WRT54G and Tomato Firmware, a New Chapter...

Readers will recall that I was running an ancient Linksys WRT54G v1.0 router with the excellent Tomato firmware to move all the data packets into and around the home network here at Casa de' G-98. I'd modified the router a smidgen by installing heatsinks on the internal ICs, adding a squirrel-cage fan, and placing a Linksys WSB-24 downstream to boost the 802.11g signal strength a smidgen. It worked flawlessly, and logged 430+ days of continuous use before I recently unplugged it.

Why, then, did I unplug it? Charter Cable performed a free upgrade from 20/2 to 25/3 service earlier this month, and I wasn't seeing anywhere near that kind of throughput on the LAN side of my network. I was lucky to see maybe 15/500k on a good day even before the free Charter upgrade, which left me scratching my head for answers. I attributed a lot of the slowdown to heavy usage of Charter Cable in my residential neighborhood, but I figured I could still see decent speeds in off-peak hours. Such was not to be the case, and I considered downgrading to Charter's 10Mbps service if I wasn't going to realize the speeds they promised. Then I received an email from Charter, telling me I needed a new DOCSIS 3.0 cable modem, please authorize their shipment of the modem to my household to better take advantage of the speed upgrade!

Well, of course I was all over that like white on rice. The new Ubee modem arrived, I installed it, and then went through the steps of having Charter provision it through their website. Once it was flashing all the lights in a happy pattern, I ran a speed test. Better, but not great. I saw 18/2, sometimes peaking at 20/2.5, but nowhere near the 25/3 advertised. Hmm...

Drilling through all the myriad posts over at http://www.linksysinfo.org/ led me to believe I was simply asking too much of the 125Mhz processor and lesser memory capacity of my early-model WRT54G v1.0. Comparisons of WAN-to-LAN speeds of all the home routers showed that the Linksys WRT54G-TM was the one I wanted, with more memory and a faster 200Mhz Broadcom CPU. I bit the bullet, bought one already flashed for Tomato 1.27 on eBay, and waited for it to arrive. Once in my hot little hands, I installed heatsinks on all the ICs that looked like they could benefit from additional passive cooling, placed the router in my basement comm rack, and let fly.

Voila'! 25/3 speeds now came up in the speed tests, although prime time network usage in my residential neighborhood can still put a crimp on that. Overclocking the WRT54G-TM to 225Mhz let the speed tests spike closer to 30/3, so for now I'll hold at that vs. overclocking to 250Mhz. (The router does feel noticeably warm to the touch at 225Mhz) My trusty WRT-54G v1.0 is still around, but it's getting transferred to my sister's home network to run on their Charter 5Mbps service, where it won't have problems compared to what I was asking it to do. In the meantime, here's what the Comm Rack looks like now, I fully expect to see another 430 or more days of uninterrupted service from the "new" router:

And the whole thing still looks pretty cool with the lights turned off, too!

Tomato Firmware - 421 days on a Linksys WRT54G

This is just a short check-up on the whole Tomato Firmware and Linksys WRT54G relationship. To date, we've had 421 days of non-stop operation on Tomato v1.19 Firmware. That's 421 days without rebooting, without loss of network connectivity, without loss of Vonage VoIP phone, etc.

That's pretty darned good.



The physical configuration of the system has remained pretty much the same as it's been for the last couple of years. The router is buried under all the other peripherals, but it keeps chugging along:


Not too shabby, in the general scheme of things!

Tomato and Linksys WRT54G - 303 days and counting!

I've stayed with Tomato v1.19 on my older Linksys WRT54G v1.0 router, but so far, no problems. The router is on a 400VA UPS, so as long as the power doesn't drop for longer than the UPS can feed the little router, I have no doubt it can go another 300+ days or longer!

Linksys WRT54G and Tomato Stability Update...

For all the folks who have been hitting this blog via Googling the "Linksys WRT54G" and "Tomato Firmware" search terms, thanks for the traffic. Sitemeter says you're popping in here from places far and wide, and fairly often, too. Hopefully, you're getting your questions answered - if not, just ask! Tonight's blog posting will offer another answer to a question that I'm certain is being asked.

Playing around with this remarkable little piece of hardware, I've been more than just a little intrigued with the capabilities offered by the router and the aftermarket Tomato firmware.

In fact, one of the main reasons I migrated to Tomato was that particular firmware's reputation of excellent stability. Some will recall that I had to install a timer switch in my previous Linksys WRT54G v5.0 configuration, because it routinely experienced the dreaded "5 day" lockup problem. Others more technical than I related that it was due to connections being cached, and when the buffer filled up, the router locked up. My fix was to have the router reboot by setting the digital timer to simply turn it off for one minute every morning at 5:00 AM. That worked, but seemed a Mickey Mouse type of solution at best.

The Linksys WRT54G v5.0 and later have a reduced internal memory capacity and therefore don't do well with aftermarket Linux code, save for maybe the DD-WRT Micro installation. Truthfully, I wanted more functionality and performance than that, so I kept my eyes open for versions older than v5.0, as well as the special WRT54GL and WRT54GS variants that had the larger memory and Linux compatibility.

Long story shortened, I found an early WRT54G v1.0 variant for fairly cheap (see the picture below with matching Linksys WSB-24 and Linksys 8-port switch nicely stacked on top), and proceeded to modify it with heatsinks on all the larger integrated circuits, as well as the internal Mini PCI WiFi card. I've since also added an external squirrel cage fan to the router's rack to move a steady stream of air around and through it as an added measure of protection from overheating.

Once the hardware mods were done, I upgraded the factory Linksys firmware to DD-WRT. While it was neat and provided oodles of functionality, it was slow, and made the little WRT54G struggle to run it. Maybe it was a setting or two in the menu that I didn't toggle, but the hardware and firmware combination just didn't seem too happy together.

Then I found Tomato, and immediately set about to change the DD-WRT installation to Tomato v1.11. All I can say is, "Wow!" This was exactly what I was looking for to make my WRT54G get up and go, and did it ever! It goes like a scalded cat, I kid you not. Now, since that time, Jon has released several updates to the firmware, culminating at present with v1.21. I've purposely stayed with v1.19, because I'm not too familiar with DNSmasq theory and application. I may switch to it later, but darned if what I already have isn't working just fine, so I feel no real urge to fix it or otherwise complicate a wonderful thing.

Just how fine? Put it this way - about the only time I ever reset my WRT54G v1.0 is when my ISP does something goofy with their network, and my cable modem goes wonky. Sometimes when that happens, I have to cold boot the cable modem, and I get a new IP address assigned as a result. Then I also have to reboot the D-Link DI-102 Packet Optimizer (another really neat and useful gadget) and the WRT-54G. Otherwise, like that damned Energizer Bunny, it keeps going and going and going...

For example, had it not been for a cable modem hiccup, tonight's screen capture from the Tomato status page would show something longer than 37 days of continuous operation. However, I'm not complaining very much - 37 days is much better than what I was experiencing this time last year, and I'm confident that it's by no means a record.


So, the next time I post about the hardware/firmware combination, I have considerable confidence it'll be somewhere after the 60-day mark. Jon did his part quite well, and I hope I've done mine - we'll see.

Stuff gets hot!

Truthfully, I can never leave well enough alone. I'm sure there's a medical description for such behavior, but it's best described as, "If something's not broken, keep fixing it until it is broken".

Such is the case with my household communications rack. Upgrading the router to Tomato firmware was the first step in what would become a vicious circle. With the firmware allowing an increase in transmission wattage, I was getting all sorts of happy trying out my Dell Inspiron 6000 notebook's 802.11g connection further and further away from the router. The Linksys add-on 7dBi antennae made things even better.

Could I just let that sit? Hell, no! So I found one of the older Linksys WSB-24 signal boosters, originally intended for 802.11b routers and wireless access points. Since 802.11b and 802.11g both transmit at 2.4Ghz frequency, it was no big deal for me to move the big 7dBi antennae to the Wireless Signal Booster's output terminals, then feed the RF output of the Tomato-modded WRT54G v1.0 router into the RF inputs of the WSB-24. Egads! Linksys doesn't want you to do this, because the extra output power more than likely violates an FCC rule or two. Honestly, I'm far enough out into the sticks that nobody's going to know the difference, and unless they have the correct WPA code to sponge signal off of me, they'll never get to take advantage of the larger wireless cloud I created anyway.

And larger it is. I'll say it's considerably larger, with excellent signal strength and connectivity no matter where I am in the house and also most parts of the property. That's what I wanted, and that's what I got.

However, when you stack all sorts of electronic devices in close proximity to each other, with each widget sucking a bit of wattage and converting it to heat as a byproduct, you have to pay the piper eventually. It may look all nice and tidy, but when things start cutting out on a recurring basis, chances are you overlooked something. The comm rack got hot, in other words - way too hot.

The fix? Sometime around Halloween and Christmas, all sorts of cute, inflatable, lit-up lawn ornaments go on sale, looking like pumpkins and snowmen. Inevitably somebody buys one, runs it for a short while, and discovers they tear open and generally refuse to stay inflated. That's too bad for the original lawn ornament owner/sister, but in my case, the little squirrel cage blower with low-voltage power supply came in darned handy. These things are really small, fairly quiet, and move a goodly amount of air with a minimum of fuss. I positioned the little fan underneath the Vonage Motorola phone voice terminal on the far left, and it now blows air to the right through the Linksys router/booster/switch stack and on towards the SnapServer 1000 and cable modem/packet optimizer stack. There's a faint "whoosh", and I can keep my hand comfortably on any component without feeling excessive heat.

That's somewhat of a "ghetto" fix, I know, but it kept everything neat and tidy, and we're no longer experiencing drop-outs caused by widgets overheating. That's the important thing, and it allows me to continue fixing perfectly functional items until I break them once again!