• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Delidded the 7960X - just a little difference ( -9 to -14 degrees )

Soldato
Joined
5 Sep 2011
Posts
12,812
Location
Surrey
My point is a cracked solder joint will still transfer heat better than paste so why even bother changing. I think the whole thing is just about cost saving and placing artificial limits on overclocking (particularly the voltage used) myself.

You'd have to show something to back that up because that's not really true (at all, in fact).


1. Yes, the CPU will run hotter, however the TIM through thermal cycling won't degrade or crack nearly as quickly. In fact, it's only likely to do so if using sub zero cooling, which is why a much, much thicker compound is used for benching when delidding the CPU. If a soldered CPU suffers micro cracking along the length of the die, depending on the size of the fracture you may as well not have a heat spreader at all.

2. Yes, obviously using a compound is cheaper than sourcing gold and indium, both rare materials. However, this isn't the only reason no matter how many times it's speculated. It's a mixture of time, money, environmental footprint and longevity.

3. Using a different thermal compound doesn't change anything to do with overclocking settings. Scaling rules, VID and board partners aren't privy to the thermal compound used, that's just a silly thing to say.

You just have to be aware of the bigger picture here. We understand the benefits here, but we're a small fraction of users, and ultimately there are real legitimate reasons for all of it. Just don't knock Intel about so much for it. If you really knew better, you'd be on an 8 figure salary.
 
Last edited:
Associate
OP
Joined
22 Oct 2006
Posts
1,035
It's arrived, not the quickest of deliveries.

Need to find a bit of time to fit the thing now.

Fitted it tonight, if anything the temps are a bit worse. Hey Ho. I'll pop the original back on tomorrow.

Nothing ventured.....
 
Associate
OP
Joined
22 Oct 2006
Posts
1,035
Pooh, could you expand a bit on your post with info on room temps, cooler / block used, paste etc please ?

Cheers :)
Hi mate

I'll do my best
Room temp was the same as pre swapping the Intel IHS for the copper one. Don't know what the actual temp was, quite warm, but it was a constant.

All other factors the same, paste, liquid metal IHS to the die and Arctic Silver 5 IHS to the waterblock which is a nickel one ( can't remember the model )

The biggest issue I had was that despite resetting the new IHS twice I could not get core 15 covered properly and it was scaling up to 100C when I ran Cinebench with all the cores set to 4.1. Looks like there was an issue with the milling of the copper IHS or something.

I took the copper one off and put the standard IHS back on after a good clean up of all the pastes and liquid metal, so pretty much a fresh start.

I made one change this time and used liquid metal between the IHS and the waterblock ( very carefully I hasten to add ) so it's now got liquid metal all the way through the interfaces.

Have to say it's now looking very good.

As a first pass I whacked the multiplier to 44 across all cores and left everything else unchanged and the temps were good. Room ambient is 29C so a bit warm.

Next job is to start per core overclocking which is a laborious pain in the ass task but the best way to squeeze the CPU to it's best performance while maintaining a decent thermal profile.

Capture2.jpg
 
Associate
OP
Joined
22 Oct 2006
Posts
1,035
Getting there

Most cores at 4.4 with a few a 4.5, edging the voltages down now.

Temps not looking too bad

Capture3.jpg
 
Soldato
Joined
10 Oct 2012
Posts
4,415
Location
Denmark
So TLDR if Intel used solder (or just sorted out their TIM application) temps even on their high tier chips would be respectable?

I'd like to imagine someone in Intel is very embarrassed that people are delidding their premium grade chips and getting results like this, but I doubt it :(

Intel are like Bethesda.. why bother when your users fix the problems for you.
 
Associate
OP
Joined
22 Oct 2006
Posts
1,035
Think I'm pretty much done now.
Running at 4.6 across all cores with manageable volts and temps. I found out the AVX offsets in the bios were a bit too small so was resulting in crazy volts & temps to get it stable. Bumped them up a couple of clicks, dropped down the CPU voltage offset to +0.010 and everyone seems to be happy.

bench4.6.jpg
 
Associate
OP
Joined
22 Oct 2006
Posts
1,035
Next job is to do some die lapping for a further few percent temperature drop. De8auer has a video on it.

I've lapped processor dies in the past. No real need to do it in this one. I'm happy with the current performance. Can't be a***d pulling the rig apart again. :D
 
Back
Top Bottom