|
|
Advertisement:
|
|
Slug-Sanding How-To |
|
Join the community - in the OCAU Forums!
|
Reader's Comments
Reader's Comments:
On your slug sanding article, you ask what the strange little logo in the middle of the slug is. It's a 2D barcode. If you take a look at pretty much any CPU these days (from AMD as well) you will see one of these little squares.
Hope this helps,
Bert Kenward.
A couple of suggestions for future sanding.
Wet sanding takes the material of a little quicker without the build-up of dust that will put scratches in the finished product that are hard to remove. I dry sand only the 1000 or 1200 paper because it polishes where the wet sanding leaves a dull surface.
If you could come up with a way to hold the 370 flat like I did for the SEPP, you wouldn't have the rounded corners (and a little less surface area to transfer heat.)
If you haven't seen my lapping attempt, http://www.benchtest.com/lap.html
Jim Fager
benchtest.com
a touch of detergent in the water or touch the cpu on a bit of wet soap occasionally will help keep the abrasive paper clear of metal particles and prevent it clogging (if wet sanding)
baz
Just thought I'd chip in with my success (?) with CPU sanding. I call it lapping rather than sanding seeing as that is actually the process.
Anyway, the system is typical slot 1 C300A on BH6 with 64MB of Micron RAM, ASUS TNT card (heavily overclocked) and a WD8.4UDMA. I have a cheapish heatsink (Vantec) that had 2 fans on it (total 12cfm) that I recently replaced with one 60mm fan from Jaycar (20cfm). The case is well ventilated and typically keeps the MB temp a few degrees C above ambient.
The CPU is a natural at 450 and 464 at 2.0v. I tend to run it at 2.1v to avoid the occasional BSOD when I shut down Win98. On previous attempts to run 504MHz I received mixed results (none good) from no POST to the occasional boot into 98 followed by instant lock up. This best result at 2.3 volts.
I lapped the CPU and now it will boot into 98 at 504 on 2.2v but often locks up quickly. At 2.3v it will boot into 98 and run Wintune benchmarks, and grind Seti@home for an hour or so or do some QuakeII Crusher timedemos but still eventually locks up.
I am happy...the benchmarks are better than any of my friends or workmates can produce!!
I am convinced that a better heatsink will help me. I am sure that the Vantec doesn't pull up flat even against the newly flat slug. I sort of confirmed this to myself by looking at the pattern of thermal compound transfer to the HS when I spread a REALLY thin layer onto the slug and bolted it up. It wasn't all over the contact patch. I am currently using more thermal goo than I would like given the 'flatness' of the slug.
I have a suggestion for your sanding article. Metho is a good liquid for wet lapping with if the room temperature is low enough to stop it evaporating too fast. And it is also good for rinsing off CPUs that may have metal particles on. It is also good for displacing water (from wet sanding operation) from between the CPU and the board on slot 1 processors.
Cheers
Brad Wilson
Thank you for your article on sanding. A few answers to some
of your questions:
1) The Celeron logo (on both sides) is a laser-encoded, machine
readable code (2D barcode according to Intel docs. It
probably gives product code, but not serial number.
2) The slug is indeed copper, fairly thick (~2mm?), nickel plated.
Again according to Intel docs.
3) I've work out the theoretical benefit of sanding at about 1'C
for each 0.001" effective average gap reduction, even filled
with good 1.2W/mK thermnal grease. 0.005 is possible,
especially adding in the heatsink side.
4) Stability testing is a sore point. I wrote `cpuburn` [Linux]
in optimized asm specifically for testing. See:
http://users.ev1.net/~`redelm.
-- Robert
|
|
Advertisement:
All original content copyright James Rolfe.
All rights reserved. No reproduction allowed without written permission.
Interested in advertising on OCAU? Contact us for info.
|
|
