Mystery solved?
My white MacBook started the random shutdown thing five days ago. I searched the net for solutions and only one course of action was being advised, and loathed: send it back to Apple. I’m not a send-it-back kinda guy so I rooted some more. A number of people were of the opinion that it had something to do with the heatsink, because that’s what Apple were tending to replace at their stores. There was also a mention of a German engineer and some wires that were too short. This gave me a sound idea what to look for despite my German being rather bad. So…
I ripped the thing apart and found the problem. Please let me stress, lest I get sued, that I do not condone, encourage or otherwise urge people to do this; I’m merely reporting what I decided to do.
I found an excellent guide to dissassembling the MacBook here: http://www.ifixit.com/Guide/86.0.0.html
I took it apart and took out the heatsink. I could see nothing wrong at first and then I spotted it. There are four wires connecting the heatsink temperature sensor to the motherboard. They’re connected as two pairs (presumably two sensors). The routing of the cables is seriously dodgy. They’re passed through a small gap and are compressed between the heatsink and a metal post. It looks like the heat from the heatsink (OK, from the processors), caused the insulation on these wires to melt. They weren’t shorting out against each other though so I couldn’t understand the problem. Then it occurred to me that the angle of the wires in their original routing meant that they would have been in contact with the heatsink itself.
I’m guessing that either physical vibration, or expansion of the heatsink due to increased temperature, caused the wires to come into contact with the heatsink, shorting out, and thus triggering the MacBook’s protection against overheating. I simply insulated the wires with some standard insulating tape and found a routing for them further away from the heatsink. Five days, no random shutdown.
My fingers are still firmly crossed.
Lee

October 21st, 2006 at 6:18 pm
It would be interesting to hear how this progresses. Maybe try the Stress Test?
October 22nd, 2006 at 7:08 pm
Just an update, tried ’stress test’ (thanks to previous commenter), have run all manner of applications, high-res video compression included, and it’s all fine. Personally, I’m fairly convinced this was my problem. Might well be for others!