A couple of weeks ago, i got a new graphics card and some RAM. The new graphics card replaced what i correctly guessed was a dead graphics card. I installed the RAM and that was that, or so i thought. Since then, i’ve been having some fairly random problems…
PC doesn’t like mornings
Initially, there were no problems. The graphics card was running fine, the RAM was running fine and in fact just to make sure everything was working properly, i ran extensive tests & kept a close eye on temperatures on day 1. The tests revealed no errors & the temperatures of everything remained stable.
However, the next morning, i got an avalanche of BSODs (blue screens of deaths). The errors they were reporting weren’t consistent and i rebooted about a dozen times to try and change BIOS settings. MEMORY MANAGEMENT, IRQL_NOT_LESS_OR_EQUAL and BAD POOL HEADER. Eventually, after multiple reboots, i got in to windows and the PC ran as normal. I put it down to improper shutdown or corrupt windows files (due to the old failed graphics card). But this wasn’t a one off.
It made no sense whatsoever and this happened daily for the past week or more. Even once i’d be logged in, i’d deliberately shut down, restart & disconnect the power to try and replicate the problems i’d faced earlier. No joy. The first couple of days i attributed it to incorrect shutdowns or corrupt files and i reinstalled windows which i thought worked, but it didn’t. That ruled out a driver / software problem.
I then began to think some of the RAM modules were faulty (despite the fact they’d all been tested extensively on day 1). Other culprits that entered my head were the power supply & the possibility of a ‘new’ bad graphics card. The power supply was something i thought carefully about when getting my new graphics card. The new graphics card consumed less power than the old one so i figured it would be ok…
Testing one thing at a time
Yesterday morning, the same thing happened. Blue screens of death, no sign of booting and infinite restart loops no matter what i tried. “ENOUGH” i said to myself. I was prepared to spend all day taking the machine apart and test every component individually and collectively until i could pinpoint the problem. I was convinced it was now a hardware problem.
I decided to test each stick of RAM, one by one, in each slot, one by one. In total, 16 different combinations & 32 boots (i restarted windows twice with each stick just to make sure they were working properly and i hand’t ‘fluked’ a restart with a bad stick). I was determined to ‘prove’ that *if* i’d got a bad batch of RAM, i could send it back having diagnosed it properly & with bullet proof evidence but i knew that without testing each stick individually, i couldn’t be 100% sure it was the RAM.
Here was my very scientific testing plan… RAM modules numbered 1-4 and DIMM slots (place where the RAM goes) on the motherboard labelled 1-4.
I then booted windows (or tried to) with each ram module in each DIMM slot twice (just to make sure nothing was fluked). Eventually, a pattern began to emerge as you can see from the x’s above… Whilst i was annoyed, i was in many ways relieved because i’d found a major problem, most likely the cause of all my previous two weeks of hassle.
Ruling out the Power Supply
The other *potential* cause of blue screens or the problems i’d been getting is with the power supply. The power supply in my dell is 375w but as i’d found out when researching graphics cards, that doesn’t mean it’s weaker than, let’s say, a 425w PSU (from here on PSU = power supply unit). In fact Dell don’t measure their wattage in the same way as others. The 375w for example isn’t ‘peak’ consumption, that’s a level it can maintain consistently, whereas a regular 425W PSU can reach a max 425W – it can’t maintain that level. Whilst i can’t ‘prove’ this, my current PSU that came with my Dell was (if we go by wattage) about 125w below the recommended wattage for my old graphics card. Not only should that spell danger, it would mean that i’d have very little room to expand the system / add more hard drives etc… and that was the set up Dell built for me and gave me. Would they put an underpowered PSU in a PC? Doubt it as it’d be a recipe for disaster with mass recalls if the PC wouldn’t boot or burned out quickly. It was probably pushing it, but they knew it would cope.
However, the fact i knew my PSU has a label of 375w when my graphics card recommended 500w always left me uncomfortable but because i’d heard others ran the same set up as me with no problems, i felt i’d be ok. However, with these errors popping up and an unstable system on my hands, i can’t afford to not get to the bottom of whatever the problem is. Stuff has to be ruled out and because i can’t rule out the PSU as the cause of some of these problems, i’m getting a new one… even though the chances are there’s nothing wrong with the existing one.
Another bit of research showed it wasn’t uncommon for my Dell PSU to get a bit flakey after a few years. With hindsight, i was thinking it could also accelerated the death of my old graphics card but for that reason, i can’t take any chances now that i’ve a new graphics card. I’ve ordered a 600w OCZ PSU which has a pretty good reputation and 3 year warranty, plus it doesn’t look like an octopus (all sorts of cables dangling out of it). It’s a ‘modular’ PSU so instead of having multiple wires and crap coming out of it which will never be used and which clutters up the inside of the case, you can just connect the cables you need. It makes for a tidier PC inside the case.
Yesterday, along with testing the RAM, i basically ripped my entire PC apart and took out the power supply, left it sitting for a few hours then i powered it up just by itself, out of the PC, using the paperclip trick (i’m learning a lot). It turned on ok, but then again it wasn’t under any load… today, it’s been running perfectly in the PC so i’ve no reason to think there’s anything wrong with the power supply however if stuff goes wrong again, i’d be kicking myself for not replacing it sooner.
Morning after the night before
This morning, for the first time since i installed my RAM & graphics card, the PC booted up straight away without any blue screens. Interestingly, i’d left out the ‘bad’ RAM module that i’d found yesterday and just ran the PC with 6GB RAM. So this would indicate the problem is indeed down to that faulty RAM module and not the PSU, not the new graphics card, not anything else.
I decided to try the bad RAM module again in the PC today, ‘for the craic’…. assuming it wouldn’t boot. It didn’t boot so i can say with 100% certainty, that RAM module is corrupt and i’ve got sufficient evidence to prove it. It’ll now be going back to where it came from and i’ll be getting a replacement. Interestingly, if i’d have wanted a refund, i’d have been charged a 15% restocking fee. Not sure if that’s legal but all i want is a replacement & i’m getting that at no cost so i can’t complain about anything other than the fact the original module they sent out (1/4 of the RAM i got) was bad.
At this rate of going i’m almost building myself a new PC without realising it. What i am realising is that it makes much more sense to build a PC yourself because generally you get 3 year+ warranty on things. In some cases it’s a lifetime warranty. With a brand name PC, you get 1 year and pay for more. Provided you can figure out what part breaks & are able to take it out / put it back in, you’ve basically got a computer with a lifetime warranty (the computer’s lifetime / until you upgrade which will probably be 4/5 years at most).
I could easily fallen in to the ‘get a new computer’ mentality but that would in effect be saying “there’s something wrong with my computer and i don’t know what it is or how to fix it”. That’s a criminal offence in my view if you’ve got a couple of degrees in IT. I see it as my responsibility to get to the bottom of problems, particularly if it’s my own stuff where i have all the time in the world to work on figuring out what the problem is.
I’ve been pretty unlucky though to have a graphics card fail on me, then order new RAM and get a faulty module. The technology gods are not smiling on me of late but on the plus side, my PC is error-free, clean on the inside & will have had the RAM, graphics card & PSU all upgraded by the end of next week.