Quote:
Originally Posted by theducks
Check all fans,
Check all (heat sink) Fins and grills
Overheats can cause erratic operation.
Inspect your MoBo for bulging Caps (the tops get a dome) |
Thanks. I've done all that. Symptoms are a random and intermittent "lights out" shutdown.
Dust build-up inside the system and on the heatsinks was minimal, but I blew what little there was out with Dust-Off. Inspected both mobo and power supply for bulging caps - none found (but caps can go bad without bulging). Replaced power supply(it was a high end unit before, it's an even more high end supply now). Replaced memory (all modules). The system has always run super cool (a high-end case with four 120mm fans plus the cpu cooler facilitates that). But I added a fifth 120m fan just for grins, even though temperature sensors showed no high temps and all heat sinks were cool to the touch. Tried different OS'es (booting from flash drives) even though it does not feel like a software issue.
Nothing has worked, dang it!
The only remaining thing on my troubleshooting list is to remove the cpu cooler, clean, and reapply fresh thermal paste. But immediately after a shutdown - I had the case open at the time - I grabbed all the heatsinks within five seconds of the shutdown. All were cool, I couldn't detect any heat at all. I checked this because thermal sensors can sometimes lie.
BUT, the only way I can force this shutdown repeatably is to start transcoding video. When the cpu temp hits about 50 Celcius, which is far under the max specs for this cpu, ... lights out. Repeatably. I checked BIOS to see if somehow a setting there got corrupted related to thermal shutdowns, but it is set to alert at 65 Celcius - and it never alerts. So this repeatability certainly points to a thermal shutdown, albeit way too early. The random shutdowns, when I am not trying to force them with transcoding, are occurring at the computers normal operating temp of 30 Celcius. So it both "feels thermal", and then it "doesn't feel thermal" (especially since the heat sinks are cool to the touch at the random shutdowns).
So fresh thermal paste is on the agenda for tomorrows troubleshooting. Can't hurt. The current paste has been on there for about four years with no heat issues, but it's time to redo it anyway. The cpu cooler itself is running fine. The system will alarm if the cpu fan ever drops below 100 rpm, which it never has, except when I have stuck my finger in there and manually stopped it to test the alerting feature.
Thanks for trying to help! If you have any other ideas for me to try, I'm all ears!!! I don't know what else to troubleshoot. I'm thinking bad cap on the mobo, even though nothing is bulging. I won't be replacing caps on the mobo, I'll just buy a new one (and new cpu, and new memory to go with it). I would prefer not to spend the hundreds of dollars that this will cost at the moment, but I may have to. The system is old, but still perfectly adequate for my needs (well, if it runs, that is).