PDA

View Full Version : AS5 Not Working As Advertised


F_L_C
10-09-05, 05:44 PM
I did a new application of AS5 about 2 weeks ago and have noticed a steady increase in idle temp. This goes against how it's supposed to work (ie as it "cures", temps should drop a few degrees).

I measure effectivness by the difference between ambient and idle temps. When I first fired it up after the application, the difference was about 13 degrees (ex. ambient = 40 and idle = 53). Now the difference has worsened to 17 degrees (ex. ambient = 40 and idle 57). So, idle temps throughtout the curing period has actually increased by 4 degrees and this translates to higher load temps of course.

Why is the AS5 not working as advertised?

keith33
10-09-05, 06:00 PM
How much did you put and what method did you use to apply it?

Omega53
10-09-05, 06:00 PM
You might of put it on too thick. Id reapply it with a thiner layer

Roadhog
10-09-05, 06:02 PM
maybe the heatsink is dusty?

F_L_C
10-09-05, 06:08 PM
I used the dab in the middle approach for large cores: http://www.articsilver.com/arctic_silver_instructions_big2.htm

This is the approach recommended by Clay from this site. It worked great for the first few days (difference between ambient and idle = about 13 degrees). But since then it has steadily increased; it seems to get worse the longer it cures.

As far as the dust, I use a 1/2HP Craftsman compressor to blow the dust away weekly. No difference.

toxikneedle
10-09-05, 06:09 PM
You might not have good contact with the core, and as others mentioned a common mistake is putting too much. You want a small ammount, a little smaller than a piece of rice.

F_L_C
10-09-05, 06:20 PM
I used a little less then the size of a BB, set the sink down, turned it a couple degrees left/right, and then tightened it down. If I used too much, then the effectiveness (performance) would've been bad from the very beginning. But this was not the case, it only got worse after several days of curing.

toxikneedle
10-09-05, 06:24 PM
Well make sure it's tight, I mean remember with AS5, worst case scenario is the temps will stay the same as the old thermal compound. Their thermal compounds are very cheap and I doubt they would out-perform AS5. Did you clean the old paste completely off?

F_L_C
10-09-05, 06:43 PM
Yeah the old paste got the full alcohol treatment. I was just stoked to see such good performance in the beginning and fully expected it to get even better. But it has almost degraded to the point at which it was before. Well, actually it's still better than the eVGA stock paste which had a difference between ambient and idle = about 20 degrees.

aZn_plyR
10-09-05, 07:03 PM
room temp?

F_L_C
10-09-05, 07:22 PM
Room temp has been constant. Matter fact, I've kept all relevant variables constant. The only change is the paste and drivers. I even rolled back to the drivers I was using when I applied the AS5, v78.05, but it didn't make any difference compared to v81.84.

Sazar
10-09-05, 08:22 PM
When I saw this thread, I thought it said ASS not working as advertised and I thought to myself, wow, why is this in the video card section, I MUST see what this is about.

Now that I see it was simply an optical illusion, I shall exit stage left :angel:

gram_vaz
10-09-05, 08:35 PM
as5 advertises a drop of 3-12c. at 'worst' you only get a drop of 3c in temp, at best about 12c.

Rytr
10-09-05, 09:37 PM
There must be another variable to have a variation of 4C+ in temps. If not an increase of room temp or case temp I would suspect the heatsink's fan.
A 3C reduction in temps is normal using AS5 vs. standard white thermal paste. I would not expect much more from just changing thermal paste. The problem is that you saw a 7C reduction which has deteriorated to 3C. To me that shows something else is effecting thermal transfer of heat. Either the heatsink has been moved/jarred and has lost good contact, the fan is failing or not maintaining peek rpm's, or the thermal paste properties are deteriorating...which I have never seen happen in such a short period of time with any of Arctic Silver's products.

acrh2
10-09-05, 11:42 PM
I've used AS5 for years and I've never seen any temperature drops between application and running for a week or so, as other have reported. I do lap and polish my heatsinks though.

rflair
10-10-05, 09:18 AM
I used a little less then the size of a BB, set the sink down, turned it a couple degrees left/right, and then tightened it down. If I used too much, then the effectiveness (performance) would've been bad from the very beginning. But this was not the case, it only got worse after several days of curing.

One thing about AS5 is that as it cures it loose some vescoisty and becomes more of a thick dry grease and will literally suction cup the CPU to the heatsink.

With that said you don't mention spreading the AS5 with a creidtcard or other hard backed surface object evenily over the whole core, slapaing a drop in the midlle and then turning the heatsink a couple of times is not an even application.

F_L_C
10-10-05, 02:21 PM
To me that shows something else is effecting thermal transfer of heat. Either the heatsink has been moved/jarred and has lost good contact, the fan is failing or not maintaining peek rpm's, or the thermal paste properties are deteriorating...which I have never seen happen in such a short period of time with any of Arctic Silver's products.

I'm not a clumsy mofo and did not bump the sink. I've got the Zalman fan connected directly to a mobo header and MBM5 shows it spinning @ 2850-3000 RPM. The only logical conclusion I can draw is that the AS5's properties have deteriorated in the 2 week curing period, if that's even possible.

With that said you don't mention spreading the AS5 with a creidtcard or other hard backed surface object evenily over the whole core, slapaing a drop in the midlle and then turning the heatsink a couple of times is not an even application.

I'm aware of the debate between the two application approaches but went with the other method because:

Our testing has shown that this method minimizes the possibility of air bubbles and voids in the thermal interface between the heat spreader and the heatsink. Since the vast majority of the heat from the core travels directly through the heat spreader, it is more important to have a good interface directly above the actual CPU core than it is to have the heat spreader covered with compound from corner to corner.

http://www.articsilver.com/arctic_silver_instructions_big2.htm