AMD cpu running hot - my story
I just thought I'd post this (new to here, btw). I build media pc's from friends and family and had a weird issue with one I just finished kept powering off after awhile. It seemed ok on idle, but record or process any video and a few min it would just shut off. Usually I just build them, and they work. So I did some searching and found this, and other forums where the issue of pc's powering off for no reason was talked about.
This directed me to cpu heat issues, as well as other heating issues and also software issues causing these "out of the blue" shut downs. And then to programs like SensorsViewPro and speedfan.
After installing these two, my cpu was running red hot. At idle it was around 57, but toss some processing at it and the thing went up to over 70.
The cpu fan was running ok, so I turned off the pc and pulled the fan/heatsink and noticed it wasn’t "stuck" to the cpu. In fact there was only a trace of gook on the bottom of the heat sink.
Looking closer, I discovered the plastic outline thingy on the mb that the sink fits onto or into was not lying flat on one side to the mb surface. There are four plastic pull tabs that release it, and I realized it was not snapped in all the way. So I re attached it, pushed in the four tabs, cleaned and regooped the cpu and attached the sink/fan.
Well the cpu temp is now running around 32 during video processing.
This was the strangest thing I ever came across and in originally putting all the parts together, I did not realize or notice the mb cpufan bracket was not flat on one of the sides to the mb. Thus, the sink was not touching the cpu. Poor thing was running un-sinked.
A little longer and I could have had a great cheese sandwich maker.
Wonder how many mysterious pc shut downs and heating problems are caused by this simple headache?
Last edited by randym431; 11-05-2007 at 01:20 PM.
Quite a few. There's as many with twitchy connectors with the RAM, flaky PSWs, etc. Heat sinks and PSW's are really the most common source of headaches on DIY and to a lesser extent OEM provided machines.
Originally Posted by randym431
I know of at least one Shuttle XPC that had a spontaneous reboot problem and nothing could be done to stabilize it. Then I tried hooking up a comparably rated normal ATX12V supply instead of the 1U supply that shuttle used with the cube- rock solid stable operation since then. While the supply's ratings are fine, this one was just below tolerances and it caused unstable operation. I am going to be getting a new 1U supply, but you can see where this stuff can cause bald-spots.
My old Athlon 64 3000+ ran for about three (!) without any cooling at all, because there was too much dusk between the small cooling-metal-plates, so the cooling-unit just blew it's cold air against a big pile of dusk, instead of cooling down the cooling-metal. (Heatsink? No idea what's it called, sorry.)
I always wondered about random freezes, shut-downs over night and other abnormalities, but I couldn't find anything, when opening the case. RAM-checks were fine, the GPU (ATI X800 by then) was pretty cool, harddisks worked like a charm and the CPU-fan was functioning properly. Some day I lost a lot of work made in GIMP, because of another random shutdown, so I unplugged all the cables and just disassembled the entire rig. I couldn't resist laughing about the huge amount of dusk, that prevented the CPU from being cooled down. After cleaning it, everything went fine. Now the box is doing it's work at my father's and he stresses it out like hell.
My props to AMD, for such a heat-resistant CPU.
Now I'm on a Intel C2D E6750, and I'll check for dusk every quarter year, just to be sure.
I just found out a few days ago that my Turion64 laptop CPU (single core, 2 ghz) has been running at 84 C under load,,,