Jump to content
Corsair Community

Why does my H110i GTX suck so bad at cooling?!?! 60c idle 100c load 4790, please help


drnict61

Recommended Posts

The problem: I replaced my broken h100i with a new h110i gtx, and I never realized until now but my temps are always too dang high!

 

Idle: fluctuates wildly from 35c to 60c, sometimes up to 70c

Burn in test/Full load: 100c lmaooooo

 

Clockspeeds: Normal. It's anywhere from 1.somethingGHz to 4.somethingGHz cause turbo boost doodoo is working fine.

Voltages: VCCIN is set to auto. core voltage WAS auto but in a troubleshooting attempt I set it to 1.25v.

 

I haven't been overclocking. I have no magic OC settings turned on in my bios.

 

What I've tried:

  • Resetting the bios to defaults
  • Setting the vcore to 1.25v
  • Using intel's stress tester. It fails cause it gets too hot.
  • Noticed the backplate was indeed incorrectly installed, so I rotated it to not sit on the screws and this made ABSOLUTELY no difference
  • Reinstalled windows
  • Reapplied thermal paste many many times
  • Google'd for days. D.A.Y.S.
  • Verified the temps in my bios
  • Checked all my drivers
  • Checked airflow
  • Checked fans
  • Checked that water is flowing. One tube is warm, one tube is cool.
  • Checked how loose the waterblock is. Ehhhh its pretty snug? Like if I ream on it it wiggles a little but not without moving the whole computer

 

THE BUILD:

  • Intel Core i7-4790K
  • Arctic MX5 thermal paste
  • Corsair H110i GTX cooler
  • MSI Krait Z97s motherboard
  • Windows 10
  • Common sense airflow in a corsair Air 540 case
  • Does anything else matter?

 

All in all, I'm incredibly disappointed. Two of my corsair case fans are dying. My old h100i didnt last 6 months. My corsair ram acts up sometimes. The case has plastic broken off of it. The h110i gtx cooler that replaced the dead h100i doesn't cool for crap. I was a corsair fanboy but I'm about to change that. My ram, case, ssd, keyboardS, fans, and cooler are all corsair and Id like to stay loyal but not for faulty products. Most infuriating is the cooler not cooling.

 

Anyone have ideas? Apologies, I noticed my builds profile was wrong, please refer to the build specs I listed in my post!

Link to comment
Share on other sites

Wildly fluctuating temperatures, particularly at stock settings and a near idle power state, are usually indicative of a contact problem between the cold plate and CPU. Steadier but high idle/load temperatures usually indicate they are not touching fully. Completely bizarre, seemingly random CPU core temps usually means you have too much TIM material and are witnessing the expansion and collapse of air bubbles in the material. I think either way you need to take off the pump block and take a peek. If the TIM never spread, no physical contact. If it is everywhere and spread heavily, even off to the sides of the socket, probably too much.
Link to comment
Share on other sites

Wildly fluctuating temperatures, particularly at stock settings and a near idle power state, are usually indicative of a contact problem between the cold plate and CPU. Steadier but high idle/load temperatures usually indicate they are not touching fully. Completely bizarre, seemingly random CPU core temps usually means you have too much TIM material and are witnessing the expansion and collapse of air bubbles in the material. I think either way you need to take off the pump block and take a peek. If the TIM never spread, no physical contact. If it is everywhere and spread heavily, even off to the sides of the socket, probably too much.

 

Could also indicate a sporadic pump, either not pumping thoroughly or at all intermittently.

Link to comment
Share on other sites

Wildly fluctuating temperatures, particularly at stock settings and a near idle power state, are usually indicative of a contact problem between the cold plate and CPU. Steadier but high idle/load temperatures usually indicate they are not touching fully. Completely bizarre, seemingly random CPU core temps usually means you have too much TIM material and are witnessing the expansion and collapse of air bubbles in the material. I think either way you need to take off the pump block and take a peek. If the TIM never spread, no physical contact. If it is everywhere and spread heavily, even off to the sides of the socket, probably too much.

 

Every time I take the block off to reapply its /perfectly/ spread across without going over the edges or being too thinly spread. The fit is tight, I can hardly wiggle it. The spread indicated that it's on there perfectly. I want to believe this to be the problem but there's nothing supporting it other than wild temps.

 

Could also indicate a sporadic pump, either not pumping thoroughly or at all intermittently.

 

I use corsair link to verify the pump is functioning at 3,000 rpm (which I believe is max) and I can feel the pump always working by putting my hand on the block. Also, one tube is hot water, one tube is cold water, and I can feel a distinct difference

Link to comment
Share on other sites

Every time I take the block off to reapply its /perfectly/ spread across without going over the edges or being too thinly spread. The fit is tight, I can hardly wiggle it. The spread indicated that it's on there perfectly. I want to believe this to be the problem but there's nothing supporting it other than wild temps.

 

 

 

I use corsair link to verify the pump is functioning at 3,000 rpm (which I believe is max) and I can feel the pump always working by putting my hand on the block. Also, one tube is hot water, one tube is cold water, and I can feel a distinct difference

 

Is the Radiator putting out hot air? If your temps are approaching 100C you should feel lots of hot air exiting the Radiator.

 

Unless of course the coupling between the CPU/Block is bad... I assume you removed the plastic protective film over the copper block before install (assuming it even had one)?

Link to comment
Share on other sites

Every time I take the block off to reapply its /perfectly/ spread across without going over the edges or being too thinly spread. The fit is tight, I can hardly wiggle it. The spread indicated that it's on there perfectly. I want to believe this to be the problem but there's nothing supporting it other than wild temps.

 

 

OK, then lets go with that as fact for the moment and approach this from a different angle. If it isn't the contact and the pump is working, then the only thing left is the voltage and the load.

 

You said using "Intel's tool". If you mean Intel Burn Test, stop using that. The name says it all and unless you have your settings dialed in, you are going to hit some sky high temps particularly if you run it with Auto voltage. Try using Intel XTU as a load tool until this is sorted. It is mild and not the most difficult for stability testing, but it useful for this and won't fry your CPU if something is off. It also loads in a smooth wave pattern, so if you start seeing jagged lines all over, we know there is some kind of contact issue.

 

When the CPU core temps are going up and down "at idle", are you sure you are really in a stepped down voltage state? Are you able to see frequency and load with the tool you are using for monitoring? The frequency and voltage may not be in a rest state. 70C still sounds rather high, but I assume this a momentary peak and not a consistent level. Are there any other power state factors that might contribute to higher than expected temperatures? (Windows Power Plan on Performance-100% up time, C-States and EIST disabled).

 

I don't see a specific cooler problem in what we have so far. If there was one, the H110i GTX temp in Link would be in the 40-50C range all the time, although this would be the moment to double check. Since the H110 GTX Temp (coolant temp) is your baseline CPU temp, even a little blip of voltage would take you to 60-70C core temperatures with a 40-50C baseline. It is possible to have a pump that turns, yet still does not move the water through the system. That type of problem does not match the physical characteristics you have described, but I think at this point double checking everything is prudent. Certainly easier than re-mounting the damn pump block again. The coolant temp (H110 GTX Temp) is also helpful in distinguishing various types of heat problems. Some information on its behavior might help (idle H110 GTX Temp, load GTX Temp) Does it rise quickly? Does it go up and not come back down after the load stops? Does is start rising the moment you boot and keep going?. Include your approximate room temperature with this information. That serves as the lowest possible coolant temp and gives it some context.

Link to comment
Share on other sites

OK, then lets go with that as fact for the moment and approach this from a different angle. If it isn't the contact and the pump is working, then the only thing left is the voltage and the load.

 

You said using "Intel's tool". If you mean Intel Burn Test, stop using that. The name says it all and unless you have your settings dialed in, you are going to hit some sky high temps particularly if you run it with Auto voltage. Try using Intel XTU as a load tool until this is sorted. It is mild and not the most difficult for stability testing, but it useful for this and won't fry your CPU if something is off. It also loads in a smooth wave pattern, so if you start seeing jagged lines all over, we know there is some kind of contact issue.

 

When the CPU core temps are going up and down "at idle", are you sure you are really in a stepped down voltage state? Are you able to see frequency and load with the tool you are using for monitoring? The frequency and voltage may not be in a rest state. 70C still sounds rather high, but I assume this a momentary peak and not a consistent level. Are there any other power state factors that might contribute to higher than expected temperatures? (Windows Power Plan on Performance-100% up time, C-States and EIST disabled).

 

I don't see a specific cooler problem in what we have so far. If there was one, the H110i GTX temp in Link would be in the 40-50C range all the time, although this would be the moment to double check. Since the H110 GTX Temp (coolant temp) is your baseline CPU temp, even a little blip of voltage would take you to 60-70C core temperatures with a 40-50C baseline. It is possible to have a pump that turns, yet still does not move the water through the system. That type of problem does not match the physical characteristics you have described, but I think at this point double checking everything is prudent. Certainly easier than re-mounting the damn pump block again. The coolant temp (H110 GTX Temp) is also helpful in distinguishing various types of heat problems. Some information on its behavior might help (idle H110 GTX Temp, load GTX Temp) Does it rise quickly? Does it go up and not come back down after the load stops? Does is start rising the moment you boot and keep going?. Include your approximate room temperature with this information. That serves as the lowest possible coolant temp and gives it some context.

 

I was using cpuid and corsair link to watch temps and voltage. Then I used the burn in tool.

 

I snagged the extreme tuning utility and ran a few tests. First screenshot is while I was running the stress test. Before running the stress test, I changes a setting called "Core adaptive mode" which changed the voltage from a constant 1.25volts (verified in the extreme utility) to the adaptive setting resulting in the voltage being much more dynamic.

 

The second screenshot is a few minutes after stopping the stress test, again noting that I had already switched to the adaptive voltage setting from the constant.

 

Using the intel utility also had seemingly more accurate information? Speaking of which, the H110iGT "Temperature" reported peaked at 74c. I'm assuming that is the block's temperature.

 

While I was stress testing, I took another few feels around the system. The block of the cooler was fairly cold, until I got really close to the cpu. One hose was pretty warn, almost hot, and the other was pretty cold. Feeling the radiator under and over it was all....pretty cold?

 

At this point, is everything expected? Or what's next?

 

Also thank you for the reply, lots of good pointers so far

 

Oh, approximate room temperature is uhhhhHHHHHh something like 70f? Also I JUST noticed its a H110i GT not a H110i GTX (theres a difference :S)

ixtu.thumb.PNG.afb525a4b7e18928164eb45c00ff02f8.PNG

ixtucd.thumb.PNG.0e0e9c467c33af0ab053c2d973846ec4.PNG

Link to comment
Share on other sites

Yes, there is a difference and when you RMA this broken cooler, Corsair will send you a H110i back.

 

74C H110 GT temp in Link? That means the coolant is not moving no matter how loud the pump sounds or feels. Normally you two hot tubes at the bottom and cold air up top, but the number is 30c out of range and evident enough. Essentially you are starting 30c higher than everyone else. Rather than fluctuating, the idle temps likely stay elevated for a prolonged period of time after load and get worse for every moment you are powered on. You need to be careful until a new cooler is in place.

 

I forgot to tell you to select the wrench in XTU and set the graph to show core 0-3. Doesn't matter for now, but that will make it easier for future use. Package temp is a weird variable sometimes but 99c is in shutdown territory.

 

When you submit the RMA info, focus on the very high coolant (h110i GT temp). That is the critical distinguishing factor. Also let them know this is a new unit. The invoice will show that, but it may be to your benefit for shipping. If this is brand new, you might also have faster exchange options through your purchase vendor.

Link to comment
Share on other sites

Yes, there is a difference and when you RMA this broken cooler, Corsair will send you a H110i back.

 

74C H110 GT temp in Link? That means the coolant is not moving no matter how loud the pump sounds or feels. Normally you two hot tubes at the bottom and cold air up top, but the number is 30c out of range and evident enough. Essentially you are starting 30c higher than everyone else. Rather than fluctuating, the idle temps likely stay elevated for a prolonged period of time after load and get worse for every moment you are powered on. You need to be careful until a new cooler is in place.

 

I forgot to tell you to select the wrench in XTU and set the graph to show core 0-3. Doesn't matter for now, but that will make it easier for future use. Package temp is a weird variable sometimes but 99c is in shutdown territory.

 

When you submit the RMA info, focus on the very high coolant (h110i GT temp). That is the critical distinguishing factor. Also let them know this is a new unit. The invoice will show that, but it may be to your benefit for shipping. If this is brand new, you might also have faster exchange options through your purchase vendor.

 

I bought this cooler 5 months ago, replacing the (failing) H100i I had, and saw somewhat lower temps and didnt bother to check up on if they were RIGHT.

 

I've been running like this the whole time and I feel bad! I never realized my temps were BAD until I took a second look and thought "I wonder what it should be at" It's never ran right, BUT again I bought this 5 months ago and I don't have a trace of purchase history, so I'm screwed yet another corsair cooler.

 

That being the case, two corsair coolers failing within their expected life is unacceptable (if we've concluded this one has failed, and in fact never worked right in the first place) and I'll be going with a different brand

Link to comment
Share on other sites

Just a warning. If you do an RMA and then find out your CPU was damaged due to the malfunctioning cooler, you will be stuck as they require the original cooler along with the rest of your system to file and complete a damage claim. I had that cooler completely malfunction and when I troubleshot it, it was pump failure. During the troubleshooting process I turned on the computer to check temp and I thought I was good. But when I did the RMA and was approved for a refund, I bought a new CPU cooler (thermaltake) and when I installed it the computer would only get to the bios for about 20 seconds then freeze. After troubleshooting and everything it was determined that the CPU was damaged. BUT since I already sent in my cooler and they won't be able to find it in their warehouse, I am out the cost of my CPU.
Link to comment
Share on other sites

Archived

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

×
×
  • Create New...