VERY strange CPU temperature problem

[10BaseT]

Hey guys,

I have a pretty strange problem that has been messing with my head for quite a while and I wondered whether any of you could give me a hand with it.
I have a computer with the following specifications;

Microsoft Windows XP Home Edition
ECS 755-A Mobo (SIS 755, AMD Hammer Chipset)
Award BIOS
AMD Athlon64 3000+ (2000Mhz)
512MB PC2100 DDR RAM
XFX GeForce FX 5700LE (256MB)
Samsung SV0643D 6GB HDD (C:\)
Seagate ST3160022ACE 120GB HDD (F:\ & G:\)
Generic CD-ROM Drive
Samsung DVD Writer

My problem is this, I built the system from scratch and everything was working fine for a couple of months, then the computer started freezing up whenever I tried to play a game. After a while it started freezing up whenever I tried to do anything. So I put the computer to one side until I had time to fix it and used my other computer for a while. Well two months down the line, I decide that I will have a go at trying to fix it, but lo & behold it is working absoloutely fine so I just start using it again. Well a month down the line and the freeze ups are back again.

I dismantled the comouter several times and put it back together again, made sure all the cards were seated properly, tested my RAM in various other machines and ran several different Memtest's on it and everything was okay, there is not one particle of dust left inside my entire machine. So I was completely lost for ideas until I thought about it being an overheating problem, so I installed motherboard monitor and it started setting off alarms left right and centre when I started up a game because the CPU was overheating. So I took off the heatsink and the thermal pad was completely shot, so I put some thermal grease on there, just enough and nice and evenly spread. That helped, Motherboard monitor was now reporting much lower temperatures and I could play my games; for a couple of hours until it started freezing again. The weird thing is that just before it crashes out Motherboard Monitors readings go completely mental, they go from their standard 51 degrees celcius when I have started a game, to 110 degrees celcius, then it will shoot down to -98 degrees, then maybe back up to 82 degrees, then -50 degrees. It will keep this up for a while until the system crashes out. The temperatures I have given above are only an example of the type of thing that happens, they dont go on the same temperatures everytime. It is very, very strange. Should have said earlier, when the computer freezes it litterally does freeze, everything on the screen stops and stays exactly the way it was, none of my input devices will respond and any sound that was playing will completely stop.

I have a new 550w PSU on order to rule out that it might be my stupid ass stock 350w PSU that is too blame for the crashes. If that does not work then I will be looking at buying the Thermalright XP-120 here to seriously cool the CPU, but apart from that I am fast running out of ideas. I have even tried CPUIdle here but that has had no effect. Please guys, I need some help here; I have hundreds of £'s worth of equipment that I can't do anything with


Thnx for the help in advance guys, this is a damn good tech board.

Tenbaset

 

[LegendBreath]

ECS.... is the worst motherboard for temperature sensors. Mine too does that. Right now the CPU says 102 celcius. And, the 3th sensor (dunno wath it is) says 80 to -75. So, the CPU temperature is really not the right one. And, BTW at 110 celcius... your CPU would MELT! I got 2 solution, if you overclocked it, then, unclock it. And, if you dind't then, get a better heatsink!

 

[gaara]

My recommendation is that you replace that motherboard. Those temperatures are just simply wrong, and I think that your motherboard has some serious problems and is the cause of the crashes.

Either that, or the the temperatures readouts your motherboard is "detecting" could be activating the thermal protection on your CPU, which shuts it down to prevent damage from heat.

I've never even heard of ECS, but they don't sound very good at all, spend a hundred bucks or so and invest a nice motherboard from DFI or Abit.

 

[10BaseT]

Would it be safe, to disable to automatic heat shutdown thing on my Mobo and see what happens. Or would that not be a wise idea????

 

[gaara]

I'm not sure if it can be disabled.

I would first check out what your real idle and load temperatures are. If your load is less than 50, I think it would be safe, although I personally wouldn't.

 

[Codeine]

Well i see a HUGE problem with your RAM yo. First you have PC2100!!! With only one 512mb stick. Your mobo is dual channel memory, you should have 2x 512mb sticks. You should also have PC3200. Your RAM is extremely slowing your system down.

That mobo is probably the worst you could of got. But still, you shouldent have those temps. What kind of case do you have?
I think you should get a new heatsink/cpu fan. Something good. I suggest the Vantec Tornado, for your cpu fan.


Over all, you need to use your dual channel ram, thats extremely important. You should really get PC3200, since PC2100 is just slowering your system down. For your mobo, if you cant afford the new 939's, then get a Abit AV8 3RD EYE. Never get a ECS.
Always make sure you get a Abit, MSI, DFI, or Asus.
Heres the fan.
http://www.newegg.com/app/ViewProductDesc.asp?description=11-999-613&depa=0