Lapped Ninja/CPU - results

It's been a while since I posted here...been a little busy during the summer months...but anyhow:

Recently, I decided to take the time to clean out the computer. It's a good thing I did, too! It was chock full of dust. The reason I decided to in the first place was because I noticed my temps creeping up and up. Also, I noticed my chipset fan had become quite loud. I bought a new fan and got to installing it. I had to take the CPU heatsink off to get my clubby hands in there. And that's when I decided...I'm gonna do some lapping.

I had already lapped my original version Scythe Ninja but was not completely satisfied with the results. This time I started with 220, moved to 320, then 400 and finished with 600. I just put the piece of glass on the table, grabbed a beer and went to it. 3 hours later (I got distracted often) I was satisfied with the results. I had also, during this time decided I might as well lap my CPU as well.

After everything was said and done, I had a reasonably shiney, but most importantly, FLAT CPU and heatsink. I go to install them and what do I see when I boot up? Temps that don't make much sense! Prior to my lapping the CPU or heatsink I had a reasonably small variation between cores ~5 or so degrees. Now I have ~10-13 degrees. Albeit, the difference is throughout the spectrum. Load temps seem to have improved, showing a smaller difference between idle and load. But it's the variation between cores that bothers me.

Is there a limit to what is acceptable as temperature difference between cores? Again, the load temps are not worrisome, but it's just the new/larger difference I'm seeing between cores has me somewhat concerned.

Try less TIM .
If your CPU and HSF are now flat it will need less TIM than it did before .

The first thing that comes to mind when there's a big difference between cores is improper seating of the heatsink.

Since you lapped with a lot of low grit paper, the heatsink might not contact the CPU very well, even with a good seating. You could try using washers on the mounting hardware to ensure good contact.

I did try to reseat it a couple of times. The first time I had not realized that I only had 3 of 4 clips holding the heatsink to the CPU. I did get the feeling, Matt, that due to the lapping I increased the tolerance between the two and was trying to figure out a way to increase the mounting pressure. I'll have to tear it all apart again and see what I can do with washers, good suggestion. I also did not decrease the amount of TIM I used. I just put a grain of rice size dab of AS5 in the middle of the IHS and put the heatsink on.

So...2 things to check when I get home.

striker85 said:

[...] I only had 3 of 4 clips holding the heatsink to the CPU.

Click to expand...

If I understand correct it's a 775 CPU? If yes do yourself a favor and forget about those pesky push-pins and buy a Scythe Stabilizer 775 or TR Bolt-Thru or Xigmatek Crossbow.

Nope, socket 939 AMD. I did do some research on modifying the mounting mechanism of the Ninja and was only able to find the replacement for the pushpins on the 775 design. The socket 939 design uses 4 simple spring clips mounting to a bracket attached to the motherboard. The easiest solution, I believe, would be as Matt mentioned and shim the bracket out with a few washers.

OK, my bad.
Indeed after lapping your temps looks like your Ninja needs a little bit more pressure.