Corsair h100i pro very high liquid temperature

[fryzen]

Hello all!

 

What I've noticed with my AIO cooler the past couple months is that the liquid temps are very high; a bit too high for my comfort. During idle, it stays in the 32-37 C range I believe. The problem is when gaming; I get up to 45 C and the fans are spinning like a plane jet. Is there something wrong with my AIO?

 

This has only happened with my Ryzen 5 3600 processor. I previously had a i5 3570k and whenever I gamed it never went above 35-37 C or so. What could be the problem?

[c-attack]

You most likely are heating up the interior of the case with GPU waste heat bringing the ambient internal temperature to that level. The coolant can't be less than the surrounding environment. Compare your coolant temp before/during gaming to your MB temp sensor, chipset, RAM temps, etc. It is likely each is increasing by the same 12-15C as your coolant. This also will be evident when quitting the game and the coolant temp stays elevated versus dropping to the original level.

 

Compare this to running a CPU only stress test where the coolant temp likely goes up about +6C in 5-10 min and then cools off within 2-3 minutes after the test stops.

 

Besides the normal measures for case heat management, you likely want to take control over the fans. Create your own curve with a more tolerable fan speed for your temperature. The default curves were not designed to accommodate additional heat sources and were based on a standard 20-23C baseline temperature where a +20C increase in coolant temp represents a massive increase (if from the CPU alone).

[fryzen]

You most likely are heating up the interior of the case with GPU waste heat bringing the ambient internal temperature to that level. The coolant can't be less than the surrounding environment. Compare your coolant temp before/during gaming to your MB temp sensor, chipset, RAM temps, etc. It is likely each is increasing by the same 12-15C as your coolant. This also will be evident when quitting the game and the coolant temp stays elevated versus dropping to the original level.

 

Compare this to running a CPU only stress test where the coolant temp likely goes up about +6C in 5-10 min and then cools off within 2-3 minutes after the test stops.

 

Besides the normal measures for case heat management, you likely want to take control over the fans. Create your own curve with a more tolerable fan speed for your temperature. The default curves were not designed to accommodate additional heat sources and were based on a standard 20-23C baseline temperature where a +20C increase in coolant temp represents a massive increase (if from the CPU alone).

 

But why does that only happen with my ryzen cpu? Besides, the mounting bracket is quite bad compared to the intel one, to be honest. The MB temps increase by about 5-10C while the cpu goes from 40-50 to 60-69. and my coolant temp does go down after quitting a game; however, it does take a minute or two.

[DevBiker]

But why does that only happen with my ryzen cpu? Besides, the mounting bracket is quite bad compared to the intel one, to be honest. The MB temps increase by about 5-10C while the cpu goes from 40-50 to 60-69. and my coolant temp does go down after quitting a game; however, it does take a minute or two.

 

Did you have the same GPU? Same fan configuration?

The evidence presented so far seems to indicate that you are pushing waste GPU heat through the radiator, as c-attack said. Let me guess - your radiator is mounted on the top of your case as exhaust, right?

 

Try what c-attack suggested and get back to use with the results.

[c-attack]

But why does that only happen with my ryzen cpu? Besides, the mounting bracket is quite bad compared to the intel one, to be honest. The MB temps increase by about 5-10C while the cpu goes from 40-50 to 60-69. and my coolant temp does go down after quitting a game; however, it does take a minute or two.

 

 

Don't get caught up with CPU temp = 70C puts more heat into the cooler than CPU temp=50C. That's not how it works. The coolant temp rise is going to be a factor of CPU watts (when ignoring other environmental factors). Your Ryzen is probably a good deal bit more efficient than the old Ivy Bridge CPU, but also probably a similar wattage. It would extremely difficult to raise your coolant from 31 to 45C with a 3700x at any overclock level without deliberately sabotaging the results by turning the fans down to near zero. That's obviously not the case here and they are maxed out.

 

If you have a bad CPU/block mount, the CPU temp will be quite high and very erratic, but it also isn't transmitting as much heat as it should into the coolant system. When this happens, the coolant temp hardly changes. This does not seem to be the situation here.

 

It is possible for the cooler to become less efficient because of flow restrictions or other internal problems that are not your fault. However, this would be evident at all times regardless of load type. To that end, try testing the cooler without the GPU load present. If there is a cooler problem, it will show here too. Use something mild like CPU-Z Bench Test. It's load is linear and it makes it easy to see the +1C coolant = +1C CPU temp relationship. You only need to run it 5-10 min. I would expect about +6C at moderate fan speed on a 240mm and your CPU. Anything like +10C or more suggests a problem. Coolant should rise slowly and drop several degrees in the 1-2 min after quitting.

[fryzen]

Don't get caught up with CPU temp = 70C puts more heat into the cooler than CPU temp=50C. That's not how it works. The coolant temp rise is going to be a factor of CPU watts (when ignoring other environmental factors). Your Ryzen is probably a good deal bit more efficient than the old Ivy Bridge CPU, but also probably a similar wattage. It would extremely difficult to raise your coolant from 31 to 45C with a 3700x at any overclock level without deliberately sabotaging the results by turning the fans down to near zero. That's obviously not the case here and they are maxed out.

 

If you have a bad CPU/block mount, the CPU temp will be quite high and very erratic, but it also isn't transmitting as much heat as it should into the coolant system. When this happens, the coolant temp hardly changes. This does not seem to be the situation here.

 

It is possible for the cooler to become less efficient because of flow restrictions or other internal problems that are not your fault. However, this would be evident at all times regardless of load type. To that end, try testing the cooler without the GPU load present. If there is a cooler problem, it will show here too. Use something mild like CPU-Z Bench Test. It's load is linear and it makes it easy to see the +1C coolant = +1C CPU temp relationship. You only need to run it 5-10 min. I would expect about +6C at moderate fan speed on a 240mm and your CPU. Anything like +10C or more suggests a problem. Coolant should rise slowly and drop several degrees in the 1-2 min after quitting.

 

Sorry for not understanding your request earlier, but you are indeed correct. Something like prime95 without GPU stress seems to increase the liquid temp by about 5-6 degrees after 4-5 minutes. I also forgot to say, my GPU is terribly long, and so it barely touches the pumps (I mounted my AIO so that the pumps are on the lower end; Gamers nexus idea). Could that be the case?

 

Did you have the same GPU? Same fan configuration?

The evidence presented so far seems to indicate that you are pushing waste GPU heat through the radiator, as c-attack said. Let me guess - your radiator is mounted on the top of your case as exhaust, right?

 

Try what c-attack suggested and get back to use with the results.

 

No, my case can't handle that, unfortunately. It is front mounted with the fans as intake.

[c-attack]

OK, most people run into this when top mounted as the GPU waste is sucked through the CPU cooling loop. That was what DevBiker was inquiring about. Front mounted is normally more protected.

 

I don't know the Phanteks cases that well. Do you have the 400 with the solid front door? Or the mesh one? If solid door, that could be a restrictive factor in getting air through the radiator and into the case. You can pop the front door/panel open and see if that makes a difference.

 

If your top fans are currently set as exhaust, try flipping them around to intake. This may help make up for the restriction at the front and creates enough pressure to force a lot of GPU waste heat out the back venting, rather than through the rear exhaust alone. This one is very case specific. A miracle in some set-ups. Inconsequential or even worse in others.

 

Regardless of anything else, get yourself on your own custom curve. Performance Tab +. You can use the either of the three shape tools in the upper right corner as a base template. Then slide the dots down to a more reasonable speed at 40C+ mark. If the case is 42C, you can't go under no matter what. There is no reason to run 2000 rpm on the front intake. It will only make you miserable. Coolant temp has a 1=1 relationship with CPU temp, so if it goes up 2C more on H100i temp, don't worry about it. That's only 2C on the CPU and a good trade for noise reduction.

Edited December 20, 2020 by c-attack

[fryzen]

OK, most people run into this when top mounted as the GPU waste is sucked through the CPU cooling loop. That was what DevBiker was inquiring about. Front mounted is normally more protected.

 

I don't know the Phanteks cases that well. Do you have the 400 with the solid front door? Or the mesh one? If solid door, that could be a restrictive factor in getting air through the radiator and into the case. You can pop the front door/panel open and see if that makes a difference.

 

If your top fans are currently set as exhaust, try flipping them around to intake. This may help make up for the restriction at the front and creates enough pressure to force a lot of GPU waste heat out the back venting, rather than through the rear exhaust alone. This one is very case specific. A miracle in some set-ups. Inconsequential or even worse in others.

 

Regardless of anything else, get yourself on your own custom curve. Performance Tab +. You can use the either of the three shape tools in the upper right corner as a base template. Then slide the dots down to a more reasonable speed at 40C+ mark. If the case is 42C, you can't go under no matter what. There is no reason to run 2000 rpm on the front intake. It will only make you miserable. Coolant temp has a 1=1 relationship with CPU temp, so if it goes up 2C more on H100i temp, don't worry about it. That's only 2C on the CPU and a good trade for noise reduction.

 

I do have the solid one and it is definitively a big factor! I do think that I can get the mesh one as an accessory. What do you think, should I? It costs about 30€ with shipping.

 

Furthermore, I did try to flip my top fan to intake, but to no avail...

[c-attack]

Tough to say on the front piece. It will let more air through. It should help with temps. It will be louder without the panel blocking the fans. That's how it goes. More open space, more energy released including noise. Same thing with noise suppression material. Blocks noise, but retains heat as well.

[fryzen]

Tough to say on the front piece. It will let more air through. It should help with temps. It will be louder without the panel blocking the fans. That's how it goes. More open space, more energy released including noise. Same thing with noise suppression material. Blocks noise, but retains heat as well.

 

Trust me, that noise will probably be 10x quieter than what I'm going through. Having my fans spin at 2500 RPM (changed now with custom) is like being near a plane during takeoff...

[DevBiker]

Hmmmm ... this is unusual.

 

Front mount with the fans as intake is usually isolated from the GPU waste heat issue. Yet ... I think that the evidence clearly points to this being the root cause of the issue here. We very often see this with a top mount as exhaust.

 

I wonder if perhaps you need to do some work on optimizing the overall airflow in the case. Here's a hypothesis to consider - the GPU is producing heat but that heat isn't getting out of the case. Instead, it is building up inside the case. This, then, is causing the air around the radiator to heat up and is increasing the coolant temperature. We see similar issues when we have restricted exhaust airflow (it's a heat column effect). Maybe look at the control variable for your exhaust fans to try to pull the waste GPU heat out of the rear of the case. Do you have any way to measure the interior case temperature? Does your motherboard support additional temp sensors (some do)? Do you have a Commander Pro (they have 4 temp sensors for these kinds of things)?

[LeDoyen]

If there's not enough fresh air intake, the front fans just recirculate hot inside air. they effectively heat up your AIO more than cool it :P

 

You can do a simple test and pull the front panel out. it pops out by pulling carefully on the bottom air intake

 

Just look at how your temperatures go once you remove it.

If it's drastically better, well, the mesh front will most likely solve your issues.