I’ve still been having some difficulties with my MacBook and my home wireless networks. When running on battery power, as well as more infrequently when on regular power, my MacBook seems to lose its wireless connection, and will not automatically reconnect: I have to do do so manually, which is a bit of a chore.
I wrote about this before, and as yet have found no resolution.
But the other day when using the network from my bedroom, I noticed that my wireless signal was quite weak, often dipping into the teens as far as quality. I thought that was pretty bad, and indeed, the wireless access point is on the opposite side of my house, through about six walls and one floor, so I thought that it might be time to consider adding a second access point on the second floor.
My current router is a Linksys WRT54GS (version 2) which I upgraded to run the third-party firmware DD-WRT. It has many improvements in QoS and other goodies compared to the official firmware, and was remarkably stable. Despite the fact that have heard that this firmware might be the source of my difficulties with the MacBook, I decided to pick up a new Linksys router as my second access point, and also install and run the new DD-WRT on that box as well. Why? I’ll get to that in a bit…
One of the reasons that the Linksys hardware is so popular is that it was based upon the Linux operating system, and the sources were made available under the GPL. This spawned a huge amount of development, including DD-WRT, which I really like. Unfortunately, the WRT54G underwent a metamorphosis, and they shifted to firmware based upon VxWorks and cut the amount of memory in each box in half. Early on, these boxes were incompatible with third party software, which made them less desireable. However, the “micro” version of DD-WRT (which doesn’t include things like VOIP) will fit in these machines, and bitsum worked out a way of updating the flash appropriately. So, with this handy table in hand, I set off to CompUSA.
Why did I need this printout? Because I wanted to be able to tell which model router they had in stock. It turns out that you can look at the serial number on the outside of the box, and using this table, figure out what version of the hardware is inside. Surprisingly, CompUSA had both v5 and v6 models on the shelf, either of which should work with DD-WRT micro, and for only $49. I hemmed and hawed, and finally decided to get the v5 hardware, mostly with the idea that since it had been out longer, DD-WRT would have had the most chance to be debugged on that hardware. I have no idea whether that was a logical choice, but that is the choice I made.
Okay, I brought it home, plugged it in, it seemed to work. I then went and did some more reading.
I knew that I could string an Ethernet cable to the new box (wherever I put it) and could configure a different local network on it, effectively cascading the routers. But I really wanted the new box to be mounted a significant distance from the old one, and I didnt really want to pull any new cable. I had vaguely heard of something called “WDS”, and looked into it a bit more. It basically allows you to mesh wireless access points together. Each node repeats or bridges traffic going to it to all the other access points in the mesh. The cool thing is that only one machine then needs to be able to access the internet, and then all the rest can. This means that I just have to plug the new access point in somewhere where it can still get a signal from my original one, and it will extend the range.
You can find instructions on how to do this here:
The long and the short of it: it works! I now have two APs “brainwagon” and “brainwagon-upstairs” and can use either one interchangeably to access my network. my signal is around 55 now in my bedroom, which is much better and I should be able to improve it a bit more with some moving around of the APs.
The bummer is that I’m still getting disconnects occasionally from my MacBook. The search continues for a resolution to that problem.