A while ago, I got a job as a video editor for a video gaming site. This job required a system with a lot more power than the one I had, so my then boss shipped me one they had previously used for a few years for the same function. It was a bit of a clunker Pentium D, but it performed a lot better than my own rig, and I also had to pay for it as if it were knew, because of it's outrageous video card (8800 ultra).
For a while, things went well, it could do the work in a reasonable amount of time, and heck, I had a dual core machine and all I had to do was work for free for a few months. Well, not the best idea I ever had, but that's besides the point. After about half a year, it started doing things. It would on occasion freeze up when playing Team Fortress 2( and no other game besides that), and sometimes, when I tried to start it in the morning, at 5AM, when my work day started, it sometimes wouldn't boot.
All the fans would be running, all the lights would be on, but until I turned it off and turned it back on again, (like the people at The IT Crowd thought me), it wouldn't boot. I didn't think much of that, mostly because I was afraid the boss would make me pay for repairs, and it continued to work, in spite of that strange behavior.
About a year after I started the job, actually, a year to the day since I got that PC, something else happened. I was playing League of Legends, having some free time on my hands, since the week before I had quit my job. As I was doing my best to bash people in the head with Mordekaiser, suddenly, the screen got corrupted, and the number of frames per second started to drop. A few restarts later, no game would run for more than a few seconds, before the screen became corrupted. I figured it was something to do with the video card, so I left it alone for a day. The next day, I didn't start any games, I just opened a video file, I was watching it, and then, the screen got corrupted again. It was definitive, the video card was borked.
I tried to take it out, and see what was wrong with it. Nothing was evident, since being a gigantic power guzzler, it had a huge plastic case, that covered the cooling elements. I popped it back in, and tried to start the PC again. It went through it's "won't boot" phase again, but this time, no matter how much I tried, it just wouldn't start. Tried for a while, and just gave up on it, since I couldn't afford a new video card, and I was on vacation. But then I got another, similar job, which sort of required a better PC. For a while I struggled with ye old PC as much as I could. I had adapted well to not being able to playback in real time anything I edit. Then, a miracle happened, I got a brand new Radeon 5850 from the "Your spell in Elemental Contest", that Stardock hosted while ago.
Things were going great, I placed the video card on a shelf, to admire it, and the day after my birthday, full of hope, I installed it. But the PC still wouldn't boot. Tried everything I could, and still nothing. A few weeks later, I passed it on to someone who had experience with fixing PCs, not just braking them. Sometime later, he brought it back, sort of working. He couldn't figure out what was wrong with it, after a while it just started to work. He brought it back, and it still wouldn't boot, guess it didn't like me. But after a while, it worked.
Since my second PC got fried over night, I developed the habit of unplugging it when I'm not using. So it turns out that after I plug it in, and switch on the PSU, I had to let it warm up for a few minutes in order to boot directly. I could try and start it up without that, and the fans were on full power, the lights were on, but no booting. If I did that about three times, the third time, the fans would slow down, the lights that didn't need to be on went off, and it booted up.
After a few more months of this, one day, while it was more or less idle, it stared to reboot. It happened once, I didn't think much of it, problems happen. It happened twice, in the same day. Then again, faster, more frequent, and then would just crash the second it got into windows. I managed to see it was somehow related to the voltage, well, at least it seemed to. CPU-Z showed some wild fluctuations in the CPU voltage, once the restart coinciding with the CPU slowing from 3.2 to 2.8 Ghz, in order to conserve power. It does that by default. I booted it up again in the BIOS, the voltage was constant, everything was constant, and it was stable. It was sort of stable even in Safe Mode, though I didn't do a proper reliance test. I decided to give it a rest, and I tried starting it the next day, and surprise surprise, it wouldn't boot. Fans on full power, all the lights, even the once that shouldn't be on, are lit, the keyboard gets power for a second when it boots, the mouse doesn't, neither did the CD Drive, and it's been doing that for two weeks now.
I have no clue what's wrong with it. The guy who sort of fixed it the fist time said it was probably a faulty connection in the motherboard, but I still hope it's something that I can fix.
If you're wondering, it's a Pentium D 935, on an Intel D975XBX2 motherboard, with 2GB of RAM, an SATA Seagate HDD, and an OCZ 650 PSU.
The best assumption on what's wrong with it that I have so far, is that it found out I plan to replace it later next year, and decided to throw a fit.
Has this ever happened to you?
the things I would suggest checking are
1 unseat and re-seat the RAM Sticks
2 unplug & re-plug all the power connectors
3 unplug & replug the expansion cards (including the video card)
if these tests do not help, then my first suspect would be to look at the top of all the capacitors on the motherboard to see if ANY of them have a domed end (they all should be flat)(if you find any capacitors that are domed, then the only recomendation is to REPLACE the motherboard),
the second would be to gently try rocking the cpu fan heatsink(if it rocks then the cpu is possibly damaged from poor cooling as the heatsink was not correctly pressed down during assembly),
and if all of theses tests and checks do not help then I would suspect the power supply as they do age(getting weaker) and the system would be trying to draw about 400-450 watts and the aging factor for power supplies is approximately 10% of initial capacity PER year.
hope this helps
harpo
It sounds a lot like a bad power supply (not delivering correct voltage) but could also be bad memory/cpu (which wouldn 't handle voltage right either) or bad capacitors.
Did you notice when you monkeyed around in it if the heat sink cooling vanes were clogged with dust?
Did someone overclock this some time in the past?
I did have one board that got a miniscule amount of rust (not even visible) and unseating and reseating cards and memory scraped it clean enough to clear up.
Slim chance of corrupted bios--but that could happen because of bad power too.
I did service a machine at a house once that had really bad power lines--lots of light dimming, etc.--and it constantly messed up their PCs until we got a backup-regulator power supply.
unplugging is a good idea however replugging would best be done after everything is unplugged apart from what you need to get it running including removing graphic card & using onboard,for my issue i even left mouse & keyboard off as well as ram,then installed one stick rebooted,then turned machine off added other stick,rebooted to see if POST screen showed full ram,replugged,mouse,keyboard,1 dvd drive & then after another reboot 1 hdd ,basically only adding one item at a time after a period of making sure all was running well after each part was reattached,however in my case the hdd had no OS so installed it,then added the rest 1 at a time .In your case perhaps leaving most unplugged & running a linux live distro & see how the comp performs,that includes leaving hdd off
To me, it sounds like the PSU is faulty, since you have power to the board and components and the lights come on and that it does boot occasionally. Judging by your experience karma isn't on your side but I would advise probably not playing games on a work machine. Also I advise in power surge protectors, They have saved my equipment countless times. The best way to test is strip to bare bones, If it has on board graphics just the cpu/heatsink, memory and power. See if it boots and then from there add each component and test etc..
Throw it away and get a real computer that is NOT someone's hand-me-down.
+1 what Jafo said
The most notable thing about the Pentium D is it runs hot and is prone to early failure due to overheating.
And, what do you know the same is true of the 8800 based card.
That machine should be scrapped.
Probably right... it'd cost a lot more than it's worth to repair it, I bet... but even with repair, you'll end up with the same problem...
Sorry for all your trouble, unacomn. Oh yes... and that employer was rather exploitive.
BDBF...you missed the part where he replaced the 8800 with a Radeon 5850.
Depending on which power supply you have, it may still be under warranty. Some of the OCZ 650's had a 5 year swap out warranty. You should check to see if yours is still covered.
While I agree that unacomn should replace it with a new system, with times the way they are he may not be able to afford to.
I have a similar situation that I am in the process of diagnosing. In my case, it simply won't power on. I am down to motherboard, processor, and/or Power Supply. I built the system myself and a friend brought over a PSU and a processor that I can use for diagnostic purposes.
What happened is this: I had relegated it to the secondary desktop status up on the second floor. It had been functioning smoothly. I had it turned off and sitting for a while. When I went upstairs to use it, It would not power on. Everything else connected to the UPS powered on. That eliminated the UPS. I unplugged the computer from the UPS, turned the PSU power switch off, waited, then turned the power switch on, and plugged it back into the UPS. The keyboard LEDs lit up briefly, like normal. I pressed the power button and nothing. Needless to say I was not happy at this.
Next I replaced the power cord and same thing. That is eliminated. Time to open the case.
Let me go into a bit of back story with this computer. I had spent 11 months in nursing home for treatment of aggressive anal cancer, surgery, and recovery afterward. After I got home, it was running smoothly and I bought some Blu Ray players and a 22 inch digital TV. Plus I bought Vue 9 Complete. And Asus P7P55D-E Pro motherboard, Intel Core i7 860, and a 1 GB nVidia graphics card, with the intent of building a new system.
Here is where things get interesting: first the system HDD went bad and I replaced that. Then the graphics card died and fortunately I had that 1 GB nVidia on hand. Then my only SATA HDD died. URGHHH!!! I got it up and running smoothly again and things proceeded normally here for a while. Now I upgrade the memory from 8 GB DDR2 to 16 GB DDR2, replace two PATA HDDs with two 2 TB SATAs, and add my internal WiFi card. No problems. Then I have it sitting up there for a while turned off. In the meantime, I have purchased a super nice Core i7 notebook and have acquired the rest of the parts for the Core i7 desktop and have it built and running. I decided to go upstairs and go online up there and WTF?!? That 16 GB DDR2 cost $320 at Newegg (ouch!). I have tried everything except replacing the PSU, CPU, and graphics card.
Keep in mind that when I press the power button, NO fans spin. Nothing. I seriously doubt that it is CPU overheating, since I have a huge heatpipe/heatsink and fan on that CPU. Only 2 possibilities come to mind: bad PSU or lightning.
Na CarGuy I saw that part but with two sources of excessive heat in that machine everythingwould say to me don't waste money on "fixing" it.
He can still see about how to scrap it including swapping out the old power supply. It would be cheaper to buy a used machine with a 90 day warranty and use the scrap parts from the old one if cost is the deciding issue.
sarissi, your suspects ARE my first choice from your descriptions, and most of the better(dearer) power supplies have over voltage protection to reduce/prevent damage from getting to the rest of the system,and it could be a ground imbalance between the motherboard and power supply (the quickest test and fix for this is to disconnect the power supply from the motherboard for a couple of minutes then re-connect
but regarding the no power up, my first check is power cord, then power switch on back of power supply, then test power supply not connected to any part of system, then test connected to system essentials ie motherboard & video card if card NEEDS extra power, then add ONE extra item each test until problems occur.
What I did re the HDDs, is replace 2 PATAs with the 2 SATAs. The only extra thing is the WiFi card and it had been in that case, with that PSU before. However, the PSU is old enough to have seen its last days. My first suspect is the PSU since it had been running fine. Then it sat with the power off for a month or so. Then nothing. The video card is not one that requires extra power. Since I had done a normal shutdown, it should have powered on when I pressed the power button.
The PSU is a 550 watt, by Antec (I think). Since I built it for doing 3D CG, I used the best parts I could afford. However, it has been through several motherboards, and is the oldest part in the case. Thus my Primary suspect and I hope that it is the problem. I built that system back in 2008.
The PSU I have on hand is a 500 watt. Good enough for minimal connections to see if the 550 is the problem. I will have to do that before Thursday.
I hope it is the PSU, since it will be replaced with a single 12V rail modular, instead of the rat's nest cabling I have now.
OK, people. If you have a system that's been left OFF for some time your FIRST point to check is as simple as the CMOS battery.
At worst it'll cost you $2 and your problems will likely be over.
Had the exact same issues with a computer once.....refused to boot....lights would come on but no-one was home....
One button battery and it was game over.
Oh, yeah....fogot.....I went through all the 'usual suspects' thinking...PSU, GPU, loose ram....fried MoBo....loose cables....communist plot....Microsoft hated me....
It was a flat battery.
Thank you for all your suggestions, I plan to take it apart later this week, but just in case it is the battery, I'm going to try that first. That actually makes sense, it didn't need to warm up, it needed to recharge.
At best, I hope to extend it's life for a few more months, until I can get the budget for a new one, a better one!. Well, just about anything I get will be a better one. Failing that, I'll wait until AMD improves the singlethread performance of it's new FX series, or until Ivy Bridge. Since I'll probably end up playing for it for the next 5 years, like I did with my third PC, I want to be sure it'll last that long.
Update.
Sad to say, it wasn't the battery. Wish it was. But at least I got to ride an escalator in order to get it, that's always fun.
Up next, taking it apart and going over it with a magnifying glass, a prospect that sometimes I find frightening, not because of my phobia of magnifying glasses, but because whenever I take apart an electronic device, the chances of putting it back together in working conditions are slim. I excel at being incompetent at tinkering with electronics. Which is ironic, since I'm not entirely sure what ironic means.
If you're going to tinker, make sure it's unplugged and it's battery (if a laptop) is out. Don't touch the RAM wafers without being grounded first.
You know, there are PC Technicians all over the place that could look at your computer for you and tell you whats wrong instead of taking a chance a totally destroying it over something that could be minor.
and the next MOST frustrating part is the power supply, and I have just had a customer computer that was behaving like the hard drive was at fault, but under testing with a NEW hard drive was still showing the same fault, so tested the computer with a new power supply and the system had fixed with both the new AND previous hdds, and do not forget that power supplies DO age and lose approx 10% of their capacity per year, so if the computer is several years old the cause MIGHT just be the power supply no longer being able to supply ENOUGH power to the computer, and get all sorts of intermittent faults like you have seen.
this is a bit weird.. i replaced my power supply a few months ago (the people checking it said the power supply was dead... as opposed to switch or motherboard)... and now sometimes hitting the power on switch does nothing. (twice? so far?) wonder if i simply didn't hit it properly or what, as a 2nd hit boots up just fine.
This is a bit old, but I'm digging it up because of some new info. I stuck an old PSU in a stripped down version of the PC today. It can't run with it since it lacks certain power cords(like the ones for the GPU), but the essentials were there, and it sort of booted up. The fans powered up properly and then slowed down, only the lights that needed to be on were on, and it looked like it would work. The most surprising part was that absolutely nothing exploded, which is a first for me when it comes to taking apart electronics, and I wasn't once electrocuted.
I'll do another test soon to check that it wasn't a fluke, and if all goes well, a new PSU should solve the problem. Unless of course the fans and lights didn't go nuts because this PSU is an ancient 400W model, which may not actually be able to power them properly.
Don't mean to dredge this up again, but it was 100% the PSU at fault. I'll be testing FE on a dual core from now on.
So if you ever have a PC that has the same problems mine did, change the PSU.
And they lived happily ever after.
The end.
I hope there won't be a sequel.
Thank YOU for letting us know the results unacomm
There are many great features available to you once you register, including:
Sign in or Create Account
