Linksys WRT54GS - which version?

John Navas hath wroth:

I'll confess to not staying up to date on Super-G versus Afterburner. Let's see what I find:

and DLink use Super-G (Atheros), Linksys uses Afterburner (Broadcom), and SMC uses something from TI.

At an average throughput of around 50 Mbps, the NETGEAR products using Super-G clearly beat Linksys' Afterburner-based offerings, but only when Super-G is running in its controversial channel-bonding mode. As the plot shows, Super-G drops out of channel bonding on a periodic basis to ensure that non Super-G products get a chance to communicate, which lowers its average best-case throughput to about 45Mbps.

What this means is that the improved Super-G uses channel bonding only when it has exclusive use of the channels involved. A bit of personal testing showed this to be the case only on Ch 6 (the center channel). It doesn't listen on the adjacent two channels.

The Tom's Networking URL's are from 2003 and 2004. This is slightly later and still shows a problem:

Page 7 for where the thruput on a Belkin F5D7230-4 droped 94% when operating near a Netgear WGT624 in Super-G mode. The Belkin was setup to hunt for usable channels during the test. It didn't find any.

Change apparent to obvious. Broadcom was more than a little obnoxious with their smear campaign. However, they were initially correct about channel bonding problems.

Digging a bit deeper into Super-G at:

Updated December 6, 2003 A Super-G WLAN running at full speed will seriously interfere with an 11g WLAN running 2Mbps streaming video even at 30 feet. The interference is essentially gone at 50 feet. I'll assume they're talking about interference on the adjacent non-overlapping channels. Yeah, I'll buy that. It's only a problem in close proximity. Add 6dB of antenna gain, and the 50ft becomes

25ft. Add 6dB of antenna gain on both ends, and the 50ft becomes 12.5ft. In other words, with the stock rubber ducky antennas, 50ft is safe. With better antennas, it's not.

No. They're designed to disable Super-G channel bonding if there's anything on the channel. The problem is that they don't seem to be listening (my guess) on the adjacent channels. Only on Ch 6. This is not exactly my idea of non-interference. Locked in channel bonding mode, they generate quite a bit of RF on the adjacent channels.

I only have some spotty experience with both Super-G and Afterburner. In both cases, I turn them off. The problem is the same as with

802.11b compatibility in an 802.11g router. The receivers can only decode the mode if it "scans" for it. If it's listening for 802.11g, it won't hear an Afterburner or Super-G (channel bonded) signal. So, it time slices the receive listening for 802.11b, 802.11g, and enhance modes respectively. I spent a few minutes trying to deduce the algorithm used and just gave up. My current policy is to disable the enhanced modes and run in 802.11g mode only (if possible). For public access systems, I'm stuck with 802.11b mode only, or possibly 802.11b/g compatibility mode. Anyway, it's for this reason that I don't have much experience with the enhanced modes. I always disable them.