How would you use thermal paste on this?

>Pea method won't work anymore.
Bullshit.

image is pretty cool

Attached: .jpg (1200x1198, 202K)

Attached: .jpg (1194x1200, 223K)

Attached: image0.jpg (4032x3024, 2.04M)

Because of the upper right placement of the CPU die itself the pea method will leave a portion of it without direct contact with the heatsink. Although the IHS is soldered this will still leave a hotspot which will affect PBO boost clocks.

The graphite pad would theoretically be the best since it transfers heat across the surface better than up and down but it's electrically conductive.

Attached: 1558986292253.jpg (691x771, 64K)

with my dick

no idea. i sand mine down on glass and 3m sand paper along with the copper of the heat sink like ive done on every processor the past 8 years. i also build high performance engines and routinely sand down heads to achieve the same precision. looks like a mirror when im done and i run cooler without paste. only a couple degrees c but still cooler.

OY VEY!

Attached: 1504702779128.png (531x720, 467K)

protip: use a diamond DMT stone, never buy sandpaper again at least for this particular task

based

pic related: 25 micron stone. Best part is they stay flat even after hundreds of uses.

Attached: Screenshot_20190719-142056.png (720x1280, 361K)

ill look into it thank ya. i have a small one i sharpen knives with but isn't wide enough for this application.
i figured i couldn't be the only one doing this. tested multiple times with and without thermal compound and its always slightly cooler. but it for sure needs to be precise.

appreciated

Can't you just do a diagonal line or two small peas instead?

shut up retard

the best and only way

Nope, you have to manually spread it via brush, finger wrapped in piece of plastic, card, whatever as long as you even spread a thin layer on BOTH the IHS and heatsink contact place. This will ensure the max amount of heat will radiate from the top right corner of the IHS to the heatsink itself. It's okay if you use a little bit too much, diamond does not conduct electricity.

Attached: slide_3.jpg (800x600, 60K)

i used pea method. 70c max load is 10c hotter than my 2700x but manageable.

of course you can you fuckin retard

This. A pea sized dot, and let it spread itself when I put on the cooler.

You're guaranteed to have a hotspot near the CPU. I recommend you try again with the method mentioned in the post above yours. PBO works by monitoring the temps of ALL cores so even if one is a few degrees hotter than the other ones it will refuse to boost more than it normally would.

>How to get air pockets
Fucking retards

is spreading with a credit card acceptable?

In this case the slight air pockets are mitigated by spreading the TIM on the IHS AND heatsink contact plate.

Yes but do this on BOTH IHS/heatsink to avoid most of the air bubbles.

>slight airpickets are mitigated
How are you mitigating the uneven transfer of heat dumbass?

You fucking spergs. It has an IHS already, that means Internal HEAT SPREADER. IT'S EXACT PURPOSE IS TO SOLVE THE NON-ISSUE YOU GUYS ARE BITCHING ABOUT.

Attached: 1556728669594.png (467x458, 431K)

>It's okay if you use a little bit too much
Wot m8. The goal is to use just enough to seal the micro fractures on both points of contact.
TIM is not as efficient at transferring heat as the metal contact points.

Why both the heatsink contact and the IHS? Don't you usually just spread it on the IHS and then let it spread onto the heatsink

By spreading the TIM on both side you fill the cracks and holes that have the air pockets you mong. The VERY FEW air pockets created by making contact and tightening are as small if not smaller than those created with the pea method.

If you REALLY gave a fuck about air pockets you'd grind 1mm off the IHS and heatsink and polish with 0.5 micron diamond lapping compound to a literal mirro finish.

Pea method if your thermal paste is easily spreadable. Line method if your thermal paste takes more pressure to spread. It's all about viscosity rating in your paste.

By spreading TIM on both sides you will trap air in between the Tim you fucking retard.
Just go away. You literally don't know what the fuck you're talking about, and you're probably just regurgitating something you saw a YouTuber do.

the same way I did for any cpu in the last 15 years. one very small drop in the middle and leave the rest to the pressure of your cpu cooler for the spread. or just use a $10 graphite pad. that will probably last you the next 10 years. it also wouldn't make more than 1°c difference to good thermal pastes.

The pea method, and even the line method for rectification shaped his literally squeezes as much air out as possible during application without having your shit sit three days in a vacuum chamber.
Do everyone in this thread a favor and bash your retarded head against the corner of a sidewalk until your ears bleed, and stop posting.

It's correct though. The peak method was only thought to be the best because Intel's dies were perfectly central and that's technically the bit that needs cooling, but even then it's not the BEST method. The best method is having paste on every part of the IHS that comes into contact with your cooler's coldplate, in order to facilitate the maximum amount of thermal transfer possible. The worries about spreading causing air bubbles were largely unfounded, since the mounting pressure on any decent cooler is great enough to ensure they're not an issue. People also used to say that if you used too much thermal paste it would hurt your temperatures, but more modern test methods have proven that to be bullshit (again because of cooler mounting pressure). At worst it'll just be a horrible mess to clean up.

common sense tells ya if you have two perfectly flat surfaces joined together you will have the best case for heat transfer. anything else between the two no matter how amazing will cause degrade to some degree that heat transfer.

Lies. The best method is whatever gets the thinnest amount of tim to spread evenly across the surface without airpockets. Currently the pea method is the best at doing that by hand.
Remember fuckwad the his is there to distribute the heat on a wider surface, so the pea method let's the his do its fucking job, and will be the most efficient at transferring heat between two surfaces.

Just stop. You're fucking dumb, and you're spreading the wrong information Linus.

I'd just use intel tears instead

So, if i combine both posts you guys are saying that i should just spread a fuckhuge amount of paste in the middle and benchpress that shit until it doesn't squeeze out of the side anymore?

Yea, but a small amount of paste on just one surface.

smallish X like I always do
fuck your spergery

>throw in trash
>shell out for INTEL

I broke a motherboard that way in the past, I don't recommend applying too much pressure.

X or dot method is the best

spread will leave only bubbles...

>with a IHS
Jesus our Savior?

an X or spread a layer evenly across the entire ihs.

Alright, everyone can stfu now. The reason why anons are pleading for the spread method is because of the placement of the CPU is NO LONGER at the smack center underneath the IHS. Heat from the upper right corner needs to travel to the center of the heatsink for max transfer. It can't do this when methods like the line, dot, x, swastika, ect don't 100% cover the entire area of the IHS/heatsink evenly.

Applying a thin layer to both the IHS and heatsink would ensure max radiant transfer from the upper right corner of the IHS into the heatsink.

Attached: pic_disp.jpg (601x751, 93K)

Attached: 1541642713524.jpg (207x253, 8K)

>applying so much pressure that you broke a motherboard.
user how do you stay alive.

It was one of those one-size fit all coolers. I had a hard time judging the tightness of the screws and it just ripped a bit and i slipped directly after into the board, completely popping that shit. I was fucking furious.

we live and learn

you must have stock in some meme grease company

this is like those ram timing charts with 1fps difference

>Applying a thin layer to both the IHS and heatsink would ensure max radiant transfer from the upper right corner of the IHS into the heatsink.
with BUBBLES and no any advantage to X method...
your source:
pugetsystems.com/labs/articles/Thermal-Paste-Application-Techniques-170/
and my source:
gamersnexus.net/guides/3013-amd-threadripper-thermalpaste-application-methods-benchmarked

Attached: Smoothly spread across CPU.jpg (800x600, 64K)

None of these even have putting grease on both surfaces listed. Probably, because it's the stupidest shit imaginable.
Also 26.5 is better than the higher temps not to mention that the chart is so vague with methods, weights, environment, and every other variable to be taken seriously.

"x" or "+" shape will cover all CPU

So for Zen2 and future Intel glue chiplets the best way to apply paste is to fingerfuck it evenly on the CPU IHS?

Lapp the CPU ihs, and the heatsink
then apply enough TIM to cover ihs.
Then place in a vacuum chamber.

Anons just STOP and think for a second: the results in were done with a CPU LITERALLY UNIRONICALLY DAB SMACK in the fucking center underneath of the IHS. This is no longer the case.

If any of you had been paying any attention to the chart it tells you the AVERAGE core temp NOT individual core temps. PBO WORKS because it reads individual core temps and boosts accordingly so as to not kill your CPU.

WHAT do you think will happen when core 0-5 are 65C but core 6-7 are 79c?

Attached: QxIK5KZwaKoVp0yz.jpg (546x442, 29K)

>Anons just STOP and think for a second: the results in # were done with a CPU LITERALLY UNIRONICALLY DAB SMACK in the fucking center underneath of the IHS. This is no longer the case.
If it mattered then why do they still make ihs plates to cover that much surface area?
Maybe it's because the ihs serves a purpose regardless of CPU placement.

>Delidding
>Ever

You retards need to go and stay go

I've don't the exact same method for every single PC I've ever built. A grain of rice worth of IC-Diamond in the center of the chip and let the heatapreader smoosh it out. Only Tim ethis was changed a little was for a Threadripper build. Use 2 grains of rice then.

Currently have an R7-3700X and it told out at 72C after 3+ hours of Prime95 smallFFT.

nu/g/ enthusiast

The IHS is soldered on BUT it does NOT perfectly radiate heat from the CPU cores to the heatsink with 100% efficiency. We know this because soldered zen+ chips dropped 5C on average when delided and liquid metal is placed between the IHS and the CPU die.

Because of this the spread method would ensure as much surface are is being used to transfer heat from a corner of the IHS into the heatsink itself to mitigate this inefficiency. It's 10X better, easier, and safer to do this than to fucking delid the damned thing.

Attached: untitled-3.png (725x351, 35K)

It doesn't matter as x, + or spread will cover all CPU...
But spread shit will leave more bubbles, while x and + not.
Look on your test, as it written that there...

a spread method would probably work best desu

>muh airbubbles
literal nonsense

i delidded my 6700k and temps dropped 15C across the board.

is written in the koran also

see
A few micron thick air bubbles will make a +1C difference AT BEST. The problem is making sure ALL the CPU cores get similar temps which is what the spread method will do.

Lower temps per core = more PBO boosting.

absolute madman

Attached: rZOaHhx.png (192x154, 56K)

Those piles of shit were glued though. Personally I wouldn't risk to delid a soldered CPU, much less a brand new one.

SHHH let us SPERG SPERG

and I see better results with X method
what am I missing?

At ultra settings, yes.

It actually doesn't matter.

>Internal
Its Integrated you retard.

reddit.com/r/Amd/comments/ccrisu/ryzen_3000_ihs_orientation_vs_the_position_of/

Even amd confirms X pattern to be the best

Yes it is you retard the point of thermal paste is to minimise air bubbles (insulation) between the HSF and the CPU. The fact that the heat is concentrated in one particular area does not mean that it's better to use more goop or less goop or apply goop in a certain pattern. You just legitimately don't understand the point of thermal paste so you think that the CPU die placement impacts the ideal way to use it.

All this changes is that it makes strategies that were always non-optimal like the "pea" method that applied paste thicker around the center than the corners even less optimal.

It DOES impact how effective certain HSF's are because their performance is going to be determined by how well they cool one corner of the processor so heatpipe placement can make a pretty large difference. We're seeing large variations in HSF performance for Ryzen 3k that don't exist for other chips.

The "spread" method has never been good because people simply are not steady handed enough to spread out paste evenly... and it's stupid to when just clamping the HSF on does a better job at spreading evenly.

The "X" method has long been optimal because it tends to improve distribution of the paste to the corners. Not just for ryzen 3k.

How about spreading it evenly with an old credit card?

>dot
>off center hotspots

Attached: 61581113_10156401813163716_6328228307397509120_n.jpg (857x701, 28K)

OY VEY

Attached: jewscared.png (354x504, 62K)

>the ihs is an insulator on the horizontal axis

Yeah that's because in this case a full coverage of the IHS is optimal.

The IO die is fuckhuge and only draws 20-24w at most with a single 8 core CCX. Meanwhile the cores can pull upwards of 80 watts at stock (non XFR/PBO) speeds.

What you want is full IHS coverage so all of the core heat spreads out to as much of the unladen IHS, and into the IO die secondarily, as possible.

A dot for every chip is going to constrain heat only to the chip at worst, and at best just barely spread that heat to the empty areas+into the IO.

So yes, X method would be superior.

And always look up the chip maker's suggestion. They have a paste method for every CPU.

These dots are the best. Reddit meme "Air pockets" will pressurize and punccture though the compund given the intense heat of the chip and the expansion and contraction of the compound.

Attached: this.png (427x261, 150K)

>Delidding ryzen
for what purpose?

Because you think that somehow you're more intelligent than the engineers that spent years designing it

i can see a reason behind doing that on Intel shit TIM but iirc Ryzen has it soldered (at least 2nd and 3rd gen).

>Heat from the upper right corner needs to travel to the center of the heatsink for max transfer
You have zero fucking clue about heat transfer and what goes on.
Necking yourself won't improve this so let's get some information in there.

Heat does't have to get to the center of a heatsink. There are many different heatsink designs but in general there is a large contact plate on the bottom. The purpose is to conduct heat all across the plate. Some heatsinks have centralized heatpipes that contact the IHS that would not be ideal in some configurations on these Zen2 processors, but most, especially the high end ones, have enough heatpipes to contact the whole surface area of the IHS, so it's absolutely not an issue.

Remember, the purpose of thermal paste is to FILL VOIDS between the IHS and the heatsink. It is NOT to be the main medium for heat transfer between the IHS and the heatsink.

Dot should remain sufficient, so long as it spreads out to cover nearly the whole IHS.