Why do my consumer WiFi routers keep failing?

Question

Over the years, I've had many of the wifi routers used in my home fail. ? Is there any reason. Since there are no moving parts it seems odd they would fail so often. I've probably gone through four in about 8 years. I've had Linksys, DLink, and Netgear devices that stopped working.

Just this weekend, one stopped working. It is a linksys 54gs that I use as a wifi access point (turn off DHCP, don't plug anything into the WAN port). There has been no lightning recently, no power outages, no changes in the house. It has worked great for a year and a half, then this weekend wifi connections started to drop for hours at a time. Sometimes sporadically, the wifi connections would work. I have Wifi analyzer software on my phone and it showed a strong signal and did not show any other networks near the same channel. When I stood right next to the device, I could not get my phone to connect (or my Wii or my laptop). However a couple of hours later, they seemed to connect fine. Perhaps there is some interference I can't find, but nothing changed in my house and I didn't see any signals from my neighbors. I tried power cycles and resets, but no luck.

Any explanations?

Answer 1

Electrolytic capacitors don't last forever. They are used in the power regulation part of the circuit and usually are the components that die first. But still, they should not fail in only a couple of years. There was a scandal a few years ago where a capacitor supplier produced faulty capacitors for a few years. I had a motherboard that died because of this, and maybe you had some routers that had these faulty capacitors.

There could be other reasons. In my personal experience, I had a router go out on me one summer because the AC/DC adapter died. It was a particularly hot summer, and some components in the adapter (maybe some capacitors) slowly went bad. With no load, it was able to reach its target of 5.2V, but as soon as any current was drawn from it, the voltage would drop and the router would behave unreliably. It was not obvious to me that this was the problem until it completely died and I bought a universal power supply and set it to 5V, and voila, it worked like new.

It could also be heat. Electronic components work more reliably when they are cooler. You should make sure that the router is not in a place where heat gathers.

Answer 2

If Wifi Analyzer shows a strong signal, it's not the radio and it's not interference - it's the device processing the signals. Some things to check for:

• Make sure you have the latest firmware
• Make sure the device isn't overheating / is stored in a cool place
• Make the sure the connection to the upstream device is working
• Try just restarting the device

I have a router that's starting to go as well. To prolong it's life, I put it on a $3 christmas light timer. This has worked like a charm - it shuts down for 15 minutes in the middle of the night each day when I'm not using it, and this has greatly increased stability. Of course, my router is 6 years old going on 7 rather than just 18 months, but the principle still applies.

Answer 3

I make the observation that consumer grade plastic "Made in China" routers like Linksys, D-Link, Netgear, etc. eventually all fail after 3 to 5 years too. Yeah with no moving parts it's hard to see why.

There are no electrolytic capacitors visible on most of them - they are entirely reliant on the 5 volt power transformer.

When they start to fail, you notice your internet goes out, so you power-cycle it and it works again...for a while.

I have tried updating the firmware, but it still acts flaky. I have tried swapping the firmware to DDWRT instead - still flaky.

It looks like something in the hardware is gradually failing, causing uncorrectable bit errors that the router CPU cannot handle. Routers boot from flash memory - they typically boot a Linux kernel and run entirely out of RAM.

Therefore I suspect that it's the RAM that's degrading over time - the frequency of device failure increases over time (i.e. the number of times you need to reset it, suggesting that the RAM bad cell count is increasing. Rebooting reloads the runtime code into RAM but as soon as it attempts to read faulty data from a RAM cell it crashes again.