Of Artic Silver and T.I.M's

I recently bought a GlobalWin SAK38, but as I was unable to get any Arctic Silver at the time, had to use the standard thermal pad. It has been established that my mobo reads temps high from the onboard sensor, so actual temps are not really important. What is interesting is the fact that the thermal interface material that came with the HSF, performs better than Arctic Silver III.

Initially my temps with the HSF straight out of the box were 41.9 C Idle and 48 C under load. With the HS lapped and Arctic Silver III applied as per their website, the temps are now 42.9 C and 50C, with rooms temps and case temps being the same. I have checked that the lapping is level and have repeated re-seating the HS with the same results. I also noticed this with my old ThermoEngine, which as I remember had a similar "pad" to the GlobalWin


I realise that Arctic Silver claim that temps will drop during a settling in period. The HSF has been applied for 18 Hrs now and I have yet to see a drop in temps. I am coming to (perhaps an erroneous) conclusion, that some thermal pads are not quite as bad as we think!!!!

Turn off your computer over night. Then turn it back on. After I applied my ASIII, temps were lower, but after I turned off the computer for the night (my fan is loud) and turned it back on, temps dropped 3 degrees C. Also, make sure you have enough, or not too much ASIII, that can effect how well it performs also.

"some thermal pads are not quite as bad as we think!!!!"

I suspect, but cannot prove (due to not having a sampling of PCMs), the same

and do recall that AMD recommends PCM, NOT GREASE for long term use (installations)
- due to the 'pumping' caused by on/off cycling and differential expansion

some threads you may find of interest
http://www.tekforums.co.uk:81/posts.php?threadId=5140
http://forums.overclockers.com.au/showthread.php?s=&threadid=67401
http://forums.overclockers.com.au/showthread.php?s=&threadid=69388

its a real can - or should I say tube ? - of worms

be cool

yep, many TIMS are actually quite good. Some are better than paste, including AS2/3.

The reason we overclockers hate them is because they can only be used one time. Every time you remove your HSF, you will need a new TIM or will need to remove it and add thermal grease. For me, at least, this makes TIMs worthless to me.

Phase-change TIMs are usually as good as AS2/3 if not better- they melt into a very thin liquid, and the HSF mounting pressure forces the TIM into cracks and crevices. When the TIM dries again, it is perfectly filling the CPU/HSF junction. AS2/3 is just too goopy to get squeezed around like that.

Of course, as overclockers we lap our HSFs to a shine- there are no cracks for a TIM to melt into. For lapped HSFs, thermal patse is almost always better. If your HSF is not machined very well and has a rough bottom, I would leave the TIM on there.

(personally, I do not think AS3 really changes at all over time. I think it is marketing hoopla, that's all. No matter how long my AS3 is installed for, my temps do not change. Not even if I do a high-heat burn-in.)

linkys for ya:
http://www.dansdata.com/goop.htm - thermal goo comparo
http://www.dansdata.com/powerdev.htm - TIM review

I could never get my temps right with AS3 until I discovered something. I was going off the pic of the processor they have on thier site as a guide. But, if you look closely this is what it says

"Here is a close up of the top of the CPU core with a fairly thick layer of Arctic Silver thermal compound applied. Lapped heatsinks would probably require a thinner layer while heatsinks with very rough bases could require a slightly thicker layer."

and

"The flatter the mating surfaces, the thinner the layer that is required. Stock processors and/or heatsinks with normal surface irregularities will require a layer 0.003" to 0.005 thick as shown below to fill the resultant gaps. (Equal to the thickness of about 1 sheet of standard weight paper.) Properly lapped heatsinks with mirror finishes will only require a translucent haze ."


Notice Translucent Haze. Thats just it. If you lap a heatsink you want a layer thin enough to see though. I did that my last couple of times and temps are much better. After reapplying like this the first time my temps dropped 4C.

Remember, in this case, LESS IS BETTER.

13oots2 said:

I recently bought a GlobalWin SAK38, but as I was unable to get any Arctic Silver at the time, had to use the standard thermal pad. It has been established that my mobo reads temps high from the onboard sensor, so actual temps are not really important. What is interesting is the fact that the thermal interface material that came with the HSF, performs better than Arctic Silver III.

Initially my temps with the HSF straight out of the box were 41.9 C Idle and 48 C under load. With the HS lapped and Arctic Silver III applied as per their website, the temps are now 42.9 C and 50C, with rooms temps and case temps being the same. I have checked that the lapping is level and have repeated re-seating the HS with the same results. I also noticed this with my old ThermoEngine, which as I remember had a similar "pad" to the GlobalWin


I realise that Arctic Silver claim that temps will drop during a settling in period. The HSF has been applied for 18 Hrs now and I have yet to see a drop in temps. I am coming to (perhaps an erroneous) conclusion, that some thermal pads are not quite as bad as we think!!!!

Click to expand...

you may be right, some of the more expensive pads work really well, the thermo engine apparently had a really good pad, but i took it off before i found out it was good.

'nother link
http://www.overclockers.com/articles502/

clamping force is the unquantified variable
tests in progress

be cool

It is actually possible to buy high quality T.I.M's so I may order some and conduct a couple of experiments. Just look for thermal pads or T.I.M's, there appear to be plenty of places that stock them. On the subject of applying Artic Silver, I have followed Artic Silvers instructions to the letter, and my temps remain the same each time I have tried. If I was doing something wrong I would expect some difference in my results.

BillA:- On the subject of Cooling Flow, I too experienced that it was a better paste than ArticSilver and Thermal Pads to start with. After using it for 3 months my CPU temps had risen by 5C, when I removed the HS my CPU core was discoloured and the paste had appeared to have undergone a colour change. It was now a darker shade, and appeared to be re-acting with the aluminium HS. I have thoroughly cleaned the CPU core with Isopropyl, but the colour change seems permenant.

re Cooling Flow
only a series of 24hr tests so far

I have found its ease of application to be far better than AS
and this yields greater consistency -> sensitivity
(mind that I'm using calibrated RTDs with 0.01^C resolution)

the results are so much 'better' (= consistency) in fact,
that I have identified several problems with my wb mounting setup

(the threaded mobo wb mounting screws were causing the wb to 'hang up' as the springs were incrementally tightened - nylon shoulder washers eliminated this; a recommended mod for all)

for sure CF is different
it has much more tack and cohesion, yet its viscosity is less
the higher tack makes cleanup more difficult, I scrub several times with Xylene

I have only copper; but if CF causes an oxide formation with aluminum, it could affect the heat transfer

be cool