- ultimate "goop" better than toothpaste and as3 -

after a million trials and errors and about three burned up durons, i know i've narrowed it down to the last three things...

1. gold foil ~60usd for a really small roll

2. the soldering method

3. the old thermometer solution


so, after this: http://forum.oc-forums.com/vb/showthread.php?s=&threadid=226879 i've now concluded that basically everything, EVERYTHING, is better than as3, pcm+ etc... (e.g. nyquill, dayquill, vaseline, first aid cream, water, lithium grease, and stuff just like it)

btw, tests were done at an ambient of 20 celcius and with the coolermaster aero 7 (fan at max), a t-bird 1ghz oc'ed to 1.33ghz and an abit NF7-S 2.0 (1.8 bios, august ver board, w/plastic clip handle)

all tests started out with a really cheap duron with 2.0v vcore mod on mb and an old celeron hsf and if it proves that its not going to catch fire then its re-done with a lapped t-bird 1.0ghz oc'ed to 1.33ghz..

all results were the best out of three mounts each for each paste and after every trial the hsf base was cleaned off and relapped with 600 grit sandpaper and the cpu soap and watered... artic silver was tried on first as the control.... temps were all taken from the board bios, v1.8...

so...

first the gold foil http://www.sisweb.com/vacuum/sis/goldmatrl.htm

the thinnest pure gold foil apparently gives about a 2 degree lead over the ceramique on the duron on an older via board, but, i ran out of it for a proper mount of the aero 7.. will have to post updated results after i get some more...

then the soldiering method

i've perfected this now down to the best it can go but its still pretty unsafe cause i lost two durons this way... you'll need a board with a functional thermal shut-down at ~100 celcius, lowest melting solder (preferably foil and you can find stuff that melts at ~105 celcius) and a blow-dryer and a thermometer that goes to 120 celcius

1. freshly lap ur cpu so its not all smooth-ish and also remove the thin plastic kinda coating over it (yes there is one)

2. lap ur hsf to 600 grit and MAKE SURE ITS FLAT (you only have ONE shot at a proper seating and if you screw up its the whole thing over again assuming something hasn't caught on fire yet)

3. apply a THIN layer of solder flux to both the cpu and the bottom of the hsf/waterblock where its supposed to mount on

4. put a piece of the solder foil in between the cpu and the hsf/wb

5. attach clips/bolts /whatever and mount the hsf/wb

6. remove the fan plug its its it

7. use the blow-dryer on the hsf/we and fire up the board and wait till it auto shuts down then measure the hsf temp all the way untill it hits 120 celcius and then quit heating it and let it cool off

8. done

this method apparently gives the best temps, ~5-6 celcius lead over as ceramique on the duron with 2.0v vcore... i also haven't tried this on the t-bird but it should work the same... it all depends on how ur board is responding to the thermal shut down thing...


and now the much easier method

while i was messing around measuring temps and all on the heatsink, i cracked my really fragile thermometer which was bad since it was mercury cause mercury was poisonous, but then with it came the ultimate method of "soldering" ... why not just use the mercury?

and so, with the lapped core and the lapped hsf, it took the longest time to figure out how to actually USE the mercury and get the stuff to sit somewhere AND STAY THERE... man it was a drag cause the stuff would just bunch up and run away if left somewhere then it hit me... drill a small dimple right above the cpu on the hsf and sit the mercury there and then mount the board upside down...

and so even with that dimply of the ultimately uneven contact, the thing beat down any other paste possibly out there and made a full 5 celcius lead over the ceramique...


so there, now i really think everything, EVERTHING is better than as3, ceramique, etc...

one word: hardcore

yes everything insane is better than as3 ;d but for the rest of us sane/wimps we use as3 ;x hehe

oh yah, did i mention that all temps were taken from the best mount and then the comparison temps taken two days afterwards?

just enough time for the ceramique to set...

and the water to evaporate from the toothpaste

Why not silver foil? Shouldn't it be better than gold anyway...

cause i couldn't find any "pure" stuff that was *that* thin... and from what i remember gold was a bit softer, which might help fill some more gaps....


and also gold foil tends to be a bit more even and since its foil oxidization would definitely be a problem... otherwise i'd have used copper or something...

There may be other things "better" than AS3/Nanotherm in terms of pure thermal conductivity, such as pretty much any water-based paste, and liquid metal, but, for a real-world application, they have significant drawbacks.

The water-based pastes will all evaporate very quickly, leaving nothing but thermally insulative powder or grit between the CPU and the heatsink. That is bad. That problem cannot be overcome without running the system below freezing point, and using ice as the thermal contact material.

Liquid metal is insanely impractical; the solder method will fry the CPU more times than it'll work, and mercury vapour is highly toxic.

What is really needed is a solid, stable, spreadable, non-toxic, non-evaporating paste that's reasonably thermally conductive... just like AS3.

mercury next to plutonium is the most toxic metal you can come in contact with.

mercury in the form of vapor when inhaled readily crosses the blood brain barrier and causes brain damage.

My brother in law removes toxic waste and many jobs are private homes where kids have learning disabilities due to vapor from a broken thermometer.

heating mercury on a die will increase the vapor concentration in the air.

I advise getting rid of all sources and get tested (hair analysis) and maybe start a program of chelation, maybe take 2-5 years to get it out of your system

"Mad as a hater"
hope you dont go too mad fafnir

Awsome, but like my friend chris jumping off of a small platform with the wind blowing ontop of a train tressle 70 feet in the air leaping over telephone wires into a shallow river where he hit bottom and only came out with a few scratches from the rocks; 'awsome' might not be completely sain. (true story )

If you ran a subzero system would the minute condensation let the toohpaste stay wattery so that it would work well?

I doubt that the condensation would get into the right spot to keep the toothpaste moist. It would eventuall drip somwhere it shouldn't and fry something after a while.

Haha! Hardcore dude Fafnir !!

Well, it seems like you've started with the mercury solution first eh? ...otherwize, how did you become so insane ?!!!

Anyways, the tin method seems to be the most reliable.
Did the tin actually WET the silicon wafer?
What tin did you use? (I've heard about indium alloy tin which melt at 75C)
Where did you get the tin? (I'm in sweden too)

I had a thread going at the Technical discussion about soldering coolers to the cpu-core before the forum got re-arranged.

What about banana's? How well do they transfer the heat.

I ask because if you've been exposed to too much mercury you might turn bananas.

heheh. Hardcore dude.

What about WATER SOLUBLE SOLDER PASTE?

nah, i think i'll be alrite,... i was a bit raxed about the mercury evaporating and stuff so i also used some prometeia seal string along with it and a toxic vapour detector in the case... either way the room was well ventilated but i ended up ripping apart the rig after needing the cpu for some other "testing" which'll proally leave me with another keychain too many...

and Pax, yah if you actually get the thing heated right the solder WILL melt right around the cpu and leave almost an "impression" when ripped apart... but i lost one proc that way cause ripping off the hsf cracked the die,... so....


but anyways, i think i'm actually on to something now that involves a waterblock that can beat the cascade... and the best thing about it is that it "provides" its own "goop" joint, you know what i'm thinking?