[conspire] socket A cpu cooler attachment/s

[Mark S Bilk]

On Fri, Aug 26, 2005 at 05:38:49PM -0700, bruce coston wrote:
>I hope someone has experience with these as trying
>mine out shows i need about the strength of 9 men to
>get it on properly, my cooler certainly looks right
>for my socket

Actually, you only need the strength of 8 men, but they all 
have to be professional wrestlers on steroids.  

The instruction sheet for installing my AMD Athlon XP CPU said 
that you must keep the cooler (heatsink/fan) absolutely level 
when installing it, otherwise the processor chip, which is 
exposed and makes direct contact with the heatsink (through the
thermal goop), can be cracked!

I tried for hours, but only had the help of 6 wrestlers.  And 
though they jumped off the top rope and landed on the cooler, 
and even snuck up behind it and slammed it with a metal folding 
chair, they could not get the locking bracket latched on both 
sides of the socket while keeping the cooler level.

If you try to latch both sides at the same time, there are two
problems.  It will be impossible to maintain the necessary 
horizontal alignment of the cooler relative to the CPU chip,
and one of the two screwdrivers used to push down the two ends
of the spring may slip and damage the motherboard, perhaps 
fatally.

Fortunately I had obtained a copy of the O'Reilly book _Building
the Perfect PC_, at 100% rebate from Fry's, which is full of 
useful advice.  It says that these coolers are indeed extremely
difficult to install, and that you have to latch one side of 
the spring in place first (which is easy), and then latch the
other.  

If I remember correctly, I had to tilt the cooler to even get 
one side latched, violating AMD's instructions.  The locking 
spring has an asymmetrical "V" shape, and the point of the V has 
to be directly over the center of the chip.  I latched the short 
arm of the V first, on the theory that the long arm would be 
easier to bend.  I think I used a screwdriver to push the long 
arm downward, and was able to latch it.  Of course the cooler was 
pushing asymmetrically against the CPU chip during this procedure.  
But I did my best to keep the cooler level; this may be essential 
to avoid disaster.

I don't think the chip got cracked, although if a few of the 
high-end bits got cracked off I might not notice, since the 
programs I'm running may not count high enough to need all 32.

Since you have to apply tremendous pressure to the cooler, cpu, 
socket and motherboard during this procedure, you have to do it 
with the motherboard not installed in the case, and with something 
supporting it directly under the socket, like a paperback book
with some soft padding on top to distribute the force.  I don't 
remember if there were any components mounted on the bottom of
the mobo around the socket, but if so you don't want the support 
to exert any pressure on them.

Incidentally, I tried using a round (twisted-pair) EIDE disk 
cable, rated for 166 MHz, on the 133 MHz disk, and got all kinds 
of errors in the written files.  It was scary.  I had to switch 
to a flat cable and start all over from the beginning, partitioning 
the drive, creating the filesystems, and reinstalling Linux.  
I read somewhere that the round cables violate the specifications.

Good luck!

  Mark