Curse

caekmaster · Oct 28, 2009, 01:55 AM // 01:55

Near the end of June, I put together a computer from scratch, and it worked wonderfully and magically. Now it's been giving me some problems and I don't know where the problem is coming from.

It started when I was playing Call of Duty 4 online, when the game minimized itself and I got an error that looks like this...

So, I checked the hardware monitor and I watched as the readings went from looking something like this:

To something similar to this, except for it was only the Temperatures that spiked, not the fans...

As you can see, the TMPIN0 max reading was above 100C. What happens is that for just a split second, the readings on TMPIN0, TMPIN1, or TMPIN2 would jump from 36-38C to above 100C, but just for a moment, and the readings would go back down. Sometimes a 4th fan would appear on the monitor, like this:

I blew the dust out of the fan, ran the computer with the side panel off to make sure the fan was spinning, and I got myself some thermal paste to reseat the heatsink on my CPU, but still the error kept showing up. So I decided to ignore it and dismiss the error and the temperature readings as erroneous. I left it for a few weeks.

That screenshot was one I took just today. I was surfing YouTube, nothing really intensive, when I suddenly got a Blue Screen and the computer restarted itself. (Now I can't just ignore this anymore) When it was running again, the fan readings would spike up much like the temperature readings, only much more often. (every few seconds) I also noticed that the temperatures for the CPU fluctuated much more often between 36C and 37C.

Could it be the chipset overheating? I really don't know how to fix this, and it's discouraging since this was my first build and now it's acting up. I haven't even overclocked anything. :/

Gigabyte MA790FXT-UD5P
Phenom II x4 955 BE @ stock settings
Corsair XMS3 4GB DDR3 RAM
WD Caviar Black 640GB 7200RPM
PC Power & Cooling Silencer 610W power supply
EVGA GeForce GTX 260 Core 216
Pioneer DVD burner
Coolermaster CM690 midtower

running Windows XP 32-bit

What should I do? ._O

moriz · Oct 28, 2009, 02:00 AM // 02:00

36C is pretty normal for a modern CPU... btw, what is "TMPIN1" actually measuring?

you can also try another temperature reader like speedfan. maybe the Asus temperature reader is screwed up somehow.

caekmaster · Oct 28, 2009, 02:21 AM // 02:21

That's part of the problem.. I think I read somewhere it's the chipset, or something. I searched all over and didn't find very much on what TMPIN actually measures...

Speedfan... well, it looks like the temperatures have stabilized themselves for the time being. No spike readings on either monitor.

...Would a lightbulb being changed in the next room have an effect on the voltage from my wall and have an effect, possibly crashing my computer?

moriz · Oct 28, 2009, 02:32 AM // 02:32

maybe the motherboard's manual can shed some light. and really, no computer is that sensitive to current fluctuations, so a lightbulb being changed in the next room almost certainly is not the cause.

End · Oct 28, 2009, 02:47 AM // 02:47

Quote:

Originally Posted by moriz

maybe the motherboard's manual can shed some light. and really, no computer is that sensitive to current fluctuations, so a lightbulb being changed in the next room almost certainly is not the cause.

...unless to much power is already being pulled...and that lightbulb is just enough to put it over the limit?...but I doubt it...

rb.widow · Oct 28, 2009, 05:54 AM // 05:54

Check your Motherboard book, espically the bios settings, im near sure the AMD boards have an auto fan speed controller built into the bios, that allows the system to slow the CPU fan down when the system is running idle, and speed it up when under more stress

( As you can see from your pic you posted, the CPU fan min is around 600rpm )

I remember awhile ago, people where reporting the controller would make the fan slow down even when the CPU was under load, making the temp spike, after turning the setting off, and making the fan run at a set speed all the time, it fixed the problem.

( Like i said this was on older boards, i thought this problem had been sorted )

If your board supports this it should be in the motherboard book,

In failing this, it could be down to a hardware fault on the board itself.

Aera · Oct 28, 2009, 07:16 AM // 07:16

I think the 4GB DDR and almost 1 GB from your videocard on 32 bit XP adds to instability because it has a max of a 4GB in total. Regardless of if that's the problem or not, you should get 64 bit Win 7 !

Also, did it happen once or does it occur often? A one time fling is nothing to be worried about really.

godis · Oct 28, 2009, 08:57 AM // 08:57

Have you cleaned your computer inside ?
Use compressed air to blow away dirt!

raddish · Oct 28, 2009, 09:54 AM // 09:54

Don't mean to sound dumb or anything, but what have you downloaded recently? I had a similar problem only a week ago, i downloaded 'Spotify' and my computer kept blue screening every few hours, then when i uninstalled it, thinking it may be the cause, the problem went away and i havent blue screened for about 5 days.

RTSFirebat · Oct 28, 2009, 11:20 AM // 11:20

Check your graphics card drivers, are you running non-official ones? Maybe worth trying some other ones.

Overclocking can also cause issues.

Finally as stated above, try removing whatever you have installed recently.

Killamus · Oct 28, 2009, 12:45 PM // 12:45

Yea, I'm going to agree with the above posters, this doesn't seem like a temp issue. Even 60*C isn't horribly bad (Although you would defiantly want to check it out if it happens consistently) - Most processors shut down at 72*C, for Intels at least. From the bad readings, it would make me want to think that it's a hardware issue, like the motherboard, but because this is recent it makes me think software. Try doing the typical virus/spyware/etc scans, and remove any recently installed programs that might have screwed with settings. I would also advocate backing up any data just in case it is hardware-related.

Also, have you checked that all of the fans are secure and are plugged in securely? If nothing else works, as an ABSOLUTE last-ditch effort, start switching out pieces (Graphics card, psu, etc) to see if it still happens.

And one last thing: Next time you get the blue screen, can you post the error message? Make sure to disable auto-restart in your control panel first.

EDIT: After doing a bit more research into this chipset, 60*C IS terribly bad, in fact, it's 2* below the shutoff point - If this high temp is actually your CPU, if course.

Kenzo Skunk · Oct 28, 2009, 01:29 PM // 13:29

Could be temperature monitoring probe failure, or malfunction with fan itself. If the components are under warranty, claim your rights.

Elder III · Oct 28, 2009, 01:40 PM // 13:40

based on what you have said and what I see in the pics posted it looks like a hardware related issue - either a faulty heatsink/cooler on the cpu (I suggest swapping it with a different one if you have a spare handy) or a bad fan controller on the motherboard. You are nearly hitting 115 which is the shut down limit on AMD Phenom cpus - they should not get anywhere near that hot in the first place of course.

First thing to do though is to download the trial version (free) of Everest Ultimate. You can monitor most hardware temperatures in the free version and you can also run a System Stability Test which will put a 100% load on the cpu. You can set it up to give you the temperatures, fan speeds and voltages and it will show you in a hurry if your CPU is on it's way to overheating and shutting down. It's a more boring stress test than playing COD4 for an hour, but more accurate and informative too.

Killamus · Oct 28, 2009, 02:47 PM // 14:47

Quote:

Originally Posted by Elder III

based on what you have said and what I see in the pics posted it looks like a hardware related issue - either a faulty heatsink/cooler on the cpu (I suggest swapping it with a different one if you have a spare handy) or a bad fan controller on the motherboard. You are nearly hitting 115 which is the shut down limit on AMD Phenom cpus - they should not get anywhere near that hot in the first place of course.

I think you're a bit off there. 42*C isn't that hot (In fact, it's exactly 20*C lower then the shutdown temp) - Several people state that 36ish is the average temp for these chips, as well as the stock heatskink failing being a common ailment to it. I would second putting in a different heatskink (Or even a different processor, if you have one) and seeing if that helps. The chip itself could even be faulty, reading the temps wrong and then prematurely shutting down.

Also, doing a stress test to figure out if a processor is actually overheating isn't one of the best ideas out there, by the way.

Quaker · Oct 28, 2009, 02:47 PM // 14:47

And don't overlook the power supply. A faulty PS can cause all sorts of havoc.

One of the joys of building your own system is getting it to work.

Edit: just noticed what power supply you have, so move 'faulty power supply' way down the list of possible problems.

Elder III · Oct 28, 2009, 02:56 PM // 14:56

Quote:

Originally Posted by Killamus

I think you're a bit off there. 42*C isn't that hot (In fact, it's exactly 20*C lower then the shutdown temp) - Several people state that 36ish is the average temp for these chips, as well as the stock heatskink failing being a common ailment to it. I would second putting in a different heatskink (Or even a different processor, if you have one) and seeing if that helps. The chip itself could even be faulty, reading the temps wrong and then prematurely shutting down.

Also, doing a stress test to figure out if a processor is actually overheating isn't one of the best ideas out there, by the way.

The point of my post was to get a different program to see if the temperatures match in both programs.... hopefully eliminating software error there. The purpose of a stress test is to be able to see what is happening. You can watch as the temps rise and you can stop the stress test at any point if the temps are nearing a danger zone. It's an extremely safe method if you pay attention to what you are doing.

Also, I never mentioned 42 celsius - i referred to the 114+ degrees that were showing - which is just about exactly the 115 that AMD states is the shutoff point for that series of CPUs.....

Spiritz · Oct 28, 2009, 04:03 PM // 16:03

Its possible the thermal paste between heatsink and cpu is causing the problem.
If you built the pc did you use enough thermal compound or if pc was premade then they may not have used enough.
Thermal compund is cheap to buy and theres plenty of guides on web to show how to replace it.
Replacing will after a few weeks show a dramatic drop in cpu temp - my pc used to play up real bad in summer . I could do a reboot and check cpu temp and be around 50c where my wifes pc can be on for days and not get over 40c an that was coz my thermal compund had literally worn out causing the cpu to constantly overheat every few days .

Killamus · Oct 28, 2009, 05:36 PM // 17:36

Quote:

Originally Posted by Elder III

The point of my post was to get a different program to see if the temperatures match in both programs.... hopefully eliminating software error there. The purpose of a stress test is to be able to see what is happening. You can watch as the temps rise and you can stop the stress test at any point if the temps are nearing a danger zone. It's an extremely safe method if you pay attention to what you are doing.

Also, I never mentioned 42 celsius - i referred to the 114+ degrees that were showing - which is just about exactly the 115 that AMD states is the shutoff point for that series of CPUs.....

Ah, I wasn't looking at those temps, I was looking at the core temps. I have no idea what those are, and neither does the OP from what I understand, so I pretty much ignored those. I assumed you meant 115F, which is ~42C, as the core temps were close to that. I never did find AMD's "Official" shutdown temp specs, but a good bit of googling turned up around 62C.