Jump to content
Corsair Community

H115i not cooling my CPU?


Cenzik

Recommended Posts

The temp on my i7-7700k when in the BIOS is typically around 37°C and about 50°C when idling on my desktop, not too sure if these temps are too high. I first experienced a problem when trying to emulate a game and my PC crashed after about 2 hours of it running. I tried the game again while monitoring my CPU temps, and they were hitting 100°C or near that on all cores. I thought it might just be an issue with the game, but I got near identical temps when I tested Rise of the Tomb Raider (both games running in 1440p). I checked the water temp of my AIO when playing Rise of the Tomb Raider and it only slowly works its way up from its idle 27-28°C to the low 30s. The RPM of the pump and fan seemed normal the whole time, just the speed at which the water rises. This was with the stock thermal paste that comes pre-applied on the H115i. I changed out the thermal paste to see if something was wrong with that, or if I had poorly installed the AIO, yet I'm still getting the same results; my CPU temps immediately jump to the 90s under heavy load. Same thing happened when I tested prime95. When playing less intensive games my temps are still reaching the upper 70s and even into the 80s. I reset my BIOS to see if this changed anything, but my temps still remained the same as before. This seems way to hot for an AIO from what I've seen on the internet, so I'm not sure if this is a problem with my AIO or with the CPU itself. Edited by Cenzik
adding more info
Link to comment
Share on other sites

...I'm still getting the same results; my CPU temps immediately jump to the 90s under heavy load.

 

You need to take a look at the Vcore actually applied during load. Any time the temp instantly jumps to an elevated level (vs slow and continual climb) you are looking at voltage or conductivity/contact. If the cooler was not functioning, the coolant temp (H115i Temp) would start ticking upwards at a brisk rate 30..35..40..45..50... etc.

Link to comment
Share on other sites

I checked my Vcore in the bios and got 1.184 volts, when idle and got 1.261 volts, and when under heavy load it would only barely go past 1.3 volts. It seems that even though 1.3 volts is a "safe" voltage, people don't recommend it. I'm not sure if the number of volts is acceptable for my CPU or not, and if not should I manually set a stable voltage for my CPU?
Link to comment
Share on other sites

That does not seem like a particularly high Vcore. On full auto with Multi-core enhancement on, you can get up near 1.40v right out of the box and that is where most people get into trouble.

 

Prime is going to be hot and unless you are using it for specific targeted runs, I don't think it is a good indicator of your set-up. It is a really fast way to bin chips. 9 out of 10 are too hot instantly. Move on until you find the one that can take it. Try running the stress test in the Bench Tab of CPU-Z. This is a linear load, so aside from the initial Vdroop it will jump to CPU temp X and hold, only increasing by +1C each time the coolant temp moves by the same 1 degree. This is a better indicator if there is a real contact problem or something else. Prime makes it look like there is always a problem.

 

For the games, are we talking about the peak recorded CPU temps or the average? Once you start heating up the case, the CPU temp will be elevated as well. A lot of factors involved, both with the hardware and the game. It's easier to sort this stuff out with controlled loads like CPU-Z and then the games fall in line.

 

If you are on full auto for voltage and settings, leave it for now until we get a better sense of where things are. Setting a specific voltage is usually good even if you intend to run the Intel specified frequency. Auto is designed to make sure the worst CPU made gets enough voltage and Z270 didn't have some of the Asus tools that would come later to regulate that.

Link to comment
Share on other sites

I downloaded CPU-Z and over 90°C on all cores almost immediately and throughout the rest of the stress test. As for the games, immediately when I begin to run them, the temps skyrocket into the 90s, however when monitoring Rise of the Tomb Raider while just playing it, I was getting around 80°C on all cores. The original spot where I noticed the issue was with the game I was emulating and I'm still getting over 90°C throughout the whole time I play it. Even though Rise of the Tomb Raider is only around 80°C it still seems a little too hot for the cooler that I have and the airflow that my PC is getting. I'll post my build in case this raises any sort of red flags as well as the temps I was getting during the stress test.

35031345_pc2.thumb.jpg.c18b02a7fe0e1b70132a83777fd6eac6.jpg

1731799331_pc1.thumb.jpg.881b8ecd326a1dc04122292982dc7e23.jpg

pc.thumb.jpg.2188f9aada3f981c6c1ee0adb4ae97c5.jpg

606700009_stresstest.thumb.png.b28bbe3437ef1ae04f1d5b4f47a6ad59.png

Edited by Cenzik
fixing wrong image
Link to comment
Share on other sites

There is nothing wrong with your physical layout, airflow, fans, etc. When the temp immediately hits extreme levels, it can only be too much voltage or bad thermal transfer. Those other factors do matter for prolonged use, but if you hit 96C in 5 seconds it's voltage or conductivity.

 

I don't think your voltage is alarmingly high or enough to create these type of temperatures. There is a slight possibility you have the worst 7700K of all time, but then you would have been dealing with this all along. That suggests the most probably cause is the contact between the CPU and the cold plate on the pump. This is a finicky business and I can't tell you exactly what's wrong. 1mm off is enough to do this. Make sure the backplate is on straight in rear and not caught on the dimple. I would loosen the block and remove it to look at the TIM spread. If there is a section that does not seem to have spread, that would suggest uneven force. Tighten the thumbscrews back down a little bit at each corner until they are all tight. DO NOT forcefully over-tighten them. Finger strength is more than enough.

 

How long has this been going on? I assume neither the H115i or the 7700K are brand new. You are still getting coolant rise, which is why I the contact issue must be very slight or we are looking at a different problem. Have you updated the BIOS recently forcing you to redo your settings? I can't see the Vcore in that HWiNFO clip, but the game temps are too high as well, so I don't think this is connected to CPU instruction type or load type.

Link to comment
Share on other sites

So I took off the block to see my thermal paste spread, and I got this (thermal paste.jpg). This is the thermal paste I reapplied from a few days ago to try and fix the problem, but this almost seems like too little. I put the block back on after cleaning off the thermal paste and adding a little more than I did the previous time. I let my PC idle for 30 minutes and get these temps (idle temps.png), which is about a 5°C improvement on all cores, maybe a bit more. I then ran the CPU-Z stress test for 5 minutes with the core temps only jumping into the 80s and cores 1 and 3 ever making it past 90°C. I accidentally copied over my screenshot of the results at the end of the test, but you can still see my max temps in this later screenshot and the average temps were only about 5°C lower at the end of the test. I then played Rise of the Tomb Raider and got significantly better temps. I tried the game I was emulating too, and it first started in the upper 70s compared to jumping straight to the 90s, but overtime it worked its way into the upper 80s and some cores would make their way into the 90s, though none of the cores remained completely in the 90s. I seem to have gotten an improvement from reapplying the thermal paste, though I'm not sure yet if I'm achieving standard temps for the setup I have going on.

984782143_idletemps.thumb.png.cd78b9b8114a3cf4388e9c70b78d24e0.png

2135597457_thermalpaste.thumb.jpg.43c25538187d1655936c77538773615e.jpg

363531465_stresstestnew.thumb.png.f5889e679549c10c700c76b6c99f94bd.png

516597779_riseofthetombraider.thumb.png.6f58b71400316060a0759613424400d8.png

772376519_persona5temps.thumb.png.4b8eaff9ecb54ac4f9e6558ee423c91d.png

Link to comment
Share on other sites

I agree the TIM spread is not quite what I would expect. It's hard to tell from this end whether that is lack of pressure/contact or the original paste dot was very small. It does look like the top right corner has a built up edge and that would tilt my guess toward lack of pressure. it looks like you put a decent amount on, but it doesn't seem to spread like it should. That normally means lack of contact/pressure.

 

Check to make sure the stand-off post is all the way into the motherboard and back plate. I had a Z370 board that was a nightmare for this on 3 different brand backplates.

Link to comment
Share on other sites

×
×
  • Create New...