4690k - First Core running much hotter.

Nexus Radical

Reputable
Jan 29, 2015
173
0
4,680
Hey guys,

Im not all that much experienced when it comes to overclocking. But i've been testing this clock for a while now. Its 4.5Ghz across all cores and the vCore voltage is 1.195v. The other cores a fine but the first core can get up to +20c hotter than the other cores. The first core is also more jumpier.

Here is some temp data monitored and recorded in a .gif:
e7ae4018740fdb195d57e0efe2a085bc.gif


Any ideas on what's causing this? If its specifically hardware related, fat chance for RMA im guessing?

CPU: I5-4690k
CPU Cooler: FrostFlow 240mm
Mobo: MSI Z97 Gaming 5
Psu: XFX XXX Pro 750W
 

Lutfij

Titan
Moderator
Are you on the latest BIOS revision for your motherboard? Are you stress testing the CPU to get those temps or is it during game play? If stress testing then what stress testing app are you using? If stress testing, those temps are an absolute worst case scenario since games wouldn't be taxing it at that range. Furthermore, mind sharing the ram's make and model as well as the OS you're on?
 

Nexus Radical

Reputable
Jan 29, 2015
173
0
4,680
Ah yes, sorry about the lack of information. I haven't posted here in a while now. Sure! :)

Full specs
CPU: I5-4690k @4.5Ghz (All cores) vCore 1.195v.
CPU Cooler: FrostFlow 240mm
Mobo: MSI Z97 Gaming 5
RAM: Corsair Vengence 4x4Gb @ 2Ghz
GPU: MSI GTX 970
HDD:
1) Sandisk - M.2 128Gb
2) Western Digital - SSHD 1Tb
Psu: XFX XXX Pro 750W
Case: Kolink Punisher.
OS: Windows 10 Pro

The stress tests were done with MSI ¦ Intel XTU. However the monitoring i did with CPUID's HWMonitor. The BIOS is running the latest version too and although im using XTU i've done my overclocks in the BIOS and not XTU.
 
Core fluctuations are normal in game use. In a stress test program like AIDA64 or OCCT they are not however and should only variate a few C from each other. Generally if one core is out of wack with the other three it means the CPU cooler mounting screws are not equally screwed down or the thermal paste is not evenly distributed. FYI CPUID is known to not be entirely accurate. The programs I referenced above have their own temp monitoring built in which I find more accurate.

With that said, your temps at a 4.5GB overclock on a 4690K are much higher than what one should expect with FrostFlow's 240L. For reference, my Noctua NH-D14 air cooler keeps my 4690K at 65C at 4.7GHz pushing 1.297v. The 240L performs 3-4 degrees better than the NH-D14. This is all assuming of course you don't have a technical problem like an air bubble in the line or the fans aren't stuck in silent mode or something.
 

Nexus Radical

Reputable
Jan 29, 2015
173
0
4,680


Well the sound rams up when i do start a test so its not that im guessing. As for air bubbles, i can't be too sure on that one but i'd feel like all cores would take a hit if that was the case. So it's worth cracking off the heat sink and checking the thermal paste? Im using MX-2 Paste BTW if that is relevant, heard it's pretty good. Right now my package temperatures are about 44c stable with the cpu only at 5% usage.
 

Nexus Radical

Reputable
Jan 29, 2015
173
0
4,680


Sorry for late reply. I just realized... I dont know how i didnt before but there is a massive crack in the copper. I think it cracked more and expanded when it was getting even hotter. idk but RMA'ing the cooler. This is the 2nd cooler i've had to RMA in a row. xD