I don't know whether this should go in Cases and Cooling, or CPUs and Overclocking, but here goes... "CPU die outlines on heatspreader".

[VirtualLarry]

Here's a thought, to aid cooling devices and integration, why not have a set of laser-etched dotted lines on top of the heatspreader, where the dies are UNDER the heat-spreader. The idea being, that those are the areas that are most critical to receive paste.

Edit: Feel free to say that this idea is stupid, and that you should always make sure that the entire heatspreader is covered with paste, and connected to a heatsink.

 

[aigomorla]

why would that matter?
Its a given pratice just to put a blob in center and apply pressure on heat sink to make a pancake which under most conditions covers the entire die, unless you want to talk thread ripper, where sometimes the die is larger then the heatsink itself.

 

[VirtualLarry]

Its a given pratice just to put a blob in center and apply pressure on heat sink to make a pancake which under most conditions covers the entire die

Yes, that's how I've always done it, the pea/grain-of-rice method, except for a HDT heatsink, in which I put (thin) lines of paste along the center of each heatpipe. But "people are saying", that with the offset-ness of the die locations on the Ryzen 3000-series CPUs, you can't use the "pea method"/squish any more, reliably. You need to either use "thin coat", or something akin to that, to make sure that the heatspreader is covered.

I guess my line of thinking about this idea, was that you could easily, at a glance, take the heatsink off, and "check your spread", and see if the "hotspot areas" were covered completely / properly.

 

[ehume]

why would that matter?
Its a given pratice just to put a blob in center and apply pressure on heat sink to make a pancake which under most conditions covers the entire die, unless you want to talk thread ripper, where sometimes the die is larger then the heatsink itself.

It would matter because a CPU might get dished in, especially by a heatsink with a convex contact surface. A rice-sized blob might not do.

 

[DrMrLordX]

I hope we aren't still using concave/convex HSFs in 2019.

@VirtualLarry

Depends on the paste. Stuff like Kryonaut and ICD7 demands spreading. MX4 doesn't. Since you brought up Ryzen 3000 . . . I would avoid Kryonaut. It breaks down above 80C and boy does R3K like to run above 80C.

 

[VirtualLarry]

Depends on the paste. Stuff like Kryonaut and ICD7 demands spreading. MX4 doesn't. Since you brought up Ryzen 3000 . . . I would avoid Kryonaut. It breaks down above 80C and boy does R3K like to run above 80C.

I use MX-4 almost exclusively now. (Have some tubes of MX-2 too, that I bought cheaply.)

 

[DrMrLordX]

Then you should be able to follow conventional spreading techniques with it.

 

[ehume]

I hope we aren't still using concave/convex HSFs in 2019. . . .

You'd be surprised.

 

[aigomorla]

Heat sinks are now bowed under most occasions as it offers a more superior mount under pressure directly under the DIE.
Some will come flat with that mirror like finish, but bowed heat sinks and bowed blocks are vastly superior.

This was discovered a while back in the watercooling world when people would artifically induce a bow by putting a o-ring inside the inlet of the block.
Swiftech goes into it in detail here, as the first block to induce this artifical bow was the D-Tek Fusion, and Apogee.

http://www.swiftech.com/apogeegt-bow-tweaking.aspx

In CPU's where the geometrical center of the Integrated Heat Spreader (IHS) coincides with the geometrical center of the water-block base plate, an artificially induced bow in the water-block base plate increases the contact pressure at the geometrical center of the IHS and can be further qualified as a pressure gradient applied to the entire surface of the CPU IHS with a maximum pressure value at its geometrical center. The observable result is an improvement in the thermal joint between the CPU IHS and the waterblock base plate in the general area directly above the CPU die(s), while the water-block thermal resistance remains approximately the same.

Now the problem with a bow is when intel changed from solder to paste.
It now allows the IHS to shift geometry or in a sense move ever so micromil and not apply even pressure on the bow.

Also there will always be idiots that delid the cpu entirely and decide to put a bowed block ontop of a naked and die, and well, 5 letter word.. I D I O T.


But bowing also makes the concave heatsink moot, as it again applied enough pressure to fix the deformity of the heat sink.

 

[extide]

So, if you are curious about Ryzen 3000 series -- Look around the edges of the IHS -- you will see that 3 sides have silicone going all the way, but one side (I think it is left side if AMD text is facing up) has a gap in the silicone in the middle. The IO die is on the side where that gap is, and then the CCD(s) are on the opposite side.

 

[DrMrLordX]

Heat sinks are now bowed under most occasions as it offers a more superior mount under pressure directly under the DIE.

Huh. Is this something that's observable to the naked eye? My NH-D15 didn't appear to be bowed, nor did my Heatkiller IV. Not sure about my NH-D14 since I lapped it.

 

[ehume]

Huh. Is this something that's observable to the naked eye? My NH-D15 didn't appear to be bowed, nor did my Heatkiller IV. Not sure about my NH-D14 since I lapped it.

As a heatsink reviewer, I saw a fair amount of them. Heatsink manufacturers improved the performance of their heatsinks in two ways: their mounts put a lot more pressure than Intel specified, and they made heatsinks with convex contact surfaces.

The OEMs that incorporated springs in their mounts generally also had screwstops that would prevent overtightening/overpressure. I tended to do a razor blade test and made in-profile pictures to demonstrate how convex or flat the contact surfaces were. In one review I had a heatsink that was moderately tightened vs firmly tightened to show what a difference pressure made.

One thing about convex surfaces and overpressure: they ruined a CPU for comparative purposes. But that does not mean the CPU was ruined as a CPU. Nor did an overpressure heatsink disqualify itself as a heatsink. In fact, I reused such a CPU and such a heatsink: the needs of a reviewer are different from the needs of a user.