Jump to content
Corsair Community

Is my Corsair H60 a defect?


Androbot10

Recommended Posts

I currently have 16gb of ram running at 3000mhz and a i5 7600k running at 4.2 ghz. I have taken the water block off, cleaned and reapplied thermal paste and even changed the fan at the rear that is on the radiator to be exhaust (recommended by someone from technical support. I do this and my CPU gets to 80 degrees celcius while playing games. The water block is fully attached and the radiator and tubes get kind of hot which causes the motherboard to get hot as well. Is it a defect h60 or am I doing something wrong. I've even let me computer break in the thermal paste. Please help me.
Link to comment
Share on other sites

Since the H60 does have a coolant temperature sensor, we need to use other data to try and figure out what's going on. You will need to be specific to put all of this into context.

 

1) Is the H60 unit functioning properly? To figure this out we need a pump speed (likely around 4000 rpm), do you hear the unit, can you feel it vibrating. My guess is all these will check out, otherwise your gaming sessions would last 2 minutes.

 

When you first boot the PC, wake from sleep, or are otherwise in a normal idle state, what is the CPU core temperature range? What is your approximate room temperature? This is the baseline CPU temperature and provides a comparison point for your load data.

 

 

2) Kaby Lake CPUs tend to run a bit warm out of the box. This can be refined through small BIOS adjustments. Run a mild stress test with no GPU component (like Intel XTU) for 10 minutes. This will remove the GPU waste heat from the equation. You should adjust the settings on the line graph prior to the test to show cores 1-4. Package temp and "CPU temp" have less value that the actual core data. The load is very wave like and should produce a relatively smooth curve of CPU core temperature. Upload the screenshot, if possible or provide the peak values. See how those compare to the average core temps during the test.

 

What was the peak Vcore registered during the test?

 

 

3) Leave the thermal paste and block alone for now. If you made a real mess of this, your idle temps would be crazy and you would hit your temp limits the moment a game launched. If it is not 100% perfect, this may show up in the XTU graph. Either way, you don't need to remount the block right now.

Link to comment
Share on other sites

Since the H60 does have a coolant temperature sensor, we need to use other data to try and figure out what's going on. You will need to be specific to put all of this into context.

 

1) Is the H60 unit functioning properly? To figure this out we need a pump speed (likely around 4000 rpm), do you hear the unit, can you feel it vibrating. My guess is all these will check out, otherwise your gaming sessions would last 2 minutes.

 

When you first boot the PC, wake from sleep, or are otherwise in a normal idle state, what is the CPU core temperature range? What is your approximate room temperature? This is the baseline CPU temperature and provides a comparison point for your load data.

 

 

2) Kaby Lake CPUs tend to run a bit warm out of the box. This can be refined through small BIOS adjustments. Run a mild stress test with no GPU component (like Intel XTU) for 10 minutes. This will remove the GPU waste heat from the equation. You should adjust the settings on the line graph prior to the test to show cores 1-4. Package temp and "CPU temp" have less value that the actual core data. The load is very wave like and should produce a relatively smooth curve of CPU core temperature. Upload the screenshot, if possible or provide the peak values. See how those compare to the average core temps during the test.

 

What was the peak Vcore registered during the test?

 

 

3) Leave the thermal paste and block alone for now. If you made a real mess of this, your idle temps would be crazy and you would hit your temp limits the moment a game launched. If it is not 100% perfect, this may show up in the XTU graph. Either way, you don't need to remount the block right now.

 

While touching the tubes, I feel them vibrating and the pump is moving at 4000+ rpm. When I first boot the pc, it maxes out at like 65 then drops to 40-45 at idle. Then on stress tests with no gpu it teaches like max 80. Then while playing games it can go even further. I think It could be my ASUS rog strictly Gtx 1070 heating up causing the heat to rise so when I make the gpu's fans at 100%, it stays below 80 degrees so like from 70-79 which is still extremely hot. I gave it 1.28 volts at 4.2 ghz for the cpu. It's about 79 degrees farenheight as I live in CA. I even put a fan right next to my computer but not difference. What is a vcore. Not sure what you mean in your 2nd point but while running Aida 64 I got like 78 max. I'll upload screenshots if you still need them.

Link to comment
Share on other sites

Just look at these pictures.

https://drive.google.com/open?id=0B6rQTM_jsys8NmFaNml2YjlkTzg

 

https://drive.google.com/open?id=0B6rQTM_jsys8X0NVVzAzUW9Vb1k

 

Sorry if you have to copy and paste the links but I'll quickly describe. Around 40 Degrees Celcius on Idle. Max 81 degrees Celcius with prime 95 for 10 minutes and lowest was 76 degrees celcius with prime 95. Ambient temp is 27 Degrees Celcius. While running games with GPU (Asus Rog Strix GtX 1070) fans at 100%, cpu gets a max temp of 79 degrees celcius and ranges from like 70-79. Its mounted correctly and I have reapplied thermal paste. Fan and pump is running at full speed. What do I do? CPU Is intel core i5 7600k at 4.2 ghz w 1.28 volts/

Link to comment
Share on other sites

Vcore is your core voltage and the amount supplied to the CPU for processing tasks. The Kaby and Skylake CPUs are somewhat famous for being overly flexible on their base voltage table. If your default value is around 1.29-1.30v (more than you need anyway at 4.2GHz), it may go even higher when pushed on synthetic stress tests. The CPU will do what ever it thinks it needs to remain stable. In this case, pour on a lot of voltage. That then translates into heat or what are often characterized as jumpy, spiky, or just high peak temp readings.

 

It sounds like the H60 is working and the pump is at the right speed. My guess is your temperatures are the result of excess voltage, but you need to find that value. Every monitoring program out there will find Vcore. HWinfo, HWMonitor, MSI's own software all should report this value. Find the normal peak value (likely around 1.30v). You can probably tame it down some by settings a specific adaptive voltage in the BIOS instead of the AUTO value.

 

 

EDIT: I see you started another thread. So 1.28v Vcore. Scroll down to the yellow-green table at the bottom of this page.

 

Compare the top AUTO line and how much voltage it pulls to what Anandtech needed for their overclocking. Huge difference. They ran 1.10v all the way up to 4.4GHz. I tend to be a little conservative when advising other people, but that sounds like plenty for a go at the base 4.2GHz. The trick is to set an adaptive voltage that uses 1.10 as the turbo mode voltage. I need to find a MSI BIOS shot since I am used to Asus boards.

Link to comment
Share on other sites

Vcore is your core voltage and the amount supplied to the CPU for processing tasks. The Kaby and Skylake CPUs are somewhat famous for being overly flexible on their base voltage table. If your default value is around 1.29-1.30v (more than you need anyway at 4.2GHz), it may go even higher when pushed on synthetic stress tests. The CPU will do what ever it thinks it needs to remain stable. In this case, pour on a lot of voltage. That then translates into heat or what are often characterized as jumpy, spiky, or just high peak temp readings.

 

It sounds like the H60 is working and the pump is at the right speed. My guess is your temperatures are the result of excess voltage, but you need to find that value. Every monitoring program out there will find Vcore. HWinfo, HWMonitor, MSI's own software all should report this value. Find the normal peak value (likely around 1.30v). You can probably tame it down some by settings a specific adaptive voltage in the BIOS instead of the AUTO value.

 

 

EDIT: I see you started another thread. So 1.28v Vcore. Scroll down to the yellow-green table at the bottom of this page.

 

Compare the top AUTO line and how much voltage it pulls to what Anandtech needed for their overclocking. Huge difference. They ran 1.10v all the way up to 4.4GHz. I tend to be a little conservative when advising other people, but that sounds like plenty for a go at the base 4.2GHz. The trick is to set an adaptive voltage that uses 1.10 as the turbo mode voltage. I need to find a MSI BIOS shot since I am used to Asus boards.

 

 

Thanks. Imma run the intel stress test that you recommended to me. I will also drop the voltage but you said to pour in a lot of voltage so I dont know what that means.

Link to comment
Share on other sites

Don't worry about the graphs. If you are running 1.28v, that's why your load temps hit those marks. I was looking for a definitive picture to guide you through the BIOS, but I don't have any MSI boards and the BIOS format is not familiar.

 

The page you want looks like this. On my Asus boards, this field is a drop down menu with adaptive and offset and other choices. Those then create additional boxes to input the voltage. MSI may not work the same way, however if you enter 1.10 into the CPU core voltage box, at the very least it will give you a fixed voltage of that same amount and likely a large temp drop. If you are at all squeamish about blue screens or crashing, up the value to 1.15v. That is still a large chunk under what you are currently running.

Link to comment
Share on other sites

Don't worry about the graphs. If you are running 1.28v, that's why your load temps hit those marks. I was looking for a definitive picture to guide you through the BIOS, but I don't have any MSI boards and the BIOS format is not familiar.

 

The page you want looks like this. On my Asus boards, this field is a drop down menu with adaptive and offset and other choices. Those then create additional boxes to input the voltage. MSI may not work the same way, however if you enter 1.10 into the CPU core voltage box, at the very least it will give you a fixed voltage of that same amount and likely a large temp drop. If you are at all squeamish about blue screens or crashing, up the value to 1.15v. That is still a large chunk under what you are currently running.

 

Here is the link to my temps using that intel software. I got max 81 on Core 3. https://drive.google.com/open?id=0B6rQTM_jsys8YmxYZ1J5TFpCbm8

Link to comment
Share on other sites

Here is the link to my temps using that intel software. I got max 81 on Core 3. https://drive.google.com/open?id=0B6rQTM_jsys8YmxYZ1J5TFpCbm8

 

No contact problem. Graph is too smooth. One core hotter than the rest, just like every other CPU out there. I have 6 cores, so I get 2 hot ones. Everything points to higher than needed voltage for your current frequency. If you were running 4.9GHz and has these temps, we wouldn't be having this conversation.

Link to comment
Share on other sites

No contact problem. Graph is too smooth. One core hotter than the rest, just like every other CPU out there. I have 6 cores, so I get 2 hot ones. Everything points to higher than needed voltage for your current frequency. If you were running 4.9GHz and has these temps, we wouldn't be having this conversation.

 

So I went to the bios and changed the volt core mode thingy to adaptive. After doing so, my pc ran into problems booting. It would turn on but nothing would post. My monitor would just say no signal and neither my mouse or keyboard would light up. What is happening?

Link to comment
Share on other sites

What value did you use for voltage? Normally if it is not enough, you would still make it to Windows and see a BSOD while doing something. Does MSI have a LED error code on that board? Assuming it is the voltage that is too low, add +0.05 to that value and try again.
Link to comment
Share on other sites

It's set to 1.125 volts. If I leave the volt core mode to auto t boots up normally. If I switch it to adaptive mode it screws up my pc and I have to use a screwdriver to reset cmos. I think it has enough volts but the adaptive mode is messing it up.
Link to comment
Share on other sites

Since the Auto mode is also adaptive, there may be a secondary issue. Unfortunately, this is probably BIOS configuration related and I don't know your BIOS. I am not even sure you are running adaptive if you input a manual voltage into that box. There is likely another setting that needs to be changed. The heat is from the excessive voltage the Auto table uses, but you are going to need to find a guide for your board to make the proper changes.
Link to comment
Share on other sites

Since the Auto mode is also adaptive, there may be a secondary issue. Unfortunately, this is probably BIOS configuration related and I don't know your BIOS. I am not even sure you are running adaptive if you input a manual voltage into that box. There is likely another setting that needs to be changed. The heat is from the excessive voltage the Auto table uses, but you are going to need to find a guide for your board to make the proper changes.

 

Well I do not think i will. My pc now idles around mid to high 30s and i can play games that peak at 69 degrees Celsius. I think it is because I took down the volts and I will take down some more if possible. The main problem I am having is that I have to turn my gpu fans on because the heat rises from the gpu to the processor making the processor heat up more.

 

Reading the manual there is 6 modes for cpu core voltage mode

- Auto: This setting will be configured automatically by BIOS

-Adaptive Mode: Sets the adaptive voltage automatically for optimizing the system performance

-Override Mode: Allows you to set the voltage manually

-Offset Mode: Allows you to set the offset voltage and select the voltage offset mode

-Adaptive + Offset: Sets the adaptive voltage automatically and allows you to set the offset voltage

-Override + Offset: Allows you to set the voltage and the offset voltage manually

These are the modes that I can use. Any recommendations or tips on all of this?

Link to comment
Share on other sites

https://forum-en.msi.com/index.php?topic=274249.0

 

https://forum-en.msi.com/index.php?topic=268370.0

 

It seems you are not the first person to run into this issue. Based on what I have read so far, it seems MSI does not have the user set adaptive mode like on Asus boards. When set to adaptive, it uses the Auto table -- period. Thus when a manual voltage is entered and this does not match the auto table, the boot lock occurs and your have to clear CMOS.

 

So, in your case you could select Adaptive + Offset. Then set a negative offset of -0.16 to reach the 1.12v target. Now this should work, but negative offsets can cause problems at low voltages (idle) and given the way MSI has set this up, I cannot predict how everything will react. You really need an experienced MSI user who knows the ins and out better than me. There likely is additional information out there about this exact issue.

 

My short term suggestion is to select the "Override" mode. This is MSI speak for manual voltage. Set the 1.12v. To counteract the loss of adaptive mode, make sure you have Intel Speedstep (EIST) turned on and enabled C-states, C1E enabled, and at least a package limit of C0/C1. This should allow frequency to drop at idle. I don't think you should take too much more voltage out under 1.12v. It still needs to support the turbo frequency of 4.2 and the gains in temperature from dropping voltage at the lower end are very small. This is a sharp contrast to the scale at the top end, where knocking off a few hundredths can take 5C off your temps.

Link to comment
Share on other sites

https://forum-en.msi.com/index.php?topic=274249.0

 

https://forum-en.msi.com/index.php?topic=268370.0

 

It seems you are not the first person to run into this issue. Based on what I have read so far, it seems MSI does not have the user set adaptive mode like on Asus boards. When set to adaptive, it uses the Auto table -- period. Thus when a manual voltage is entered and this does not match the auto table, the boot lock occurs and your have to clear CMOS.

 

So, in your case you could select Adaptive + Offset. Then set a negative offset of -0.16 to reach the 1.12v target. Now this should work, but negative offsets can cause problems at low voltages (idle) and given the way MSI has set this up, I cannot predict how everything will react. You really need an experienced MSI user who knows the ins and out better than me. There likely is additional information out there about this exact issue.

 

My short term suggestion is to select the "Override" mode. This is MSI speak for manual voltage. Set the 1.12v. To counteract the loss of adaptive mode, make sure you have Intel Speedstep (EIST) turned on and enabled C-states, C1E enabled, and at least a package limit of C0/C1. This should allow frequency to drop at idle. I don't think you should take too much more voltage out under 1.12v. It still needs to support the turbo frequency of 4.2 and the gains in temperature from dropping voltage at the lower end are very small. This is a sharp contrast to the scale at the top end, where knocking off a few hundredths can take 5C off your temps.

 

I did not get a lot of that. Should I leave it or set it to adaptive + offset. And If so what is my offset value? Should it not be positive.

Link to comment
Share on other sites

No, a postive offset would add voltage to base value. So a +0.10v offset would cause you to run at 1.38v. We want to reduce voltage and it appears MSI makes you do this through a negative offset. So, 1.28v (base) - 0.16 = gets us the desired load 1.12v. However, the problem when at idle, the voltage may also drop -0.16. So when it is stepped down at 0.75v and 1.2GHz resting, it will take -0.16 off that too. That's a problem. Some BIOS versions tell you it will not take it below the table value at minimum. MSI is unclear.

 

For this reason, I suggest you set a manual 'override' setting instead, with the 1.12v. This will get you your temp reductions until someone more experienced with MSI can tell you a better way to set the adaptive without tanking the low end.

Link to comment
Share on other sites

https://forum-en.msi.com/index.php?topic=274249.0

 

https://forum-en.msi.com/index.php?topic=268370.0

 

It seems you are not the first person to run into this issue. Based on what I have read so far, it seems MSI does not have the user set adaptive mode like on Asus boards. When set to adaptive, it uses the Auto table -- period. Thus when a manual voltage is entered and this does not match the auto table, the boot lock occurs and your have to clear CMOS.

 

So, in your case you could select Adaptive + Offset. Then set a negative offset of -0.16 to reach the 1.12v target. Now this should work, but negative offsets can cause problems at low voltages (idle) and given the way MSI has set this up, I cannot predict how everything will react. You really need an experienced MSI user who knows the ins and out better than me. There likely is additional information out there about this exact issue.

 

My short term suggestion is to select the "Override" mode. This is MSI speak for manual voltage. Set the 1.12v. To counteract the loss of adaptive mode, make sure you have Intel Speedstep (EIST) turned on and enabled C-states, C1E enabled, and at least a package limit of C0/C1. This should allow frequency to drop at idle. I don't think you should take too much more voltage out under 1.12v. It still needs to support the turbo frequency of 4.2 and the gains in temperature from dropping voltage at the lower end are very small. This is a sharp contrast to the scale at the top end, where knocking off a few hundredths can take 5C off your temps.

 

did not work when I tried it. I think imma just leave it because my games never break 70 degrees celcius with my gpu running at 70%. My stress test never made my cpu break 65 degree celcius. Unless you have any other tips, thank you for your help. I set the volts at 1.115. But whemn I stress test it runs the cpu at 1.22 volts. idk y.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...