Jump to content
Corsair Community

is my h115i faulty or is it just how liquid cooling works?


max_one

Recommended Posts

Hello,

I have noticed a strange behaviour with my newly acquired Corsair H115 RGB platinum and I thought about asking directly for the community's advice since I am new to liquid cooling.

 

Basically, what happened is that I tried to run a small stress test on my CPU with a software called AIDA64 and as soon as the load on the cup reached 100% various softwares (core temp, AIDA 64 itself, etc....) warned me about the CPU temp (cores and packages) skyrocketing to 70°C.

All of this happened while the H115I radiator's fans kept their initial RPM (around 600 i think). After five minutes of stress test running, given that the CPU still had 70°C or more and the radiators fan were not reacting, I stopped the test fearing damage to the CPU.

 

After doing some research and one more short test I've noticed that the radiator fans are programmed in iCUE to react in according to the temperature of the coolant.

While I agree that for a liquid cooler reacting to the coolant's temperature is the most appropriate behaviour, what concerns me is that coolant itself gained just a couple of degrees (from 28 to 30) while the CPU was way over 70°C.

Is it normal? Or maybe is my unit faulty? I don't know much about liquid cooling but I do remember that when I was using air cooling I never reached those hight temps on the CPU, I am worried that any demanding and stressful task I might do in the future could possibly damage my CPU because the coolant does not react fast enough to the CPU temperature changes.

Is that how things work with all liquid coolers or do I have a faulty unit?

Are there some test I can do to make sure everything works correctly?

 

Thanks

Link to comment
Share on other sites

Yes, this is normal. The liquid has a higher heat capacity (specific heat) than the CPU. So it takes more energy to warm it up.

And 70C is fine. Even 80C is OK. 90C is where you start throttling. And the CPU will throttle and shut down before you damage it.

Also, do a search for some of c-attacks posts on this. He explains is very, very well.

Link to comment
Share on other sites

That's how cooling works...

 

Two parts to the cooling process. Stage 1 is conductive. Pins on the motherboard supply voltage that passes into the CPU on demand. This creates heat both at the pins and inside the CPU. The only way to move the heat elsewhere is through physical conduction. That through the CPU lid, TIM, and then the cold plate of whatever CPU cooler type you have bolted on. There are no settings for this that affect thermal transfer and it is a relatively fixed rate based on the physical properties of the substances involved. Decreasing or increasing the voltage is the only thing under your control. Delidding the CPU is a method that addresses cooling on this end.

 

Stage 2 is what you do with the heat once you get it away from the CPU. Whether it's an $10 blower, an air tower, or massive water cooling system, you take the heat and you move it somewhere else. With small air coolers, if you don't move the heat away relatively quickly, the heat builds up and transfers back across to where it just came from and the CPU temp increases further. The big advantage water coolers have is they can hold more heat at one time with a lower penalty to the CPU and you effectively can move it to the location of your choice - directly to an case outlet, etc. However, all of them are the same -- if you don't move the heat elsewhere, there will be a further increase in CPU temperature.

 

The oddity most people new to water cooling run into is they can see the second stage of the cooling for the first time. With an air tower, you have no idea how much heat is there, how efficiently it is being conducted, or much else. All you see is end CPU temperature. The mistaken assumption is that all of the CPU's heat can be transferred through the cold plate and into the water system. Sorry - that would work out great, but it is not possible with any materials on this planet. The limitation is always going to be at the CPU socket where the heat is created. On some motherboards, this is your 'package temperature' although the nomenclature is not always consistent. You can't stop the pins from heating up the CPU as the voltage is carried through. That part is inevitable and why extreme overclockers resort to liquid nitrogen. They can't solve this problem either, but they can lower the socket temperature to negative whatever Celsius so the CPU can take a +80C rise when the voltage kicks on.

 

When you launch AIDA's stress test (or any other) and you see your idle CPU temp go from 35 to 70C in that next instant, that is stage 1 heat management and the conductive part. That differential is directly tied to the Vcore and the load presented. If you keep watching the CPU temperature (particularly on Linpack or other steady load CPU tests), you will see the CPU temp tick up +1C every 30-60 seconds for a short period of time. That is the second stage of cooling and each +1C in CPU temp should correspond with a +1C rise in coolant temp (H115i Temp). It will not go up indefinitely, the but the increase is directly related to radiator size, fan speed, cycle rate (pump speed) and the controls you find for the cooler. Expected max rise for a 9900K and a 280mm is probably around +8C. That is the value you are attacking with various pump and fan speeds. However, remember the same is true in reverse -- minus 1C coolant only equals -1C CPU temp, so an extra 200 rpm to go down 1C may not be worth it in terms of noise. For most people, the effective range for 140mm fans on a 280mm will be 750-1100 rpm. Past that you can still grab back another 1-3C, but it will be loud. The fans are linked to coolant temperature because that is the only element it can control. You can blast the fans all you like, but that won't cool the underside of your CPU or prevent voltage from heating the pins. Until you start to warm up the coolant, effectively there is no more extra work to do and the fans remain the same. Heat is being conducted into the cooler, but the 280mm surface size and current fan speeds are enough to keep the coolant from rising further. If I were to swap it out for a 120mm radiator and weak fan, you would see a more dramatic rise. If I were to connect your cold plate to 10m external cooling panel, you likely would never see any rise at all.

 

 

 

You were using an air cooler on that 9900K and the temps never reached 70C? In the same conditions (Vcore and AIDA)?

Link to comment
Share on other sites

Thank you for your answers guys ! I appreciate you did not just dismissed me with a “that’s how it works” as Corsair customer center did.

 

As for the Air Cooling Reference , I was referring to my previous pc , with a q9200 processor , known for its high temps which never went over 65 under air cooling . I get it is not comparable with the 9900k , but It still made me doubt .

 

Thanks again !

Link to comment
Share on other sites

×
×
  • Create New...