PC Randomly Shutting Down

[petermitchell44]

That's great news Peter, I'm glad it's working for you now!

Hi Ian Can I start a new thread on a more fundamentally serious problem .

Over the last couple of years I have been increasingly plagued with my PC either freezing on shutting itself down.Currently it will do either maybe 3 or 4 times a day or sometimes it will go OK for a few days in a row.

I have investigated everything I can ,all to no avail

I have reinstalled the operating system(XP home edition)
I have extra temperature sensors and there is no indication that anything is running warmer than normal.I have infact reseated the CPU heatsink
One by one I have eliminated various software items
I have tried 4 different browsers
I have tried 4 different mouse
The problems do not seem to be related to the % utilisation of the CPU
Fanspeed analysis indicates that my harddrive is healthy
The tower has been thoroughly dedusted

I basically do the same thing with my PC each day, trading the stockmarket using ADVFN dynamic charts which use Java.The Java is by no means perfect at loading up the charts but I dont think this is the source of the problem as sometimes the PC will freeze within a few seconds of booting ,before I have even opened a browser

There is never any clue in event viewer

I dont know the technicalities of what actually happens when a PC freezes its as if either something is just told to stop running or something clogs up. I just have a feeling that the L1 cache is not clearing itself down( I dont know whether that makes any sense ) but I doubt this thought as I've already said it can freeze within a few seconds of booting.

My local computer shop says they can find nothing wrong but with a 4 yr old PC things are likely to start wearing out.If that was the case you would think that there might be some indication as to the location of the ageing componant.

Sorry this is a bit rambling .I feel its probably something obvious that I am missing. I am reluctant to abandon this PC befor identifying the fault. I welcome any comment. Peter

 

[V_R]

Hi Peter.

You say you have done a fresh install of windows, as in Formatting (wiping) the harddrive and reinstalling Windows, and it still did it.

As you said you have cleaned out the pc etc, That says to me its possibly a hardware issue.

Your sure its not overheating?
Try running this program wile your using the pc and keep an eye on the temperature. Just to rule out if it is a heat issue.

 

[petermitchell44]

Hi Peter.

You say you have done a fresh install of windows, as in Formatting (wiping) the harddrive and reinstalling Windows, and it still did it.

As you said you have cleaned out the pc etc, That says to me its possibly a hardware issue.

Your sure its not overheating?
Try running this program wile your using the pc and keep an eye on the temperature. Just to rule out if it is a heat issue.

Hi VR thanks for responding.

I have never run a games programme so I dont know how to respond to your suggestion to run the programme in order to check CPU temperatures.However I can comment on CPU activity and temperatures as follows:-

All fans are allways running at 100% speed.

With a fresh boot up from cold after the CPU has settled down from the boot up activity the resting CPU voltage is 1.34 v and temp 40 c ,for an ambient of 24 c.
With CPU activity at 50% its temp rises to 48 c and already the CPU voltage regulator has started to kick in reducing the voltage to 1.33 v ( presumably in responce to rising CPU temp ).
At 100% CPU activity its temp rises to a max of 60 c and voltage is being fully regulated down to 1.31 v. However .when the CPU again settles down to resting activity (and the voltage has returned to 1.34 v ) I notice that the CPU voltage will still periodically spike down to 1.31 v with no visible rise in CPU activity or temp ( It may be that the CPU activity and temp monitors cannot respond to extremely short duration excursions)

This morning within 2 minutes of boot up ( whilst loading java driven charts ) the system froze with the following readings :-CPU usage 36% voltage 1.31 v temp 50 c.

My system uses an Intel Core Duo and most of the time CPU activity is somehow restricted to 50% on any one activity.I need to simultaneously open a number of activities in order to get more than 50% CPU activity.

Whatever the CPU activity level ,it is always shared more or less equally between the two CPUs

Is there a CPU temp trip or does the CPU voltage regulator have a lower trip level at which it will stop ( freeze) CPU activity or even shut down the system if CPU temp continues to rise after max voltage regulation has been achieved.If either of these do exist it could be that a malfunctioning CPU temp monitor is causing the system to freeze or shut down in responce to an errorful CPU temp signal.

As the day progresses and the ambient temp increases 2 degrees to 26 c ,with a CPU usage of 50% the voltage regulator is fully active holding the voltage down at 1.31 v to limit the CPU temp to 55 c with a corresponding reduction in running speed which means that most of the time my PC is running slowly because the voltage regulator is holding the CPU temperature down unnecessarily

It does surprise me that the voltage regulator never allows the CPU temp to rise above 60 c ( this seems particularly low)

 

[petermitchell44]

Hi VR I have replied to you but I cannot see my reply on the site .Can you see my reply .Thanks Pete

 

[V_R]

Yep i see it just fine.

I also split this thread just so it makes more sense as its a different problem to the Outlook issues you has previously.

That program Core Temp is just for monitoring the temps so its fine to leave running as long as you like.

First up the temps dont seem an issue as they are well within tolerance of that CPU. Maybe a tad on the warm side but nothing that would cause concern or the issues that your having in my opinion.

The 50% bit your talking about sounds like the norm, just the way the OS reads the CPU load over the two cores. Not an issue.

Im still thinking a hardware issue maybe.... Need more info on the system if possible, what motherboard is it?

Try this....

http://www.piriform.com/speccy

Install that, and run it.
Then can you click on file at the top > 'Save as Text file'
Then attach that text file to this thread via the 'manage attachments' button below the reply box.
*But delete the windows key line out of the text file first for obvious reasons.
(it will look like this: Serial Number: XXXXX-XXXXX-XXXXX-XXXXX-XXXXX)

 

[petermitchell44]

Hi VR here is the spec.Will you be replying Thanks Pete

 

[V_R]

*Bump to remind me later when i have more time*

Anyone else got any ideas?

 

[petermitchell44]

Hi VR I sent the information you asked for.Will you eventually respond.I tried bumping you previously without result. Is there an established method of bumping.Thanks Pete

 

[Me__2001]

I'm with V_R, definitely sounds hardware related hence the lack of anything in the event viewer. Could be overheating or something more difficult to pinpoint

Looking through the spec you've posted i'm fairly certain that the processor isn't throttling back during use. I think it has to be 70C+ before anything starts happening with the Prescott's

You could try a stress test on the processor with Prime95 to see how it responds, Other thing to check would be the graphics by doing something similar but i'm not sure on the software to do that

 

[petermitchell44]

Hi to Me _2001. Thanks for taking up my cause.I ran Prime 95 through the night,infact its still running and with no problems reported.I am using my PC as normal whilst Prime95 is running and I notice that whilst the cpu is running flat out with a max temp of only 63 C and Vcore is controlling between1.312v and 1.301v there does not seem to be any slow down in performance.With the cpu running at 100% wouldnt one expect cpu temp to be more than 63 C.I'm quite keen to install a better Graphics card and a new 500GB harddrive .Before I do this can you think of any other areas where I might look for the fault.As it will sometimes shut down and reboot when I am away from it ,and its just sitting there ticking over,I did wonder whether background programs (Symantic scanning or Windows updater) might be triggering the shutdown.Thanks a lot Peter

 

[floppybootstomp]

If I may comment I can't glean any clues from all the information provided thus far.

I don't see that memtest has been run, that would be worth doing. Memtest is a free memory testing software facility which you can download then burn to a CD or load to a usb memory stick. You then boot from it and let it do at least two 100% runs testing your RAM.

It should take about 30-40 minutes to complete and usually shows any faulty memory modules fairly quickly.

Other than that, as others have said, likely a hardware fault.

Although the CPU should run happily at 70C the ideal operating temperature for any CPU is less than 50C. I have know CPU's to crash/freeze on 60C before and although I'll admit this is unlikely to be the cause it may be worth observing how high the temps are when the machine crashes, if possible.

To eliminate temperature as a possible cause of crashing leave the side of the computer case off and run a desk fan blowing directly into it. If after a few days of heavy use the machine hasn't crashed, consider buying a more efficient CPU cooler.

Other than that, a hardware fault could reside almost anywhere, the memory, motherboard and CPU being the usual culprits.

 

[V_R]

Ah yes good shout on the Memtest FBS.

 

[muckshifter]

without any "error" messages to go on, you is peeing into the wind on this one

I've had a look at the spec posted and I wouldn't be at all surprised if your "problem" stems from a driver issue, although, I to won't rule out some obscure hardware problem ...

IMHO, Intel, while supposedly able to manufacture a reasonable CPU, have squalidly tit expertise in building graphics cards, and in your case, you ain't got one, it's an onboard chip.

The saving grace is that the MB does have an AGP slot (0.8 V / 1.5 V AGP 3.0 connector supporting 1x, 4x, and 8x AGP) , not sure if the expense in acquiring one would "fix" the problem.

As far as I can tell, drivers for your motherboard & on-board-graphics, are at "EOL" 2007 (end of life) though you can check to see if they have anything later than you are using.

http://www.intel.com/p/en_US/support/highlights/dsktpboards/d865glc

http://downloadcenter.intel.com/Sea...s&ProductProduct=Intel®+Desktop+Board+D865GLC


good luck!!

 

[Me__2001]

I'm guessing its not overheating if it ran prime95 fine for that amount of time. As suggested memtest is the next thing to try

If that fails to find anything then i'm not sure, power supply possibly?

 

[floppybootstomp]

I'm guessing its not overheating if it ran prime95 fine for that amount of time.

A valid point young Lochinvar, you is correct

 

[petermitchell44]

Hi Guys .Thanks for all the attention.I ran Memtest successfully about 4 days ago .I am continuously monitoring temperatures and am surprised by how low they are.There is never any suggestion that the failures are more prevalent when the cpu is being driven hard or when temperatures are higher ,and very often freezing or rebooting will occure when the system is just sitting there doing nothing .If anything the system failures are more frequent when the PC is just sitting there idling.I have been running Prime95 nonstop for the last 24 hrs and the system has not failed once.

I think as a first step I will get a reasonable Graphics card.

It seams to be a bit like a motorcar thats got the tickover set too low.Nothing drastically wrong but it keeps slowing to a standstill.

Ive run S.M.A.R.T and that indicates the harddrive is in good order.Are there likely to be harddrive faults that S.M.A.R.T would not identify.

Keep up the good work Many thanks Pete

Addition.

Hi Guys .I've just noticed something very peculiar

When my PC starts booting up and displays the full screen Intel logo.I notice that sometimes the logo says " Pentium 4 inside " and sometimes the logo says " Celeron D inside " .
Does this indicate a fundamental mishap in the start up programme .Does it mean that the Bios is confused as to whether it is starting up a Pentium 4 or a Celeron D

 

[Alf]

My bet is that the PSU is failing...

 

[petermitchell44]

Thanks Alf . Is there anyway of checking it.The voltages seem to be ok.

Does anyone know how my PC at start up can alternately tell me I have 'Pentium 4 inside' and then 'Celeron D inside'

 

[muckshifter]

ah, now then

the program we had you run, Speccy, clearly states you have a Intel Pentium 4 Prescott

are you seeing this BIOS from first boot-up, or when you have to re-boot?

P4s (prescott) have a very neat way of running hot, they have what is called a thermal throttle. Thermal throttle will engage at a pre-set temp, you cannot alter that, if the temperature doesn't drop to a safe level, the computer is turned off, in order to protect your CPU.

go give Throttle watch a try out.

as always, every scrap of information is important, what you may think as trivia could be the cause of the "problem" in the PC

 

[petermitchell44]

Hi Muckshifter thanks for getting involved.

I dont see the Bios when I start up I just see the normal Intel advertising logo for a few seconds before the Windows XP logo appears.

Sometimes on start up the Intel logo will say 'Pentium 4 inside' and sometimes it will say 'Celeron D inside'

I have set Throttle Watch running together with Prime 95 which is fully loading the 2 cpu according to Throttle Watch and Task Manager.

On Throttle Watch there is nothing showing up on theTM1 Throttling Chart

The cpu Frequency Chart is around 1500MHz but it occasionally spikes up to the top of the chart which is supposed to be 5000MHz ?

Whilst running Prime95 ,simultaneously with my normal activity the cpu temp never goes above 63c and the Vcore cycles between 1.312v and 1.301v.I assume this is the normal workings of the voltage regulator and not any Throttling back.

When working normally ,with Prime 95 turned off, I am normally continuously monitoring cpu usage and temperature and am aware that when the system trips out it is never associated with high cpu usage or high temperature.

A few days ago I ran Prime 95 continuously for 3 days, simultaneously with with my normal activity.It never tripped out once !