Sign in with
Sign up | Sign in
Your question

PC shuts down immediately after POST

Last response: in Systems
Share
December 14, 2007 12:08:57 PM

Hello all,

My homebuilt PC, which has worked fine for a couple of years, suddenly isn't working after I cleaned out the pounds of dust from inside it.

During POST process, the BIOS/RAID controller seems to identify the hard drives properly (I'm running two pairs of RAID-1 mirrored SATA drives) ...and I think it's not the drives anyway because of the last bullet-point below.

System goes through the BIOS part of the startup without problems but then either shuts down or reboots almost immediately:

* If I try to boot into windows, it shuts down after a couple seconds of the "blinking light" moving from left to right on the first Windows screen.

* If I try to boot into safe mode, it shows a long list of drivers on the screen, the last of which I believe is mup.sys, and then says at the bottom center, "press ESC to cancel loading VAX347B.sys", and then crashes. I doubt that VAX347B is the problem, however, because of the following:

* If I disconnect all hard drives from the system and boot from a Windows XP CD, it shuts down or reboots when screen says "Windows is inspecting your hardware configuration."

During the cleaning process, I removed the CPU fan, cleaned it and replaced it. Someone suggested to me that if it weren't back on exactly right, that could cause the problem so I took it off again and put it back on and the computer then worked...for two days...after which all the symptoms returned. I don't believe it's overheating because the crash will happen 15 seconds after starting a cold computer (though I suppose that doesn't mean the computer couldn't think it's a heat issue.)

When I was restarting the PC after the original cleaning, I thought I heard my UPS make an odd noise as if turning on the PC caused a sudden (bad) power draw...and I thought I smelled something burning. It was very faint. I moved around some cables and tried again and that issue went away, but that certainly could have caused a problem...I just don't know what problem.

When I boot into BIOS setup, the computer will run without stopping. When I go into the BIOS monitor sections for voltage and for temperature, all readings seem to stay within very narrow and normal ranges.

Any thoughts on this would be GREATLY appreciated!

Please respond to group and, if you can, to me directly at rossputin (at) gmail.com

More about : shuts immediately post

December 14, 2007 12:34:10 PM

The first thing the comes to mind is a bad HDD or RAM. It may be the PSU, but the fact that it's perfectly stable in BIOS is strange. You can also try reseating the RAM and securing all connections (could be something as simple as a loose connection). *edit* read that the Windows installer crashed with no HDD attached.

Other things to try:
- Try the system with another PSU
- Try with different RAM
*edit*
- You can also boot into your BIOS and check your CPU temp to see if it's the culprit. If it idles over 45-50C, I'd say something is definitely wrong.

If something was burned, you should be able to remove your PSU / board and see or still smell the burned area. In my experiences - the culprit is almost always the PSU.
December 14, 2007 12:37:18 PM

Quote:
If I disconnect all hard drives from the system and boot from a Windows XP CD, it shuts down or reboots when screen says "Windows is inspecting your hardware configuration."


So not hard drive related. At first it sounded like a corrupt windows install or a bad HD but not if the above is happening. Unless you have multiple problems.

Download memtest86 http://www.memtest86.com/ and run it from a floppy or CD. If you get errors then we can assume you have a basic hardware failure.

1. Unhook EVERYTHING un-needed for booting. All drives and all usb devices. Remove all extra PCI cards etc.

2. Also be sure to reseat everything very carefully (RAM, video card, power cables) and check and dbl cchk all connections.

3. You also want to take this MOBO out of the case and rebuild it barebones.

4. Did you smell the burning and recognize it as such before the problem started? Or did you reconstruct that memory after the problem?

5. Try removing all but 1 stick of RAM during this testing process.

If you are getting errors still in a barebones setup. It could be RAM, PSU or MOBO. Not likely CPU so long as it is definitely not overheating.

Search for burned spots on MOBO.

Is the PSU fan spinning?
Related resources
December 14, 2007 12:37:53 PM

Thanks for the reply. The reason I sort of doubt it's RAM or HDD is that the memory test from BIOS works fine, that the thing sometimes crashes at very slightly different points, and that the same problem happens with no HDD connected to the system!

Could be PSU, though the BIOS test makes me think it's unlikely. If I can find PSU to borrow for testing, I will....
December 14, 2007 12:41:53 PM

The BIOS memory test is not very thorough at all. You will need to run a real memory diagnostic to rule it out. Memtest86.

December 14, 2007 12:44:40 PM

Same applies to PSU. They can fail partially. Rebuild system out of case barebones and then run memtest.

Since you likely caused this problem during the cleaning it's likely something is either knocked loose or else you fried something on MOBO.
Reseat everything to test the former.
December 14, 2007 1:06:35 PM

does it work if you bypass the UPS?
December 14, 2007 2:08:25 PM

I agree with most of the people here that it could very well be moemory.Ive seen bad memeory tell me that my video\sound\chipset is bad or i have no OS.If you have more than one stick of memory run it for 15 to an hour(at least for 1 pass) most windows xp boot issues that are caused by faulty memeory are usally picked up in the first few minutes, run each stick indivudally, helps to determine which one is bad.
December 14, 2007 2:35:49 PM

You did not mount your cpu fan correctly causing it to overheat, this would make sense you then remounted fan ,possible it overheated enough to cause damage to cpu ony way to fix is to replace cpu.

check your cpu if it looks like this:





its fried
With that much dust you may also have damage to capacitors:


examine motherboard esp capacitors:

good ones look like this



Notice that they are perfectly flat on top.


bad capacitors look like this:


and bulging like this is also BAD:




Did you put new heatsink compound on fan and clean old away before remounting?

this is bad and needs to be cleaned:



you can either use artic silver or good heatsink compound.
December 14, 2007 6:40:26 PM

techguy911,

I have a feeling you're right. I'm going to try the memtest thing. But I really didn't clean off the cpu well before remounting the fan. It never occurred to me that it could be very important...

I'll do the inspection as you suggest.
December 14, 2007 6:47:25 PM

You need to clean and reapply thermal compound every time you remove your heatsink.

Quote:
Did you put new heatsink compound on fan and clean old away before remounting?

this is bad and needs to be cleaned:



That pink crap in the picture is almost impossible to remove w/o putting scratches in the metal. It's almost like rubber or glue. :p  (I had an old ECS chipset HS with that on the bottom and was scraping it off one day.)
December 14, 2007 6:51:55 PM

any advice for getting off old heatsink if it doesn't come off very easily?

CPU is a pentium 4 (model #640, I believe)

if there is no obvious damage like picture a couple posts earlier, how can I tell if I've damaged the CPU?
December 14, 2007 8:59:23 PM

I have my own computer store have been working as a tech over 25 years this is what i use to clean heat sinks:

http://www.arcticsilver.com/arcticlean.htm




works wonders the heatsink is so clean you can eat off of it :lol: 

I have seen cpu's that have fried that have a crack underneath or hole, only other way to test is to try it on a board that you know that works.

December 14, 2007 9:26:01 PM

ok, I took off the CPU fan and cleaned off all the old heatsink compound. The lowest level of the stuff was dry and caked on. I also took out the CPU and cleaned the stuff off there. It was quite a mess. I do not see any obvious damage or anything that looks like burn marks on the CPU or on any capacitors.

I'm going to get some more heatsink compound from a neighbor.

Any advice on the "best" amount of heatsink compound to use when I reattach the fan?
December 14, 2007 9:52:42 PM

Not too much too much is BAD:

like this is good:

December 15, 2007 12:08:44 AM

there are several reasons why you should not remove a CPU heatsink and then put it back on without cleaning off the old compound and applying new compound.

1) air bubbles - when you place a heatsink on a CPU, the thermal compound spreads out perfectly with no air bubbles if you do it right. If, during the placement of the heatsink, you lift at any time, air pockets form. Air is a powerful insulator and will significantly degrade the heat transfer.

2) thermal compound setup - the thermal compound's conductive matrix requires many thermal cycles and many hours of use to setup and reach its maximum heat transferring ability. once this has happened the change is permanent and the compound cannot be reused (but you would not want to reuse it anyway because of air bubbles).

read these instructions from artic silver before you re-attach your heatsink.

now would be a good time to upgrade your CPU. If your CPU is 2-3 years old, you could probably get something orders of magnitude better for very little.

my guess, just like everyone else reading this thread, is that your CPU is damaged.
December 15, 2007 2:27:27 PM

OK, it gets stranger.....

I ran Memtest86 and the computer crashed. So then I'm thinking bad memory....until:

My configuration is 4x512MB DDRAM.

When I put any one single DRAM in any slot, Memtest runs fine.

But when I put any two DRAMS in any two slots, i.e. slots 1+2, slots 2+4, slots 3+4, slots 1+3, Memtest causes the computer to shut down.

Could that be CPU? Or is it now sounding like a motherboard issue?

What should I try next?

Thanks again to everyone whose taking the time to try to help me. I really appreciate it.
December 15, 2007 4:02:09 PM

I notice that my BIOS (ABIT uGuru utility) says my CPU temperature is around 177-179 degrees F, or around 80-81 C when it hasn't been on for long and hasn't been doing much. From my online research, that seems really warm for a computer having only been on for a minute or two and not running any programs. I don't know how accurate that uGuru utility is, but I doubt it's off by 50 or 75 degrees....

I'm quite confident that I reattached the heatsink properly, after cleaning and applying a thin layer of thermal compound.

I downloaded and burned the "Ultimate Boot CD". Anything I did crashed the computer, and as I tried rebooting and getting into new tests repeatedly, the computer shut down sooner each time....leading me again to think that you guys are right that I damaged the CPU. I did this with only one DRAM in the system.

Still, it's very strange that Memtest worked fine with 1 DRAM and crashed with more than one, no matter which DRAMs I used or in which slots.
December 15, 2007 4:21:22 PM

What make/model RAM do you have and what MOBO?

80 - 81c is quite high for an non-overclocked CPU not under heavy load but then again the p4 could run VERY hot and temp reading utilities can be inaccurate. You might still be in the high end of normal range, considering too that you are running stock cooling (it is stock right?)

The p4 came in varying cores and each had differing temps. We need a p4 guru to address this. I would NOT assume you have a bad CPU at this point. I would assume otherwise. CPUs are very hardy. MOBO's are not.

Keep digging. Make sure your voltages are set correctly in BIOS, for both CPU and RAM.
December 15, 2007 4:30:08 PM

The memory is GEIL 512MB PC2-4300
I think model is GX21GB4300DC

Mobo is Abit AA8XE.

But it's getting much worse at this point:

I just tried re-seating the CPU heatsink again, in case I hadn't done it right...but I had.

Now the computer won't boot at all. Doesn't even show POST/BIOS stuff on screen.

ABIT motherboard post code shows 8.3 then 8.9 then 9.0, getting to 9.0 much faster than it should. (Not sure if it showed other codes but just showed them faster than I could see.)

I tried clearing CMOS but it didn't make any difference.

So, now I'm totally dead in the water....

Help!!!!
December 15, 2007 4:33:33 PM

Do you have another power supply you can try?

I'd try that first. If you are sure the PSU is good then try another CPU IF you have one. But CPUs almost never go bad. MOBOs go bad all the time. It's probably the MOBO.
!