FalconRock II + 3200+ = reboots, then dead proc

[engineer2001]

I just tried to change out my stock retail fan on my Athlon XP 3200+ with a Spire FalconRock II heatsink to get better temps (it was running mid 60's to low 70's C under load and mid 40's C idle). :hotbounce When I swapped the heatsinks, I cleaned the processor die and new heatsink using cleaner that I know works for processors (cleaned many of them in the past using it). I applied a T725 thermal pad that I bought at Micro Center, so I wouldn't screw anything up. I can't seem to put thermal grease on right - it gets everywhere. Anyway, on to my problem...

When I attempted to boot the computer after the heatsink replacement, it went into the part of BIOS where it detects the processor and drives, and I got video just fine (a second or so into booting), then it rebooted about 7 times really fast (about two times a second). I cut it off fast, took off the heatsink, and everything looks fine. I cleaned it and put it back on with another pad, and now nothing works - it just runs the case fans and heatsink fan. I switched back to stock HSF, and it is the same. I tried the proc in another computer, and it won't work. I tried another proc in this computer, and it works. That means I fried the processor somehow. :dead:

How did a new heatsink kill a processor when it was running fine? Anybody have any idea what happened? I would like to avoid frying another $150 processor, if possible. I don't think it's a thermal issue, as it only ran about a second. Is it possible the MSI motherboard I'm using shut the computer down repeatedly and restarted quickly until it popped something because the fan speed was too low on the new Falconrock II? I don't think MSI boards have FOC (fan off control), do they? Any ideas about what happened that can keep me from doing it again with the new 3200+ I'm buying will be appreciated...

 

[zephead]

i suspect an electrical fault in the new cpu cooler, assuming you didn't miss anything during installation.
test the new heatsink/fan for electrical faults with a multimeter.

 

[engineer2001]

Nope, I just tested it, and it tests out fine. None of the pins are shorted to one another, and none are shorted to the heatsink body. The aluminum framed fan on it runs fine as well (at about 2000 RPM). Is it possbile that the processor was damaged by heat in a second or two, like that? Maybe the thermal pad wasn't pressed hard enough before I popped off the release tab (the blue plastic cover that comes on the pad), and it wasn't thermally connected well to the heatsink, as there could have been an air bubble in there. Is a couple of seconds long enough with a semi-good heatsink seat to damage a processor?

 

[dgower2]

Heatsink Grease Shorting CPU?

I've read stories of miniscule specs of heatsink grease, that people inadvertently got on their CPUs, shorting the CPUs. You mentioned that you got it everywhere. I would inspect the burnt CPU very closely to see if that's a possibility.

 

[zephead]

it wasn't on long enough to die of heat, unless the heat from a short circuit caused it? just how much thermal paste did you use?

 

[engineer2001]

None - I cleaned the paste off the FalconRock II and put on a T725 phase change pad. The FalconRock II came with a huge square of white paste, and I knew better than to use all of that (plus, it had a black speck of dirt in the middle of that square of paste that I couldn't get off).

I have used paste in the past on my old Athlon Thunderbird, and I ended up getting it everywhere and having to clean the processor. I see the exact same temps on all three systems I have tested with using cheap thermal paste/grease and a T725 thermal pad. The pad is easy to put on, is perfect every time if you do it right, and easy to clean off. Plus, it doesn't void your warranty like grease does. So, that's what I use now.

This is the exact pad:
http://www.microcenter.com/single_product_results.phtml?product_id=0178823

It's a 5-pack of T725 thermal phase change pads (the pink ones). They are a snap to put on. You just pull off the clear back, press it on the heatsink with your thumbs, and pull straight out on the white paper tab. The release tab (white paper part with blue, plastic pad on end) pops right off and leaves a perfectly smooth, pink pad on the heatsink.

On this time, I think I may have pulled it sideways a bit and made a small air bubble in there, though. I smooshed it out, but it may have still had some air between the heatsink and pad. I should have cleaned it off and started over, but I thought it would be ok. I still don't think thats what did it, though. It doesn't look burned at all, and it was only going 3-4 seconds tops.

 

[Samstoned]

those are supposed to work as the cpu gets hot the pad melts and you get a nice thermal conduction?
my guess is the pads may have been old stock and the chip was smashed
or cracked when the clips where added
I use my hand and hold that typ down as I turn system on let it melt a little first
shut dwn and add clip and fan
most cases my customers let me use my own copper base thermal conducter
a small drop inthe middle all you need
put loads together not one burn out yet crossing fingers now :chef:

 

[zephead]

the pad you linked is much larger than the contact area of the athlon xp. do you know if the pad contacted the cpu anywhere other than the regular contact area?

 

[engineer2001]

No, it just contacted the rectangular die on top. It is not cracked or damaged in any way - it looks brand new after I cleaned it again. I am assuming at this point that it is totally dead and that I somehoe got an air bubble under there or shorted it out somehow. I'm getting a new processor to replace it, and it should be here soon (it's already shipped). I think I'll use the stock fan and heatsink again and not take a chance on this $10 heatsink frying ANOTHER $150 Athlon XP 3200+.

Thanks for all the help, guys!