i7-4790K Thermal Throttles on Stock and Overclocking issues

Equilibrium23 · 17 Jan 2021 at 13:13

Hi,

I have posted the same on reddit and am struggling to get some advice so thought I would post this here as well. Reddit post: https://www.reddit.com/r/overclocking/comments/kystwp/i74790k_overclocking_was_stable_now_isnt/

Essentially I got a used i7-4790K and intended to socket swap it with my old i5-4690K. To cut a long story short (more detail can be seen in the reddit post) the CPU now won't even properly stress test using OCCT Small Data Set 8 Threads - it just instantly thermal throttles to 100 deg. I have manually set the settings in the BIOS to stock and ensured FIXED mode is on (not adaptive). Now I need to stress that I have had this chip run at 4.7GHz at 1.29V with a temp of 68 deg. I am unsure why this is happening, the CMOS has been cleared, optimised defaults loaded before any changes are made. I have been at this for a couple days now so any help would be highly appreciated.

RIG:
i7-4790K
MSI Z97 Gaming 5
Coolermaster 212 Evo w/ Arctic Silver 5 + Noctua Fans
16GB 1333MHz RAM

drakulton · 17 Jan 2021 at 13:15

How long ago was it running at 4.7? Has it been delidded?

Equilibrium23 · 17 Jan 2021 at 13:18

Not delidded no. Was running for a 10 minute stress test on that occasion. Prior to that I had it at 4.7GHz and was running fine and was able to play warzone for numerous hours. It was after that point I stupidly did a GPU OC and that opened a number of issues. I have fixed that GPU and reverted that but still no luck. Seems like every time I attempt an OC, different behaviour is displayed.

Equilibrium23 · 17 Jan 2021 at 13:37

https://pasteboard.co/JK3UI0h.jpg
https://pasteboard.co/JK3V6CB.jpg

Please see the 2 images above for the settings I am using to replicate "stock" or default.

Hakanese · 17 Jan 2021 at 14:00

I don't know what your vccin is, but 1.8V I think that might be your issue

Equilibrium23 · 17 Jan 2021 at 14:04

Vccin is the same as reference voltage or the voltage into the motherboard. I have been reading that this is fine and that it needs to be within 0.4V+ of the Vcore value.

Mr Mcgoo · 17 Jan 2021 at 14:10

what voltage is the cpu running at when at stock settings. Last resort try reseating the cpu and cooler

Equilibrium23 · 17 Jan 2021 at 14:13

Voltage at stock is 1.056 as in the image. I have reseated the cooler about 4 times now. Like I said it runs fine, decent temps and then randomly just throttles after being turned off. Could it be the motherboard?

Hakanese · 17 Jan 2021 at 14:14

So your issues happened after a gpu oc, did you do anything else? Like move the cooler?
What are your temps now at normal use?

Mr Mcgoo · 17 Jan 2021 at 14:19

so it happens after you turn you system off and back on again?

Equilibrium23 · 17 Jan 2021 at 14:20

Yes I believe so. No nothing else all remained the same. Normal Use playing warzone are 50/54 degrees.

Hakanese · 17 Jan 2021 at 14:22

I think you have your core and ring voltages the wrong way round, not too sure

Aretak · 17 Jan 2021 at 14:25

Strange. Could be a motherboard issue, but one other possibility is that the thermal paste under the IHS could be on its last legs. Intel were still trying to con people into believing that solder was too expensive during the Haswell era, so it's (really quite crap) paste under there. If the results you're getting have been changing after thermal cycling, it could be that poor contact is being made and the heat is shifting things around, causing the mixed results. Delidding it and reapplying (or switching to liquid metal) might be something to consider.

Hakanese · 17 Jan 2021 at 14:27

Normally if your temps jump through the roof on a benchmark, my 2 initial thoughts are, it's either too much power from an overclock, or the cpu cooler isn't making decent contact/fans not spinning/failed

Equilibrium23 · 17 Jan 2021 at 14:32

So when I was running the stress test max power was hitting around 115W and fans were spinning at max RPM (1700). The flucation in results is what is really confusing me.

Hakanese · 17 Jan 2021 at 14:35

Do you have Hwinfo? you can check the voltage draw, but I'd just reseat the cpu cooler just tick the box and get if off the list.

Aretak · 17 Jan 2021 at 14:42

Hakanese said:
Do you have Hwinfo? you can check the voltage draw, but I'd just reseat the cpu cooler just tick the box and get if off the list.

He said earlier he's reseated it four times already. I'd find it hard to believe somebody could screw up mounting a cooler that many times in a row. The voltages in the screenshots look fine too and 115W peak is nothing outrageous for a 4790K. Still think I'm right, personally.

Mr Mcgoo · 17 Jan 2021 at 14:46

i'm leaning towards a motherboard fault. if having reseated the cpu cooler hasn't worked then that maybe your issue. have you tried re-applying the thermal paste for safety?

Equilibrium23 · 17 Jan 2021 at 14:51

So everytime I have reseated the cooler I have replaced the paste as well. the fact that sometimes it works fine and others it doesnt just screams something is not working correctly. Any other ideas?

Mr Mcgoo · 17 Jan 2021 at 14:54

have you still got you old 4690k and see if that gives the same issues your getting with the 4790k?