Computer shuts down by itself

[georgeg4]

I was dual booting Win Vista Home and Pinguy Linux . All of a sudden my computer started to shut down when I tried to boot linux . Now it has spread to windows When I try to boot it shuts down as soon as windows starts loading . I did manage to redo Windows in one partition but the main one will bnot boot . But even the one that I reinstalled ( That I am using now ) I have to reboot several times to get it to take. I seem to remember a virus that acts like this but I dont remember the cure . Can someone help me

[georgeg4]

Sorry I couldnt seem to be able to post this thread and I hit it a couple of times and I dont know how to delete the repeats

[zlim]

The fact that it does this in both Windows and linux, probably rules out a virus. I'd say heat is the problem. Open up the computer and be sure all fans are spinning and the thermal paste on the CPU hasn't dried out.

[Corrine]

Hi, Georgeg4. I deleted the extra posts for you.I believe you are remembering either the Blaster or Sassar worms. Both ran on Windows XP and Windows 2000, not to my knowledge on Windows Vista.Are you getting the NT Authority Shutdown notice? If so, when you do, click on Start then select Run, type in what is shown below in bold but do not click OK yet.shutdown -aWhen you see that shutdown message appear, return to the run box and click OK to launch the shutdown -a command. This will abort the shutdown. ~~~~~~~~~~~~~~~~~~Since the problem is occurring with both Linux and Windows systems, I agree with zlim and doubt it is malware. However, you could try scanning with an antivirus software. ESET has a Beta program for antivirus software for Linux but I the ESET Online scanner is just for the Windows operating system.Avast has both Linux and Windows versions: avast! Linux Home Edition avast! Free Antivirus

[georgeg4]

Thanks for your help zlim and Corrine I will try and see what happens . The fans seem to be working but and I blew the dust out of them but that didnt seem to help and now the computer will not boot up at all , I am using an old backup computer now . If I find the problem I will come back and update for others if they have this problem in the futureThanks for your help zlim and Corrine I will try and see what happens . The fans seem to be working but and I blew the dust out of them but that didnt seem to help and now the computer will not boot up at all , I am using an old backup computer now . If I find the problem I will come back and update for others if they have this problem in the futureP.S. Thanks for deleting the extras Corrine

[b2cm]

Disconnect the hard drive and run a live CD. If you can run the live CD, that should tell you that you have a hard drive problem (most likely a partitioning or file system issue).

[Tushman]

Thanks for your help zlim and Corrine I will try and see what happens . The fans seem to be working but and I blew the dust out of them but that didnt seem to help and now the computer will not boot up at all

If it happens with both Linux & Windows, you more than likely have a hardware issue. Do you have a power supply tester handy? If not, uninstall the power supply and take it to a computer shop in your area. If you have a MicroCenter store near you, they will test it for a nominal fee. $5-$10 at the most. Likewise, if you have a multivoltage meter, you can test the power connectors on the power supply and see if it's kicking out the proper voltage.

[RichardKR]

I agree that heat is a likely culprit. However, you might want to go a step further and consider that the heat is causing a weak power supply to fail. I've pretty much trained myself to check the Power Supply when there are illogical problems or problems without a seeming cause. Good luck.

[Tushman]

I agree that heat is a likely culprit. However, you might want to go a step further and consider that the heat is causing a weak power supply to fail. I've pretty much trained myself to check the Power Supply when there are illogical problems or problems without a seeming cause. Good luck.

Indeed excessive heat can cause a weak power supply to fail. In fact, people don't realize that keeping your PC in a hot running environment does shorten the lifespan of the power supply greatly. However, in this case it doesn't seem like it's contributing to the problem. According to what the OP is saying, the unexpected shut down / reboot is happening as it's booting into the operating system environment. There simply isn't enough time during that duration for the heat to build up inside the case to the degree that it would cause the PC to shut down. However, as Zlim mentioned, if the thermal paste has dried up, the CPU could be initiating a shut down command due to dangerous heat levels. In my experience, the easiest thing to start the troubleshooting process is to swap out the power supply or test each of the power connectors with a multi-voltage meter.George, one other thing to try (if the previous suggestions are not feasible) is to disconnect everything and see if you can get the system to POST and stay on for a while. In another words, disconnect all the drives, and leave only the CPU/motherboard and auxiliary power connector. Turn on the PC and see if the computer will post. If it does, leave it on for several minutes. You will know if it's a thermal paste issue if the PC shuts down in that condition. However, if the PC stays powered on for several minutes, you will not likely find anything wrong w/ the thermal paste. Swapping out or testing the PSU would be your next best option then. Edited September 15, 2011 by Tushman

[burninbush]

George, one other thing to try (if the previous suggestions are not feasible) is to disconnect everything and see if you can get the system to POST and stay on for a while. In another words, disconnect all the drives, and leave only the CPU/motherboard and auxiliary power connector. Turn on the PC and see if the computer will post. If it does, leave it on for several minutes. You will know if it's a thermal paste issue if the PC shuts down in that condition.

On most pc's you'll also need at least a video board to complete a POST -- if not on the mobo itself. And a keyboard is also useful. And a monitor won't add any heat issues to the pc. Why not just hit the key during bootup to enter cmos setup -- then go directly to the screen that monitors various items like cpu temp and system temp and fan speeds, etc. It'll be running maximum frequency mode in that condition, and you can easily monitor temps and probably also see the various voltages displayed. That should be the first stop in any unexpected shutdown situation. If that all still looks fine for [say] 15 minutes, then you can go ahead and boot with some confidence that the power supply and cpu temps are not the issue. I have seen a case where the user had set the overtemp shutdown item so low that it wouldn't say up for more than a minute or two -- that one was easy to fix.

[Tushman]

On most pc's you'll also need at least a video board to complete a POST -- if not on the mobo itself. And a keyboard is also useful. And a monitor won't add any heat issues to the pc.

You are spelling out the obvious. It goes without saying you need some type of video output connection to know for 100% certain if the computer is able to successfully POST or not. You cannot rely on beep codes because certain models, especially OEMs, do not offer a beep when posting (silent POST). So yes, a monitor is required and I would think someone like George who has been tinkering with PCs long enough, should know that by now.

It'll be running maximum frequency mode in that condition, and you can easily monitor temps and probably also see the various voltages displayed. That should be the first stop in any unexpected shutdown situation.

The BIOS temperature sensor located on the MB is not reliable to any significant degree. Use it only as a rough estimate of the temperature - not as an exact science. The temperature diode used on motherboards is cheap and will wear out over a period time. This is not news. A motherboard manufacturer selling their boards for $100 is not going to spend a whole lot of time or money on manufacturing a motherboard with quality parts that give high accuracy readings. Regardless of what the motherboard costs to produce, the temperature diode is the least of their worries and was not designed to provide high accuracy temp readings.In my experience the most common cause of unexpected shutdowns (or reboots) is overheating and/or a faulty power supply. Anyone that has worked as a tech in the field will know to test the PSU with a multivoltage meter first as it is the most common parts to fail in a computer. From there, you can start looking at other possible causes such as overheating. Edited September 16, 2011 by Tushman

[Tushman]

I should also say that monitoring CPU temperature while the PC is running @ idle is not useful. Not useful in the sense that trying to diagnose a overheating problem should be done more thoroughly besides just seeing how hot it gets while at idle. The temperature difference between idle and while running under load can vary significantly which is another reason why you should not be relying on just the BIOS temperature reading alone. This is why I told George to try this method only as a last resort if none of the other options are feasible. Has anyone mentioned testing out the power supply yet? Hmm.... sounds vaguely familiar. Geez, I wonder why no one has mentioned this!

Edited September 16, 2011 by Tushman

[georgeg4]

Thanks all for all those excellent replies I have not had time to do them yet but I will definately report back what I find . Right now I am using temprarially an Apple I-book

[jolphil]

Hello,I agree with the fine suggestions mentioned..The only thing i might add is that I know as others mentioned, power supplies can and do cause various problems..I have a PS tester that the LED lites green when the various voltages are correct and red when out of limits..It also has ports that voltages can be monitored with a DVM.. The problem is that at least my tester only presents a very light load and not nearly what the actual computer loads present..Ultimately substitution is the final solution..I personally prefer a single rail supply of a reputable brand.. I just had a case where CPU heating was not the problem but mainboard Chipset heating was..Usually the Northbridge heats because some boards use skimpy heatsinks..Generally if you remove the component case side and blow a fan into the Mainboard, you will get some indication if you have a CPU/Chipset heating problem..On a final note with power off remove the power connectors, and inspect for burned or overheating power connectors.Burned connectors present a high resistance and can cause intermittent problems..GoodluckJolphil

[burninbush]

The BIOS temperature sensor located on the MB is not reliable to any significant degree. Use it only as a rough estimate of the temperature - not as an exact science. The temperature diode used on motherboards is cheap and will wear out over a period time. This is not news.

I was really thinking about the cpu temp -- which uses an on-chip diode to measure the temp, can't get any better than that. I would be interested to hear your theories about silicon diodes "wearing out" with time. What better system would you use to measure "BIOS" temperatures? (and what does a mobo manufacturer know anyway?) Unless you are checking the ps voltages with a digital storage scope, you will not get any better measurement of voltages with a voltmeter than what the mobo monitor already provides. Short of gross failures, it will be transients [spikes, dropouts] that cause cpu resets -- you will not see those with a voltmeter, either on the mobo or external. That is to say, if you find a pc that won't even power up, a voltmeter would be an appropriate first tool -- but that isn't the situation described in this thread. Testing a switching power supply is not a simple matter; sadly, swapping in a new supply of higher wattage is often the best [i.e., cost effective] remedy in those cases where you can't figure what else to do.

[Tushman]

Unless you are checking the ps voltages with a digital storage scope, you will not get any better measurement of voltages with a voltmeter than what the mobo monitor already provides. Short of gross failures, it will be transients [spikes, dropouts] that cause cpu resets -- you will not see those with a voltmeter, either on the mobo or external. That is to say, if you find a pc that won't even power up, a voltmeter would be an appropriate first tool -- but that isn't the situation described in this thread. Testing a switching power supply is not a simple matter; sadly,

digital storage scope? You really are confusing matters greatly. Just what does a digital storage oscilloscope have to do with testing a power supply unit? Talk about overkill. We are talking about testing a possibly defective power supply unit burninbush. We're not sending a rocket into orbit here.I don't know what kind of experience you have working as a tech, but testing a power supply is commonly done with a power supply tester or multi-voltage meter. Having worked as a tech both in the field and in corporate environments, this a very appropriate tool to use for testing a possible defective PSU. I have trained other tech how to how to use such a simple device in just a couple of minutes and it will give accurate readings enough to know whether the PSU is defective or not. No need to get fancy schmancy and use a "digital storage scope". Completely unnecessary.

swapping in a new supply of higher wattage is often the best [i.e., cost effective] remedy in those cases where you can't figure what else to do.

Before you start talking about "upgrading" the power supply to a higher rated one, let's hear what George has to say about his current power supply. It may not even be necessary to begin with.Generally speaking you should only upgrade the power supply to a higher wattage if you have upgraded a component inside the case (e.g. newer video card, another hard drive, etc.). Otherwise, there is no need to spend the extra money for 650 watt PSU versus a 450 watt PSU if your PC doesn't require one.I think burninbush you have come up with off the wall suggestions for what is a very standard diagnostics procedure. For the vast majority of situations where a random shutdown/reboot is involved, overheating and/or a defective power supply is at fault. I thought the suggestions that were given by members in thread before you jumped in were quite sufficient. Talk of using a digital oscilloscope and upgrading the power supply only convolutes and detracts from the troubleshooting process that George has to go through. Edited September 17, 2011 by Tushman

[burninbush]

digital storage scope? You really are confusing matters greatly. Just what does a digital storage oscilloscope have to do with testing a power supply unit? Talk about overkill. We are talking about testing a possibly defective power supply unit burninbush. We're not sending a rocket into orbit here.

Well, let's count the cards here; George's pc does POST -- which means that it's fundamental voltages are ALL PRESENT, hellooooo.... Why a storage scope? From the symptoms George is reporting, it appears that his machine =could= be getting reset by power supply faults -- i.e., transients from a noisy power supply. You will NOT see that kind of fault on any voltmeter. If you were to take the candidate machine to an engineer, and charge him with characterizing the system voltages, then he'd first reach for a digital storage scope, that being the only way to capture microsecond transients capable of causing a pc reset. Of course, it would be way cheaper to just throw in a newer & better power supply than to hire the engineer, or even to rent the scope; thus my comment re that being the economical choice. Otherwise, please tell me how to test for ps noise with a voltmeter. As I wrote previously, that tool is only useful for a gross failure test, applicable when the pc will not POST -- which is clearly not the situation here. My experience is: 17 years in the electrical engineering department of an Applied Materials company (Etec Systems), MSEE from Berkeley, retired 2001.

[georgeg4]

WOW what have I started here :thumbsup: :"> and maybe even a But seriously I have a few other facts I wasnt w=aware of before , First after sitting for awhile the computer posts with no problem but starts to act up more and more the longer I try to work on it >>>> definately an overheating problem IMHOSecond in the cms it shows a fan speed from the power supply but the Service unit shows as 0 speed and I assume that is the cpu fan >>>>> Yet both fans are running , I can see that visuallyThird as Corrine suggested I checked the paste on the CPU and though it is there I dont believe there is enough And to be honest I dont know where to get some or if there is a certain amount neededFourth All the voltages seem to be OK even the Battery is 3.1 VDC Fifth I feel like throwing the darn thing out even though it has 2 gigs of ram and a substantial HDD but it is 6 yrs old and has been left continously on excluding power outages for 6 yrs and last but not least thank you Tushman I whole heartedly agree with you about that scope , I believe I could buy quite a good computer with the money I would have to spend for that

Edited September 18, 2011 by georgeg4

[zlim]

#3 pea sized amount, some say to spread with a credit card, other say just to squish the two things together and move it a bit to spread the paste.Arctic silver:http://www.google.com/products/catalog?oe=...ved=0CE0Q8wIwBw

[Tushman]

Third as Corrine suggested I checked the paste on the CPU and though it is there I dont believe there is enough And to be honest I dont know where to get some or if there is a certain amount neededFifth I feel like throwing the darn thing out even though it has 2 gigs of ram and a substantial HDD but it is 6 yrs old and has been left continously on excluding power outages for 6 yrs

George, I know you're feeling frustrated but there are yet some simple options left.#1. You should be able to buy thermal paste from any local computer shop. If there are no local shops nearby, then you purchase it online. New egg has some for $6.http://www.newegg.com/Product/Product.aspx...N82E16835100016Directions for applying Arctic Silver 5http://www.arcticsilver.com/methods.htmlMake sure that you completely remove the old thermal paste before you apply it. You can find thermal paste remover on New egg's website. However, if you don't feel like spending the money for it, you can use pure alcohol (or isoproply 80%). Do not use paper towels or any lint based material for removing it; that will only increase the chances of impurity with the contact surface between the HSF with the CPU. I recommend that you be very thorough when removing the old thermal paste.#2In addition to applying some fresh thermal paste, check the motherboard for any bulging capacitors. Since your computer is 6 years old, this would definitely be worth checking.#3Like I have been saying all along since the beginning, if overheating is not cause of the problem, more often than not, it is a faulty power supply. They're SUPER CHEAP. Go out and buy a new one for $50 and be done it. You can always return if it doesn't cure the problem.

and last but not least thank you Tushman I whole heartedly agree with you about that scope , I believe I could buy quite a good computer with the money I would have to spend for that

You're tellin' me....pretty crazy stuff! But hey, if someone wants to spend a few hundred dollars to test a $50 power supply - let them waste their money. Who knows it might even stimulate the economy! Edited September 19, 2011 by Tushman

[striker]

Oh...smashing an ant with a sherman tank! Why not just take a flamethrower and get rid of it? :">

[georgeg4]

Thanks again for the info Tushman ,It will probably be awhile before I get the paste and power supply but if I get it fixed I will surely post back and let you know And striker Thats a great suggestion but where would I get a sherman tank ?

Edited September 19, 2011 by georgeg4

[Corrine]

Third as Corrine suggested I checked the paste on the CPU

'Twasn't me. Was zlim.

[jolphil]

All the replies to George4's shut-down problem has merit and whether it is a heating problem or a power supply problem is in fact hard to resolve..The logical suggestions that are cost effective and easy to do should be tried first of course..I think burningbush's reply did not suggest that George4 should go out and rent or purchase a digital storage scope but rather indicate how hard it is to truly test the switching mode power supply used in computers..Yes, the ultimate way is thru the use of an oscilloscope as the computer is booted and put under severe load..G4 certainly is not going to rent or purchase one..But his reply is most likely based on his engineering experience ..If there is any doubt that a SMPS is complex, this author, went thru extremes to show what the typical power supply is subjected to under various computer uses..It is a long article and I expect most will not read it entirely but at least look at the method used to test a SMPS and the charts showing how voltages vary under booting and load conditions..A DVM will not generally solve shut down problems most times.While the article has a different title it does offer insight to power supplies...Here is the article:HereAfter looking at this article I will guarantee you will never look at a SMPS the same again..Good luck G4jolphil

[Tushman]

A DVM will not generally solve shut down problems most times.jolphil

Not solve shut down problems???Where in the world have you worked at? What kind of real life experience are you basing these statements on? If you're making this statement based on your one little computer at home, or because you worked your neighbor's PC - I'm sorry but that's not enough qualification to make that kind of goofy comment. I'm talking about long term experience where you have touched/ worked on/ diagnosed more than just 2 computers.Both of you guys are out to lunch on this matter. Neither of you guys have worked a day in your life as a tech. Not in any professional because if you have, you wouldn't making these kind of illegitimate claims. In any computer shop I have worked at, a multivoltage meter has always been a regular tool [standard tool] for testing power supplies. Any decent multivoltage meter (and it doesn't have to be "digital" Jolphil) should be able to tell you whether the power supply is kicking out the proper voltage or not. I don't know how many times in the last 8-9 years that I have used such a tool to diagnose a possibly faulty power supply. Too many to count - it has NOT ONCE ever let me down in giving a faulty reading.Both of you guys are out to make this sound like it's so "complicated" to test a power supply. I think I've read enough jibberish to last through the year. Edited September 21, 2011 by Tushman

[georgeg4]

Tushman I still agree with you about the power supply and I do have an extra one on hand , But it is an older one so I will try it and see if thats the problem and if so I will get a new one , But I have to put the project on hold for awhile so I will post back here to let you all know how I make out . Again I appreciate greatly all of your suggestions and help BTW Corrine and zlim sorry I got your suggestions mixed up

[Guest LilBambi]

It is great that so many are also trying to help George.HOWEVER, such help should never include being mean/nasty in response to the others who are also trying to help. Our goal should be to work together and there is no room for ego in that.To those who seem to feel the need to be mean/nasty, please focus on simply helping George.If there are any more nasty comments toward others who are trying to help, we will be forced to close this topic and George will not get the help he needs.I don't think anyone wants that to happen.Thanks for your cooperation.

[burninbush]

HOWEVER, such help should never include being mean/nasty in response to the others who are also trying to help. Our goal should be to work together and there is no room for ego in that.To those who seem to feel the need to be mean/nasty, please focus on simply helping George. >LilBambi+++++++++++++++Well ... I'm sorry my comments caused so much controversy. I feel impelled to point out that in #16 and again in #18 I asserted that replacing the power supply was the most economical way to proceed -- if you didn't know what else to do. That would be the situation in which you suspect the power supply, even though it's voltages all measure within spec, as appears to be the case with George's computer. In general, an IBM-form computer will not POST unless all required voltages are present and within tolerance, so there's little point in measuring voltages on a computer that POSTS. In any event, you can read those voltages from the hardware monitor page in cmos setup -- along with the critical value of cpu temperature, whether or not the fans are running, etc. Case temperature [aka BIOS temperature, detected by a diode placed somewhere on the mobo] will not be very relevant unless the case is closed and the pc has had time to normalize. If you encounter a computer that won't power up at all, then a voltmeter can inform you that it is because the power supply is dead, or that the problem is not the power supply. But that is not George's case, as was clear in his very first post at the head of this thread. To the engineer who might face a problem like "our product is failing because of that power supply you spec'ed" -- you will want to know how to find out if the real problem is the power supply, why it's killing your machine, and for that there is no better tool than the storage scope. If the engineer is lucky, the scope will prove that the failures are somebody else's fault, not due to his power supply choice. Consider, if a critical voltage goes below margin for only a microsecond, with a 2ghz cpu clock that might be long enough for 500 machine instructions to go random, with a cpu reset as the likely outcome. You will never see fatal glitches like that on a voltmeter of any sort. By now it's well known that a marginal pc power supply can cause erratic and random system failures. So I suggest, again, that if you have this situation and you can choose between replacing the suspect power supply with an exact duplicate, or you can move up from a 400-watt supply to a 500-watt supply -- experience tells me to choose the one with the higher capacity. This is almost never a bad decision. @ George -- suggest, again, that when you next power up your computer, go immediately to the cmos page that monitors your hardware. Watch the cpu temp for a while -- if it settles to a reasonable value [50C or less ?] then you can move on to further testing. If it doesn't stop climbing, then you probably have a damaged cpu chip, or else someone forgot the silicone goop. That cmos page should be the first stop for anybody building a new computer -- could prevent you from melting an expensive new cpu. If you have further concerns about temp issues, then there are utils [fanspeed comes to mind] that will put a continuous readout of temps in the system tray area. And I think everybody should keep at hand an alternate way to boot the computer -- a Bart'sPE or a linux livecd is a good choice -- it's possible that the reason booting windows fails is just software corruption.

[Guest LilBambi]

So true burninbush!I think we have all seen strange erratic behavior from marginal power supplies. And that is sage advice. Thanks!

[jolphil]

Well said BB, I choose not to respond to Nastyjolphil