Did I apply too much arctic silver?

S

Spare-Flair

Supreme [H]ardness
Joined
Apr 4, 2003
Messages
7,471
Without thinking about checking the latest instructions on the website of how to apply AS (ie: the single line in the middle between the 4 cores), I went ahead and applied it as I used to on my old AMD 64 CPUs.

I used a credit card to paint and scrape an opague layer on the heatspreader of my Q6600 but I also put a good size drop of silver on the middle of the heatsink baseplate as well before putting it on.

I probably have about 2-2.5x the silver as the recommended application (single line).

Should I do it over? Not happy about Prime95 temps now that I am using a decent temperature monitor (Realtemp and thermistors as well). On small fft test, it goes to 80° C after 3 hours but the CPU is a hot voltage hungry B3 stepping at 1.425V

Now I'm not sure if this is the compound's fault or because my Panaflo 120mm medium flow (83 CFM) is getting starved of air since it is very thick and only about 1 inch away from the rear 120mm intake fake (case has reverse airflow).
 
you put it on the cpu and the heatsink? Yeah, way too much. Pick one and stick to it.
 
there are videos on youtube that will show you how much and where to put it.

i wound't bother trying to explain or read how to do it. just watch the vids.
 
A small blob in the middle is all you need. The compressive force from the heatsink mounting will take care of the spreading.
 
It's meant to fill in the gaps left by imperfections in both the IHS and the heatsink, not act as a layer between them.

So you really don't need all that much.
 
Zero82z said:
A small blob in the middle is all you need. The compressive force from the heatsink mounting will take care of the spreading.
Click to expand...

That is what I do, works perfectly every time

64302-bigthumbnail.jpg
 
lol if you use too much your temps will actually be worse then they were before
 
Thanks everyone. Reapplied with the line method and temps are 10° C cooler than the paint method.

Now my problem is C1E somehow was broken in between mounting and remounting the heatsink!!!
 
Zero82z said:
A small blob in the middle is all you need. The compressive force from the heatsink mounting will take care of the spreading.
Click to expand...

+1

This is how I do it too and works the best and is far easier to apply than the credit card method that some do (I used to do it that way but not any more since I saw the pressure from the HSF makes it smooth out into a layer anyway). Credit card method is more work for no gain. But my Nctua HSF said to use 4-5mm blob and my testing finds that even that small amount is a bit too much. I use about 2-3mm dab on top/middle of the CPU and gives good thin layer. For you Americans, get a ruler with metric scale. ;)

BTW, what's in that Articlean 1+2 shit they are trying to hawk in that .pdf file? They are claiming 99% Isopropyl alcohol is just adequate compared to their Articlean product. I don't buy that for 1 second and 99% Isopropyl is all you need and is just as good or better than their expensive product.
 
Last edited:
I actually did pretty much what you did (very thin layer) and my temps are fine as far as I'm concerned. Depending on ambient, my i7 920 idles at 36-42C and load temps are around 75C. I doubt they would drop 10C if I used the drop method.
 
Zero82z said:
A small blob in the middle is all you need. The compressive force from the heatsink mounting will take care of the spreading.
Click to expand...

This is indeed what i do, works well enough i suppose.

Question, my core 1 is typically a couple (~2-3 degrees) hotter and I was wondering if maybe the goo was spread out erratically (i did shift my HSF alot before it was clamped down) and that is why core 1 is a little hotter
 
Tbolt said:
This is indeed what i do, works well enough i suppose.

Question, my core 1 is typically a couple (~2-3 degrees) hotter and I was wondering if maybe the goo was spread out erratically (i did shift my HSF alot before it was clamped down) and that is why core 1 is a little hotter
Click to expand...
Temperature variations between cores can be caused by many different things. However, it's really not something you need to worry about.
 
The easiest answer to this question is just taking your heatsink off and looking at it. The shit ain't magic, observe and make a determination.
 
In the process of removing it the compound will smear. Though it's not static safe I have place a piece of saran wrap over the compound, installed and removed the heat sink. Here is a video with another interesting way to see how the compound spreads.

And in my limited experience I find the pee or line method did not make a difference on AMD dual cores. I never used the paint method on devices with a heat spreader. Also this is in regards to heat sinks with solid flat bottom surfaces.
 
Last edited:
Line method made a huge difference. Load temps after hours of Prime95 are now 65° C instead of 80° C

System idles 10° C cooler as well.

I originally used the paint with credit card method plus a small drop on the heatsink and personally I think that caused 2x the amount of compound that was needed.
 
Last edited:
Zero82z said:
A small blob in the middle is all you need. The compressive force from the heatsink mounting will take care of the spreading.
Click to expand...

yep


all these other "methods" are an utter waste of time small blob around the size of a grain of rice in the middle and go, doesn't matter what cpu or how many cores, the only time this method varies is on older cpus without metal heat spreaders where the core/die is a much smaller area, same method applies just use less
 
I actually dont agree with the blob method.

I take some AS, put a plastic bag around my finger (keeps the paste off your finger/prevents body oil from mixing in), apply the AS to the tip of my finger/plastic bag and use my finger to spread an extremely thin layer across both the chip, and the heatsink - this insures complete coverage of the entire surface - it also insures the thinnest possible layer while maintain complete coverage - the blob method does work, but it does not insure that the entire surface is covered with thermal paste.
 
Elipsn said:
I actually dont agree with the blob method.

I take some AS, put a plastic bag around my finger (keeps the paste off your finger/prevents body oil from mixing in), apply the AS to the tip of my finger/plastic bag and use my finger to spread an extremely thin layer across both the chip, and the heatsink - this insures complete coverage of the entire surface - it also insures the thinnest possible layer while maintain complete coverage - the blob method does work, but it does not insure that the entire surface is covered with thermal paste.
Click to expand...
The important thing is to have good coverage at the center of the heatspreader, which the blob method accomplishes perfectly. And if you use enough paste, you can get full coverage of the entire heatspreader as well, although it takes a bit of trial and error to figure out the right amount to use.

The main flaw with the method you use is that if the gap between the heatspreader and heatsink is not perfectly even, you will end up with pockets of air between them in areas where the gap is larger than the amount of paste you used. This can result in much higher temperatures if it happens above the center. The blob method ensures that the paste completely fills up any gap, which will result in the best possible thermal transfer across the entire covered area.
 
Elipsn said:
I actually dont agree with the blob method.

I take some AS, put a plastic bag around my finger (keeps the paste off your finger/prevents body oil from mixing in), apply the AS to the tip of my finger/plastic bag and use my finger to spread an extremely thin layer across both the chip, and the heatsink - this insures complete coverage of the entire surface - it also insures the thinnest possible layer while maintain complete coverage - the blob method does work, but it does not insure that the entire surface is covered with thermal paste.
Click to expand...

I thought the same but I tested with my latest Noctua HSF install (Noctua recommends the blob method in their instruction booklet) and when I took the HSF off after using the blob method I could see there was complete coverage. But I also saw some paste had squeezed out of the edges of the spreader too and is why I recommended 2-3mm blob instead of the 4-5mm blob that Noctua instructed to apply.

One thing I read that is good to do is to rub some paste into the HSF bottom and top of CPU heat spreader as that fills in any pits. I never did that on the Noctua install but used to do it and then the blob after. I'm wondering though if just rubbing paste into the HSF bottom and top of spreader and no blob at all would do an even better job. I'll have to test that theory out one day. Of course the HSF bottom and CPU heat spreader would need to be perfectly flat for it to work well.
 
^ LOL

I like my credit card method. I just do a really thin, even layer over the entire CPU (and just the CPU) and then smash down the heat sink. My temps are basically the same doing that as the blob method (which I use to do), but at least with the credit card method I know the entire surface has an even amount, where as the blob the edges a lot of times do not see any compound.

Note: Don't actually use your active credit card. ;)
 
SixtyWattMan said:
There is no such thing as too much.
Click to expand...

LOL, I have had a few customers who thought that. In fact a while back I built a 1U server for a customer. This was back in the P3 days. It had dual P3 1GHz and I didn't have any heatsinks that would fit and he said he would take care of it. I trusted him being a network admin.

Two days later he comes back claiming no video and no post. I hook it up and sure enough. I crack it open and start looking for blown caps. Then I start to zero config to find out what's busted and start with proc 2. I try to remove the heatsink and as I pull on it, it won't come off and it is actually pulling up the board so I slide the damn thing off instead. I find the thermal grease is covering the proc and it's layed on pretty thick. I've never seen this much used, ever.

I grabbed a known good proc out of another machine I'm working on and bang, it starts right up. So, too much will simply insulate the proc and cause it to overheat and possibly ruin the proc and voiding your warranty.

maximus87 said:
^ LOL

I like my credit card method. I just do a really thin, even layer over the entire CPU (and just the CPU) and then smash down the heat sink. My temps are basically the same doing that as the blob method (which I use to do), but at least with the credit card method I know the entire surface has an even amount, where as the blob the edges a lot of times do not see any compound.

Note: Don't actually use your active credit card. ;)
Click to expand...

This is what I do as well. Just a thin layer it perfect for procs with the heat spreader. If you have one with an exposed die like an older proc or laptop proc then just go with enough to equal the size of an uncooked grain of rice.
 
Blazestorm said:
It's meant to fill in the gaps left by imperfections in both the IHS and the heatsink, not act as a layer between them.

So you really don't need all that much.
Click to expand...

x2. You shouldn't need any more than a dab the size of the pencil eraser :) Put it in the center and mush it down. It will automatically fill the gaps
 
maximus87 said:
Note: Don't actually use your active credit card. ;)
Click to expand...

Yea, Coolermaster actually included a plastic card for that reason in one of their kits I bought and still have it but the blob method is so much less work and less messy so prefer that way now.
 
let me know how that works for ya. Seems like a good concept (rosewill brush on stuff)
 
PolygonGTC said:
If you have one with an exposed die like an older proc or laptop proc then just go with enough to equal the size of an uncooked grain of rice.
Click to expand...

2 grains of rice with a heat spreader is about right but I prefer to use metric so say 2-3mm.
 
Last edited:
dr.stevil said:
x2. You shouldn't need any more than a dab the size of the pencil eraser :) Put it in the center and mush it down. It will automatically fill the gaps
Click to expand...

Pencil eraser size is still too much IMO. 2-3mm is closer to what it should be, maybe 4-5mm as Noctua says but I use 2-3mm. Get a ruler and look at it to see what 2-3mm is, it's way smaller than the average pencil eraser.
 
Remember what it is that you're trying to accomplish. Thermal paste is designed to fill in the tiny imperfections in the surfaces of your CPU and heatsink and conduct more heat than air bubbles would. http://gutski.net/jefflinsales.com/Tech/Thermal paste application.htm has a very basic, exaggerated diagram of this.

If you were to lap both your heatsink and your CPU perfectly flat, they would make perfect contact and you would not want any thermal paste at all. The direct metal to metal contact transfers heat way better than any TIM. If you use too much thermal paste, you'll actually push the parts away from each other and reduce the amount of direct metal to metal contact, thus reducing the heat transfer. If you don't have enough paste, you'll be left with some air in the middle, which will also reduce the heat transfer.

If you use anything other than a blob in the middle, you risk trapping air bubbles inside. With the blob, the clamping force in the middle presses the paste outward, pushing the air out in front of it. The line and X methods should be similar, but the different sections of the lines could get pushed toward each other, trapping air bubbles near the center.

Regarding the brush-on stuff, I've been using a bottle of Zalman STG1 for a few years now. It seems to work ok, and it's quick and easy to put a very thin layer on.
 
I have a slightly off topic question.. I have a tube of Arctic Silver 2 that I've been using for years (since AS2 was the shizzle). Would I see much difference getting a new tube of the AS5 stuff?
 
robble said:
I have a slightly off topic question.. I have a tube of Arctic Silver 2 that I've been using for years (since AS2 was the shizzle). Would I see much difference getting a new tube of the AS5 stuff?
Click to expand...

Thermal grease does expire, usually by becoming hard. If it hasn't hardened it should be fine. As for going from AS2 to AS5 I would guess you would see more than a two degree difference.

To be honest, two years, putting new grease in there is never a bad idea. I say do it.
 
The grease hasn't been in between the cpu and heatsink two years.. only about a month actually (new build). It's just the tube is many years old - and there has been 3 revisions of it since it came out.
It is mainly the AS2 vs AS5 that is my question. My temps are ok now so I'm not too worried about it but I was wondering if it would be worth getting some of the "newer technology" stuff.
 
You must log in or register to reply here.
Back
Top