To monoblock or not to monoblock?

[tistou77]

You used LM on the underside of the waterblock too to cover the size of the die?*

Yes of course
Tested several times

[Abaidor]

Something is wrong then or defective......

What waterblock are you using.....

[JustinThyme]

Something is concave or convex.......I always do the spin test. for convex and straight edge for concave. Usually no worries with dies but blocks can be a different story. Ive lapped them in the past. Do you know which cores and can identify them on the layout. to see of they are in the middle or outer edge.

[Brighttail]

Something is concave or convex.......I always do the spin test. for convex and straight edge for concave. Usually no worries with dies but blocks can be a different story. Ive lapped them in the past. Do you know which cores and can identify them on the layout. to see of they are in the middle or outer edge.

Where does one find this information? O Specifically for the 7900x CPU.

[Abaidor]

Something is concave or convex.......I always do the spin test. for convex and straight edge for concave. Usually no worries with dies but blocks can be a different story. Ive lapped them in the past. Do you know which cores and can identify them on the layout. to see of they are in the middle or outer edge.

I guess that lapping the CPU block removes the nickel plating too though....does this mean no LM then after lapping?

[tistou77]

Something is wrong then or defective......

What waterblock are you using.....

Ek Velocity

I had a Koolance 390I but too convex, which I changed by the Ekwb

Something is concave or convex.......I always do the spin test. for convex and straight edge for concave. Usually no worries with dies but blocks can be a different story. Ive lapped them in the past. Do you know which cores and can identify them on the layout. to see of they are in the middle or outer edge.

That's it for the hottest cores (@stock), I do not know physically where it's located on the die (core 2, 4, 12, 13, 14 I would say)

[LiveOrDie]

Ek Velocity

I had a Koolance 390I but too convex, which I changed by the Ekwb



That's it for the hottest cores (@stock), I do not know physically where it's located on the die (core 2, 4, 12, 13, 14 I would say)

almost 30c difference that is abnormal i would be looking into that.

[Abaidor]

Ek Velocity

I'll be damned if the Velocity is concave/convex - I already bought 2X of them because when I bought the first one the D-RGB version was not out, so I got a second one (D-RGB)....LOL we're addicted with this hobby and the "excuse" that this machine makes me money does not hold for many of the parts purchased in my case..

Anyway I will put the second one to another PC (Gaming) I am building most probably with an 9900K....


Anyone else using the Velocity and facing issues with direct die?

[tistou77]

For my part, I think of a mounting problem with Velocity
Otherwise the WB is slightly convex but much less than the Koolance

The hotter cores on my screen would match what location on the die ?

I hesitate to take the frame of ICE
Or I expect more tests of 9980XE (heating, etc ...) and no problem of direct die

[Abaidor]

For my part, I think of a mounting problem with Velocity
Otherwise the WB is slightly convex but much less than the Koolance

The hotter cores on my screen would match what location on the die ?

I hesitate to take the frame of ICE
Or I expect more tests of 9980XE (heating, etc ...) and no problem of direct die

The Velocity comes with springs for the 2066 socket based on their manual and the ones I have here. I am not dismantling my build to test it before the direct die frame is here so I rely on literature and the parts in the box right now.


Since I am personally interested in this and I was suspecting contact issues I was "studying" it a bit and if there is an issue here is what I suspect. Look at the draft diagram I modified below from the manual.




With that said I am not sure how safe it is to let the block "descend" freely on the bare die but that's why we have the direct frame which is there to prevent the block from crashing it.


***EDIT #1***

My guess is that EK calculated the tolerance of the mechanism with the IHS in place and maybe left a bit of headroom to account for the convexity/concavity of the IHS. With the IHS removed that tolerance might not be enough thus preventing 100% contact. This might or might not be the case and I am soon about to find out. The solution will be to either use other struts / thumb nuts and be REALLY careful about the force applied while tightening which would be best found through trial and error. I mean going really slow and measuring temps instead of brute forcing it.


***EDIT #2****

ICEMAN states it clearly on their product description