How evenly is your thermal paste spread out on your processor?

doesn't matter, the die is under 1cm wide in the middle anywayd

does this really deserve it's own thread?

>>59227148
>implying the entire lid isn't heated

>>59227159
Would you prefer another ryzen thread?

Or how about hype for vega/1080 ti?
Your want an anon to tell you how amd btfo Intel or vise versa from cherry picked benchmarks?

>>59227131
That looks awful.
You either use the dot method, or the line method.
For Intel CPU's I recommend the line method a little more.

File: pic_disp.jpg (81KB, 800x600px) Image search: [Google]

81KB, 800x600px

X-shape

https://www.pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/

Use a phase change thermal interface.
Unless you fuck up majorly installation should be perfect.

Put a dab on the CPU and spread it in a thin layer with your finger for even converage.

>>59229319

Fucking tard.

Use a business card for even coverage!

how do i delid a cpu and solder a cpu cooler directly to it? seema like the best choice

>>59227131
you dont use pea sized dude. you use lentil. it really shouldn't have made it all the way to the edges man

I always spread it with a latex-gloved finger, then toss the glove. While seating the heat sink and fitting the tensioning X-brace, it tends to spin and slide a bit, which works it into the heat sink. Even if there's a tiny bubble or two, there's no way they add up to the lost heat transfer area shown in >>59227131.
Blobbers BTFO

>>59229369
>>59229319

hope you enjoy air bubbles.

this shit is not needed on modern systems, the actual heat producing portion of the cpu is so small you only need coverage above it.

a small round, slighly larger than a bb sized blob is ideal

>>59230005
I'd think you would have to put the exact right amount of raw solder on the die, then set the IHS on it and melt it from the top by heating the IHS.

>>59230060
knowing G you guys would do that with acid core solder

>>59230055
Functionally wrong, the uncovered corner gets hot, right? If it has no heat transfer medium you have less area transferring heat. This translates into hotter cpu die temps. Period, end of story. You're dumb.

>>59230065
>G
GTFO
T
F
O

>>59230065
phone poster?/g/ isnt one person

>>59230080
>corners
top kek

>>59230080
the small amount of slight gain you probably wont get is not worth the risk off too much paste

fuck the paste is only there to ensure more direct contact, to fill in the unevenness. itd be better to have metal to metal contact. but that is functionally unattainable

I hope you can understand why the risk of too much is not worth a full spread. 100% of the heat is coming out of a very tiny die this is all that matters

>>59230139
thats why he said

>you guys

>put a teeny little rice grain dot on
>get shit temps
>take cooler off and see 90% of my paste is on one side of the chip
>put a bigger blob and make sure it's right in the middle
>STILL shit temps
>take cooler off and see spots in the middle where there's 0 paste
i dont even fucking bother being careful anymore, thermal paste isnt hard to clean anyway

>>59230189
Most of that heat transfers into the IHS. If only 95% of the IHS is making contact with the heat sink via the goop, you just introduced a heat bottleneck didn't you? Nice job moving the goalposts though. You went from saying the die transfers its heat straight up and only straight up to saying the goop can cause problems if it leaks onto all the shit around the processor. Now, that's true, but do you think if you're designing a CPU socket that maybe you'd see this problem coming and make the socket able to accept a little too much goop?

>>59230080
Conduction isn't convection; heat does not rise, but rather it follows the negative temperature gradient. If the corders are hotter than the center then they will loose heat through the paste just as efficiently as the parts touching it. And do you seriously think the RATE of heat flow is gated by the area of conduction? What, do you also think that the steam coming out of a boiling pot is hotter than the edge of the pot?

>>59230269
wait until you get some under the chip and think the world is about to end then and there

>>59230312
There is no paste on the corners in this fictional example we're discussing. It's only above the cpu die. And yes, I'm saying the rate of heat flow is affected by the area able to transfer heat. You goddamn mongolian.

>>59230346
no ive done that with an LGA socket too
>try sticking a paper towel on the pins and pushing them down
>one of them starts glinting because the light is hitting it differently
>i bent it of course lmao!!!!~11
>press it down again and it straightens itself
everything turned out ok

>>59230370
>>59230312
>both agree
>still arguing
exquisite

>>59230388
I remember my first time building a PC I got paste under the CPU

I almost felt like calling newegg to amazon to pick my shit up I'm about to hang myself

luckily I found out rubbing alcohol and lint free cloth fixes everything up

>>59230418
I don't think we agree at all.

>>59227131
I spray it on with an ultra thin mist

Use mercury

>>59230678
It'll work great until it cooks off and makes you crazy.
>tfw your gold turns grey

>>59230785
At least it hits two birds with one stone: gets rich and kills himself. Both good things.

>>59230849
That's the life I'm after.

Also to go insane. It's no fun going insane when you're poor.

>>59227167
Not with unsoldered lids.

This isn't difficult. Method depends on the viscosity of the paste you're using.

Thinnest
DOT
LINE
FIVE DOTS
Thickest

>>59230189
> itd be better to have metal to metal contact. but that is functionally unattainable

Vacuum weld would do it.

Only needs direct contact between two of the same metals in a vacuum.

>>59227131
i just use the intel stock fan that had thermal paste preinstalled. been running it 24/7 for 2 years without any problems.

>>59231357
Same but for about 4 1/2 years. temps are getting hotter than you would want but aren't alarmingly bad considering stock cooler, going to have to apply paste this year I think though.

>>59231357
>>59232528
for the love of your ears just buy the cheapest air cooler you can find, it'll be cooler and quieter than intel's

>>59232639
Yeah, out of order priorities and making purchase decisions when I shouldn't have lead to corner cutting. I have been looking to getting something these days though just to squeeze some more life out of it.

Weird question but what is the actuation force required to unmount/mount a heatsink? I ask because I have extensive tendinitis issues and have a very very shaky and basically non existent grip but no one to help me with it these days. Last time I installed a new gpu just pushing the clips down on the power connectors was really difficult and I almost couldn't (but as able) to do it.

>>59232639
Except my intel fan isn't loud at all because I haven't overclocked?

>>59231137
>doesn't know how heat transfer works
>t.underage

I bet you guys are mad I use the spread method.

Stay jelly.

>>59227131
According to Carey Holzmen, it doesn't matter how you apply thermal compound. You apply it how way you like, its will still spread evenly when the heat sink applies pressure.

>>59232821
If you wanna make a mess of your socket and it's surrounding area on your motherboard go ahead.

>>59234279
This^

>>59232639
my intel fan is quiet. it will only be audible if i compile something that takes over 5 minutes. its the hdds that make most noise in my computer.

>>59227131
well how much was on the heatsink?