Jump to content
Sign in to follow this  
lujate

PC will not stay running

Recommended Posts

I was playing last night, my pc just shut off with no warning. I hit the power button, it came on and had a power light and was hitting the disk drives as expected. Before the display even came up, it shut off again. I opened up the case and used a flashlight to look for obvious flaws but everything looked good. I turned it on with the case still open. All the fans can on and this time I actually got the my Windows desktop. I logged in to the Lobby in order to check on the character I had been playing when it had died. The computer seemed to be running fine. I left it running and went to bed. Got up this morning and it had stopped.

 

I am no hardware expert, but my #1 suspect is the power supply. What do you hits guys think?

Share this post


Link to post
Share on other sites

Had some similar troubles way back and for me, it was indeed a PSU, which was at wit's end. Some time things like that could result from overheated processor or gfx-card. Maybe you could test it with programs like Hyper-pi (for processor) and Furmark (for GFX), to see if big load on either of 'em resets the comp. Both are easy to find Googling. It could be them, seeing it refused to start correctly right after the shutdown. For some troubleshooting you could check the bios to see, if some of the voltage displayed are either too high, or low. That could tell if it's the PSU. ex. +12v is something like +10 or so.

Share this post


Link to post
Share on other sites

Had some similar troubles way back and for me, it was indeed a PSU, which was at wit's end. Some time things like that could result from overheated processor or gfx-card. Maybe you could test it with programs like Hyper-pi (for processor) and Furmark (for GFX), to see if big load on either of 'em resets the comp. Both are easy to find Googling. It could be them, seeing it refused to start correctly right after the shutdown. For some troubleshooting you could check the bios to see, if some of the voltage displayed are either too high, or low. That could tell if it's the PSU. ex. +12v is something like +10 or so.

 

Good call. Newer BIOS should also have some temperature monitoring if the motherboard supports it, so this would be a good way to check on the CPU/chipset temps without actually booting the OS. If they climb and aren't stable, there's a heat issue. I too have run into this sort of thing before, so I concur with your assessment.

 

I guess we need more info from the OP to know for sure, but it def. sounds like either a heat issue or a power issue.

 

Oh, just thought of one other thing. lujate - if you have access to a UPS, those will often "clean" your power for you by normalizing dips and spikes. It's also possible that your power feed is just unreliable.

Share this post


Link to post
Share on other sites

Had some similar troubles way back and for me, it was indeed a PSU, which was at wit's end. Some time things like that could result from overheated processor or gfx-card. Maybe you could test it with programs like Hyper-pi (for processor) and Furmark (for GFX), to see if big load on either of 'em resets the comp. Both are easy to find Googling. It could be them, seeing it refused to start correctly right after the shutdown. For some troubleshooting you could check the bios to see, if some of the voltage displayed are either too high, or low. That could tell if it's the PSU. ex. +12v is something like +10 or so.

New power supply did not help. Bad motherboard? :(

 

It ran for about 30 minutes, now it will not start. When I can get it running again, I will try those utilities you suggested.

 

 

I have it on a UPS.

Share this post


Link to post
Share on other sites

is the battery good in the UPS? I've had that go bad on me at work with our Print Engine. After checking out the computer and all is well we traced the power back to the UPS and it was bad.

Share this post


Link to post
Share on other sites

As I sat here staring at the open case, I asked myself if there were any non essential components I could remove. Two things came to mind, a RAM DIMM and my graphics card. They are the only two components that have been added in over a year. I pulled them both and started it up. The PC has been running now for about 45 minutes. It is running my Linux partition instead of my Sacred partition (which runs Windows 7 :) ). I guess I should start Windows and see what it does.

 

is the battery good in the UPS? I've had that go bad on me at work with our Print Engine. After checking out the computer and all is well we traced the power back to the UPS and it was bad.

Good question. The UPS is quite a few years old. I pulled the plug from the wall and gave it about 15-20 seconds. My PC stayed up fine, so I assume it is OK.

 

Edit:

The plot sickens. I restarted my pc and loaded the Windows partition. It came up to the desktop with a low resolution (it had never used the integrated vga adapter). I went in to change the resolution and it went belly up. I tried again and it made it to the Windows splash screen. So I tried Linux. It made it to the login screen but also died before making it to the desktop.

 

Edit:

I started out with 2 RAM DIMM's. I removed the newer one before. That did not solve the problem, so I swapped DIMM's and now running only the new one.

 

I ran Hiper-pi with the defaults (1 million digits, normal priority, 2 processors). At one point, the entire screen went white and I had the spinning wheel pointer, but it finished without issue.

Edited by lujate

Share this post


Link to post
Share on other sites

Diagnosing pc problems over the net is always a tough call.

 

First thing I always recommend is to check your RAM and the best way I have found is to use memtest. It is a well known and trusted piece of diagnostic software and has helped me narrow down pc faults in the past.

 

You will need to burn a bootable disk and run it at least overnight. The necessary files are on the site I linked to above.

 

If just one fault appears it means your RAM is faulty....though not necessarily your actual RAM stick.

Let me explain.

It could well be your PSU supplying the RAM is faulty, it could also be one of your RAM channels on the motherboard is faulty.

But...if your pc does fail a memtest test then we have a good starting point.

Because by swapping RAM modules around and restesting overnight you will eventually, after a few days, be pretty confident whether it is an actual RAM stick.

 

So do yourself a real big favor......run memtest overnight. If you don't get any faults then we can be pretty confident that we can rule out your RAM as being the culprit and it also helps to suggest that your power supply and motherboard might also be okay.

 

Oh and one other thing. Do a complete virus/trojan horse/spyware scan on your pc. Any one of those buggers can cause a pc to crash.

 

And is a complete reinstall of your operating system out of the question?.......sometimes corrupted system files can also make a pc crash.

 

But do the memtest first ;)

Edited by stubbie

Share this post


Link to post
Share on other sites

Virus is not very probable since it must have affected Linux too.

Check all connections at the motherboards: cards, cpu, ram, cables, ... All have good contact?

 

Haarriss(german) = very fine rupture in material

A conductive trace on the motherboard could have a Haarriss. As long the whole motherboard is cold the rupture may be closed, at higher temperatures it may open--- at 8 and more layers motherboards: repair not possible= way more expensive than a new motherboard. A test if such a failure might be the reason can be done with electric cooling spray. Working for few minutes everytime the spray is sprayed might indicate it.

 

Spannungswandler(german) = dc-dc-converter?

I saw a few destroyed motherboards by bad dc-dc-converters or however they are called in english. They have no fan but are normally placed in the airflow of cpu-fan or power supply-fan and located close to the cpu. They can get really hot if they are not cooled -up to a temperature which would melt tin. Once such a temperature is reached a Haarriss in tin may happen or they get destroyed.

Pictures of motherboard at different CPU speeds: CPU mid, dc-dc-converters lower part of images:

 

1211382475234.PNG

 

http://www.pcgameshardware.de/aid,644498/Waermebildvergleich-CPU-und-Spannungswandler-bei-Uebertaktung/CPU/Test/bildergalerie/

 

If a card, a cable from an additional harddisc or whatever stops the ventilation of these converter: they can get extremely hot, even destroy a motherboard.

Edited by chattius

Share this post


Link to post
Share on other sites

stubbie, I will add the RAM test to my to-do list. I am relatively certain it is not a software issue for two reasons. One, it starts up cold and runs for a while before dying. Then it will not even get as far as connecting to the monitor. Two, I have both a Linux and a Windows partition and the problem affects both. I doubt a software problem would do that.

 

 

I thought I had it. My pc was up last night and I had Sacred running. Gameplay was a little choppy, but it was up. I left the machine running and went to bed. This morning, it was off and would not come up. I immediately pulled the graphics card and the newer DIMM, and it came right up. So my suspect de jour is the graphics card and that is where I will start tonight.

Edited by lujate

Share this post


Link to post
Share on other sites

It sounds to me like a component heating up and failing. You are on the right track to remove items until the problems disappears. Then add them back one at a time until the problem returns. But be sure to go slowly, adding components too quickly in a heat or time related failure with give you a false indication. Stubbies overnight test sounds like a good plan. My PC's temps vary by 20C from idle to full Sacred2 play. So that may also be a contributing factor. The load test programs the guys mentioned do come in handy for this type of diagnosis.

 

My best guess is a faulty ram stick. I have had that happen before. It may not surface until everything warms up. But that same logic can apply to the GFX card as well.

Share this post


Link to post
Share on other sites

Thanks for all you input guys. I am trying to isolate the faulty component. Here is my list of components:

  1. Power supply
  2. Graphics card
  3. New RAM DIMM
  4. Old RAM DIMM

 

I have pulled the graphics card and I am going to run a memory test overnight.

Share this post


Link to post
Share on other sites

Well, my PC has been up for 10.5 hours and the memory test came out good. I think (got my fingers crossed) that I have identified the bad component.

Share this post


Link to post
Share on other sites

Well, my PC has been up for 10.5 hours and the memory test came out good. I think (got my fingers crossed) that I have identified the bad component.

 

Ok.. Lemme guess - video card...

Share this post


Link to post
Share on other sites

Well, my PC has been up for 10.5 hours and the memory test came out good. I think (got my fingers crossed) that I have identified the bad component.

 

Ok.. Lemme guess - video card...

Yes. I have the box and the receipt, and I am going to see about getting it replaced.

Share this post


Link to post
Share on other sites

Well, my PC has been up for 10.5 hours and the memory test came out good. I think (got my fingers crossed) that I have identified the bad component.

 

Ok.. Lemme guess - video card...

Yes. I have the box and the receipt, and I am going to see about getting it replaced.

Good thing you held onto them :)

Share this post


Link to post
Share on other sites

Good news: I managed to get an in-store exchange.

Bad news: Sacred will not start with the new card.

 

 

I started out with an NVIDIA 8400 GS. It ran Sacred just fine for 8 whole days before giving up the ghost. I returned it to the store and they agreed to an exchange. Unfortunately, they did not have any more of that same card. I guess they are clearing out the 2010 cards to make room for the 2011 cards (I did not even know graphics cards had model years :Just_Cuz_21: ). They only had one other card in the same price range, an ATI Radeon HD4350. I checked the wiki from my Droid and it recommended a Radeon HD3870 or higher. The ATI card was marked $20 higher but I went with it. It must have been on sale, because it only cost me $5.

 

I got home, read the instruction while having dinner and installed the card right afterwards. Per the instructions, I uninstalled all of the NVIDIA programs before inserting the ATI card into the motherboard. Windows automatically installed the new card without prompting for the CD. I ran the CD anyways and installed an update using the ATI software.

 

The moment of truth came...and I got this message box:

sacred2.exe - System Error

The program can't start because PhysXLoader.dll is missing from your computer. Try reinstalling the program to fix this problem."

 

Is this something I have to install manually off the ATI CD? Do I have to download it?

 

FYI, here is my shortcut:

"C:\Program Files\Deep Silver\Sacred 2 - Fallen Angel\system\sacred2.exe" -skipopenal -nocpubinding

Share this post


Link to post
Share on other sites

"Just when it looked like he was about to recover, he felt the icy hand of death upon him."

 

I had it on all evening. Just before bed, I went to check the forum and it died on me. I pulled the new graphics card, but that did not really help. I now have the power and HDD lights on constant and a light on the motherboard itself. The NIC has no lights and the monitor is dark. Looks like it just went over the cliff. I am now stuck with the dilemma of whether to replace the motherboard on a 3 year old pc, or cough up the money for a whole new computer.

Share this post


Link to post
Share on other sites

"Just when it looked like he was about to recover, he felt the icy hand of death upon him."

 

I had it on all evening. Just before bed, I went to check the forum and it died on me. I pulled the new graphics card, but that did not really help. I now have the power and HDD lights on constant and a light on the motherboard itself. The NIC has no lights and the monitor is dark. Looks like it just went over the cliff. I am now stuck with the dilemma of whether to replace the motherboard on a 3 year old pc, or cough up the money for a whole new computer.

 

Ack.. Ah well.. That sucks hard boiled rotten eggs. You might want to check the capacitors on the mobo... One or more of them could have blown out and caused your problems. Look at the top of the capacitors. They should be flat on top to slightly depressed. If they've got a dome on top, they're toast.

 

I suppose it depends largely on what you've got in the box now. If the technology was at the top of the heap back then, it's likely to be somewhere near the bottom of the proverbial barrel after 3 years. If it wasn't top of the heap back then, then it truely is the bottom of the barrel.

 

The question to be asking is what do you expect it to do now and in the forseeable future? Were you happy with the performance of say, Sacred 2 over the past week or so? If so, then a motherboard swap may be just the ticket. On the other hand, if it was just passable and you figure it could have been better, then in that case, it might be better all around to just invest in a new motherboard/CPU/RAM combo.

 

Or if you don't want the hassle - it might be best to buy a complete system.

Share this post


Link to post
Share on other sites

Sorry to hear the bad news with the pc lujate.

Do as wolfie2kX suggests and check the MB capacitors.

I do disagree though that a top notch 3 year old pc is now at the bottom of the barrel. Mine is 3 years old and runs all the games I have and those that I want to buy fine.

 

Unless you are a pc wiz and are confident that you can work out what the faulty part is I would now suggest putting the pc into a repair shop.

Pay them $50 to check it and tell you what is the problem.

Then have them repair it. If it fails again you will at least have a warranty for their work so it wont keep costing you money.

 

Because you are unable to definitely locate what hardware is faulty it will cost you a lot of money slowing replacing parts yourself.

 

Time to decide. Repair the pc yourself(if you know for sure what to repair), give it to a repair shop to fix or buy a new computer.

 

Good luck.

 

Oh one thing....what does the motherboard manual say about the light you mentioned? It might give you a hint what is wrong. ;)

Share this post


Link to post
Share on other sites

Sorry to hear the bad news with the pc lujate.

Do as wolfie2kX suggests and check the MB capacitors.

I do disagree though that a top notch 3 year old pc is now at the bottom of the barrel. Mine is 3 years old and runs all the games I have and those that I want to buy fine.

 

Unless you are a pc wiz and are confident that you can work out what the faulty part is I would now suggest putting the pc into a repair shop.

Pay them $50 to check it and tell you what is the problem.

Then have them repair it. If it fails again you will at least have a warranty for their work so it wont keep costing you money.

 

Because you are unable to definitely locate what hardware is faulty it will cost you a lot of money slowing replacing parts yourself.

 

Time to decide. Repair the pc yourself(if you know for sure what to repair), give it to a repair shop to fix or buy a new computer.

 

Good luck.

 

Oh one thing....what does the motherboard manual say about the light you mentioned? It might give you a hint what is wrong. ;)

 

 

For what it's worth - Just because I said it was near the bottom of the barrel doesn't mean it's obsolete by any stretch. What I was saying and meaning - it is the bottom of the barrel as far as what is being offered on the market. Intel and AMD both tend to release bigger, badder, (hopefully) faster gear all the time. The stuff that was hot last year isn't quite as hot an item as it was when it was first announced. Heck.. My current system is on the verge of being 6 years old. It is, by all accounts, 'obsolete' tech (socket 939 mobo, Athlon 64 X2 4200+ chip and 3 GB of RAM). And yet, it still plays Sacred 2 just fine.

 

As I said in the 3rd paragraph - if it was working fine and did everything you wanted to do, then there's no reason to toss it - just replace the fubared part(s) and move on with life. On the other hand, if it could have done a bit better, then an upgrade to something newer, faster and better might be in order.

Edited by wolfie2kX

Share this post


Link to post
Share on other sites

I feel for you, Lujate... Looks like you comp is starting to go to smithereens. I remember having similar troubles. Really hard to find what caused it. Though, for me it was the PSU all over. Failing HDs, crashes while playing Sacred 2, etc.

Mebbe the repair shop could be a nice choice. Never done that myself, guess I'm lucky to have couple of good irl-friends who have helped. Can't say I'm a "pc-wiz" or anything.

About the Sacred error: whenever you get your gear back up, just install the PhysX-drivers. They don't actually do anything. Nvidia only. But it could do the trick. Weird thing though, I've only ever used ATI's card and they never gave me such errors...

Share this post


Link to post
Share on other sites

The lights could be a very good indication of what the trouble is. I remember on my old Dell , it had 4 lights on the back panel. Depending on which ones were lit, it would tell you where the issue was.

 

Plus unless you are familiar with electronics I wouldn't attempt any repair. Especially if you do not know how to 'read' a capacitor to find it's rating. Capicitors come in many flavors: film, radial, ceramic, tantalum and electroylytic to name a few. And some are polarity sensitive. Add to that, many will be surface mount types. If you do not have the correct repair equipment you will do even more damage to the mobo. Installing a differently spec'd cap will also potentially cause serious problems.

 

Capacitors block direct current (dc) and pass alternating current (Ac), thus they're widely used in filtering, smoothing and shaping various signals. The wrong type/size cap will adversely affect things like signal shaping/detection, audio and dc regulating.

 

I would do as already suggested and find a repair shop if they charge a small fee for checking out a pc. (At my shop we repair 2-way business radios - and charge only a small fee ($25) to evaluate the equipment and then suggest either repairs or replacement - or the customer can simply pick up his equip and walk away)

 

Or depending on finances, I would buy a new mobo and power supply.

Share this post


Link to post
Share on other sites

Well, yea, I'd advice not to fiddle with caps either. Friend of mine has done some maintenance, cause he works at an electronics company. And has about 13 years of work experience. Plus he has most of the compatible parts aviable.

Though, those lights could help. Just check the mobo manual if aviable.

Share this post


Link to post
Share on other sites

If I believed in signs from the Divine, the supernatural or "the other side", I would say they definitely do not want me to play Sacred.

 

I was 100% satisfied with my pc performance until my problems started this week. I had every intention of keeping it until Sacred 3 comes out. I will throw my cpu in the trunk of my car and take it to the shop during lunch.

 

Assuming they can fix it, any ideas about the ATI card?

 

Edit:

I visually inspected the motherboard with a flashlight and my untrained eye saw nothing that looked wrong. I do not have a manual for the motherboard. The light is next to the graphics card slot and was always on.

Edited by lujate

Share this post


Link to post
Share on other sites

There's no good reason to be replacing caps on a motherboard. Unless you a.) know what you're doing, b.) have nothing better to do c.) have nearly infinite patience and d.) the soldering skills of an industrial robot, it's far too much of a hassle to do it yourself. Not to mention you can generally wind up spending far less money on a brand new board than it would cost to fix it - not to mention it would take much less time to just swap it out. Motherboards are usually fairly cheap. This is the era of commodity hardware - finding a new board shouldn't be a problem.

 

There's also the potential issue of something else being wrong with it beyond the capacitors being toast. The caps are there to protect other components on the board. If they failed, something else could have gone frizzle fry as well. The only reason to check the caps on the board is to see if it's toast.

 

About the only good thing about taking it to a shop is that the shop may have another motherboard of similar vintage that works properly and can be used to test the CPU and RAM sticks.

 

And even then... You pays your money and takes your chances. Some shops are best avoided like the plague. Fry's Electronics (a chain in California and some other states in the US) is notorious for particularly craptastically bad service. The horror stories I could hijack this thread with... Eh.. Maybe around Halloween when we can all use a good scary story...

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×