i7-8086K Strange Temps, anyone else?

JarJarBinks

JarJarBinks

Limp Gawd
Joined
Jan 13, 2008
Messages
271
I have a i7-8086k @ 5Ghz with a Fractal Design Cooler S36-BK (360MM Water Cooler). I wanted to ask about some stuff because a lot of stuff just doesn't seem right. It pulls 1.296v according to CPUz. And according to Intel power gadget, I'm using upwards of 194W of power and a temperature of 96C when running prime95 (In-place FFTs). The temperature just doesn't seem accurate though, the air blowing off the radiator isn't abnormally hot, the tubes flowing from the water block do not feel hot but I can feel the liquid flowing. And the system never becomes unstable either while video encoding for several hours or playing games for 15 hours or running Prime95. Prime95 Blend Test stays at 70C on the dot. Should I be concerned about these temperatures are they accurate? When Intel says the max temperature is 100C does that mean I can run it at 96C for 8 hours at a time and not experience any problems over say 3 years? Idle temp is 37C but I live in a hot climate and my air conditioner is poor. I have read that the 8700K and this processor run hot it just doesn't seem like it actually is. Mobo in the signature. Thanks for any facts or opinions on the matter.
 
Delid it. Else back off the clocks a little.

I wouldn't run it that close to max temp personally.
 
I am considering deliding but I don't have any experience with liquid metal TIM that is suggested when deliding, plus its normal load temps are usually never higher than 70C. Is 70C a problem? Do you know a one-stop site that I can buy a delidder for that processor, the TIM, and then a replacement copper lid like I have seen?
 
JarJarBinks said:
Is 70C a problem?
Click to expand...

No.

JarJarBinks said:
Do you know a one-stop site that I can buy a delidder for that processor, the TIM, and then a replacement copper lid like I have seen?
Click to expand...

RockitCool

The bare-cooper IHS is pretty pointless unless you're looking to squeeze every last ounce of performance from your CPU. Replacing the internal paste with metal TIM gives a massive improvement (10-20C), while swapping the stock nickel-plated (also copper) IHS will net you single-digit temp drops.
 
Kyle posted up an article about it not very long ago. It should be easy to find. I think it was the 8700 which is essentially the same.
 
Ok, thank you, everyone. Would it be a good idea to put solder mask on the CPU so that if some of the liquid metal makes its way onto the processor it can't short anything or would the solder mask hurt the CPU?
 
JarJarBinks said:
I am considering deliding but I don't have any experience with liquid metal TIM that is suggested when deliding, plus its normal load temps are usually never higher than 70C. Is 70C a problem? Do you know a one-stop site that I can buy a delidder for that processor, the TIM, and then a replacement copper lid like I have seen?
Click to expand...

SiliconLottery.com has a service with a super fast turn around for delid and TIM replacement. They'll even OC it and give you the max stable clock/voltage and other relevant info. When they install the new liquid metal TIM, they use liquid e-tape to seal everything off so no concerns about shorting, and then glue the IHS back on. I did a delid on my 6600K, and any longer, I would just pay for this service. It was nerve wracking to do this on my own. You have a limited edition processor... I would let them do it and that way if they break it during the process, they are going to have a spare to swap out with.
 
JarJarBinks said:
Ok, thank you, everyone. Would it be a good idea to put solder mask on the CPU so that if some of the liquid metal makes its way onto the processor it can't short anything or would the solder mask hurt the CPU?
Click to expand...

If some of the liquid metal makes its way onto the processor? I don't think you understand how this works... You apply liquid metal to the CPU die. That is "the processor". Do you mean the CPU package/PCB? There's not much risk of that. The amount of liquid metal used in this process is too small for the liquid to run off the die, as long as you do it right. If you sit there and glob on liquid metal like it's some generic thermal paste and then just slap the lid on (without spreading the liquid metal into a thin layer as you're supposed to) then you could have this problem, but if you're doing that you haven't done your homework.

Bottom line: watch a guide on the procedure and you'll be good.
 
I got it, I watched enough videos that I feel confident in doing it myself. I just have never used liquid metal and I know there are parts on the package that can become bridged so I wanted to know if using solder mask would be a good way to make sure no liquid metal bridges anything. Also, is it recommended to use the liquid metal on the outside of the IHS too or is that too risky?
 
JarJarBinks said:
I got it, I watched enough videos that I feel confident in doing it myself. I just have never used liquid metal and I know there are parts on the package that can become bridged so I wanted to know if using solder mask would be a good way to make sure no liquid metal bridges anything. Also, is it recommended to use the liquid metal on the outside of the IHS too or is that too risky?
Click to expand...

Use it between cooler and ihs as well.
 
So I got the stuff in and attempted the DELID process. Everything seemed to go perfectly. And temps are definitely 25C lower at load.
42Q7XkH.jpg

UEFeCBY.jpg
CZJcO3J.jpg
I7GhxHX.jpg
 
Looks great.

Just remember with the any of the gallium based "liquid metal" compounds to keep them far, far, away from aluminum.



Copper is better, but it may occasionally "dry out" as the gallium is absorbed, pitting and staining the copper, and need to be reapplied.

Liquid gallium at ~30°C also slowly reacts with copper to leave the copper surface pitted. This chemical reaction with copper is a concern...
Click to expand...

From: https://www.osti.gov/servlets/purl/811932/

And:
...liquid metal will literally eat into the copper until the gallium is gone, and the resulting copper-gallium alloy is a silver-ish color. Yes, - in case you are wondering - the gallium in liquid metal reacts this way despite the fact that there are other metal stabilizers present in LM
Click to expand...

From: http://forum.notebookreview.com/thr...-works-why-it-fails-and-how-to-use-it.809332/

According to that, it does it to nickel too, just much slower...

But hey, it's not all bad "The CuGa alloy obviously can't match pure copper for heat conduction but it's still way better than solder or thermal paste..."
 
I'll report back my experiences. I have it on copper on both the IHS and Water Cooler. I saw a video after 9 months of running liquid metal on an i7-7700k and while there was some discoloration on the copper the liquid metal had not dried out in that amount of time and temps remained exactly the same. I have heard other people say that as long as you don't remove the IHS/Heatsink the temperatures will remain stable. Once you remove those you gotta reapply to get the connection but if it dries out while connected the heat transfer still remains stable. My system runs 24/7 with a very mild overclock at 1.325v and probably only spends 10 hours a week on full load. I kept the application of liquid metal as thin as possible and I put solder mask on all the CPU package contact points, forgot to take a picture of that. This is my first experience so I will see what happens. Because I will unlikely pop the IHS in the next 2 years if temps remain constant.
 
In one of the posted links, someone easily cleand the lm stain after almost 2 years by using a product made specifically for cleaning thermal contact surfaces. There is even a video posted of someone else using polishing compound (for cars) to remove the stain. Ive even seen silver jewelry polish (metal polish) used. Basically something a bit stronger then isopropyl can clean the stain off the ihs.
 
I copper topped my 8700ks, cheap, fun and totally worth it.

Gallium will do nothing that you care about to pure copper in the meaningful lifetime of a delidded cpu.
 
You must log in or register to reply here.
Back
Top