Do Routers Die?

Like any other electronic equipment, sometimes they fail.

Nothing lasts forever....

Yep. Lots of ways to kill electronics. My most recent disaster was to place an Efficient 5260 DSL modem on top of an APC Smart-UPS 400. The UPS is a flat and I pile everything on top of it. You can sorta see it in:

the AC power was going on and off for the last few days as mother nature watered the area. The big torroid switching transformer in the UPS was about 5mm from the filter coils in the Alcatel. I'm not sure what it fried but the 5260 is now comatose.

I've also killed off electronics with major doses or RF energy, failing the drop test, sloppy wiring, wrong wall wart, water damage, and my favorite, dropping a 19" monitor on top of the plastic case. However, these are obviously un-natural causes and can be avoided with a little care and common sense. Electronics does die and sometimes deteriorate without my help. Static electricity is always a problem. The epoxy-b packages often leak where the leads enter. Water vapour gets in, and rots the aluminium chip traces. Circuit boards may be soldered slightly warped, putting components under stress. These will eventually crack. Marginal and borderline component rating often cause premature failures. The well known Low-ESR leaky capacitors has cause widespread damage. For example:

So, what blew up and what makes you think it's dead?

No! Routers do not die, however, they do eventually fail to operate

Of course - the law of entropy wasn't repealed for them you know!

Common causes - component failure, power-supply overheat/failure, firmware damage due to bad upgrade, firmware or hardware failure due to invalid update etc. I've had one die due to the power supply failing and pushing out too high a voltage.

Why? One of the most common causes that happens *without* users playing or fiddling is that the caps in the circuits dry out and eventually

*pop*. Caps start dying from when they are made.

YUP, you guessed it; electronics has a 'shelf life'! Normal operation should give years of use, but as you know; sh.t happens. Thermal cycling of equipment speeds up the drying process. Just leaving it on speeds up the drying process. Not turning it on doesn't speed up the drying process, it just drys at it normal rate depending on relative humidity...

One of the major non-truths is "power cycling kills equipment". In terms of the electronics, all caps have 'surge' ratings, and if the equipment is/was designed correctly, turn-on ( even at just before peak ) should not harm the equipment as the surge limiting resistors ( or active devices ) protect the stuff from the inrush current.

Soooo, if you buy electronics and are planning on putting it on the shelve for a *long* time, don't.

later

todh

Yep. With normal electrolytics, it takes about 20-30 years for the oxide layer to deteriorate. I've sorta watched people at the Computer History Museum resurrect some of the old "computer grade" electrolytic filter cazapitors from apparent death by going through the same process as was used to create the insulating electrolytic dielectric layer during initial manufacture. I don't know much about it, can't seem to find anything with Google, and only know that it takes about a week with a programmable power supply. Raising the dead is possible.

Sh.t not only happens, it's often manufactured. Many of the dead electrolytics I find were simply defective. See:

Sigh. There is no electrolyte loss or outgassing involved. I can demonstrate if you want, but that's not the failure mechanism (trust me). What happens is that the insulating oxide layer plated onto the surface of the aluminium plates breaks down or disappears. The common term "drying out" is a misnomer. Were it true, then one could rejuvenate a dead cazapitor with an injection of whatever electrolyte was used. There may be some outgassing for capacitors carrying lots of ripple current and literally boiling the electrolyte, but normal low power use will not cause the rubber seal to blow or vent.

I beg to differ slightly. Cazapitors only conduct current during power transitions. Powering on and off isn't much of a transition, but applying raw AC ripple to a capacitor is. In such a circuit, the capacitor is conducting perhaps 120 times every second resulting in considerable power dissipation during each cycle. That's why cazapitors get hot. The same thing applies to the CPU power filter caps on the typical motherboard. They're trying to filter literally hundreds of amps of ripple voltage at 1.8VDC. They get hot.

Old electronics isn't exactly a great investment. Such things usually end up in my storage area, or in the e-Waste pile. Want some klystron microwave radios and test equipment?

I go by the "you get what you pay for" theory. I've had the cheapie Linksys BEFSR41 routers just "flake out", stop passing traffic for whatever reason, stop giving out DHCP addresses, etc. I also had a bad run of the cheapo Linksys 8 port 10/100 switches. Ports go bad/flaky for no obvious reason. I had so many of those units go bad, I'm giving the Netgear switches a try (they're cheap too but at least they are in nice solid metal boxes )

-- Paul