[Previous entry: "HTML O' the day"] [Next entry: "Overread"]

04/10/2004: "Yet more WiFi bullshit"

Says here that WiDeFi claims to be performing the ultimate WiFi parlor trick: simultaneous transmit and receive.

"We can receive and transmit simultaneously. This allows us to support the full rates available to [Wireless LAN]," WiDeFi's CEO James Proctor writes in an email reply to questions. "The way to think of it is as the equivalent of a Hub for time division duplexed wireless systems (802.11 802,16, etc.)."

Now kids, simultaneous rx/tx (on the same channel) is easily a 120dB problem. For the non rfnerds in the audience, 120dB is a 1 followed by 12 zeros. Yes, 1,000,000,000,000:1. A trillion to one isolation problem.

Not in our lifetimes, as the saying goes.

Now, maybe they're not re-transmitting on the same channel. If this is the case, then there isn't much about this that is special. A couple channel filters and two ordinary WiFi cards can do the same thing. If you were going to have a volume play, then you'ld want to put the channel filters in some monolithic IC block in order to reduce costs.

Further, the spec sheet says that not only will the chip only transmit at 10dBm (10mW). Now, most 11b chipsets transmit around 15dBm (32mW). (Some specialized designs transmit 23dBm (200mW) or more.) Even the most lame 802.11g (or 802.11a) card typically transmits at least 10dBm at the top speeds, especially in the proprietary rates higher than 54Mbps.

Further, because the AP can typically afford a slightly more expensive RF section than the clients, (the clients are extremely cost-sensitive) it makes sense to have the AP transmit with a 2-3dBm more power than the typical client card.

So I don't think this "repeater" chip is going to succeed. At least not in its current form.