LFS on PII 233 with no fan in heatsink

Peter B. Steiger wypbs001 at bornagain.com
Wed Feb 5 12:56:08 PST 2003

on 03 Feb 2003, Rainer Peter Feller sez:
> *sigh* where are the time when CPUs didn't even need passive cooling ?

Oh, this is a good place for my melted CPU horror story from last summer... 
I had been running RH 7.0 successfully on an old AMD K6 266 for a year or 
so, and finally had the cash to buy some upgrades.  So I got a barebones 
kit with a Duron 1.6GHz on a chaintech mobo and a 60GB drive with plans to 
install a new Red Hat on the new drive and start over with a clean slate 
(the old system started on 6.2 and was gradually upgraded to 7.0 through 
lots of poorly planned mistakes on my part.)

Meanwhile, Mrs. PBS's desktop system, a relatively new 1.0GHz machine of 
some unknown flavour lost its cooling ability and melted her hard drive 
among other things.  When my barebones kit arrived, I figured I would use 
it to test out the wife's parts to see what I could salvage - her burned 
system and my new one were the only Socket A / ATX systems in the house, so 
I couldn't test her equipment on anything else.

So anyway, I popped her DIMMs into the new system and decided they were 
still good, and then I looked at her motherboard.  I figured, it's too much 
trouble to remove my whole motherboard and put hers in the case, so why not 
just test the CPU first?

Remember several paragraphs ago I mentioned that my previous systems have 
all been older 586/686 types with less, uh, critical cooling requirements.  
When I tinker with the hardware, I break all the rules and just have the 
guts of a system laid across a desk with little or no static protection 
apart from touching my doorknob before proceeding... I just lay the 
heatsink and/or fan loosely across the CPU while I power on to make sure 
everything works.  After all, how much can happen to a CPU in the minute or 
two that I have the system powered on, right?

So just like always, after removing the fan and heatsink from both our CPUs 
(when did they get so BIG, I idly wondered) and put her CPU into my 
motherboard, laid the heatsink and fan loosely across the top of her CPU, 
and powered up.

Everything worked great, and I congratulated myself on saving $200 by 
salvaging her CPU... for about 10 seconds.  That's when the CMOS setup 
screen suddenly went blank, and nothing I could do would power up my 
motherboard again.  I dropped her CPU (WOW, was it hot), put mine back in, 
clamped down the heatsink like a good boy, and still couldn't get life into 
my new motherboard with my new CPU in it.  Nor could I get her old 
motherboard to work with either CPU.

After a week or so of getting nowhere, I gave up and ordered a replacement 
motherboard for my poor CPU that had never seen the light of day.  By this 
time I had read up on the fact that modern CPUs can literally melt in 10 
seconds without adequate protection, so I was prepared with a tube of 
Arctic Silver.  I put a thin coat on the CPU and the underside of the 
heatsink, clamped it down on the replacement motherboard, powered up, 
and... nothing happened.

>From what others have told me, it worked this way:
Wife's motherboard may or may not have already been damaged, but putting 
her CPU into my new motherboard (from the barebones kit) without full 
protection caused CPU and motherboard both to melt - and I can see the 
bubbles of silicon where this happened.  In TEN SECONDS!  Then when I put 
my melted CPU into my second replacement motherboard, it sent bad signals 
all over the new  motherboard, frying that as well.

It was another four months before I could talk the Mrs. out of enough cash 
for yet another complete barebones system.  In the ever-growing pile of 
junk parts I have a Duron 1.6Ghz that may or may not be fried, a Chaintech 
motherboard that may or may not be fried, and an ATX case that I'm pretty 
sure works.  But I'm afraid to test any of them because if they are fried, 
I don't want them to do any further damage!

All told, this lesson in CPU thermodynamics cost me about $300 in parts.

Peter B. Steiger
Cheyenne, WY
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-chat' in the subject header of the message

More information about the lfs-chat mailing list