Arctic Silver on P4? Shop says it's bad?

[gramboh]

Hi there,

I just placed an order for a new machine today, Asus P4P800-UAY with P4 2.6C and an SLK-900U. I asked for Arctic Silver to be used on the heatsink (I'm paying them a small fee to build it) and the salesperson told me that Intel/Thermalright advise against using AS3 on the 900U/P4 because it can actually increase temps. This sounded way off base to me, considering how much I've read about it here.

Can anyone advise me if this is true? He said he would check with the system builders and call me about it tommorow before the work on it, any advice would be appreciated. He said the 900 comes with some kind of thermal paste but I can't verify if that is true or not. I don't want to use those thermal pads because I know they aren't as good as AS3 and are messy.

 

[Corporate Thug]

wtf? i have never heard that before and i doubt that there is any legitamacy behind it

 

[Duvie]

Hmmm....Well since the IHS from the surface looks the same with the p4c models I can assume it is like the p4a and b models and therefore I think a hundred or so Anandtech users have seen performance contrary to that....

 

[4x4expy]

I wonder if these guys know the proper method of applying AS3. If not, they may have gotten the results they spoke of. If they build computers on a daily basis, maybe they're a bit too proud to read instructions for a simple tube of thermal paste.

 

[WarCon]

What are they suggesting? A thermal pad? Or just generic thermal compound? Anyway, I have tested the pad (on the retail heatsink) versus AS3 and there are several degrees C difference at stock. If overclocking is something you may want to do then you will notice even more of a difference as you increase the vcore even slightly.

If they just wanna use a "thinner" consistency paste you will probably only be at most a couple degrees different and it is alot easier to apply correctly as most will automatically squish out. Only thing I would really consider, is how is the paste they wanna use been tested for longevity. I trust AS3 to be in the same condition years later as shortly after I put it on. I can't say that for cheaper pastes.

Make them show you where Intel suggests that you shouldn't use Arctic Silver paste. Only warning I ever remember hearing about was from AMD and that was for entirely different reasons (but I would still use it on an AMD over most other pastes).

P.S I just searched Thermalright and Intel's site for any reference to Arctic Silver and seen none, so I would be very curious for them to be able to produce something tangible. Also you might suggest that they visit Arctic Silver's site for proper installation instructions for their thermal compounds.

 

[Thor86]

I use AS2, and it works great! I have no idea where your shop got such nonsense. But then again, they may use/prefer something else that they have in stock, and not AS3. :frown:

 

[MercenaryForHire]

1) Yes, the SLK-900U comes with a tube of Thermalright paste. It's actually only a little less potent than Arctic Silver.
2) Remember that 98% of Technicians/Experts know as much about cooling and system tweaking as my ass knows about putting up stovepipes. And that's not a whole lot at all.

- M4H

 

[Quackmaster]

Originally posted by: MercenaryForHire
1) Yes, the SLK-900U comes with a tube of Thermalright paste. It's actually only a little less potent than Arctic Silver. 2) Remember that 98% of Technicians/Experts know as much about cooling and system tweaking as my ass knows about putting up stovepipes. And that's not a whole lot at all. - M4H

too funny There is also the possibility that the shop is just covering it's a** so they don't have to deal with RMAs for stuff under warranty.

 

[Zarick]

actually if you read amds white paper it will tell you the same thing. AMD says the best thing to use is thermal tape. I would have to think that both AMD and Intel have tested this and somewhere there is validity to it.

 

[Quackmaster]

Originally posted by: Zarick
actually if you read amds white paper it will tell you the same thing. AMD says the best thing to use is thermal tape. I would have to think that both AMD and Intel have tested this and somewhere there is validity to it.

Are you friggin' kidding me? Have you ever looked around at the overclocking community and seen the thousands of folks who have proven otherwise?

Normally I would be inclined to agree with you, but since when have Intel or AMD ever encouraged overclocking. I would guess that 90% or more of people who use anything but the stock interface material (it's a slim pad nowadays, not thermal tape) are overclocking. So between the fact that Intel and AMD know most people are buying aftermarket stuff to overclock (lets not forget they are BOTH AMD and Intel are increasing measures to stop us. Intel has even been given an "anti-overclocking" patent.), and the fact that a lot of people will just plain screw up in applying the aftermarket materials, what do you think Chipzilla and AMD are gonna do/say?

If I sell you a sportscar with a full bumper to bumper warranty including the transmisson and you tell me you want to put an aftermarket supercharger or turbo in the vehicle, do you really think I'm gonna want to encourage you knowing I'm gonna likely be cleaning up your mess if you screw it up? Think about it.

 

[rogue1979]

Artic Silver applied properly will outperform any generic thermal paste and most certainly a thermal pad. If you have any doubts, try both yourself. Then when you obviously have lower cpu temps with AS3 you can then determine whether or not that is better.

 

[pspada]

Just one question: when the shop said "it's bad", did they sound like Mr. Mackie? "AS3 is bad, don't do AS3, boys, it's bad."

 

[Zarick]

I think it would be interesting for a review site to do a comparsion of all different options. Maybe AS3 is simply a percieved improvement.

Personally I used artic silver 3 and saw absolutely no improvement in temperatures on my t-bird vs thermal tape.

 

[gramboh]

Well I talked to the guy from the shop on the phone, after doing a bit of research myself (look at this page http://www.overclockers.com/tips00357/ ) Apparently it is just a warranty issue, although according to that site, and Intel's document, there is no explicit limitation on what thermal paste/tape you can use on their retail boxed processors. I think the person who told me it doesn't perform well was just trying to discourage me due to the warranty issue, since once I brought this up he agreed with me about the real reason, although he said it shouldn't be a problem if you had to warranty the CPU as long as you can clean it well enough that Intel won't know (I assume this is possible).

Thanks.

 

[Mingon]

they dont recommend as3 because of phase change problems when going from room tempreture to 95c - because that happens all the time

 

[WarCon]

Originally posted by: Zarick
I think it would be interesting for a review site to do a comparsion of all different options. Maybe AS3 is simply a percieved improvement.

Personally I used artic silver 3 and saw absolutely no improvement in temperatures on my t-bird vs thermal tape.

You have to be kidding me, on the small core of a T-bird is where you will see a marked difference between any paste and a thermal pad. The huge transfer area of a P4's heatspreader minimizes some of the difference between thermal compounds, but the difference between a pad and thermal compound is obvious there also.

You either installed the thermal compound wrong, have no clue how to measure your temps or are lying. The pads are simply used because mostly they are moron proof (almost) and because they are wax (which really sucks as a thermal compound but is better than air) they don't evaporate over time. Why AMD recommends them is they help prevent core crushing and chipping for the lesser experienced installers primarily, not because they provide a better thermal transfer.

Anyway................:disgust:

Edit: For Mingon, huh? From specs on webpage "Extended temperature limits: ? 40°C to >180°C" - for the Celcius challenged thats ? 40°F - 356°F , so I am kinda hoping you are being sarcastic.

 

[zodder]

Edit: For Mingon, huh? From specs on webpage "Extended temperature limits: ? 40°C to >180°C" - for the Celcius challenged thats ? 40°F - 356°F , so I am kinda hoping you are being sarcastic.

I think that's a pretty safe bet.

 

[Mingon]

found the quote

[/quote]AMD recommends that only pre-attached thermal pads (Phase Change Material) be used with their CPUs. This helps cushion the CPU during heatsink mounting and also helps prevent overclocking. (The pads often make the chip run too hot to overclock.)

So neither Arctic Silver nor any other thermal compound is recommended by AMD for long term use and technically voids the CPU's warranty. Overclocking the chip or using a heatsink other than the stock heatsink on a retail chip also voids the warranty.

The only grease approved by AMD is G749 which is approved for short-term testing to verify that a CPU works. AMD recommends the G749 grease for short-term use because it is made by an ISO 9001 certified company and is non-metal based.

AMD also says that thermal grease is subject to "pump-out".

Since AMD would have no control over a user's selection of cooling components, AMD needed to assume the worst for their tests; a hot chip, a poor heatsink, adverse ambient conditions and mediocre thermal grease. Under these conditions, grease pump-out could definitely be a problem.

However...

In a recent Intel test, they only experienced grease pump-out when the junction temperature was cycled from 0C to 100C. (A 100C range) When the junction temperature was cycled from 0C to 80C, (An 80C range) there was no pump-out observed. Since the junction temperatures on computer systems using quality heatsinks and thermal greases will only cycle from 20C to 60C at the most (A 40C range), there is little expectation of pump-out occurring.

But AMD needs to make one general recommendation for a thermal interface material that will provide adequate performance in all standard situations. (Non overclocked, stock heatsink, etc.) PCM is pre-attached to the heatsink. It cannot be forgotten by the first time user or the assembly line worker and as I said, it provides a cushion as the heatsink is mounted.

Look at it this way, if you applied the same criteria to selecting a car, everyone would be driving a low power, fuel efficient station wagon with rain tires and foot-thick rubber bumpers all the way around

So as the starting point is always room termpreture as long as you dont A) live in the artic and B) have a chrome orb on your xp 3200 then you should be OK

 

[ChampionAtTufshop]

under normal *cough*operating*cough* conditions cpus will not jump in temerature much

what place is this in calgary??
dont tell me techtronics lol
first, and last, time i went there, i asked waht the dpi for a crt monitor they had there was.....the guy responded with "i'm sorry i dont know your "tech" babble. If you find out what your "tech" babble is in actual terms, i can help you out"....he actually quoted the "tech" as i did in the sentence lol

next, i asked him what the pixel reponse rate was for an lcd....he responded with "you are confused once again, the refresh rate as it is called on this monitor is 60Hz" lol

i ignored his stupidity, to my dismay, and asked him if the particular lcd had a digital interface or an analog interface or both ....he said "it has the usual interface videocards ouput. what is with this digital and analog business?" (looking at me with a "wtf are you talkin about dumbass" look on his face...)

i didnt even bother with him....i just walked out lol

 

[Zarick]

sorry warcon I saw no difference.
I have done it twice now and still get the same temps.
It could be however, that my hs/f is just crappy, but from what I have found by searching.. the temps on my proc are pretty normal for a t-bird.

 

[WarCon]

How are you applying the thermal compound? Are you following the instructions for application found at Arctic Silver's website?

 

[gramboh]

HAHAHA Techtronics, that is hilarious, an employee actually said those things? Believe it or not this is Memory Express. They are using the AS3 now, the issue wasn't ever performance, the guy just said that because he knew about the warranty issue. Once I explained to him that I was also aware of this he said it's fine as long as I understand that if I have to warranty the chip I will have to clean it very well so Intel can't see what was going on. The strange thing is since I am using a non-retail heatsink, that would probably void it anyway (not to mention overclocking). To me, a CPU decreases in value so quickly that the warranty is irrelevant, unless the chip is DOA or dies within a few months (never heard of this happening unless it was due to overheating or core cracking).

 

[Quackmaster]

I could go through all my saved webpages and bookmarks to dig up some thermal compound comparisons, but that's a pain in the a** for something as ridiclous as this issue being debatable. Anyone who thinks AS3 or anything else doesn't help temps, fine. Use toothpaste for heatsink compound (I did it once when I was in a pinch. No, this is not a joke. A mechanic buddy of mine taught me that trick And yes it sucked, but worked better than nothing.)

If you have doubts I suggest heading to overclockers.com, OCtools, overclockers australia, Spodes Abode, etc etc. I've seen too many reviews that back up the the "theory" that aftermarket compounds are better than the stock crap. I've also proven it to myself time and time again in personal usage.

I will say that AS3 and some other metal based compounds can possibly have adverse consequences on Athlon bridges, or mobo contacts if it connects two circuits. But otherwise, it r0x0rs.

 

[Jeff7181]

All CPU manufacturers say not to use AS3 just to cover their ass. AMD's excuse is that as temps increase it thins out and pumpout occurs, leaving no crap between the core and heatsink to properly transfer heat. I don't buy it... we're not talking about a layer even the thickness of a piece of paper... we're talking about filling near microscopic valleys... you'd have to heat up AS3 pretty high to get it to drip... it's fairly thin to begin with, but it doesn't thin out much even as it's temp increases.

 

[KF]

Zarick, your experience conforms to mine, and probably most people. I could not be sure there was a 1 degree C difference after perhaps 8 hours in use. After a couple of days, the results of any stress tests seemed identical. Very few of the web sites that have done a comparison have done it in a way that the results could be considered anything but accidental. Usually they do nothing to be sure the pad has had time to thin out. When you remove the HS you can see whether it has or not. It is obvious they believe before they have done the test that phase change pads must be worthless. Therefore, without a blind test, the result is suspect.

There is nothing conclusive in the physics of the matter that would say AS is measurably better than a phase change pad. The materials involved are too thin to say definately the temperature drop must be significant. Here is the problem: suppose AS were as tremendous as 10 times better than a phase change pad, but the actual tempereature drop was 0.1 for the pad and .01 for AS. AS would be super-duper, but it wouldn't be of much use.