What thermal paste method do you use on your home PC?

[mysticjbyrd]

I am getting a new cpu in Friday, and it got me thinking about this old quandary.

I remember my first time like it was yesterday. It was awkward, it didn't last long, and yes my finger was involved. I obviously didn't know what I was doing, and went with my made up version of the spread method. After my first time I wised up, and began using the pea method.

Although, today I noticed in the Arctic Silver 5 instructions that it states to use a line method with non-mobile Intel CPUs.

http://www.arcticsilver.com/intel_application_method.html#

I also stumbled across this interesting video series, although not complete yet.


https://www.youtube.com/watch?v=ibvs0xrDXag

TLDW
The results for the pea, line, and spread method were roughly the same. I would say within margin of error easily.


Anyways, enough about me. I am curious about what method you guys prefer, and why.

 

[Zodiark1593]

Not sure why the spread method is used at all, too much work vs letting pressure do the work.

 

[master_shake_]

spread method leaves air bubbles.

small pea in the center and leave it.

 

[Deders]

Spread is excellent for when you have bare heatpipes, otherwise a dot should do.

 

[MongGrel]

spread method leaves air bubbles.

small pea in the center and leave it.

This has been preferred a long time now.

 

[.vodka]

Pea/rice for IHS/square die application. Line for rectangular die application. No manual spread on both, I let the heatsink mounting pressure do the spreading.

 

[pcgeek11]

I have been using a pea sized blop and spreading it with my finger ( using a finger cot ) to cover the top of the CPU and then seat the heat sink/fan on it.

Used this method since they came out with the heat sinks and fans on CPU's and never had a heat problem...

I don't think it matters much.

 

[Mojoed]

I don't think it makes any difference at all.

On my old 3.4GHz OC'd Q6600 CPU @1.33v, my idle temps were in the upper 50's using the stock cooler.

Idle temps dropped to 28c after cleaning and reapplying AC5 in a circular pattern using a regular lunch baggie over my finger.

I have used this method on scores of PC's since I first bought Arctic Silver about 15 years ago and I have never, ever had any system regardless of OC have any temperature issues whatsoever.

I believe Arctic Silver puts forth these recommendations to minimize people from just applying it with their bare fingers where the natural oils could potentially fill the microscopic pits in the metal and actually adversely affect temperatures.

 

[NTMBK]

Whatever Dell ships with :awe:

 

[bononos]

I am getting a new cpu in Friday, and it got me thinking about this old quandary.
........
TLDW
The results for the pea, line, and spread method were roughly the same. I would say within margin of error easily.
......

AS5 is a very viscous and does not spread easily like most newer pastes so the dot method might not be sufficient to get the right spread. The instructions also call for hazing the surfaces to lower the break in period.
There are newer pastes like pk-3 which are less messy and easier to apply so its just better to get those instead of AS5.

 

[pcgeek11]

Much ado about nothing.

 

[john3850]

I always used the hazing method with good results with stock or unmodified surfaces.
When using the hazing method with AS5 on polished waterblocks with extra pressure I started adding a dot extra.
I believe the more polished the surfaces the better the temps. My problem is getting the AS5 to stick on polised surfaces.

 

[Charlie98]

With my 212+, and it's uneven contact surface, I spread to fill the gaps in the surface of the heat sink, then put a pea on the CPU and lock it down.

 

[john3850]

When I needed a water setup for my 3770k I put a 212+ on my wifes unused 2500k.
I always thought about adding a thin cooper shim between cpu and hsf on the 212.

 

[Yuriman]

You have to spread liquid metal, but between the IHS and heatsink I use the dot/grain of rice method. Using a pea sized blob of paste sounds like way too much (and wasteful), since the core is only a small fraction of the size of the IHS.

 

[Phynaz]

With my 212+, and it's uneven contact surface, I spread to fill the gaps in the surface of the heat sink, then put a pea on the CPU and lock it down.

Same here.

 

[MrTeal]

You have to spread liquid metal, but between the IHS and heatsink I use the dot/grain of rice method. Using a pea sized blob of paste sounds like way too much (and wasteful), since the core is only a small fraction of the size of the IHS.

I use CL Ultra so I spread as well. It doesn't really spread to cover a bare die on a GPU if you don't.

 

[ctk1981]

TIM Mate 2 from swiftech worked as good as anything else I've used and was reasonable. Got a tube when I bought my water block, I had to do a cross rather than line with the TIM since I'm using a 5775c now.

 

[bystander36]

Not all conditions work best with the same method. Review sites who test this, often try different methods, and on different CPU's, get the best results from different methods. The pea method is the most common to be best, but sometimes an X or line works best.

 

[Erithan13]

Just built a i7 6700 rig and definitely have a cooling problem, this is my first experience with backplate mounted coolers (Arctic Cooling 11i) and with applying paste manually (last build had it pre-applied to the cooler). Used the pea method for the paste and think I may have overdone the amount so I'll be looking at it again tomorrow. I was very hesitant to tighten down the cooler mounting screws as I don't know at what point it might damage the cpu, is this something I should be worrying about or is it difficult to dangerously overtighten them?

I had this feeling the cooler wasn't down tight enough and this seems to have been confirmed by the temperatures (Idling 35C, peaks at 60C with seemingly very little load on it, core temperatures are also jumping about wildly several times per second, classic symptom of bad installation from what I've read. Don't have the guts to prime95 or anything stressful until it's sorted).

 

[DigDog]

i used the spread method for my E6600 / 212 combo.
opened it after it ran 24/7 for 7 years: perfect footprint. (artic silver 5)

 

[Yuriman]

Just built a i7 6700 rig and definitely have a cooling problem, this is my first experience with backplate mounted coolers (Arctic Cooling 11i) and with applying paste manually (last build had it pre-applied to the cooler). Used the pea method for the paste and think I may have overdone the amount so I'll be looking at it again tomorrow. I was very hesitant to tighten down the cooler mounting screws as I don't know at what point it might damage the cpu, is this something I should be worrying about or is it difficult to dangerously overtighten them?

I had this feeling the cooler wasn't down tight enough and this seems to have been confirmed by the temperatures (Idling 35C, peaks at 60C with seemingly very little load on it, core temperatures are also jumping about wildly several times per second, classic symptom of bad installation from what I've read. Don't have the guts to prime95 or anything stressful until it's sorted).

I can't speak for Skylake, but it's normal for Ivy Bridge and Haswell to idle in the 30's and jump instantly to 70-90c when put under the right loads. There have been several posts per week on here for the past 3 years with users concerned about this very thing. Normal behavior, so long as it's not throttling.

You don't want so much pressure you're warping your motherboard, but 30-60lbs is common (if I remember correctly). Most mounting bracket screws bottom out, making it so you can't over-tighten them. Otherwise, stop just before the springs are completely compressed.

An uncooked grain of rice, from my experience, is all you need. Remember that thermal paste is meant to fill in the space caused by microscopic valleys and ridges on the mating metal surfaces. You want it to be able to spread nearly to the edge of the IHS in an extremely thin film. If it's squishing out over the edges, you've used too much.

 

[JimmiG]

Always been using the dot/pea method, but I noticed the stock Intel heatsink has its paste applied in three parallel lines. Don't know if that would work better?

Back in the day I used to obsess over the TIM application, taking off the cooler, cleaning and re-applying several times to get it "just right". Temps were virtually identical between applications though.

Between the bare die and IHS I obviously used the spreead method when applying the Coollaboratory Liquid Pro.

 

[Zor Prime]

I put a line on and spread it out really thin with a razor blade and if I don't have one of those on hand a credit/debit card. I might take a good 5 minutes or so making sure everything is good. If you've got thermal paste squished outside of the thermal contact zone you're doing it wrong. Haven't ever had thermal trouble ...

 

[DrMrLordX]

You have to spread liquid metal, but between the IHS and heatsink I use the dot/grain of rice method. Using a pea sized blob of paste sounds like way too much (and wasteful), since the core is only a small fraction of the size of the IHS.

When using CLU, I use a spread method particular to that paste: I use the enclosed brush (or similar), though first I cut the bristles down to a nub so that they don't pick up any paste. The amount I use between the IHS and HSF base is maybe 1/4 what I would use for other pastes.