Flat out not impressed with this 115i RGB Platinum

[RecyclopsReveng]

I've been using Corsair AIO's for a couple of years now. I've owned several different models. A couple were replaced due to problems and a handful were replaced just because I wanted to upgrade.

 

I went out a couple weeks ago and I bought the 115i RGB Platinum because I wanted to upgrade from a 240mm to a 280mm rad. First of all, this thing has a ridiculous amount of cables. I understand that they all serve a purpose but there's just so many plugs. The first issue I ran into was that my RGB lights would NOT hold a profile no matter what. I was stuck with the RGB puke out of the box for 3 days. I've very familiar with the iCUE software and no matter what I did it would not set to solid green like I wanted. I uninstalled and reinstalled the software about 15 times, and after the 15th install it finally did what i was supposed to do. Now every few days it resets itself and I have to go back in and change it.

 

Now my second issue with this cooler was the constant trickling sound coming out of it. I don't know if they short filled these loops or what, but I could constantly hear the liquid trickling at the top of the cooler. The only way that I could make it stop would be to mount it in the top of my case. However, if I were to do that then I would no longer be able to run a push/pull configuration on my fans. I ended up toying with the speeds of everything, and setting the pump to the preset "Balanced" seems to make it stop. Which super sucks since I'm all about maximum performance, so not being able to run my pump on "Extreme" sucks. At first the trickling sound would happen once a day, which was fine, but it got to the point where it was 24/7. I could hear it over my headset. So now I have my pump set to balanced, I have it mount as high up in the front of my case as I can and it seems to have stopped it.

 

Now my 3rd issue is the temps. It doesn't seem like this cooler handles higher temps very well. And it also seems like temps with these coolers vary tremendously. A friend of mine has the exact same CPU, with the exact same overclock, on the exact same cooler, and his temps are 20C Idle/55C Load playing Apex legends, yet mind is 25C Idle/78C Load on a 7700K @ 4.9GHz. I was getting better temps on my H100i V2.

 

The only issue here I can't deal with is the temps. How is it that I'm running 30C hotter than a friend on the same setup? What are all of you getting for temps?

[c-attack]

1) Can you elaborate a bit as to what was happening with the RGB colors on the pump? This is not a common complaint and you aren't a novice, so if there is some kind of bug, this is report one and we need to figure it out. Solid green should have been easy as a solid or static color.

 

2) This should be solvable. Trickling usually comes around when an bubble is blocking flow (a large one). On the front mount position, are the hoses up or down? It should work either way, but I am trying to visualize. Lift up the front of the case while the pump is running. If you get the angle vs gravity right, the bubble will release. In the interim, Extreme vs Balanced should be a literal 0.0C difference on a 7700K at any OC level. Pump speed is not a significant on short length, low to moderate restriction AIO systems.

 

3) I can't compare data for Apex Legends. Run something controllable and easy like CPU-Z stress test under the Bench Tab. It is linear so the initial load temp holds, plus any coolant rise. We want to look at the difference between the coolant temp and the 100% load CPU temp. Most CPUs have a predictable differential at a specific voltage.

What voltage for the 4.9C?

 

We are looking to see if there is a mounting issue. If so, temps should be very jumpy at the desktop (when voltage comes on) and too high at load. Also, the coolant temp will not go up with the load. That is something you can check when running CPU-Z (without GPU interference). I would expect a 4-6C rise in 10min for your CPU and approximate settings.

[RecyclopsReveng]

1) Can you elaborate a bit as to what was happening with the RGB colors on the pump? This is not a common complaint and you aren't a novice, so if there is some kind of bug, this is report one and we need to figure it out. Solid green should have been easy as a solid or static color.

 

2) This should be solvable. Trickling usually comes around when an bubble is blocking flow (a large one). On the front mount position, are the hoses up or down? It should work either way, but I am trying to visualize. Lift up the front of the case while the pump is running. If you get the angle vs gravity right, the bubble will release. In the interim, Extreme vs Balanced should be a literal 0.0C difference on a 7700K at any OC level. Pump speed is not a significant on short length, low to moderate restriction AIO systems.

 

3) I can't compare data for Apex Legends. Run something controllable and easy like CPU-Z stress test under the Bench Tab. It is linear so the initial load temp holds, plus any coolant rise. We want to look at the difference between the coolant temp and the 100% load CPU temp. Most CPUs have a predictable differential at a specific voltage.

What voltage for the 4.9C?

 

We are looking to see if there is a mounting issue. If so, temps should be very jumpy at the desktop (when voltage comes on) and too high at load. Also, the coolant temp will not go up with the load. That is something you can check when running CPU-Z (without GPU interference). I would expect a 4-6C rise in 10min for your CPU and approximate settings.

I've included a pic of the orientation for your reference. I tried flipping my case in ever which direction to move the air bubbles around, it's a big one. I currently have my case being raised in the front by a couple stacks of quarters to keep the bubble in the radiator and not the pump. As far as temps, if I run the CPUZ stress test I get up as high as 85C at 100% load after about 9 minutes. I'm running at 1.35V 4.8GHz. The only thing that I can think of is that there is a massive friggin bubble in this thing that's working it's way around the loop. And as far as the RGB issue is concerned, what was happening (now it happens but not consistently) is I'd set a color for my RGB's and after I applied them they'd just reset themselves to the stock config. I could apply it 100 times and it'd just turn itself back. Eventually it'd apply and stay and they've been good for a little bit now but it still happens every once in a while.

[RecyclopsReveng]

1) Can you elaborate a bit as to what was happening with the RGB colors on the pump? This is not a common complaint and you aren't a novice, so if there is some kind of bug, this is report one and we need to figure it out. Solid green should have been easy as a solid or static color.

 

2) This should be solvable. Trickling usually comes around when an bubble is blocking flow (a large one). On the front mount position, are the hoses up or down? It should work either way, but I am trying to visualize. Lift up the front of the case while the pump is running. If you get the angle vs gravity right, the bubble will release. In the interim, Extreme vs Balanced should be a literal 0.0C difference on a 7700K at any OC level. Pump speed is not a significant on short length, low to moderate restriction AIO systems.

 

3) I can't compare data for Apex Legends. Run something controllable and easy like CPU-Z stress test under the Bench Tab. It is linear so the initial load temp holds, plus any coolant rise. We want to look at the difference between the coolant temp and the 100% load CPU temp. Most CPUs have a predictable differential at a specific voltage.

What voltage for the 4.9C?

 

We are looking to see if there is a mounting issue. If so, temps should be very jumpy at the desktop (when voltage comes on) and too high at load. Also, the coolant temp will not go up with the load. That is something you can check when running CPU-Z (without GPU interference). I would expect a 4-6C rise in 10min for your CPU and approximate settings.

 

Now just for craps and giggles, I took the pump/block off my CPU, replaced the paste and re-sat it on the CPU. There isn't any effect from that. Temps are the same.

[c-attack]

On the CPU-Z test, what was the coolant temperature when you started/finished? Based on your listed normal 25C idle, coolant temp should be right there 23-24-25C at the start. If you are seeing an instant load temp of 80C or so, that's a 55C differential and pretty high for CPU-Z on your CPU. CPU-Z is also light, so that would make more strenuous tests untenable.

 

How much did the coolant go up during the 9 minutes? I am trying to discern a CPU thermal conductive issue (for whatever reason) from a cooler functionality issue. Large coolant temp changes would suggest a problem with the cooler. Most everything else has to be worked on differently.

 

The RGB stuff... I wonder if your cooler is disconnecting/reconnecting with iCUE. This is a reported issue for some people. See if the cooler is still listed in iCUE the next time it happens. You would also get the classic UBS removal (device disconnect) sound from Windows -- if enabled. I don't have a solution for that, if it is the issue. However, you can mask the effect by setting your hardware lighting for the pump and fans to green. the fan and pump speeds should be saved to the device automatically. Turn off the Windows device removal sound. If it happens, it will then neither change the color or alter your cooling.

[RecyclopsReveng]

On the CPU-Z test, what was the coolant temperature when you started/finished? Based on your listed normal 25C idle, coolant temp should be right there 23-24-25C at the start. If you are seeing an instant load temp of 80C or so, that's a 55C differential and pretty high for CPU-Z on your CPU. CPU-Z is also light, so that would make more strenuous tests untenable.

 

How much did the coolant go up during the 9 minutes? I am trying to discern a CPU thermal conductive issue (for whatever reason) from a cooler functionality issue. Large coolant temp changes would suggest a problem with the cooler. Most everything else has to be worked on differently.

 

The RGB stuff... I wonder if your cooler is disconnecting/reconnecting with iCUE. This is a reported issue for some people. See if the cooler is still listed in iCUE the next time it happens. You would also get the classic UBS removal (device disconnect) sound from Windows -- if enabled. I don't have a solution for that, if it is the issue. However, you can mask the effect by setting your hardware lighting for the pump and fans to green. the fan and pump speeds should be saved to the device automatically. Turn off the Windows device removal sound. If it happens, it will then neither change the color or alter your cooling.

 

So for the RGB issue, yes that does happen. I will get the disconnect sound and the lights on the block with flash really quick. Sometimes it goes back to green, some times it goes back to RGB.

And which plug needs to be in to get a coolant reading? I'm not getting one right now.

[c-attack]

USB has to be connected to get a reading. However, as long as that is the set control variable, the pump controller will still adhere to that internal sensor, even without software.

 

There are a lot of Platinum specific complaints on the USB disconnect issue in the cooling and iCUE forums.

[RecyclopsReveng]

USB has to be connected to get a reading. However, as long as that is the set control variable, the pump controller will still adhere to that internal sensor, even without software.

 

There are a lot of Platinum specific complaints on the USB disconnect issue in the cooling and iCUE forums.

 

The USB on the mobo? That's plugged in. I still don't have a reading however.

[RecyclopsReveng]

USB has to be connected to get a reading. However, as long as that is the set control variable, the pump controller will still adhere to that internal sensor, even without software.

 

There are a lot of Platinum specific complaints on the USB disconnect issue in the cooling and iCUE forums.

 

Or is the coolant temp the "Temp" option in iCUE?

[c-attack]

Full name should be "H115i Platinum Temp". Anything motherboard based should start with Asustek _____. Where you need to watch out is if you also have a Commander Pro. Those temp probes are simply Temp 1-4 (or whatever you rename them to be).

[RecyclopsReveng]

Full name should be "H115i Platinum Temp". Anything motherboard based should start with Asustek _____. Where you need to watch out is if you also have a Commander Pro. Those temp probes are simply Temp 1-4 (or whatever you rename them to be).

 

Okay so after running AIDA64 CPU Test for 5 minutes, I have the following results:

Coolant temp:

Min: 25.0C

Max: 26.2C

CPU Cores:

1: Min 28 Max 70

2: Min 30 Max 68

3: Min 31 Max 69

4: Min 29 Max 66

 

So those temps are absolutely fine by me. A max of 70 is a-ok. If I hop into a game right now though I will hover at almost 90C which I don't understand why there's that drastic of a difference.

[c-attack]

That is a more reasonable and expected coolant to core differential for your settings. The one thing I am looking at (albeit through a magnifying glass) is the coolant rise for 5 minutes. I would expect more than a 1.2C rise. I just ran AIDA blend (CPU/FPU/cache/RAM) for 5 minutes. On my custom loop, dual 360 radiator system, that was a 1.9C rise at all four measuring zones. Now my fans are strolling along at a meek 900 rpm, so I could be lower. However, I am more interested if you were blasting your fans at 1500 rpm+. That might keep it down. However, keeping an AIO coolant temp (measured in the block) at anything below +3C on load is next to impossible since the sensor is directly post CPU block. I am wondering if you are really getting full heat transfer.

 

I don't understand the AIDA vs game differences either. Asus rolled out system wide BIOS updates this past week to address those lovely Intel flaws. If you updated, is it possible you missed any of your normal settings? SVID Behavior (best case), IA/DC Load line voltage if you run Adaptive voltage, anything like that?

 

Of course if the CPU is not transferring heat through the cold plate, that would explain most things. Try a longer test - AIDA or CPU-Z is good since they are mild. Try for 10 minutes. That should be plenty of time to get the coolant to +6C. In order to reduce another variable, set the fans to a fixed rate of 1200 rpm. That is decent cooling, but will still allow coolant rise. We want to see some coolant rise. If you can run the stress test with the fans at minimum and the coolant doesn't go up, you know there is an issue.

[DevBiker]

So ... what happens if you orient the fans as intake?

One of the things that we've seen around here are that some cases have restricted exhaust flow at the top and the heat from the radiator can't escape. Over time, this builds up, creating a "hot zone" at the top of the case. This then limits the cooler's ability to do its job. And it's something that only rears its head after an extended load.

Also, that 1080Ti will put off some heat. If that radiator is configured as exhaust, it's pulling the GPU heat through it. Again, that'll impact the ability of the cooler to do its job. And it's something that you'd only see when gaming but not when doing a CPU-only test.

So ... give it a shot. Flip the fans to intake.

[Zetta1021]

I have a similar issue to you.. or I should say my brother does.

We both own the H115i Platinum we bought them at the same time exactly 1 year ago this month and were originally cooling an i7 6700k which worked fine enough.

 

Recently we both upgraded. Same Z390 Motherboard, same ram, same CPU (i9 9900k).

 

We match settings doing a 5ghz All core OC.

My temperatures max out around 77c. His go to 95c+

 

We cannot find a reason why his cooler seems to do such a poor job.

He recently RMA'd it.The RMA process I expect to be extremely lengthy so I don't expect to see it returned to him for at least a month, probably more since the holidays are here.

 

In the mean time he threw in an old Kraken X61 with after market fans that are inadequate for keeping coolant temperatures down. When in a stress testing it did much better than H115i, He was hanging around 84-85c and with very high coolant temps too.

 

Once he gets his H115i back I'll update here if I can remember.

This experience will probably turn me away from CoolIT manufactured units in the future and just stick to Asetek.