The Corsair User Forums  

Go Back   The Corsair User Forums > Corsair Product Discussion > Cooling

Reply
 
Thread Tools Search this Thread Rate Thread Display Modes
  #1  
Old 10-03-2016, 09:11 AM
vhelskud vhelskud is offline
Registered User
vhelskud's PC Specs
 
Join Date: Sep 2016
Posts: 4
POST ID # = 874042
vhelskud Reputation: 10
Default H100i V2 High Temps

I just bought a new H100i V2 yesterday. The temps weren't improved that much compared to my Corsair H60. The idle temperature of CPU Package is fine 30~35c. However, whenever I start playing or do something like using multiple tabs of Chrome, it would rise up to 60c+. I tried to monitor the temps by using my computer for a while and then look at the maximum temperature achieved of CPU Package, I saw it reached 70+. I currently use the h100i v2 as pull configuration, the fans were positioned at the top of the radiator, because It wouldn't fit inside and will hit the motherboard heatsink
Reply With Quote


  #2  
Old 10-03-2016, 09:36 AM
red-ray red-ray is offline
Banned
red-ray's PC Specs
 
Join Date: May 2014
Location: England (GMT+1)
Posts: 7,152
POST ID # = 874045
red-ray Reputation: 81
Default

Assuming the H100iV2 is connected to the CPU_FAN header you need to disable fan control in the BIOS. Have you done this?

Attach the CL4 [Home] tab screen as a .PNG file so we can see all the fan/pump speeds and temperatures.

What is the room temperature?
Reply With Quote


  #3  
Old 10-03-2016, 09:45 PM
vhelskud vhelskud is offline
Registered User
vhelskud's PC Specs
 
Join Date: Sep 2016
Posts: 4
POST ID # = 874132
vhelskud Reputation: 10
Default

Quote:
Originally Posted by red-ray View Post
Assuming the H100iV2 is connected to the CPU_FAN header you need to disable fan control in the BIOS. Have you done this?

Attach the CL4 [Home] tab screen as a .PNG file so we can see all the fan/pump speeds and temperatures.

What is the room temperature?
This is the Corsair Link image sir, however, I don't know how to calculate the room temp. And you're right, h100iv2 is connected to CPU_FAN and usb 2.0 on the motherboard. Also, the fan control in the bios is disabled. BTW I'm using pull configuration. My nephew played 1 match of CSGO and I saw the temperature reached 65c
Attached Images
File Type: png corsair.png (412.3 KB, 2152 views)
Reply With Quote


  #4  
Old 10-06-2016, 09:03 AM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 874437
c-attack c-attack Reputation: 107
Default

Generally, there are two types of problems with installing the cooler.

1) The first is a problem with the coolant flow within the unit. When this is a problem, your CPU temps will appear normal on cold boot, but slowly and consistently rise while sitting idle at the desktop. This is best observed through the H100i v2 Temp reading in the last box on the Link main page. This is the water temperature. Any kind of load will cause a sudden rise in water temp and it will remain elevated even after the load ceases.

2) The other type of common issue is the physical contact between the cold plate on the pump block and the CPU heat spreader. This can be a bad application of Thermal Interface Material (usually erratic or random CPU temp readings), but more often is the result of less than optimal physical contact between the two. Sometimes the pieces don't quite line up right and the fit has to be perfect. The orientation of the backplate is a common cause. With all of the above, your CPU temps will be consistently high or abnormally spiky when presented with low loads, like opening a browser or other program. However, the H100i v2 Temp (water temp) will remain mostly unchanged, even when presented with actual sustained load, as the heat is not being transferred into the water system.


In your screen shot, the water temp is about 34C. Normally, your water temp will idle about 4-6C over your room temperature, although this is case layout and environmentally dependent. So, if this was a rather warm day in the house or the screen shot was taken after a long period of running, then that may be a valid explanation. If it was 23C in the room and you only had the browser open, that is something to look at further.

I also noticed your fans are only turning at 900 rpm with a 34C water temp. I would expect much faster. Did you change the fan speed dependent variable to CPU package temp? If so, put it back on the H100i group (water temp) for now. Are you using a custom fan curve?

Last edited by c-attack; 10-06-2016 at 09:06 AM.
Reply With Quote


  #5  
Old 10-10-2016, 11:08 AM
vhelskud vhelskud is offline
Registered User
vhelskud's PC Specs
 
Join Date: Sep 2016
Posts: 4
POST ID # = 874913
vhelskud Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
Generally, there are two types of problems with installing the cooler.

1) The first is a problem with the coolant flow within the unit. When this is a problem, your CPU temps will appear normal on cold boot, but slowly and consistently rise while sitting idle at the desktop. This is best observed through the H100i v2 Temp reading in the last box on the Link main page. This is the water temperature. Any kind of load will cause a sudden rise in water temp and it will remain elevated even after the load ceases.

2) The other type of common issue is the physical contact between the cold plate on the pump block and the CPU heat spreader. This can be a bad application of Thermal Interface Material (usually erratic or random CPU temp readings), but more often is the result of less than optimal physical contact between the two. Sometimes the pieces don't quite line up right and the fit has to be perfect. The orientation of the backplate is a common cause. With all of the above, your CPU temps will be consistently high or abnormally spiky when presented with low loads, like opening a browser or other program. However, the H100i v2 Temp (water temp) will remain mostly unchanged, even when presented with actual sustained load, as the heat is not being transferred into the water system.


In your screen shot, the water temp is about 34C. Normally, your water temp will idle about 4-6C over your room temperature, although this is case layout and environmentally dependent. So, if this was a rather warm day in the house or the screen shot was taken after a long period of running, then that may be a valid explanation. If it was 23C in the room and you only had the browser open, that is something to look at further.

I also noticed your fans are only turning at 900 rpm with a 34C water temp. I would expect much faster. Did you change the fan speed dependent variable to CPU package temp? If so, put it back on the H100i group (water temp) for now. Are you using a custom fan curve?
I can't remember the room temp and the time I took the screenshot. By the way, I currently work as a writer now and I usually have more than 10 tabs open. I tried to restart my computer and I reopened all the closed tabs all at once. The CPU usage shot up to 100% on all threads and cores, my CPU reached 67c or close to 70c. is that normal? and I noticed that my CPU voltage is sometimes reaching above 1.315v
Reply With Quote


  #6  
Old 10-10-2016, 04:55 PM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 874948
c-attack c-attack Reputation: 107
Default

It's is starting to look like a contact issue. It's fairly normal to have a noticeable CPU spike when opening programs, particularly if using Windows 10 and Google Chrome. However, 70C is out of bounds and larger than expected temperature spikes are usually the result of less than ideal contact between the cold plate and CPU. You don't get abnormally high spikes when there is a problem with the cooler, just consistently rising temperatures the longer you are up and running.

I can't help you with the expected voltage range for your CPU, since I skipper over that series. However, the first place to start is with the backplate and the mounting.
Reply With Quote


  #7  
Old 06-26-2017, 09:52 AM
KitMan KitMan is offline
Registered User
KitMan's PC Specs
 
Join Date: Jun 2017
Posts: 6
POST ID # = 909489
KitMan Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
It's is starting to look like a contact issue. It's fairly normal to have a noticeable CPU spike when opening programs, particularly if using Windows 10 and Google Chrome. However, 70C is out of bounds and larger than expected temperature spikes are usually the result of less than ideal contact between the cold plate and CPU. You don't get abnormally high spikes when there is a problem with the cooler, just consistently rising temperatures the longer you are up and running.

I can't help you with the expected voltage range for your CPU, since I skipper over that series. However, the first place to start is with the backplate and the mounting.
I think I'm having the same problem but I want to confirm it with you

I'm idling at around 40C CPU Package (pc running for 2hrs + after encoding) and like vhelskud was saying, when opening Chrome temps spike to 60C. Is 60C ok? (Water Temp 36.5C)

But my main concern is when I'm encoding, CPU temps go to 85-90C. Is that normal for AIO? The Water temps slowly rise but wouldn't that cool the CPU more then what I'm getting?

For the Backplate and Mounting, will I need to redo the whole thing or can I just tighten the top strews to get better contact?

Last edited by KitMan; 06-26-2017 at 03:44 PM.
Reply With Quote


  #8  
Old 06-27-2017, 05:23 PM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 909668
c-attack c-attack Reputation: 107
Default

You have to be careful with package temperature. Sensor placement is often different for specific CPUs, so it is hard to turn that into a universal value. My guess is your core temps are cooler than this at idle, although voltage, power plan, and C-state options will affect this as well. Can you provide some basic frequency/voltage information?

The Chrome spikes were pretty common at the time of the original post. However, mine have long since gone away with Microsoft, Google, or whoever making the necessary adjustments. I still have some games that spike on launch (higher than their actual load temps), but that is about it. If you have a genuine contact problem, every program will spike on launch. It is helpful if you can see your core activity/frequency when this happens to differentiate between a genuine load event and poor thermal contact. I can't be certain at this point, but if you had a definite physical gap between CPU and cold plate, your render jobs would be untenable within seconds.

Absolutely DO NOT take a screwdriver and crank down as hard as you can. Hand tightening should be enough or if you can really reach or get your hands in there, a one-quarter gentle turn with a screwdriver is plenty. If you have a subtle contact problem, it is likely related to a misaligned backplate, irregular motherboard thickness, or something else more elusive. Tightening down will not help.

It is hard to distinguish based on this information alone. What is particularly helpful is the coolant delta or change in H100i v2 Temp from start to peak on things like a render or other load. Run a mild stress test like Intel XTU for 10 minutes. It has a wave pattern of loading that is easy gauge. Before you begin, note the H100i v2 temp. Take it again after the test has finished. Also, before starting the test, click the little wrench on the line graph. Configure the graph to show all 4 cores individually. This can be helpful in detecting a contact problem. The graph should be relatively smooth, but may be overly spiky if there is a gap.
Reply With Quote


  #9  
Old 06-28-2017, 01:16 PM
KitMan KitMan is offline
Registered User
KitMan's PC Specs
 
Join Date: Jun 2017
Posts: 6
POST ID # = 909779
KitMan Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
You have to be careful with package temperature. Sensor placement is often different for specific CPUs, so it is hard to turn that into a universal value. My guess is your core temps are cooler than this at idle, although voltage, power plan, and C-state options will affect this as well. Can you provide some basic frequency/voltage information?

The Chrome spikes were pretty common at the time of the original post. However, mine have long since gone away with Microsoft, Google, or whoever making the necessary adjustments. I still have some games that spike on launch (higher than their actual load temps), but that is about it. If you have a genuine contact problem, every program will spike on launch. It is helpful if you can see your core activity/frequency when this happens to differentiate between a genuine load event and poor thermal contact. I can't be certain at this point, but if you had a definite physical gap between CPU and cold plate, your render jobs would be untenable within seconds.

Absolutely DO NOT take a screwdriver and crank down as hard as you can. Hand tightening should be enough or if you can really reach or get your hands in there, a one-quarter gentle turn with a screwdriver is plenty. If you have a subtle contact problem, it is likely related to a misaligned backplate, irregular motherboard thickness, or something else more elusive. Tightening down will not help.

It is hard to distinguish based on this information alone. What is particularly helpful is the coolant delta or change in H100i v2 Temp from start to peak on things like a render or other load. Run a mild stress test like Intel XTU for 10 minutes. It has a wave pattern of loading that is easy gauge. Before you begin, note the H100i v2 temp. Take it again after the test has finished. Also, before starting the test, click the little wrench on the line graph. Configure the graph to show all 4 cores individually. This can be helpful in detecting a contact problem. The graph should be relatively smooth, but may be overly spiky if there is a gap.
Thanks for the replay!

For your first paragraph, I've looked into my C stats on my mobo and noticed it was set to Auto. so my mobo was not using them basically. I've since enabled them and see that my cpu cores have gone from ~38C to to ~29C on idle. Frequencies are Base 4Ghz turbo 4.2Ghz so the stock frequencies and Vcore before the C-stats was I think 1.36V. Now it's between 1v to 1.2v(all idle values)

As for Chrome spikes, I'm getting 50% load and 55C spike temps for the cores. Other programs - iTunes no spike, Spotify did spike the same as chrome, firefox spikes but not as bad as chrome.

For the Intel XTU, Start temp 34.5C Finishing 39.5C.(H100i Temps)
Cores start 30C, at load Lowest 52C highest 70C. The temps were spiking between them values but one core was hotter than the others.
Reply With Quote


  #10  
Old 06-28-2017, 06:22 PM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 909824
c-attack c-attack Reputation: 107
Default

Coolant delta under load is more or less as expected and there are no other signs of malfunctioning cooler. Since not all of your programs are CPU temp spiking, I do not think there is a contact problem at this time. What seems most probable is your Skylake when on Full Auto voltage was/is asking for too much Vcore under certain types of load. This has been a common complain since launch. Like the last several generation of 4 core Intel CPUs, you likely want to set a specific voltage in the BIOS, rather than leave it on default Auto. The default setting is too compliant and is designed to maintain maximum stability by piling on voltage to cover even the worst CPU ever made. Nearly everyone can get by with less.

I am sure we have a lot of other 6700K owners who can contribute to this and there are a lot guides out there. You likely can run your Vcore between 1.175 and 1.25v when at the stock 40 multiplier (or really 42 since Asus will set all 4 cores to the turbo value). You do not have to use fixed/manual voltage. You may want to check where your Vcore tops out before undertaking the steps below. If is is already under control, then there is no real need to do this. The 1.36v is often enough to run 4.7 or 4.8GHz, so that is crux of the issue. Everybody has one hot core (at least), so that is not concerning, nor is it a solvable problem.

For Asus, F7 to Advanced BIOS, then drop down to the Vcore line in Extreme Tweaker column. Replace Auto with "Adaptive" from the drop down list. it will create two more boxes below. Enter 1.20 (or whichever value you choose) into Additional Turbo Mode voltage box. Type 'auto' in the offset box. "+" should be the offset direction. You are going to want to test for stability before doing any rendering.

There is also likely some type of Load Line Calibration (LLC) setting in your BIOS under power management or power settings. You can google up a more thorough description, but this controls how much voltage swing and the levels allowed. The AUTO setting is very loose and would allow large changes in voltage that would be visible as temperature spikes. Possibly too loose so for rendering. A more medium value like 4,5, or 6 might tighten things up. I also wonder if it will calm down some of the program launch spikes.
Reply With Quote


  #11  
Old 06-30-2017, 07:50 AM
KitMan KitMan is offline
Registered User
KitMan's PC Specs
 
Join Date: Jun 2017
Posts: 6
POST ID # = 909980
KitMan Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
Coolant delta under load is more or less as expected and there are no other signs of malfunctioning cooler. Since not all of your programs are CPU temp spiking, I do not think there is a contact problem at this time. What seems most probable is your Skylake when on Full Auto voltage was/is asking for too much Vcore under certain types of load. This has been a common complain since launch. Like the last several generation of 4 core Intel CPUs, you likely want to set a specific voltage in the BIOS, rather than leave it on default Auto. The default setting is too compliant and is designed to maintain maximum stability by piling on voltage to cover even the worst CPU ever made. Nearly everyone can get by with less.

I am sure we have a lot of other 6700K owners who can contribute to this and there are a lot guides out there. You likely can run your Vcore between 1.175 and 1.25v when at the stock 40 multiplier (or really 42 since Asus will set all 4 cores to the turbo value). You do not have to use fixed/manual voltage. You may want to check where your Vcore tops out before undertaking the steps below. If is is already under control, then there is no real need to do this. The 1.36v is often enough to run 4.7 or 4.8GHz, so that is crux of the issue. Everybody has one hot core (at least), so that is not concerning, nor is it a solvable problem.

For Asus, F7 to Advanced BIOS, then drop down to the Vcore line in Extreme Tweaker column. Replace Auto with "Adaptive" from the drop down list. it will create two more boxes below. Enter 1.20 (or whichever value you choose) into Additional Turbo Mode voltage box. Type 'auto' in the offset box. "+" should be the offset direction. You are going to want to test for stability before doing any rendering.

There is also likely some type of Load Line Calibration (LLC) setting in your BIOS under power management or power settings. You can google up a more thorough description, but this controls how much voltage swing and the levels allowed. The AUTO setting is very loose and would allow large changes in voltage that would be visible as temperature spikes. Possibly too loose so for rendering. A more medium value like 4,5, or 6 might tighten things up. I also wonder if it will calm down some of the program launch spikes.
Thank you for all the help you're giving me with this!

So I've set the Vcore to adaptive but I couldn't change the Additional Turbo Mode Voltage from Auto. Also changed the LLC to 5.

Now the Vcore is going between 0.9V to 1.2V on idle but still a spike from chrome. Cores to go 50C with a load of 18% and Vcore still goes to 1.36V with that spike. Spotify with load of 45% but levels down the load slower so maybe normal load and not a spike?
Reply With Quote


  #12  
Old 06-30-2017, 08:15 AM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 909985
c-attack c-attack Reputation: 107
Default

I think the temperature spikes are normal for the programs and appreciable for those 1.30+ voltage blips. That is the cause.


Image of Adaptive Setting enabled in BIOS with two drop down boxes.
http://www.tweaktown.com/image.php?i...eview_full.png



If the voltage is still overriding the specified adaptive levels once you have it set right, you might also take a look at the IA load line calibration. In the Kaby lake guide, Asus suggests setting both IA values to 0.01 to prevent the heavy voltage overshoot.
http://www.tweaktown.com/image.php?i...eview_full.png
Reply With Quote


  #13  
Old 07-02-2017, 04:14 PM
KitMan KitMan is offline
Registered User
KitMan's PC Specs
 
Join Date: Jun 2017
Posts: 6
POST ID # = 910194
KitMan Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
I think the temperature spikes are normal for the programs and appreciable for those 1.30+ voltage blips. That is the cause.


Image of Adaptive Setting enabled in BIOS with two drop down boxes.
http://www.tweaktown.com/image.php?i...eview_full.png



If the voltage is still overriding the specified adaptive levels once you have it set right, you might also take a look at the IA load line calibration. In the Kaby lake guide, Asus suggests setting both IA values to 0.01 to prevent the heavy voltage overshoot.
http://www.tweaktown.com/image.php?i...eview_full.png
Still can't change the Additional Turbo Mode.
Attached Images
File Type: jpg BIOS.jpg (1.78 MB, 514 views)
Reply With Quote


  #14  
Old 07-02-2017, 04:24 PM
c-attack c-attack is offline
Registered User
c-attack's PC Specs
 
Join Date: Jun 2014
Posts: 9,225
POST ID # = 910196
c-attack c-attack Reputation: 107
Default

You can't click in the box to the direct right of "Additional Turbo Mode CPU Core Voltage" and type in 1.25? There will be no drop down menu and if you want to return it to Auto, you must type "AUTO" into the box.

"Total Adaptive voltage by CPU" means it is still following the pre-programmed table and that would explain the voltage blips.
Reply With Quote


  #15  
Old 07-03-2017, 01:36 PM
KitMan KitMan is offline
Registered User
KitMan's PC Specs
 
Join Date: Jun 2017
Posts: 6
POST ID # = 910262
KitMan Reputation: 10
Default

Quote:
Originally Posted by c-attack View Post
You can't click in the box to the direct right of "Additional Turbo Mode CPU Core Voltage" and type in 1.25? There will be no drop down menu and if you want to return it to Auto, you must type "AUTO" into the box.

"Total Adaptive voltage by CPU" means it is still following the pre-programmed table and that would explain the voltage blips.
I feel like a right idiot

All set and looks like everything is good. Just doing a 30min XTU test and core temps are look good (50-62C and H100i 41.7C) and Vcore staying at 1.22V (idle 0.9-1.1V).

If I wanted to OC, would I have to change the Vcore much? maybe oc to 4.5Ghz?
Reply With Quote


Reply

Tags
h100i v2

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 03:01 AM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2019, vBulletin Solutions, Inc.