Sign in with
Sign up | Sign in
Your question

Problem with computer startup probably caused by temperature

Last response: in Components
Share
February 3, 2007 4:40:01 PM

Hi,

For at least the last 4 years I have been experiencing a very strange problem with the startup of my computer.

After pressing the on button on the case, the computer begins to run, but usually fails to boot the operating system properly during the first try. It freezes either at some point even before it begins to boot the system, or during the boot.

When I have luck (and it is currently the warmer season - read below), it even reaches a point where it resets itself at some point during the booting process.

This goes on and on (it takes normally one to five restarts) until the operating system gets properly booted.

I am using Windows XP and at first, I thought it would be connected somehow with the OS, but now I am almost sure it is not since, as I have said, the first hassle often begins even before the start of the booting process.

However, I am 97% certain it is connected with the temperature of the place the computer is in. When it is generally warmer, there is less of the hassle and the computer boots up properly faster (usually one minute) than when it is colder (4-5 minutes at most). Of course - we have the heating installed, but the computer is few centimetres near the outter wall, so the weather has some, although small, effect on the temperature at that place.

I am so certain with this observation because I have had already four years to prove it day by day and I do not remember it has ever failed.

I have tried to search the Internet for the solution, but this problem seems to be so uncommon that I have not found even a starting point from which I could move on, so I would be very grateful if you could provide me even with a reasonable clue.

My configuration is:
Motherboard: ASUS A7N8X-X (nForce 2 chipset)
CPU: AMD Athlon XP 1500+
Memory: 512 MB RAM (V-data)
Graphic card: ASUS V9560/TD (GeForce FX 5600)
February 3, 2007 5:23:49 PM

You can try reseating the memory, and power supply connector, but your indications point to a power supply voltage, or Vcore (which is created on the motherboard) that isn't stabilizing at a nominal level until warmed up. First, try unpugging all unnecessary peripherals such as redundant hard drives, optical drives, and USB devices in order to unload the power supply as much as possible. Second, try another power supply. A last resort would be the motherboard.

Hope this helps. 8)
February 4, 2007 6:27:52 AM

I have to agree that the PSU is suspect.
Related resources
February 4, 2007 6:59:48 AM

Thank you for your replies.

Quote:
You can try reseating the memory, and power supply connector, but your indications point to a power supply voltage, or Vcore (which is created on the motherboard) that isn't stabilizing at a nominal level until warmed up. First, try unpugging all unnecessary peripherals such as redundant hard drives, optical drives, and USB devices in order to unload the power supply as much as possible.


I will try it tomorrow since the PC is heated up now.

Also, the day after tomorrow, I will try to get to the BIOS after the first startup as fast as possible to get the tension values at that moment and then I will check them.

Quote:
Second, try another power supply. A last resort would be the motherboard.


Actually, the original power supply that came with the PC burnt itself 2 or 3 years ago, when the problems had already been there, and there was not even a small change in the startup behaviour, so I think that should not be the cause.
February 8, 2007 6:15:29 AM

Here are the results.

I tried to unplug all unnecessary peripherals with no visible effect.

I got to the BIOS a few seconds after turning the PC on and the tension values were all as they should be.

The only value that was not on it's warmed running level were the rotations of the CPU fan. During the run it is at about 4500 RPM but immediatelly after I got into the BIOS it was on 2500 RPM and during few seconds it climbed to 4500 RPM.

Maybe it is normal - it could be caused by the electric induction after the power on, but that's only a guess.

I have also forgotten to mention the buzzing sound lasting a short time after the startup - could it somehow be connected to the fan?
February 8, 2007 6:27:09 AM

Quote:
The only value that was not on it's warmed running level were the rotations of the CPU fan. During the run it is at about 4500 RPM but immediatelly after I got into the BIOS it was on 2500 RPM and during few seconds it climbed to 4500 RPM.

Maybe it is normal - it could be caused by the electric induction after the power on, but that's only a guess.

I have also forgotten to mention the buzzing sound lasting a short time after the startup - could it somehow be connected to the fan?


If a fan has a bad bearing, then it's sometimes described as buzzing, squealing, or chattering. It's typically heard on power up for perhaps 2 to 10 minutes until the bad bearing settles down. 8)
February 8, 2007 7:15:58 AM

You are falling into the fallacy of post hoc ergo proctor hoc: After this, therefore because of this. The human brain is a wonderful device - it allows us to see connections (causes) where none exist. Anecdotal evidence is NOT evidence, it is anecdotal!

So no - the temperature of your room has ZERO effect on your booting problem. It should boot faster if colder in fact, but you'd need to make it about -40 degrees C before you noticed your boot dropping significantly.

I have two ideas for you: the first is a bad motherboard, and I would check the tops of all the capacitors for bulging and/or leaking. Secondly, I'd check the voltage-under-load of the PSU. I think it is possible you have voltage drops on the 3.3 and 5 volts rails.
February 8, 2007 7:56:23 AM

Quote:
I am using Windows XP and at first, I thought it would be connected somehow with the OS, but now I am almost sure it is not since, as I have said, the first hassle often begins even before the start of the booting process.


If you have been having the same problem with two diff power supplys then i would doubt it could be your PSU.

I have had slow boot times before, lockups, freezes and i discovered that this was because i had a dodgy stick of RAM. Try a diff stick of ram if you are able to or run MemTest to test your memory.
February 11, 2007 1:09:15 PM

Quote:
If a fan has a bad bearing, then it's sometimes described as buzzing, squealing, or chattering. It's typically heard on power up for perhaps 2 to 10 minutes until the bad bearing settles down.


I suppose this cannot be checked any other way than trying another fan?

Quote:
I have two ideas for you: the first is a bad motherboard, and I would check the tops of all the capacitors for bulging and/or leaking. Secondly, I'd check the voltage-under-load of the PSU. I think it is possible you have voltage drops on the 3.3 and 5 volts rails.


I have checked all the capacitors I could see by eye as thoroughly as I could and I have checked through camera those I couldn't check by eye and they have all looked OK.

As for the voltage under load, it is as it should be - the +3.3 value has a maximum deviation of 0.06, the +5 has a maximum deviation of 0.01 and the +12 has stably 12.6 V.

Quote:
I have had slow boot times before, lockups, freezes and i discovered that this was because i had a dodgy stick of RAM. Try a diff stick of ram if you are able to or run MemTest to test your memory.


I think I have checked my memory during a night (10h) some time ago and the results were OK. I will let it run for at least 30 hours this time to make absolutely sure.
February 11, 2007 1:30:42 PM

Quote:
I suppose this cannot be checked any other way than trying another fan?


You can momentarily stop individual fans, CPU, case, PSU, GPU, etc, with the eraser end of a pencil to see if the sound stops. 8)
February 11, 2007 2:45:25 PM

I would recommend two things to help narrow down the list of what is wrong, but they would take two cd's. Or you can just pick one step and go with that.

-Download and burn to cd a 'live linux cd'. Live linux cd's don't install linux on your computer, they boot completely off the cd (I've used the knoppix Live CD with a lot of luck). If it crashs, then there is something most likely wrong with your hardware. If it runs fine for a few hours (and/or for a cold boot), then the problem might be with windows.

-If it seems that hardware is culprit case. download and burn to cd the 'Ultimate boot cd' (ultimatebootcd.com). Download the basic version, that will work fine. It includes several utilities that were traditionally only available on floppy. You can test your ram (Memtest+ v1.65), processor (mersenne prime test), and hard drive (to make sure it functions correctly and has no bad sectors). (Be careful with the boot cd, you can easily overwrite your hard drive if you desire)

Of coure, this is assuming that your power supply is functioning correctly. Your motherboard/power supply can cause all of the above to fail, but no direct way to test them. Ask when in doubt (even if you get mixed results) The guys here are pretty knowledgeable about hardware.
February 11, 2007 2:56:21 PM

Quote:

If it seems that hardware is culprit case. download and burn to cd the 'Ultimate boot cd' (ultimatebootcd.com). Download the basic version, that will work fine. It includes several utilities that were traditionally only available on floppy. You can test your ram (Memtest+ v1.65), processor (mersenne prime test), and hard drive (to make sure it functions correctly and has no bad sectors). (Be careful with the boot cd, you can easily overwrite your hard drive if you desire)


Actually, I have ran all the test you recommend (Memtest86, Prime95, HDD test) and a few more with no failure in any of them.
April 28, 2007 5:21:01 PM

In the recent 3 or 4 days a new change in behaviour has occurred.

Somewhere in winter the computer had ceased turning itself off after the Windows were shut down and it had to be turned off manually from that time.

Now it has again begun to shut itself down after the Windows.
April 29, 2007 1:38:07 PM

I have checked it against the manual to the motherboard and they looked right.
!