High Temps for i7 4790k & H80i?

[itzhalo]

Hello Corsair Forums. I just recently purchased an i7 for the holidays and decided to use my H80i that I have had for almost half a year by now. I hooked it up and noticed some of my temps.

 

Before using washers:

Idle: 42-52c

Load: Up to 80c

Average when gaming: 65c and sometimes reaches 70c most of the time.

 

Using washers:

Idle: 35-40

Load: 68c

Average when gaming: 48-58c, depending on game could sometimes reach around 67c but doesn't stay for long at all. Never has touched 70c at all yet and hopefully will stay that way.

 

I have noticed that my temperatures do spike. It will sometimes jump from 35c to 44c and then come back down right away. I just was wondering, are these temps alright for this cooler? I am on the quiet setting. I have tried remounting numerous times with varying amounts of thermal paste to test if I wasn't using too much or too little and the temps were pretty much the same. I have everything in bios set to default except XMP enabled, turned windows power mode to balanced, and enabled C1E support and C-State. I am using the fans as an exhaust (Fan>Radiato>Fan). According to corsair link the pump is working (2268 RPM).

 

**I made sure that the backplate has aligned with the notches as well! :cool:

 

HWMonitor during simply using chrome with fans mounted as exhaust (Fan>Radiator>Fan):

http://s26.postimg.org/kadn72d89/screenshot_37.png

 

 

If you guys need more info I can post more! :cool:

 

Specs:

i7 4790k Stock

MSI Z97 PC Mate Motherboard

Corsair H80i Cooler

16GB 1600mhz G.Skilll Ripjaws

Corsair CX750m PSU

EVGA SC GTX 970

[itzhalo]

Good news, temps are still staying below 70c :cool: yet idle temps are still 34-38 and it fluctuates a lot. Should those temps be a little lower?

[Akisj]

Intel i7 4790k here too and MSI Z97 Gaming 7 mobo,Thermaltake GT10 using H80i GT

 

Idle temps:

http://s12.postimg.org/vlybhxhct/Capture.png

image hosting without registration

 

 

You definately can go lower.

[itzhalo]

Intel i7 4790k here too and MSI Z97 Gaming 7 mobo,Thermaltake GT10 using H80i GT

 

Idle temps:

http://s12.postimg.org/vlybhxhct/Capture.png

image hosting without registration

 

 

You definately can go lower.

 

Thanks for the reply!

 

Exactly what I was thinking. I have tried almost everything:

-Remounting and applying other thermal paste

-Using different power settings (C1E, C-state, balanced power option setting in windows).

-Tried turning off turbo mode

-Tried using both intake and exhaust on the fans (only around 3c difference)

-Monitored temps with the washers on the back plate of the motherboard and off

-Using optimized defaults in bios

 

This cooler worked perfectly on my FX 8350 (Originally got it for that) so idk why it wouldn't be able to cool the 4790k just fine. Almost anything I do makes the temps spike in the high 40's. Overclocking is just going to have to wait till I can get I different cooler or go back to my old 212 evo :confused:

[Akisj]

Well at summer i do get 70-75C at full load.Here in Greece we get some really high temps but the case helps a lot GT10 uses 3x200mm fans and i also installed another 120mm fan at the bottom hitting motherboard chipset and graphic card.When i first installed the 80i and get the high temps i did whats you did start changing thermal paste,fans everything.Its all about room temperature and case.

[c-attack]

I know you have been really thorough in your efforts, but the elevated base line temperature and sudden temperature jumps are classic signs the block is not making full contact with the chip lid. If all 4 cores are jumping in unison, you can be pretty sure. Bizarre background OS tasks might use one core. When you have made a mess of the TIM, various cores jump all over the place, but rather than spikes, it's more like a random temp generating machine.

 

I am going to assume all the AMD hardware bits were swapped for the Intel parts and focus on the washers and backplate. You are not supposed to need extra washers. There are some boards out there (mostly Asus) that have caused some issues, but I don't remember MSI being one of them and Z97 usually isn't the culprit. If it's possible with your case, press down on the block with two fingers to see if your idle temps drop. It may not work if the posts are physically restricting the block because of hose tension and the angle. Your actual achievable idle temp is going to depend on room temperature, but it should be within room temp + (7 to 9C). There is a guide up in the stickies section on the Cooling main page. You may have already been through it, but if not, see if something looks out of place.

[itzhalo]

I know you have been really thorough in your efforts, but the elevated base line temperature and sudden temperature jumps are classic signs the block is not making full contact with the chip lid. If all 4 cores are jumping in unison, you can be pretty sure. Bizarre background OS tasks might use one core. When you have made a mess of the TIM, various cores jump all over the place, but rather than spikes, it's more like a random temp generating machine.

 

I am going to assume all the AMD hardware bits were swapped for the Intel parts and focus on the washers and backplate. You are not supposed to need extra washers. There are some boards out there (mostly Asus) that have caused some issues, but I don't remember MSI being one of them and Z97 usually isn't the culprit. If it's possible with your case, press down on the block with two fingers to see if your idle temps drop. It may not work if the posts are physically restricting the block because of hose tension and the angle. Your actual achievable idle temp is going to depend on room temperature, but it should be within room temp + (7 to 9C). There is a guide up in the stickies section on the Cooling main page. You may have already been through it, but if not, see if something looks out of place.

 

Sure thing. I'll give the sticky a quick look as well. I'll give it one last go without the washers and everything as well but I tried pressing on the water block to see if my temps would go down a bit but no luck, guess it is really restricted. I'll post back with everything when I am done.

[itzhalo]

Alright I just give up. Temps are no different without the washers and seem actually a bit worse. I'll just have to get a different cooler because I was hoping to do some overclocking. Thanks for the help everyone.

 

Update: Put the washers back on and here is what I mean by the temp going all over the place:

 

 

Besides that I am completely out of ideas for this cooler.

[XT440]

Mount the CPU block with the Corsair logo either read normally or up and down NOT sideways. This happened to me. After I changed it idle temp dropped with 10°C. This because mounting the CPU block does not make full contact with the CPU if Corsair logo read sideways. Skip the washers.

 

I hope this will help. Good luck.

[itzhalo]

Mount the CPU block with the Corsair logo either read normally or up and down NOT sideways. This happened to me. After I changed it idle temp dropped with 10°C. This because mounting the CPU block does not make full contact with the CPU if Corsair logo read sideways. Skip the washers.

 

I hope this will help. Good luck.

 

Wish the issue was a simple fix like that. Every time I have mounted the water block on the CPU it has always been right side up :/

[XT440]

Wish the issue was a simple fix like that. Every time I have mounted the water block on the CPU it has always been right side up :/

 

Ok. Check the condensers around the CPU socket area and see if there are any scratches on them. If yes they probably are in the way for the CPU block to make full contact with the CPU when mounted. On my motherboard I can only install the CPU block right side up or the condensers getting in the way. Even if I mount it correct it's still very tight between the condensers and the CPU block.

 

You can also try this. Apply thermal paste big as a rice corn, mount the CPU block and tighten the screws as usuall and unmount. Look at the CPU and CPU block and see if the thermal paste has covered it all. If there are spots the CPU block does not make full contact.

[itzhalo]

Ok. Check the condensers around the CPU socket area and see if there are any scratches on them. If yes they probably are in the way for the CPU block to make full contact with the CPU when mounted. On my motherboard I can only install the CPU block right side up or the condensers getting in the way. Even if I mount it correct it still very tight between the condensers and the CPU block.

 

Yeah just did. The only ones that I thought would be in the way are these:

http://s26.postimg.org/t0fblnk3d/1111455.jpg

 

Yet before I installed the second fan on the other side of the radiator nothing was in the way of it.

[XT440]

Mine are close as yours. I had the same issue as you and the solution was to mount it the right side up. Have you tride to mount it other ways? I am sure something is getting in they way. Have you core enhancement enabled in BIOS? FAN RPM?

[itzhalo]

Mine are close as yours. I had the same issue as you and the solution was to mount it the right side up. Have you tride to mount it other ways? I am sure something is getting in they way. Have you core enhancement enabled in BIOS? FAN RPM?

 

Just took this pic, I have always mounted it like this:

http://s26.postimg.org/5q5r3px15/20151230_220014.jpg

 

As for core enhancement I do have that enabled. Here is a picture of my settings in corsair link, I use the quiet setting:

 

http://s26.postimg.org/shetq4i2h/screenshot_50.png

 

only 4% load on the CPU and at 38c :(:

 

Also here is what I have under the "CPU Features" setting:

 

http://s26.postimg.org/mj10fvz3t/20151231_012432.jpg

 

 

Yet I would like to ask, what motherboard do you have in your build? So if I am correct mine is mounted like yours is? I just don't know if changing the way it is mounted will change anything considering it is equal on the left and right side. I moved over my ram so the pump wouldn't be right up against it but besides that I have no clue what to do...

[itzhalo]

Update: Just purchased some MX-4 online. Will apply and see if temps lower at all. :cool:

[c-attack]

Unfortunately, a TIM change might yield 1-3C at best and is not the issue here.

 

In your LINK screen shot, your water temperature (H80i Temp) is at 32C. Was this after any kind of test run or load activity? What was the room temperature (roughly)? Is it now Summer or Winter for you?

 

Going back to the temperature spikes... Is it all four cores that jump? Or just one? You might be able to tell in HW Monitor. The free versions of AIDA64 or OCCT run line graphs which make it very easy to tell.

[itzhalo]

Unfortunately, a TIM change might yield 1-3C at best and is not the issue here.

 

In your LINK screen shot, your water temperature (H80i Temp) is at 32C. Was this after any kind of test run or load activity? What was the room temperature (roughly)? Is it now Summer or Winter for you?

 

Going back to the temperature spikes... Is it all four cores that jump? Or just one? You might be able to tell in HW Monitor. The free versions of AIDA64 or OCCT run line graphs which make it very easy to tell.

 

The screenshot I took was after a while of gaming and rendering that I have done for the day. As of now, I have played some games and rendered 2 videos and my H80i temp currently is 35c which I am guessing is pretty hot. Its been almost 25 mins after I am done rendering and it is still around that temp as well.

Edit: To be quite honest, I believe my water temps are always around 32-35c. I monitored it after 30 mins of just browsing the internet and it dropped to only 34.5 :confused: :confused:

 

My usual room temp averages between 22-23c (pretty warm) and it is winter but we have had temps sometimes around 15c outside where I live.

 

Here is what I gathered from MSI afterburners graphs:

http://s26.postimg.org/5nqgumird/screenshot_54.png

Reached a max of 66 throughout the whole day of me using it (Roughly 6hrs). If you want info from another software like the ones you posted above I will gladly post them. Also the thermal compound I have been using is a bit questionable. I have used this compound that I purchased at my local bestbuy (http://www.bestbuy.com/site/insignia-thermal-compound/3641012.p?id=1219092666136&skuId=3641012) and it really isn't a "name brand". I just figured it was getting old because its been a almost a year and a half since I used it (last used on my old 8350). If it doesn't make much of a difference I just want to assure that I have a good thermal compound on my CPU.

[c-attack]

Actually, I think 35C water temp after some GPU load is pretty reasonable for a rear mounted cooler. With that rear mount above the GPU, the radiator is always going to absorb some of the waste heat being flushed out of the case. The water temp can be very slow to cool, particularly when it's in the main exhaust path. It's not that the system isn't working, but it is very likely all the air around it is also around 32C so you won't see a large temp drop after the load stops without doing something drastic like blasting the fans and taking the door off. Both of those steps are unnecessary. Your drive and motherboard temps also suggest the average case temperature is in this range. All of this appears to indicate the internal cooler components are working properly.

 

As for the core temp graph, I don't see anything remarkable. All four cores certainly appear to be in step with each other. If you were gaming or rendering during this time, I would say great, your cores are working together for maximum performance. If you were on the desktop that whole time typing this out and slapping the pictures together, then I am not so sure. My earlier comment about the cores jumping in unison was meant for low level desktop stuff like opening a browser or some other application that should only require 1 core to complete. When you see all 4 cores jump doing something simple like that, it usually indicates a contact problem. Also, make sure you didn't unintentionally leave the Windows power state on the performance setting with 100% up time. I've done that several time this week alone.

 

I can see you have enabled most of the power saving features and I am assuming your are at the stock speed (4.0/4.4). It still feels like this is too warm. The running 38C average is about what I would expect on a high overclock with manual fixed voltage. I suppose one more way to check is to get the core temp monitor up as soon as you boot. In a cold state, your water temp should be the same as the room temperature and the cores just above it. It the cores go straight into the 30's with power saving features and automatic (adaptive) voltage on, it would suggest there still is a contact problem. Unfortunately, I am not sure what else to suggest as a fix. It is something of a hands on problem and you've already done the research on the most common issues.

[itzhalo]

Actually, I think 35C water temp after some GPU load is pretty reasonable for a rear mounted cooler. With that rear mount above the GPU, the radiator is always going to absorb some of the waste heat being flushed out of the case. The water temp can be very slow to cool, particularly when it's in the main exhaust path. It's not the system isn't working, but the it is very likely all the air around it is also around 32C so you won't see a large temp drop after the load stops without doing something drastic like blasting the fans and taking the door off. Both of those steps are unnecessary. Your drive and motherboard temps also suggest the average case temperature is in this range. All of this appears to indicate the internal cooler components are working properly.

 

As for the core temp graph, I don't see anything remarkable. All four cores certainly appear to be in step with each other. If you were gaming or rendering during this time, I would say great, your cores are working together for maximum performance. If you were on the desktop that whole time typing this out and slapping the pictures together, then I am not so sure. My earlier comment about the cores jumping in unison was meant for low level desktop stuff like opening a browser or some other application that should only require 1 core to complete. When you see all 4 cores jump doing something simple like that, it usually indicates a contact problem. Also, make sure you didn't unintentionally leave the Windows power state on the performance setting with 100% up time. I've done that several time this week alone.

 

I can see you have enabled most of the power saving features and I am assuming your are at the stock speed (4.0/4.4). It still feels like this is too warm. The running 38C average is about what I would expect on a high overclock with manual fixed voltage. I suppose one more way to check is to get the core temp monitor up as soon as you boot. In a cold state, your water temp should be the same as the room temperature and the cores just above it. It the cores go straight into the 30's with power saving features and automatic (adaptive) voltage on, it would suggest there still is a contact problem. Unfortunately, I am not sure what else to suggest as a fix. It is something of a hands on problem and you've already done the research on the most common issues.

 

Thanks for the info. Should I be on the balanced power plan all along? Also I will post about the temps with the better TIM.

[c-attack]

For most people the balanced profile is appropriate. You can fiddle with all the Link state, HDD, media, etc. settings all you want, but it puts the CPU in a 5-100% operational state. In the high performance plan, your minimum CPU activity level is 100%. This is counterbalanced by any enabled EIST and/or C-state power savings measures and looks like the graph below. I unintentionally left it in high performance mode (again) after running some Pi150 benchmarks. The graph is me shopping on Amazon. You can see the normal core activity at the end when I put it back in balanced mode. One keeps it up and then drops it when it doesn't sense a load. The other keeps it down and raises it when it does sense a load. Obviously one is more energy efficient and produces less heat. High performance mode can be useful for some older games or programs that for whatever reason try to confine themselves to one core all the time and of course it helps for optimal (and consistent) benchmarks. However, I only mentioned it because it can raise your idle core temps by several degrees and throw off any meaningful comparison to others.

 

http://www.mediafire.com/view/xqlibirclilcn7h/Performance_Mode_and_C-states_enabled.png

[itzhalo]

For most people the balanced profile is appropriate. You can fiddle with all the Link state, HDD, media, etc. settings all you want, but it puts the CPU in a 5-100% operational state. In the high performance plan, your minimum CPU activity level is 100%. This is counterbalanced by any enabled EIST and/or C-state power savings measures and looks like the graph below. I unintentionally left it in high performance mode (again) after running some Pi150 benchmarks. The graph is me shopping on Amazon. You can see the normal core activity at the end when I put it back in balanced mode. One keeps it up and then drops it when it doesn't sense a load. The other keeps it down and raises it when it does sense a load. Obviously one is more energy efficient and produces less heat. High performance mode can be useful for some older games or programs that for whatever reason try to confine themselves to one core all the time and of course it helps for optimal (and consistent) benchmarks. However, I only mentioned it because it can raise your idle core temps by several degrees and throw off any meaningful comparison to others.

 

http://www.mediafire.com/view/xqlibirclilcn7h/Performance_Mode_and_C-states_enabled.png

 

Ah alright, so I should put the power plan to high performance but keep the minimum CPU level to 5%?

 

Just booted my PC on after a good 7hrs of it being off, here are the graphs and the water temp form when it first booted. I would also like to add I haven't done any gaming or rendering to make it warm up:

http://s26.postimg.org/6syb589dl/screenshot_55.png

 

Here is another:

[c-attack]

There's no need to change it. Use the balanced plan for everyday stuff and you can load the High Performance for renders or gaming if you want it. It may or may not shave some time off the run. It will create a little more CPU and VRM heat.

 

The 28C water would be OK if were about 25C in your room. However, given the 21C drive temp, perhaps not, although the motherboard temps are all in line with the water. Those mobo sensors can be very unreliable. Drive sensors are a little more consistent. Is the HDD in the bottom of the case behind the front intake fans? The low CPU temp in the graph does just hang above the water temperature. This is normal behavior and it mostly looks in line. I don't really see the vertical spikes that usually come with intermittent plate contact.

 

Are there any other environmental factors that might contribute to higher than expected case temperatures? Is the PC under a desk or wedged between the desk and wall? Is the rear of case pushed against a wall? From the photo, it looks like standard top and rear exhaust. I assume there is a fan or two up front bringing air into the case. Your specs have a single 970 listed and I have a hard time believing it is responsible for the case temps at cold boot, especially when it's bottomed out at 35C.

[itzhalo]

There's no need to change it. Use the balanced plan for everyday stuff and you can load the High Performance for renders or gaming if you want it. It may or may not shave some time off the run. It will create a little more CPU and VRM heat.

 

The 28C water would be OK if were about 25C in your room. However, given the 21C drive temp, perhaps not, although the motherboard temps are all in line with the water. Those mobo sensors can be very unreliable. Drive sensors are a little more consistent. Is the HDD in the bottom of the case behind the front intake fans? The low CPU temp in the graph does just hang above the water temperature. This is normal behavior and it mostly looks in line. I don't really see the vertical spikes that usually come with intermittent plate contact.

 

Are there any other environmental factors that might contribute to higher than expected case temperatures? Is the PC under a desk or wedged between the desk and wall? Is the rear of case pushed against a wall? From the photo, it looks like standard top and rear exhaust. I assume there is a fan or two up front bringing air into the case. Your specs have a single 970 listed and I have a hard time believing it is responsible for the case temps at cold boot, especially when it's bottomed out at 35C.

 

You're a wizard. My computer is actually on the floor and against a wall I guess you can say. I moved it down there to save some space because my mousepad is huge (steelseries QCK+). My room temp varies a lot considering the room I am in holds heat very well for some reason and it can be well around 25c (or even higher) for the most part. How my PC is setup is I have 2 intake fans in the front, an exhaust in the back (Using H80i as exhaust), and since I kinda decided "hot air rises" I have two in the top exhausting air (kind of a stupid decision since I could have put a H100 up there instead). The reason why I have the H80i as exhaust was because I wanted to move air across the case and when I had it as intake there was only a 2-3c difference and it made my other components slowly rise like my HDD temp was running almost 40c at the end of the day. Besides me telling you my life story, I guess I will see how the temps will be once I apply this MX-4. If they are still around the same I will see what I can do with moving it back up to the top and sparing some room! :p:

 

As for the 970 goes, I use the MSI auto adjust fan feature so I can idle at around those temps constant. Only have ever reached a high of around 60c gaming on it appose to not using it I would idle at around 46c+ and reach almost 70c :cool:

[c-attack]

It is starting to look like this is environmental and not an installation issue. You are always going to run a little warmer with the H80i as exhaust in the rear slot, but that doesn't mean that isn't the best placement for you. As you already discovered, installing to the front as intake doesn't get you much more on the CPU temps, but does have a negative effect on the rest of the system. Since the heat is additive to everything else, I would predict even worse internal temperatures now that it is Summer if you front mount. One possible other mounting location might be as exhaust directly behind the 5.25 drive bays in the roof. This is assuming you have a traditional tower. It moves the H80i out of the GPU waste heat area, lets you run the rear exhaust fan at whatever speed you need, and still sends CPU waste heat out. The drawbacks are it might be noisier if you are above the case (H80i fans), the mount is difficult and may not be possible depending upon your case/board/RAM height, and if you have a windowed case you may not like the look. However, not all is lost. You seem to be staying 13-15C below the magic 80C threshold most people avoid and are doing so in a less than optimal environment. This is the worst case scenario and you are still fine. You probably can add a small overclock and still be OK. I ran an H80i in the rear slot for years on an i930 that was pushed right to voltage limit. I ran warm for games, but I never even came close to a thermal shutdown. This is a workable set-up.

 

You don't have to compromise your seating arrangement to manage this. The desk and the wall aren't going to move, but if you can leave a little space in back between the case and the wall that may help with heat soak during game and renders. You probably have a fan or two in the house. You can put a larger model behind you or even dedicate a small desktop fan for the job. You are not going to be actively cooling anything with this, but even at low speed a little 4 inch desk fan placed in front of the case blowing past it may be able to disrupt the heat pocket forming around and behind the case. You don't need a torrent of air, just enough to displace the air around the case. Another thing that might help (and certainly with dust) is to get the case a little off the ground. Even 4 or 5 inches on some books or box creates more air space around it and should help with the dust coming in the back venting.

 

Don't worry about the 970. I just wanted to make sure you didn't upgrade to a 980Ti and hadn't updated your specs. It would account for all of this on it's own. If you do make a GPU update to heavier card, you might need to go to a roof mounted system.

[itzhalo]

It is starting to look like this is environmental and not an installation issue. You are always going to run a little warmer with the H80i as exhaust in the rear slot, but that doesn't mean that isn't the best placement for you. As you already discovered, installing to the front as intake doesn't get you much more on the CPU temps, but does have a negative effect on the rest of the system. Since the heat is additive to everything else, I would predict even worse internal temperatures now that it is Summer if you front mount. One possible other mounting location might be as exhaust directly behind the 5.25 drive bays in the roof. This is assuming you have a traditional tower. It moves the H80i out of the GPU waste heat area, lets you run the rear exhaust fan at whatever speed you need, and still sends CPU waste heat out. The drawbacks are it might be noisier if you are above the case (H80i fans), the mount is difficult and may not be possible depending upon your case/board/RAM height, and if you have a windowed case you may not like the look. However, not all is lost. You seem to be staying 13-15C below the magic 80C threshold most people avoid and are doing so in a less than optimal environment. This is the worst case scenario and you are still fine. You probably can add a small overclock and still be OK. I ran an H80i in the rear slot for years on an i930 that was pushed right to voltage limit. I ran warm for games, but I never even came close to a thermal shutdown. This is a workable set-up.

 

You don't have to compromise your seating arrangement to manage this. The desk and the wall aren't going to move, but if you can leave a little space in back between the case and the wall that may help with heat soak during game and renders. You probably have a fan or two in the house. You can put a larger model behind you or even dedicate a small desktop fan for the job. You are not going to be actively cooling anything with this, but even at low speed a little 4 inch desk fan placed in front of the case blowing past it may be able to disrupt the heat pocket forming around and behind the case. You don't need a torrent of air, just enough to displace the air around the case. Another thing that might help (and certainly with dust) is to get the case a little off the ground. Even 4 or 5 inches on some books or box creates more air space around it and should help with the dust coming in the back venting.

 

Don't worry about the 970. I just wanted to make sure you didn't upgrade to a 980Ti and hadn't updated your specs. It would account for all of this on it's own. If you do make a GPU update to heavier card, you might need to go to a roof mounted system.

 

Thanks for all the info. If all else fails after the new TIM I will use most of your suggestions! Should be coming in tomorrow or the day after. Will post to see if temps have improved at all or better yet if the cores stop jumping like they usually do.