the 2.4 GHz band. This phone put out a very strong signal (-50 dBm) that frequency hops all over the band. As can be seen on the “waterfall” display (between 50-80 secs.) it disrupts the WiFi data transfer on channel 1 whenever it’s transmitting in channels 1-4.

Applying the Tools for a Better Install

With a little practice you can use Wi-Spy to learn to predict path loss and thus WiFi usability in any part of a house. By measuring the drop in signal strength (path loss) through stud walls—with and without insulation—concrete walls, floors, furniture, people, and so on, you can more accurately determine where and how many WAP’s to install.

Two factors contribute to path loss: distance and absorbing materials. An RF signal amplitude decreases roughly with the inverse square of the distance from the source. Measuring path loss requires accurately measuring a constant amplitude signal. You can set up a data transfer between a WAP and a laptop (for example), not the laptop with Wi-Spy installed. This might be simply a large file transfer from something on the wired network, through the WAP to the laptop. This creates a steady 802.11g signal on a single channel so you can use the WAP as the signal source. Use the clearest channel you can find. The transfer should take long enough to make the measurements. Place your Wi-Spy laptop near the WAP so there is as clear a line of sight between the laptop and the WAP as possible. Keep your body clear of the antenna. If possible, use a USB extension cable and mount the Wi-Spy “dongle” on a camera stand so you can keep it away from the laptop and you. Measure (and record) the average peak amplitude of the signal. Somewhere between -30 and -50 dBm is a good value. You can assume this is the strongest signal anything in the house will receive. Now move your laptop so there are walls or other obstructions between the WAP. Measure the new amplitude value. The difference between the two measurements is the path loss made up of distance and material absorption.

You can also use the tool to do an accurate site survey during house construction to check for interfering sources and thus the quietest channel. Once the customer’s network is installed you can determine what the performance will be in every part of the house and adjust WAP location, antenna orientation or type, even adjust output power. But the RF “environment” of a house will definitely change over time. So when the customer complains of poor performance in the future, you can use the tool to reanalyze the site for interfering sources. Even if you can’t find the exact cause, you can at least find the quietist RF “real estate” for the customer.

Other TIPS

Other things you can do to improve the installation include:

* Use a commercial grade WAP. If you can buy it at Best Buy or other consumer electronics outlet, don’t. My favorite is the Cisco Aironet 1200 series.
* Test the WiFi performance of the equipment you install at minimum signal strength.
* Try to remove interfering devices in the home. Get rid of 2.4 GHz phones and other 2.4 GHz devices. Use DECT phones that operate in the 1920-1930 MHz band.
* If you find a neighbor using a particularly offensive device (such as a 2.4 GHz wireless video transmitter), offer to replace it with something on another band (or a cable!).
* Remind customers that if they are not using the wireless connection of a laptop, to turn the radio off. WiFi cards not “connected” to a WAP are constantly transmitting all over the band.

The Bottom Line

WiFi is just not a reliable medium for bandwidth intensive applications. There are too many things you have no control over: the performance of the streaming client whether it’s a computer or dedicated device, it’s location, interfering sources outside or inside the house, path loss due to things in the home (people, furniture, etc.), number of simultaneous users, and so on. However, there are many things you can do to improve performance especially for non-bandwidth critical applications such as web surfing, portable touch screens, etc. to insure the customer has the best possible bandwidth over the widest area of the home. You should be able to get reasonably reliable bandwidth for medium bandwidth intensive applications such as streaming SD video in selected areas of the home if you take the time to use the tools and do the planning.

Questions, comments? Email Grayson at: grayson@trainingdept.com

More Tip-of-the-Month articles like this one are available for download at www.trainingdept.com/html/support.html