Test Results, Continued
Figure 12 shows the spectrum use for three frequency bands with the phone on-hook (not active), while Figure 13 shows the off-hook (active) condition. The top plots in each screenshot show spectrum use (from left to right) the frequency bands for 5.725 to 5.850 GHz, 5.150 to 5.350 GHz and 2.4 to 2.5 GHz. These correspond to 802.11a channels 149 - 161 and 36 - 64 and 802.11b/g channels 1 - 14. By the way, the right side of the plots are truncated since Cognio's software doesn't allow the plots to be fit to the screen and instead relies on screen resolution higher than the 1024 pixels wide that my notebook has.
Figure 12: On-hook spectrum use
(click image to enlarge)
The bottom plots in each screenshot show the "swept spectrogram" plots for the corresponding bands above them. This tool very nicely shows signal activity over time in the band being monitored, with signal presence indicated by the greenish-yellow color. Both screenshots show activity covering pretty much all of the 5.725 to 5.850 GHz band, with some intermittent activity also between 5.150 to 5.350 GHz. Figure 12 shows no signals in the 2.4 GHz band when the phone is on-hook, but nice, strong signals once the phone is active in Figure 13.
Figure 13: Off-hook spectrum use
(click image to enlarge)
I should note that for all three bands, the spectral usage is constantly changing, but the basic pattern of heaviest use in the 5.725 to 5.850 and 2.4 GHz bands stays the same. Since those are some pretty strong signals in the 2.4 GHz band, I figured that any 802.11b/g WLAN was doomed to be knocked off the air. But my experiment proved otherwise.
For the test WLAN, I set up one of my notebooks with a NETGEAR WAG511 Dual-band card that happened to be handy and configured the card to disable use of its 802.11a side and just act as an 802.11b/g card. I then set up a Linksys WRT54G 11g router for it to associate with and IXIA's IxChariot in the notebook to continuously run its standard throughput script to simulate WLAN activity.
With my test WLAN running, I then took the phone off-hook and watched the IxChariot throughput plot, waiting for it to die as the phone blasted its overpowering signal into the 11b/g band, just a few feet away from the notebook. But, to my surprise, the throughput plot kept running nice and steady, with not even a glitch to show that the phone was active!
I repeated the test, moving the WRT54G among channels 1, 6 and 11 and even forcing it to 11b-only mode to see if I could detect interference with 11b's lower data rates. But in all cases, the WLAN kept on working with no indication that I could see that the phone was in use.
I searched through both VTech's website and the PDF manuals that were on the CD that came with the BTS, and finally found one reference on this page to the product being "Wi-Fi" friendly. I can only conclude that VTech has implemented some sort of frequency-hopping scheme that steers the phone clear of any in-use WLANs.