Txtmstrjoe's "2 cent mod" - Overclock.net

txtmstrjoe · 06-26-07

As some might know, I recently decapitated an Opteron 170 (LCB9E 0652WPMW) in a mad experimental fit. In my haste I neglected to run it through a battery of tests, specifically in terms of thermal output. I didn't even bother to run it with its IHS on!

Anyway, I made it into a permanent convertible, then mounted it into my sig rig. On went the lapped Thermalright Ultra 120 Extreme, and I didn't bother to reset the clock settings on the board. The only change in the settings I did was to knock the VCore down to stock settings (1.35V, or 1.150V in the BIOS).

Despite running stock voltages, the thermal readings seemed unusually high. With the IHS off, it was running a little hotter than the CCBBE 0610DPMW Opteron 170 that hitherto was my sig rig's CPU. Even at an insane 1.3V VCore (13+ hrs Orthos stable, no less!), the new LCB9E 0652WPMW was still a hot runner. It made me wish that I kept the IHS on for one run, just to have a basis for comparison.

I inspected the HSF mount, and though nothing was particularly awry, I noticed that the Ultra 120 Extreme had a little more "give" in its mount compared to when a lidded chip was in the socket. I wondered if the now looser mount accounted for worse thermal transference. The mounting bolts were at their limits, too, so the retainer system couldn't be tightened any more.

So I got to thinking: What can I do to make this mount tighter? I hit upon the idea of using shims between the heatsink base's top surface and the retainer bars, so make up some of the space that used to be the IHS.

I didn't want to use plastic, because of the obvious vulnerabilities to heat. Plastic shims might melt. So I thought about using metal for shims.

Being a cheap bastard, I thought that my store of pennies might be suitable, so I got to testing this.

So I disassembled the CPU cooling system:

txtmstrjoe · 06-26-07

Then I found two pennies, lay them on top of the HSF's base, then tested the fit of the retainer bar:

txtmstrjoe · 06-26-07

It all fit well enough, so I put a very thin coat of AS5 on the naked die, then remounted the Ultra 120 Extreme back onto the CPU.

The fit of the cooler was definitely much tighter; it had almost no wiggle now. The shims definitely improved the tightness of the mount.

Would it translate to better thermal dissipation, though?

The screenshot taken on 6/20/07 was before my "2 cent mod."

In contrast, the screenshot taken on 6/25/07 is after the mod.

At similar ambient temperatures, there is a noticeable reduction in Orthos full load temperature.

Hope this thread demonstrates that even pennies are worth something.

DuckieHo · 06-26-07

Are they pre-1982 pennies? If not, change them out and you might see a further reduction in temperature. Current pennies are zinc with a very thin layer of copper.

DesertRat · 06-26-07

Quote:

Originally Posted by DuckieHo

Are they pre-1982 pennies? If not, change them out and you might see a further reduction in temperature. Current pennies are zinc with a very thin layer of copper.

lol. he's just using them as shims(very ingenuitive no less). They have nothing to do w/ actual active participation in heat dissipation.

alk · 06-26-07

Quote:

Originally Posted by DuckieHo

Are they pre-1982 pennies? If not, change them out and you might see a further reduction in temperature. Current pennies are zinc with a very thin layer of copper.

Lol. He's only using them as spacers between the top of the HSF base and the bracket (Check pics). Adding an extra few mm^2 to the non-active cooled part of the Heatsink isn't going to drop temps by any noticeable ammount.

EDIT: DesertRat got there first!

Mr. Jules · 06-26-07

Wow, good thinking! Nice decrease in temps for 2 cents, i'm suprised just a little more pressure can change that much!

DuckieHo · 06-26-07

Quote:

Originally Posted by DesertRat

lol. he's just using them as shims(very ingenuitive no less). They have nothing to do w/ actual active participation in heat dissipation.

I already know... just making a comment. It may help the heat path and it will only cost... $.02 to change them.

DesertRat · 06-26-07

Quote:

Originally Posted by DuckieHo

I already know... just making a comment. It may help the heat path and it will only cost... $.02 to change them.

true... I tried looking for some pre-1982 pennies around the house once and it took hrs jsut to find like 3, lol. actully i wasn't really trying, BUT ANYWAY!

txtmstrjoe · 06-26-07

Quote:

Originally Posted by DuckieHo

Are they pre-1982 pennies? If not, change them out and you might see a further reduction in temperature. Current pennies are zinc with a very thin layer of copper.

Hmmm... similar to some heatsinks? (Zalman?)

Maybe I should lap the pennies.

Actually, I didn't bother to check the age of the coins. But I just used them as shims.

Thanks for the idea, though.

		Overclock.net - Overclocking.net > Cooling > Cooling Experiments
Txtmstrjoe's "2 cent mod"

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)