Jump to content
Corsair Community

Need Help with High CPU Temperatures on My Corsair iCUE H100i ELITE LCD XT


Recommended Posts

Hi everyone,

I'm facing an issue with my CPU cooling setup and would appreciate any insights or advice. Here are my current specs and hardware:

  • CPU: Intel® Core i9-13900K, 3.0 GHz (5.8 GHz Turbo Boost) socket 1700 processor
  • Motherboard: GIGABYTE Z790 AORUS ELITE AX, socket 1700
  • Cooling: Corsair iCUE H100i ELITE LCD XT water cooling
  • RAM: Corsair 64 GB DDR5-5600 Kit
  • Storage: SAMSUNG 970 EVO Plus, 2 TB SSD
  • GPU: ASUS TUF Gaming GeForce RTX 4070 Ti 12GB
  • Case: Corsair 4000D AIRFLOW Tempered Glass midi tower
  • Fans: Corsair LL120 RGB LED PWM fan - 3 Fan Pack
  • PSU: Corsair RM1000x (2021) 1000W

Here's the issue I'm experiencing:

  • Idle Temperatures:

    • CPU: 50-60°C
    • Coolant: 35-38°C
    • Environmental Temp: 15°C
  • Gaming Temperatures:

    • Coolant: 50°C
    • CPU: 90-100°C
    • Environmental Temp: 15°C

This issue has only started happening recently. Previously, my CPU would run at 80-90°C while gaming without any problems.

Additional Details:

  • The AIO is front-mounted and sandwiched between two pairs of fans, one pair pulling and one pair pushing. Please see images below. 
  • One fan at the back pushing air out and one fan at the top pushing air out.
  • I changed the BIOS from optimal to spec enhanced and undervolted the CPU to -0.075v. This helped a little bit.
  • GPU runs fine, max 75°C really pushing it.

I'm concerned that these temperatures are higher than they should be. The AIO used to keep my CPU at around 80-90°C during gaming, but now it's hitting 90-100°C. Do I need a bigger AIO, or could there be another issue at play?

I can also try to top-mount the AIO if that might help.

Thanks in advance for any help or suggestions!

 

Taken at idle, just writing this post, running coolant temp 38°C

WhatsAppImage2024-06-20at15_36.30_7cd941f7.thumb.jpg.dfa2d947f6912b226e5dc227de86281a.jpgWhatsAppImage2024-06-20at15_36.29_a5dd7e9d.thumb.jpg.ff710ab4e23d0f39bde4c6c1df0be84e.jpg

Link to comment
Share on other sites

The idle coolant temp seems a bit warm for front intake.  However, the change when gaming is about normal.  If you have a front dust filter in place, try removing it and see what effect it has on both idle and load temps.  Filter + radiator is a decent amount of flow resistance and necessitates moderate to high fan speeds, even with push-pull.  

 

Your coolant to CPU temp differential is about right, but the one that seems a little high is the idle.  Typically Intel CPUs do not idle 15C above the coolant.  Try loading up another monitoring tool like HW Monitor or HWinfo to get a good look at actual core clocks vs voltage vs temperature.  Your voltage should sit down below 0.80v and the clocks in the 1-2 MHz range when on the desktop.  If they are staying up in the 4-5.5 GHz range with higher 1.2-1.40v Vcore, then something is keeping it loaded.  Lots of BIOS changes lately and not all of them are good or understandable in what they do to behavior.  

 

If you want to check basic AIO performance, use something like CPU-Z and it's stress test in the Bench tab.  It's linear and not overly taxing so it won't bang you off the power limits.  It should hold at a relatively steady temperature when you start the test.  I expect a similar +50C differential between your current coolant temp and the average core temp, but that is what you are looking for.  Also check the Vcore voltage level. It will be lower than what you see in games -- something like 1.15-1.24, but that is the number you want to know for comparison purposes.  

 

Are you saying it's 15C in the room?  Normally there is an expected flood of posts this time of year with temperature concerns and we slide into Northern Hemisphere Summer and room temps go up a lot.  At 15C intake air temp, you really should not be at 35C idle coolant temp after you boot up.  

  • Like 1
Link to comment
Share on other sites

Posted (edited)
44 minutes ago, c-attack said:

The idle coolant temp seems a bit warm for front intake.  However, the change when gaming is about normal.  If you have a front dust filter in place, try removing it and see what effect it has on both idle and load temps.  Filter + radiator is a decent amount of flow resistance and necessitates moderate to high fan speeds, even with push-pull.  

 

Your coolant to CPU temp differential is about right, but the one that seems a little high is the idle.  Typically Intel CPUs do not idle 15C above the coolant.  Try loading up another monitoring tool like HW Monitor or HWinfo to get a good look at actual core clocks vs voltage vs temperature.  Your voltage should sit down below 0.80v and the clocks in the 1-2 MHz range when on the desktop.  If they are staying up in the 4-5.5 GHz range with higher 1.2-1.40v Vcore, then something is keeping it loaded.  Lots of BIOS changes lately and not all of them are good or understandable in what they do to behavior.  

 

If you want to check basic AIO performance, use something like CPU-Z and it's stress test in the Bench tab.  It's linear and not overly taxing so it won't bang you off the power limits.  It should hold at a relatively steady temperature when you start the test.  I expect a similar +50C differential between your current coolant temp and the average core temp, but that is what you are looking for.  Also check the Vcore voltage level. It will be lower than what you see in games -- something like 1.15-1.24, but that is the number you want to know for comparison purposes.  

 

Are you saying it's 15C in the room?  Normally there is an expected flood of posts this time of year with temperature concerns and we slide into Northern Hemisphere Summer and room temps go up a lot.  At 15C intake air temp, you really should not be at 35C idle coolant temp after you boot up.  

Hello sir, thank you for the quick response.

Here are the HWMonitor stats after leaving the system idle on the desktop for about 5-10 minutes. During this time, I had the following applications running:

  • iCue
  • Discord
  • Brave (2 tabs: one Google and one Corsair forum)
  • Outlook
  • HWMonitor
  • WhatsApp
  • Steam
  • Pushbullet
  • PowerToys
  • MedalTV
  • Google Drive
  • IDM
  • Various Intel, Nvidia, and Corsair software

The system stays around 1.3-1.4V and 4.9+ GHz with 6% utilization. Most of the utilization is spent on these applications.

[spoiler]

image.thumb.png.5a11f82dbf1a8b370c8cc67adfaa3257.pngimage.png.7d401e0634bb78d2dfa0eac30a224d50.pngimage.thumb.png.68f9ac269f052fd9e2f42aafb3372ce6.png

[/spoiler]

When running the CPU-Z stress test, the score idles at 14650-15000 in the blue bar on the CPU multi-thread in the bench tab. The CPU temperatures would go up to 100°C instantly. Please see the attached screenshots for reference, including the uptime. Coolant temp would go to 50-52°C

[spoiler]

 image.thumb.png.1c41f561f06e3d5ce754975466000c5a.pngimage.thumb.png.148d44f830e7cd8641c0cac9bfd4889d.png

[/spoiler]

[spoiler]

HWMonitor_Nq4YPcrF0S.thumb.png.a3b98a33f008fe0a53c2be55b17b3670.pngTaskmgr_lSaIMAix9b.thumb.png.b171f12c016ef0aaa26133512ec5550f.png

[/spoiler]

And yes, my room is indeed 15°C at the moment( I like it cold haha). Maybe 20°C if unlucky, but when booting the PC, it goes to 32°C>35°C>38°C and then idles between 36-38°C coolant temp.

I'm looking forward to your thoughts on these observations. Thank you again for your help!

Best regards,

 

EDIT: I just realised spoilers don't work here, apologies for the large images. Please pretend the spoilers are there to mark which images are related to one another.

Edited by Screems
Link to comment
Share on other sites

Adding to the above, I just played a few matches of MW3:

  • Coolant: 50°C
  • CPU: 80-90°C

Image upload seems broken, sadly. 

Please see these links:

The images were taken after 3 games. Reset everything before the game. So the maxes it hit where game related.
CqXp7aA.png

QbZPhHD.png

 

Link to comment
Share on other sites

There are a couple of things bothering me in the details above.

1) Idle power consumption.  Something used 98% of the power limit on the desktop.  That should not happen.  There are some programs that use AVX instructions to load grab a bunch of information from your system.  3D Mark suit does with its SysInfo program and there are a few games that use a similar mechanic.  However, you still should not hit 100C at that level.  It could be a single one off event, but the additional information suggests it's a general trend.  

2) The CPU-Z test confirms there is an issue in one of two places.   It's either too much voltage/power or bad CPU contact with the AIO cold plate.  High idle temps can be a result of bad contact, but in your screen shot I can see the VID (CPU requested voltage) had a minimum value of 1.28v.  That means the CPU was being kept loaded the entire time.  For comparison, I opened HWMonitor when I started writing this and my 13900K VID is hanging around 0.80v with CPU core temps at 30C in a 25c room.  CPU package power has not exceeded 95W and I have deliberately opened and closed a number of similar programs to provoke it.  

 

So lots of tech media talk as of late about Intel and MB makers playing it a little aggressive with power limits on Raptor Lake CPUs.  That's us.  All the MB makers scrambled to put out a new BIOS that followed the new guidelines and the general result is everyone turned over their apple cart.  Some are too dampening for power levels.  Others don't do what is expected.  Have you recently updated to any of these new BIOS versions in the past two months?  

 

 

I also am concerned the coolant temp hits 32C as you cold boot.  If you've had it on all day, that's not out of the question.  When you turn the PC on after being off for 1-2 hours in a 15-20C room, it should take it 20-30 min to warm up to 30C or so without additional influences.  The original picture tells me the PC is not under the desk or in a cabinet, so I don't have a good explanation.  Normal things I would do to test this are not viable until the power is under control.  If you are banging off the temp limit with CPU-Z, then it's hard to test.  One semi-useful other test is to sit at the desktop with HWMonitor open to the core temps.  Change your AIO pump speed from Quiet to Extreme.  Do you see the core temps take a sharp plunge after doing so?  It should not affect them at all with a normal 15-25W CPU load on the desktop, but if it does it suggests there is some flow restriction in the CPU block.  That's  a hardware problem.  

 

The image below is just for general comparison.  That's me on the desktop opening an closing programs.  I am running a higher set of frequencies than you with the e-cores up to4.7 and up to 5.8x4 on the P-cores.  This is why my VID is higher than yours at 1.51 with light work.  However, the temps are decidedly lower.  My coolant temp is 26-27C in these shots, so that explains some of the temp differences.  However, its the power and clock numbers that are unusual.  Several cores never dropped below their non-turbo max the entire time (5.1-5.3 GHz).  

Screenshot (424).png

Link to comment
Share on other sites

Hi!

Thanks for the swift reply.
iCue is running everything at extreme. 

I tried resetting my Bios back to default and now the call of duty main menu gets my PC to 100C......

HWMonitor_73CvxZgPg6.png.f9657581178d660f8e7eddf197245002.pngHWiNFO64_sOHuMdwVGZ.png.56e5898c4334b166c8e572630ced0582.png

Link to comment
Share on other sites

I was talking to my partner about this issue, which is driving me nuts.

I have not had ANY issues for months, and all of a sudden yesterday it stats running 100C+
 

Could the pump just be faulty???

Link to comment
Share on other sites

Ok, so just yesterday and this...  that helps.  You didn't forget you updated the BIOS this morning, so we can chase that another time and lets pursue this as an AIO problem.

 

Try the Quiet to Extreme speed change on the pump in CUE.  Look at the CPU core temps when you do this.  Use HWMonitor because CUE tends not be as accurate or poll quick enough for this.  Do the CPU temps take a noticeable drop when you upshift the pump?  Normally I would tell you to do this as full load and there should be no change even at 253W, but that's not a viable test right now since you are getting downclocked instantly on starting CPU-Z.  

 

Obviously the pump is running or you would start to overheat the moment you power on, but you can get gunk build up in the CPU block cooling fins.  This causes the water to linger in the fins for too long and it picks up too much heat per unit of water.  This causes higher than expected liquid temps (your sensor is right there and we are seeing that) and the difference is passed on to the CPU.  Is mid to upper 30s also your normal coolant temp on the LCD?  Or is that suddenly up as well?

 

Just to be thorough, you could redo the TIM application on the CPU and repaste.  Maybe it got bumped recently.  I'll guess this is the original installation AIO from when you got the CPU and MB, so maybe 1.5 years old?  That's about when you should redo the TIM if you use the PC most everyday.  However, it does not explain the coolant temps at all and a typical CPU contact problem has high CPU temps for everything and low coolant temp.  I don't see that here.  

Link to comment
Share on other sites

Posted (edited)
17 minutes ago, c-attack said:

Ok, so just yesterday and this...  that helps.  You didn't forget you updated the BIOS this morning, so we can chase that another time and lets pursue this as an AIO problem.

 

Try the Quiet to Extreme speed change on the pump in CUE.  Look at the CPU core temps when you do this.  Use HWMonitor because CUE tends not be as accurate or poll quick enough for this.  Do the CPU temps take a noticeable drop when you upshift the pump?  Normally I would tell you to do this as full load and there should be no change even at 253W, but that's not a viable test right now since you are getting downclocked instantly on starting CPU-Z.  

 

Obviously the pump is running or you would start to overheat the moment you power on, but you can get gunk build up in the CPU block cooling fins.  This causes the water to linger in the fins for too long and it picks up too much heat per unit of water.  This causes higher than expected liquid temps (your sensor is right there and we are seeing that) and the difference is passed on to the CPU.  Is mid to upper 30s also your normal coolant temp on the LCD?  Or is that suddenly up as well?

 

Just to be thorough, you could redo the TIM application on the CPU and repaste.  Maybe it got bumped recently.  I'll guess this is the original installation AIO from when you got the CPU and MB, so maybe 1.5 years old?  That's about when you should redo the TIM if you use the PC most everyday.  However, it does not explain the coolant temps at all and a typical CPU contact problem has high CPU temps for everything and low coolant temp.  I don't see that here.  

The issues kept stacking up and it is becoming very late here so I have shut it off for the night. I will be sure to test the software and let you know tomorrow.

As far as I could see, iCue updated automatically last week, but other than that, I have not touched, updated or changed anything in the iCue software or my bios. Then yesterday afternoon, it ran fine, playing multiple games of COD. In the evening, it started to hit 100 and I got a warning message.

When running a windows memtest, which should be the lowest of the lowest usage? The coolant is 34C? Room temp: 18C

The PC was assembled 7/7/2023, so we arent even a year in.

Its just so weird to me how I had no issues and like the snap of the finger it overheats like crazy.

Edited by Screems
Link to comment
Share on other sites

I was unable to sleep and the memtest resulted in no issues.
This is not even 1 minute after boot up: instant 97C and throttling to then go back down to 55C
3 minutes after boot it idles at 60C, nothing running besides some random apps like Discord and Outlook.

image.thumb.png.f9a9870c2f34c06d5501021cf098a018.png

Link to comment
Share on other sites

I tried running the iCue software with quiet and extreme, it made 0 difference. To confirm, I kept it at extreme for another 5+ minutes, not a single change.

399khmN.gif

Link to comment
Share on other sites

I just booted the PC from cold after leaving it off or several hours. Coolant started at 27C and within the first minute rose to 38C, just booting the PC, doing nothing.

Link to comment
Share on other sites

Hi,

Try either...

Mounting the radiator the radiator the other way up with the hoses on the bottom, or, top mounting the radiator.

This ensures the pump is moving fluid and not air (if there's a lot of air in the system)

If this doesn't help, then the next step I'd try is to remove the pump from the CPU, clean both surfaces completely to remove all residual thermal compound then either reapply new thermal paste or use something like the Grizzly Kryo-sheet.

Hope this helps,

Mondsy

Link to comment
Share on other sites

6 hours ago, Screems said:

I just booted the PC from cold after leaving it off or several hours. Coolant started at 27C and within the first minute rose to 38C, just booting the PC, doing nothing.

OK, so I think this is what you use in your Support Ticket with Corsair.  It takes about 300W of continuous power into the AIO at medium fan speed to raise the temp 10C.  That would take about 8-10 minutes of continuous power.  When it happens in 1 minute, it means the AIO has a liquid flow issue or an extreme airflow issue.  Another way to confirm this and rule out all other software interaction is when you cold boot and stop in the BIOS.  You can't see coolant temp, but the CPU temp will do the same thing 40..45...50...55...60... all way too fast.  

 

On the Quiet-Extreme pump cycle, it's not so much whether it ultimately becomes cooler or warmer, but that there is a sudden temp change when you alter the speed.  That's the sign the liquid is lingering in the CPU block for too long and there is some type of obstruction.  Given what's happening with the coolant, you should it change the value on the LCD top too if this is the reason.  It might instantly drop 2C in coolant when you change to extreme from quiet at desktop power levels.  It should not.  If you can't demonstrate this, don't worry about it.  The first paragraph is the heavy evidence something is wrong internally. 

 

You can leave the mounting alone and will be doing this when you get a new unit back.  If there was a contact issue, you wouldn't be able to raise the coolant temp like this and certainly not at idle power levels.  Problem is elsewhere.  

 

5 hours ago, dmonds said:

Mounting the radiator the radiator the other way up with the hoses on the bottom, or, top mounting the radiator.

This ensures the pump is moving fluid and not air (if there's a lot of air in the system)

Typically an air blockage will make a lot of noise.  Gurgling, popping, something.  It would be a bit odd for it to suddenly develop out of nowhere without moving the case around, taking the radiator down for maintenance, etc.  However, all you need to do to test this is lift the front end of the case up with the pump set on the extreme.  Hold it up for 10-15 seconds.  If there are air bubbles, this will shift them, you'll hear it, and then likely a temp drop.  

Link to comment
Share on other sites

Hi everyone,

Sadly things are not getting better and the temps are just super high on the smallest tasks. 

Trying to run a PC backup got me to 40c coolant and 76C cpu.

Leaving it to idle for 20 minutes with barely anything open got it back to 36c coolant and 55-65C CPU.

Tilting the PC to try and get air out had no results.

Booting up basic kid games (putt putt, lol) threw the temps up to 95-100C for it to eventually crash the pc due to overheating.

I use this PC for work as I am a video editor and do marketing. So I decided to request an RMA since this is seriously impacting my pc’s performance.

Thank you guys so much for all the insight and help over the last few days.

Link to comment
Share on other sites

That's the right thing to do.  It's got some type of partial blockage, but not quite to the point where it goes into full shutdown in 60 sec like with zero flow.  Support is going to ask you a lot of the same questions we did.  Lead with the "I power on in the morning and the coolant temp goes up 20C in 1-2 min in a 15C room."  That should nudge them in the right direction.  You can add in there are no strange noises, you tried lifting the case to check for air bubbles, but no matter what you do the coolant temp rises up to 50C.  That should keep things on a straight path to what you need -- replacement.  Also, if you can ask for the Advance RMA.  This is where they ship you the replacement unit first on credit card hold.  Then you swap it out and send the old one back after.  That should cut the down time in half.  Unlike most partial blockages, your system is already in the near unusable state.  Can't stay like this for long.  

  • Like 1
Link to comment
Share on other sites

20 hours ago, c-attack said:

That's the right thing to do.  It's got some type of partial blockage, but not quite to the point where it goes into full shutdown in 60 sec like with zero flow.  Support is going to ask you a lot of the same questions we did.  Lead with the "I power on in the morning and the coolant temp goes up 20C in 1-2 min in a 15C room."  That should nudge them in the right direction.  You can add in there are no strange noises, you tried lifting the case to check for air bubbles, but no matter what you do the coolant temp rises up to 50C.  That should keep things on a straight path to what you need -- replacement.  Also, if you can ask for the Advance RMA.  This is where they ship you the replacement unit first on credit card hold.  Then you swap it out and send the old one back after.  That should cut the down time in half.  Unlike most partial blockages, your system is already in the near unusable state.  Can't stay like this for long.  

Hi!

About an hour after submitting the RMA, they accepted it, no questions asked. I did link to this forum post so that might have convinced them to just get it moving. Sadly, I am currently not in a position to do an Advance RMA. Where I live, credit cards are not often used, so I am also not in possession of one.

I was able to get my old laptop running for now and got my boss so far to cover some extra travel expenses to get to the office and use a PC there, sure it's a hassle, but it's the best solution for now.

Once again, thank you so much for your assistance. It is very much appreciated.

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...