Water cooling could require two loops
Water cooling could require two loops
PC Specs First:
CPU: i7-6700k
GPU: MSI 1070 Gaming X
MOBO: MSI Z170a M5
I’m considering running a single custom loop, but it seems there aren’t any water blocks for my motherboard to cool the CPU. So perhaps I can focus on cooling the GPU and RAM with a custom loop, then try an AIO cooler for the CPU?
Does that sound unusual?
My computer is only a year old, so I’m hesitant to buy a new motherboard yet—it’s been really good for me. However, I plan to overclock eventually and need better temperatures. Right now, my CPU cooler is a Cryorig H7, which is excellent, but it sometimes spikes the CPU temperature to around 70°C when loading new areas in Fallout 4.
Am I overly worried about cooling? Or is there a better approach?
Any feedback would be appreciated.
P.S. I have a mid case and intend to upgrade to a full-size this year, which should improve airflow a bit.
The temperature of your second GPU being liquid cooled remains much lower than any GPU in an air cooled setup, and many of those can run continuously for months.
Connecting a Y-connector doesn't worsen the restriction; it actually reduces it, though it may lower the flow rate.
I tested a CPU paired with two GPUs using SLI on a single D5 board for several years without encountering these problems.
There are numerous waterblocks available for cooling CPUs. I own the same CPU you mentioned, and it is watercooled in a single loop alongside my GPU. You're looking for blocks that match the 1151/1155/1156 pin configurations—they all fit the same dimensions.
Additionally, you won't need much RAM to be watercooled, particularly with DDR4, which naturally stays cool.
There are many waterblocks available to cool CPUs. I own the same CPU you mentioned, and it is watercooled in a single loop with my GPU. You're looking for blocks that match the 1151/1155/1156 pin configurations—they all fit the same size.
You won't really need RAM to be watercooled, particularly DDR4, which tends to stay cool naturally.
Sorry for the earlier vague explanation; I should have been more specific in my first post. I've been checking EKWB waterblocks and noticed some reports that the blocks don't sit properly because the caps are too close to one side. That seemed unusual since AIOs generally don't have such issues with my motherboard.
This relies on it - certain blocks could align differently based on the motherboard layout. EK offers a dedicated site to check compatibility with your parts. Generally, reputable setups, like EK's, consider these variations to ensure broader usability.
This varies – some blocks may position differently based on motherboard layout. EK offers a dedicated site to check compatibility for your parts. Most reputable setups, including EK, consider these variations to ensure broader compatibility.
I faced the same problem initially when searching for a CPU block for my motherboard; they didn’t list any suitable options. I reached out and they mentioned the cap issue.
In reality, I believe I could purchase a solid AIO for the CPU and then use its own fans for GPU cooling (like the MSI 1070 gaming X), which appears adequate. This approach also helps lower ambient case temperatures, potentially improving GPU performance.
Yeah boosting to 70c isn't too bad, particularly when you're pushing the clock speed. Are you doing that? If so, what speed and voltage are you using? You should have a good cooler if you're providing enough airflow into the case. An AIO closed loop could be useful if its fans are drawing cool air from outside the case. The same applies to the air-cooled GPU—just make sure it's getting enough cool air in. How are your GPU temperatures looking when things get tough? If your GPU is getting hot and all the exhaust is blowing into the case, it might be affecting the CPU's cooling.
Based on my experience, four custom-built water cooled PCs have shown that dual cooling loops are the best approach. My first two systems used a single loop where the CPU output connected to the first GPU in an SLI configuration and the first GPU's output fed into the second GPU. Most major PC magazines at the time supported this setup, and it still does today. However, in my case, this method led to the failure of the second GPU in both initial machines. Although theory suggests it shouldn't matter since liquid can remove heat equally from each device, in practice it does affect performance. Unlike magazine builds, I typically use my gaming machine regularly for several years before any rebuild or upgrade. The second GPU is the third component in the chain, and its input temperature will always remain above the combined heat load from the CPU and the first GPU. Given these conditions, it makes sense that it eventually failed.
My third and fourth water cooled builds employed separate dual loops—one for the CPU and another for the GPU. The GPUs were still in SLI mode, so I used a 'Y' adapter to split the GPU input lines, ensuring each GPU received the same initial temperature from the radiator or pump. The outputs are then merged back together using another 'Y' adapter, feeding the combined input back into the respective GPU loop's radiator. A Lang D5 model provides sufficient pressure to allow this split, as the water pressure is divided among the active splits or taps. One of these machines has operated for nearly four years without any temperature-related issues.
Therefore, I suggest you make your own decision, but this would be my advice.
The temperature of your second GPU being liquid cooled remains much lower than that of any GPU in an air cooled setup, and many of those are kept running for extended periods.
Connecting a CPU with a Y-connector doesn't add restriction; it actually reduces it, though it also lowers the flow rate.
I tested a CPU paired with two GPUs using SLI on a single D5 board for several years and didn't encounter these problems.
Michael Piazza shared his insights based on personal experience with four custom-built water-cooled PCs. He noted that dual cooling loops are generally effective, citing past setups where a single loop failed to protect the second GPU. He emphasized that while theory suggests equal heat extraction should matter, real-world performance still plays a role. He compared his own setup, which used separate loops for CPU and GPU with a split adapter, to previous configurations and highlighted its reliability over nearly four years without temperature-related failures.