CPU temps jumping after applying thermal paste

[JumBie]

I just got some Arctic Silver Ceramique 2 and decided for the first time to try the line method on my i5 3470. After having some trouble mounting the fan back on the CPU I finally got it on and booted into the computer. I am running HWmonitor right now and all the cores are different temps and jumping up and down. Temps range from 34-53. The cores may all be the same temp for a split second but then they all starting randomly jumping into different temps. Did I mess something up?

EDIT: Also wanted to add that using Intel stress test all cores max out at 70 degrees after about 10 minutes of 100% load.

 

[crashtech]

How does this compare with the temp readings taken immediately before replacing your TIM?

 

[coolpurplefan]

I don't know if this is the case but I did have experiences where I put either a small amount or a larger amount of what I thought was the size of a real pea. When I put a real pea-sized amount on a CPU, I had better temperatures. When I put a smaller amount thinking: "Well, this might be the amount of a pea" when I knew it was probably the size of a grain of barley, I had higher temperatures. The only thing I can guess is this: The small amount of thermal grease may raise the height of the heatsink but leave a large space around the thermal grease where the CPU and heatsink has no contact through any thermal grease. So, this area might have been allowing the CPU to run hotter.

I was using Arctic Silver Ceramique (because I wanted to have the ability to take off the heatsink later if needed). (I also bought the Coolermaster Gemini M4 to make sure I'd have the ability to twist it off if I needed to take it off. Some heatsink designs won't allow you to twist if out of place once it's set. And the reason I bought a heatsink with heatpipes is just to make sure if I get a fan failure, that my CPU won't fry before I notice the fan failure.)

 

[JumBie]

How does this compare with the temp readings taken immediately before replacing your TIM?

35 degrees IDLE, 62 degrees full load. Temps weren't jumping around either. I tried the line method again on another i3-2100 I have laying around, same thing happened. Same temps and they are jumping all over the place. Guess I am going to clean them off and use the dot method instead.

I don't know if this is the case but I did have experiences where I put either a small amount or a larger amount of what I thought was the size of a real pea. When I put a real pea-sized amount on a CPU, I had better temperatures. When I put a smaller amount thinking: "Well, this might be the amount of a pea" when I knew it was probably the size of a grain of barley, I had higher temperatures. The only thing I can guess is this: The small amount of thermal grease may raise the height of the heatsink but leave a large space around the thermal grease where the CPU and heatsink has no contact through any thermal grease. So, this area might have been allowing the CPU to run hotter.

I was using Arctic Silver Ceramique (because I wanted to have the ability to take off the heatsink later if needed). (I also bought the Coolermaster Gemini M4 to make sure I'd have the ability to twist it off if I needed to take it off. Some heatsink designs won't allow you to twist if out of place once it's set. And the reason I bought a heatsink with heatpipes is just to make sure if I get a fan failure, that my CPU won't fry before I notice the fan failure.)

I am just going to stick to the method I know and hopefully that solves the problem, pea sized dot.

 

[cretinbob]

you used too much. that will actually act more like an insulator than a conductor. It should be a ridiculously thin film.
You should also use the edge of something like a credit card (like) to spread it out across the heatsink. Don't just glop it on , a pea sized lob is too much btw use about half that, and slap the heatsink on.

 

[homebrew2ny]

you used too much. that will actually act more like an insulator than a conductor. It should be a ridiculously thin film.
You should also use the edge of something like a credit card (like) to spread it out across the heatsink. Don't just glop it on , a pea sized lob is too much btw use about half that, and slap the heatsink on.

Exactly right, I use to use a razor blade and leave practically just a translucent film for best results.

 

[tweakboy]

with finger spread ,, tim evrywhere on CPU. .and fill in your sig

 

[JeffMD]

I seriously doubt there is a coating so thick that it is inhibiting its ability to transfer heat because the heatsink will be applying enough pressure to squeeze down to an ideal thickness (and squeeze out the sides any access paste). The pinhead application method is still the best method to apply paste because it lets the surfaces do the spreading. If you use a razor then you risk leaving thin spots that you can't visual see and thus creating air pockets or zones of poor heat conduction.

With the latter said, if you have a difficult time with attaching a heatsink you also risk doing the same thing, repeated joining and unjoining of the paste covered service will create air bubbles and uneven zones. Any time I attach the heatsink I make sure to line it up on the first try, and once it is down I will buckle it down. If I have to pull it off for any reason I will apply another bit of paste, letting the pressure during mounting spread it out.

 

[dave_the_nerd]

Does it still take paste a few days to "settle in" and get best temps? I thought that used to be a thing.

 

[crashtech]

Most TIMs should not be manually spread out unless the directions specifically call for it; a proper sized dot should be used which smashes out to an air bubble-free thin layer under pressure and heat.

 

[escrow4]

Most TIMs should not be manually spread out unless the directions specifically call for it; a proper sized dot should be used which smashes out to an air bubble-free thin layer under pressure and heat.

Really? I used a dot in the middle then spread it around in a layer that actually is rather thick actually (can't see the IHS). Never had issues.

 

[bononos]

I haze the surfaces before applying the dot/line method. I think it might help prevent some hot spots especially on thick pastes like AS5.

Does it still take paste a few days to "settle in" and get best temps? I thought that used to be a thing.

Ceramique2 takes 2 hours to cure according to skineelabs's graphs.

 

[crashtech]

Really? I used a dot in the middle then spread it around in a layer that actually is rather thick actually (can't see the IHS). Never had issues.

As long as it works for you. Having a thin haze of the fresh TIM one is going to use on both surfaces before applying the dot and mounting can be helpful, but imo it should be transparent (mostly buffed off).