I've been using AS3 for HSF but I ran out , so , i used the white compound ( I believe Stars ) that came with the HSF , noticed that temps Dropped by 4-5* C . When the new tube of AS3 came in I throughly recleaned and applied AS3 and the temps went back up . I experimented SEVERAL times
with diffferent thickness of compound ( from HAZE thin to fairly heavy , all with same results , tried again with the white stuff and temps went back down.
WHAT could be the possible cause , is this white stuff (says Stars on pack ) be better ???
or what am I doing wrong ?? ( yes I know to use a paper thin layer )
using it on XP2100 & Volcano 9
Well, I'm waiting on my processor right now, and i was planning on using Artic Silver 2, but I also have a packet of stars that came with my volcano 9. I'm thinking mabye stars works better with copper base heatsinks, or something like that, all metals are different. But since we have the same hsf i think i'll use the stars too and tell you how my results are, the thing is i don't know if my processors coming, its been taking a while. Hey, by the way, is the volcano 9 very loud? I haven't decided if i'm going to use the adjustable speed or use the temp control speed. If its loud i'll just use the adjustable and keep it low, I don't want it to be loud. Its only 17dBa at 1,300rpm.
The XP2100+ is a hot running sucker I've tried 4 different cooling solutions to keep it where I'm satisfied with the temps, and right now believe it or not I'm running an older Volcano 6, which has outperformed the other solutions, using ASIII compound.
Details, Details, Its all in the Details, If you need help, Don't leave out the Details.
I have an XP2100 running with a 143Mhz FSB on a MSI 745 Ultra - with ASIII and an Akasa Silver Mountain 2Q + quiet case fan settings, the difference between core and case temps under SiSoft Burn-in is only 8 degrees Celsius (36 vs 28). I don't how accurate those readings are, but the difference was 16 degrees Celsius with my old stock HSF and thermal gunk (same stress test and quiet case fan speeds).
I fiddled around with the ASIII quite a bit too; it is funny stuff. However, I settled on an almost transluscent layer on the cpu die and an almost transluscent layer on the heatsink; both achieved by working the ASIII into each surface with a finger-inside-a-plastic-bag. Oh, I also gave each surface a good clean with undiluted isopropyl alcohol.
I like the Volcano 9 , using it manual adjust , @3100 rpm
it very quiet , CPU temp 52* load (prime 95) temp , with ambient temp 26* c , using the stars compound , with AS3
the temps are 56 - 58 , after each test HSF was throughly
cleaned with alcohol . ( yes the XP2100 is a hot one )
gonna try this test on my other machine, another very hot one 1.33 t-bird sitting under a slk-800
maybe there is something to the copper re-action to different compounds , and contrary to EVERYONES belief
that Artic silver may not be the best for every HSF
You won't be the first person to report better results with less expensive heat sink grease. I use the stuff in the little tubes from Radio Shack and get excellent results with it (XP 1800+, Spire Big Rock, temp = 46c).
There is one trick I use on AMD CPUs. If you rub your finger over the heat island you will find it's rather rough feeling. A magnifying glass revealed this to be some kind of grit on the metal surface (from laser etching the logo?).
What I do --VERY GENTLY-- is polish the heat island surface with very fine steel wool, until if feels smooth. Then I clean the entire CPU with alcohol and a cotton swab to make sure both grit and steel wool are completely removed. This has lowered CPU temps by 4 and 5 degrees more than once.
Another thing I commonly do, especially with copper bottom heatsinks, is to "lap" them on a flat surface with extra fine emery cloth. Put the emery grit side up on a FLAT surface and rub the heat sink across it a few times. This will ensure there are no bumps or lumps in the metal surface that has to contact the CPU. Following the lapping process I take a dab of heat sink grease on a piece of tissue paper and polish the contact area of the heat sink to a nice gloss. This has taken another couple of degrees off the running temps, more than once.
After cleaning and polishing, I apply the grease by putting a small blob in the middle of the CPU heat island and tapping it with my finger, creating a smooth thin layer that just covers the top of the island. Any excess is then removed from the CPU surface with a cotton swab.
Next install the CPU and Heat Sink and try it out...
The real trick is to get the best contact between the two metal surfaces... Cleanliness and smoothness of these surfaces appears to be far more important than most people think.
results from thermal compound test , first I dont recall Tom's running Current comparision tests of different thermal grease/compounds with the copper base HSF , if not , maybe it would be interesting read , I tried this test first out of necessity on XP2100 and volcano 9 , AS3 vs. Stars white compound . Stars won
Next we tried this on 1.33 t-bird under a SLK800 w/panaflo fan.
Quess what , Stars white compound also dropped the load temps by 3-4* C.
this isn't scientific, but very interesting , i have always heard that AS2-3 was the BEST and I know most ALL folks are using it ,but there may be Better compounds out there .
Note , with each test , the heatsinks/CPU's were THROUGHLY CLEANED .
Starting load temp w/1.33 t-bird 42* c w/AS3
( has been like this for months now )
load temps with Stars 38* c
same ambient temp and conditions
I have installed many HSF , worked at local comp. repair shop. and always used AS , the boss insisted on this
but there may be a better solution for Todays high temp CPU's and copper based HSF