When you work with computer hardware on a regular basis, building machines and swapping components in and out, you start to see patterns of hardware failure. Generally most hardware is reliable, but you’ll see more failures with components related to movement, power, and heat. Hard drives, system fans, power supplies and optical drives seem to have the highest failure rates, but video cards are failing more often than they used to as they run hotter, faster, and need more cooling. Things you don’t see spontaneously fail very often include CPUs, RAM, motherboards, and cables – sure, they can fail, but they usually require “help” from the owner or local power outlet.
Imagine my surprise when this week I had not one, but two bizarre failures of hardware that I’d never expect to fail. I booted up my Fujitsu P7010 laptop, a little 10.6″ screen job, that has 1 GB of RAM (2 x 512 MB) that I installed over a year ago. It’s run fine since then, but this week when I booted it up Vista wouldn’t load – I’d get a blue screen of death and an error related to memory. Because I had been running this hardware for so long (the laptop is a bit over two years old now) hardware failure was the furthest thing from my mind. I ran the Vista repair process (which is quite impressive) because that’s what the OS instructed me to do, but that didn’t work so I figured I’d run memtest86 and see if something was wrong. Below is a photo of what I saw (hint: red is bad).
I ran the test two more times, each with one of the 512 MB memory modules installed, and isolated it to one of the chips. I phoned Kingston, and was happy they had a lifetime warranty. The warranty process was amazingly fast and painless: no waiting on hold at all, the technician listened to me explain what I did and immediately agreed it was bad memory. A few minutes after that I had spoken to an RMA tech who took my credit card number so they could cross-ship me new RAM immediately, and as long as I return the defective RAM within ten days they won’t charge my credit card. Oh, and they also gave me their FedEx account number so I wouldn’t have to pay for return shipping. You never want RAM to go bad, but when it does happen, you can’t ask for anything more than a company that’s willing to make the replacement as fast and painless as possible.
Ok, so RAM can and does go bad. But when is the last time you heard of a coaxial cable spontaneously failing? Yesterday I was having massive trouble with my Internet connection, and I blamed it on my Netgear WPN824, which has been flaky since the day I bought it (mostly giving me random network failures that persist until I reboot it). I swapped in a D-Link DI-624, power-cycled everything including my cable modem, and assumed everything would work fine. I’ve swapped those routers in and out several time, with a Netgear router thrown in for good measure, because I can’t never seem to find one that’s stable 100% of the time. At any rate, my connection kept failing, so I looked at my SHAW cable modem with a critical eye. I have a 10 mbps downstream connection that’s been amazingly solid for the past two years (it was a bit rocky before that). Sure enough, I checked it and it wasn’t online. I’ve had a cable modem connection since 1995, so I’m very familiar with how they behave and know when something isn’t working right. I called SHAW tech support, he confirmed that it was disconnecting and re-connecting from the network, so I power cycled it one more time, and saw it lock in – and thought my problem was solved. I went back to work, only to lose my connection again. I went to my wiring area and watched the cable modem – it would lock in, all lights lit, then it would lose the network connection and everything would start blinking. And it would repeat that cycle every minute or so. What was going on?
I phoned SHAW tech support again, they said it must be a local problem with the wiring or the modem itself, and they’d have to book a service call – the best date they could do was a week down the road. I told him that if that was the case I’d be ordering DSL the next day – that rattled him, although I was mostly bluffing trying to get more immediate service (although I do still wonder about running cable and DSL together and bonding them together for a redundant and faster connection). I was suspicious of the cable modem rather than the household cable connection, because my SHAW digital phone (basically a private VOIP service) was working perfectly, as was my cable TV. My loving wife Ashley volunteered to drive to her parents house, borrow their cable modem, and bring it back so I could try a known good modem. I did, and it didn’t work. At this point I was starting to get really frustrated, and Ashley said “Did you try changing the cable that goes into the wall?”. I gave her one of those “Don’t be silly, that couldn’t possibly be the problem” looks, but having nothing to lose, I gave it a try – and it worked. What the hell…?
The cable in question is five years old, having been installed when we moved into our house. It’s plain old copper cable – how could it possibly fail? I don’t know how, but it did – and thanks to my wife, looking at a problem with a fresh set of eyes unfettered by assumptions, I’m back online.
Note to self: always check the obvious points of failure, even if you think they couldn’t possibly fail.