Blogs What to do when hardware goes wrong

Never buy cheap garbage, and add 40% to the wattage for PSUs that BT recommends in their roundups.

 

Many years ago my HTPC would randomly blue screen. Sometimes I would need to hit the reset button a few times to keep it from blue screening at the Windows start up. I wrote down all the error codes and began my Google hunt. Eventually I came up empty handed and figured I could just deal with it. 6 months later it became more of an issue and again I looked without luck. One day I was doing my Google hunt and came across a similar issue and the fix was that an IDE cable was bad. So I started swapping out IDE cables without much success. It wasn't until I replaced the IDE cable on the CD burner (that should tell you how long ago this was) that it fixed the problem. So for over half a year of dealing with blue screens it turned out to be a $.25 IDE cable going to the optical drive. That to date has the been the most oddball fix I've come across.

 

The very 1st thing i was taught as a service tech on electronic test equipment back in the late 80's.

1. check the bloody power supply 1st -ALWAYS

no point looking anywhere else for a fault, as a single dodgy voltage rail will skew the working of something else.

ie a video card that might be a bit off can be replaced, but if the 2nd vid card is more "tolerant" of the dodgy voltage rail, you'll incorrectly assume the fault cured.

 

Good article, reminds me of when I was just starting out as a techie, my first build just wouldn't start properly. It appeard to be POST-ing but nothing ever came up on the screen. took me about a month of fiddling and a new motherboard until I discovered the VGA to DVI adapter I had was flaky.

My greatest realization over the years has been to think of the computer not as just the hardware components, or even the components and their connections, but as the truly complex beast that it is. Components have several pieces on them, and each of those can fail independently at times, learn what everything does and you can pretty easily sort out what's causing the problem.

 

Just earlier this week I had a bit of a freakout, thinking I'd killed a Dell at work. Normally not a big deal, but I had to perform a bit of surgery on the case and it would have still been under warranty (had to pull the drive cage and swap out the PSU to put in a discrete VGA card.) It started giving a "power switch cable failure" error on boot...then the power button stopped responding completely. Turns out the bizarro 5-pin plug had popped off the headers completely...and the card was way too wide to fit the case, so had to swap in an older, smaller card, meaning I didn't have to spend three hours butchering the case in the first place. Stupid Dell

 

I'm not so sure random crashes are rare nowdays but it could be that i have a magnet for broken PCs. Take for example what happened with intel motherboards. Although they modified the chipset to solve the issue with sata ports, i think sandy bridge is very weak when it comes to memory stability and it caused problems for lots of people (yes, i'm one of these pelople -_-).
Anyway, IMO the most important thing when you're fixing a PC is to find a way to reproduce the issue or you won't know if what you did was effective. I'll give you an example. I once had to fix an almost new PC that suffered from random catastrophic failure (data lost, SO lost, BSOD...). The funny thing is that every time i re-installed the SO it would work like a charm no matter how long or hard i stressed it but, after returning it, it would broke again for no aparent reason. As i couldn't find a way to rerpoduce the problem, I spent 2 whole months looking for it and it happened to be the damned memory. I found that If i were to power up the PC after 2 days of being completely off, the memory would show lots of errors (checked in memtest). It would work fine after a restart, though. It was not until i left it alone for more than 2 days that the error wold come back again... and that was why constant stressing was ineffective as a test. Once i could reproduce the problem it was easy. I should have replaced the memory first of all things but i didn't have a spare kit at the moment T_T

 

cry...

 

Exactly the same thing happened to me. It even passed a PSU tester on every output. Took weeks before we narrowed it down to the PSU.

I know these things can happen to any manufacturer but by god I swore that would be the last Be Quiet PSU I ever bought.

 

just lost an R4E ..... isn't it ironic if you hold shift whilst typing 4, it is $

 

Before you do a build for family or friends do consider how much time you have to support it. If you are not going to be on hand that often it might be better to find a Dell or similar to meet their needs and makeing time to get it setup properly with the AV and any other junkware replaced/removed.

 

when i doubt, trow it at a brick wall

 

I am very lucky since one of my friends is a computer expert. Whenever my computer has problem, he can handle it well.

 

I understand completely. I won't go near Gigabyte motherboards no matter how good the reviews are. Only takes 1 issue to put you off.

 

I had the very same problem with the same PSU - spent 6 weeks problem testing to find out it was the PSU at fault - I yelled at antec directly, directing them to my forum posts about the issue and laying out all the testing I had done.

They responded by sending me a replacement, AND one of their 3d sound systems by way of apology, so thumbs up for their customer service (http://www.amazon.co.uk/Soundscienc...T0PW/ref=sr_1_1?ie=UTF8&qid=1350501375&sr=8-1)

Give this guy an email:

Stefan Geler
Customer Support & Service Coordinator

T +49 40 226 139 22 | Skype antecsupport.eu
Fax +31 10 437 175 2 | E-mail sgeler@antec.com

 

What do I do when hardware goes wrong? **** myself, say "Oh ****" a lot, and leave it.