Hand lapped my 5950x

lobstar

Limp Gawd
Joined
Nov 24, 2004
Messages
371
I still have to clean it up a little bit but I'm really happy with the visual result. I got a small sheet of glass from Lowe's and used 3m wet/dry sand paper from 400 to 1500 grit with a bit of water. It took about 3 hours total while watching twitch. I'll report pack once I get this under my Optimus Foundation water block.
1617513596023.png
 
looks pretty good. checked with a razor?
Yeah! It's really close to perfect. I could probably give it another hour on 2000 grit but I'm not sure how much improvement I can really get. I was thinking of covering the surface with ink and giving it another go to see which area loses the ink first just to confirm. I think I'm in diminishing returns territory as it is though.
 
So so so any results? I am curious before I install a 5950 in my loop.
I am waiting on my Optimus GPU block before I assemble the system. I'm six weeks into my 1-2 weeks shipping so who knows when it will arrive.
Are you going to lap the block too
I have an Optimus Foundation AM4 block which comes machined flat and polished.
 
I am waiting on my Optimus GPU block before I assemble the system. I'm six weeks into my 1-2 weeks shipping so who knows when it will arrive.

I have an Optimus Foundation AM4 block which comes machined flat and polished.
You can't drop this on us and then tell us to wait. Lol.
 
Yeah but I thought that would generally be useless because we don't know if he has a better binned cpu to begin with.
I understand your viewpoint but I don't really care about the data as much. I've never done this before and hate tearing down my loop. Personally, I just wanted to give myself the best opportunity with the least effort. I'm in the camp of 'there's lots of existing data' for me to make comparisons with. I always end up with absolutely terrible luck in the silicon lottery.
 
I understand your viewpoint but I don't really care about the data as much. I've never done this before and hate tearing down my loop. Personally, I just wanted to give myself the best opportunity with the least effort. I'm in the camp of 'there's lots of existing data' for me to make comparisons with. I always end up with absolutely terrible luck in the silicon lottery.
Hopefully not this time!
 
I did hand lap a CPU long ago, I may have even posted the results on here :) but it was long long long ago.
 
Ahh, it was the water block, it wasn't flat at all, so I did lap it to reduce the waviness.
(2008, apogee GT)

1617715348411.png
 
Last edited:
I decided to check the waterblock with calipers while the cold plate was torqued to 1.8Nm. Measuring from each edge using the Optimus sticker to orient the measurements: North - 21.43mm, East - 21.43mm, South - 21.44mm, West - 21.42mm. That was fully assembled so I understand some variance could come from the top however the cold plate does distort under torque while pressed against the gaskets so I wanted an 'in-use' sort of measurement. The fork of the calipers rests completely flat against the surface of the material so I'm confident it will create a good enough junction between the cold plate and the heat spreader of the cpu.
 
Temps between CCDs are super close with initial testing. Less than 1°C difference at various loads. Still tweaking and playing since getting it setup earlier this evening.
 
Water was at 30C during the tests. I have the AC going so might be able to pull another degree or two off that number.
Edit: Mid run with a data reset. Water at 28C. I'll try again if the water comes down below 25C.
1620148343684.png
 
Last edited:
Water was at 30C during the tests. I have the AC going so might be able to pull another degree or two off that number.
Edit: Mid run with a data reset. Water at 28C. I'll try again if the water comes down below 25C.
View attachment 353141

my water at quiet settings stays around 29-31 C idle, and 35 load(GPU) cpu load doesn't move the temps at all.
 
Back
Top