CPU Spike to 100C

Joined
Apr 2, 2017
Messages
20
Hi everyone, I picked up a used 4790K today and put it under my evga CLC280. Its idling around 29C which is absolutely normal, but then when I run AIDA64, the CPU will spike to 100C! The max voltage goes up to 1.35V @ 4.6ghz OC.

I also found out this CPU was de-lidded and the guy applied thermal paste instead of conductonaut.

I am speculating that I have either applied too much thermal paste between CPU and cooler or the thermal paste used in the de-lid is insufficient.

Anyone have any thoughts?
 
Do a re-mount of the cooler before anything else.

It's unlikely you've applied 'too much' thermal paste between the CPU and cooler as long as the mount is solid/secure and decent pressure any excess will typically be pushed aside anyway.
 
I would not exceed 1.3v with that gen CPU for daily use and be satisfied with whatever clock you can achieve with that voltage. That said, I once forgot to remove the plastic covering my heatsink and got the similar result. I didn't realize it until I remounted it. If you water cool, delid your proc, reattaching the IHS with liquid metal and use decent TIM I would increase that limit of voltage to 1.35v with no worries for daily use. Use a 240 or 280 rad no less for just a CPU with that TDP and voltage.
 
ive got here a 4790k @ 4.6 with 1.28v under a h110i(280mm) and it peaks at 65c, its not delidded either. so start with lowering the voltage and make sure your clc is seated properly.
 
So I reapplied thermal paste, reseated the cooler and the CPU is fine running at 4.6 w/ 1.28v.

Ran Aida64 again. The CPU still spikes up to 100C. I even set the pump and fans to 100% and it still goes up to 100C.

Could it be the de-lid job (from the guy i bought it from) was botched? Like I mentioned, he used thermal paste instead of conductonaut...

In gaming though, it doesnt go above 65C.
 
So I reapplied thermal paste, reseated the cooler and the CPU is fine running at 4.6 w/ 1.28v.

Ran Aida64 again. The CPU still spikes up to 100C. I even set the pump and fans to 100% and it still goes up to 100C.

Could it be the de-lid job (from the guy i bought it from) was botched? Like I mentioned, he used thermal paste instead of conductonaut...

Quality thermal paste will still see a temp drop vs the intel stock TIM. It's possible he didn't use enough to make proper contact, less so to do with what type of paste he used.

Is his re-lid 'glued' back down around all sides? Is it something you could re- delid easily?
 
Quality thermal paste will still see a temp drop vs the intel stock TIM. It's possible he didn't use enough to make proper contact, less so to do with what type of paste he used.

Is his re-lid 'glued' back down around all sides? Is it something you could re- delid easily?

He did use compound to re-seal the HIS back on.
 
yup look at the de-lid next. TIM could be cooked or lid isn't glued right or shifted or too much RTV so the die doesn't contact correctly.
 
Attached is what greeted me when I took off the IHS.

I took all the paste off the die and the IHS and scraped off the silicone sealant as much as I could. I ended up putting some MX4 I had on the die then plugging into my mobo and having the socket hold down the IHS.

The CPU is now running at 4.6ghz with 1.18v. Gets to around 60C on load.

My question is whether there is any substance that can remove the remaining sealant (it’s proving very stubborn against my little scraper) and any solution to use on thermal paste that appears to have gone onto the pcb. The little pin things on the left, I can’t get the paste out of there with a cotton swab.
 

Attachments

  • 8D30B506-CCA2-4BE5-936A-A700F05F270A.jpeg
    8D30B506-CCA2-4BE5-936A-A700F05F270A.jpeg
    516.4 KB · Views: 0
nice....
you can use a soft bristle tooth brush to clean the tim out. to get the RTV off you can carefully scrape it off the pcb with a razor.
 
Back
Top