While testing some new Centrino 2-based notebooks I came across a strange problem, making me wonder if there is a possible, yet strange bug with Intel's implementation of Wi-Fi on Centrino 2.

First let me warn you, it is probably very unlikely that anyone will encounter this same problem, unless the same environment is created with Wi-Fi and WEP. While testing new notebooks from both Fujitsu and Lenovo, I tried connecting to a Linksys (and then a Netgear) access point. The access points had 64 bit WEP enabled and a hexadecimal key of 10 zeros (0000000000).

I have been using both of these access points (a Netgear WN802T and a Linksys WAP200) for some time to test various products and have used the current settings with a bunch of Vista and XP Notebooks, an IPOD touch, several smart phones and even a Nintendo Wii, and everything always worked without a hitch. Imagine my surprise when I tried to use both a Lenovo SL400 and a Fujitsu T5010 and could not fully connect to the network. I was able to connect to the access points, but the access points refused to send any packets back to the notebook computers (or more likely, the notebook computers refused to receive incoming wireless packets). So, even though I had a wireless connection, I was unable to get an IP address via DHCP or access the network.

My first instinct was to flash the access points with the latest firmware, but upon checking, both units did have the latest firmware available. I verified that other devices still worked with the access points and then turned my attention to the new notebooks. I attempted to manually set the IP information, but the units would still not accept any transmitted packets. My next step was to turn off WEP and then attempt to connect; both units worked fine.

I then changed the WEP key from the hexadecimal series of zeros to a pass-phrase and then attempted to connect; the new notebooks worked fine. I tried various other encryption schemes, such as WPA, and the notebooks were able to connect.

All of that indicated that the problem only existed when hexadecimal zeros were used with 64 bit WEP. I next tried 128 bit WEP with 26 zeros and guess what, both of the new notebooks exhibited the same problem as with 64 bit WEP. Now as far as bugs go, this is very, very minor, but I can't blame the problem on operating systems, other hardware and even using a security scheme that is far from recommended. So there must be something unusual with Centrino 2 and WEP using hexadecimal 64- and 128 bit keys populated by zeros.