Liquid Cooling Flow
Liquid Cooling Flow
No worries, sending flow from component 1 to component 2 doesn't really change temperatures since the fluid in the system stays relatively consistent instead of heating up significantly after moving between components. The temperature difference between water just before entering a CPU under load could be around 56 degrees, while the water exiting might be 57-58 degrees. That's a pretty minor variation.
sending the flow back to a radiator after a component can help maintain better temperatures, though it might make the system look overly complex. however, if it leads to a complicated design with more tubing and direction changes, it’s not worth it and would need a significantly stronger pump to handle the increased workload.
Brantyn Gerik is discussing optimizing cooling system layout and efficiency.
Yes, it's definitely possible, but keep in mind you'd be reducing the heat output from both your CPU and GPU with just that 1 280mm rad. My advice would be to install the first rad at the top, then add the extra 140mm rad on the back. The main goal is to make the largest rad at the bottom for intake—using fans to push air through the radiator into the case, and possibly placing a radiator either front or back as an intake, while keeping the top area for exhaust instead of another radiator.
I'm just starting out with this. Reading through the thread has been helpful as I prepare for my first project. I'm studying plumbing in college, and from a technical standpoint, some concepts are confusing, such as why there isn't a manual or auto air admittance valve for priming. I'm considering using compression fittings connected to 10mm microbore copper tubing, with soldered joints. It's straightforward to achieve precise bends using a handheld bender. It could give me a somewhat industrial appearance, similar to a Mamod engine. I'm also thinking about adding a separate flow and return run. Without a circulator, the liquid will naturally rise when it cools down. Once it cools, it should fall back down. This approach seems logical to me—it aligns with the natural flow direction. In other words, let the pump assist in directing the coolant as it prefers. Where the pump appears less important, it's because it's a closed system; otherwise, the reservoir would need to be elevated above the top of the radiator. The main concern is placement: can the pump handle the heat directly from the source? I don't think that's an issue. Regarding the pipe layout, it seems you want both coolant blocks to receive water at similar temperatures, not one getting pre-heated. A separate flow and return system would be ideal, similar to what you see in central heating. There should be a temperature drop between the flow and return paths, which would indicate how much heat is being dissipated by the radiators. The downside of this method is the absence of valves, making it difficult to balance pressure. On the plus side, it might reduce strain on the pump—like breathing through a straw compared to two straws.
A lot of mixed information appears in this thread, but some parts are useful.
Flow direction or sequence doesn’t affect temperatures much. Make sure the pump doesn’t pull in air from the reservoir or elsewhere.
Dividing flow is a way to split parallel flow; many do this for tighter blocks like GPUs, but it’s not essential for high-flow parts such as radiators where restrictions are usually lower.
Coolant temperatures should remain consistent at any point in the loop—if you logged temperatures at various positions, they’d likely be similar. Coolant usually balances itself as a whole unit and won’t form localized 'hot spots' unless there’s restricted flow between otherwise continuous sections (like a poor parallel section).
Serial flow, where one component connects directly to the next, is the typical setup for most people, including myself. I’ve tried running dual GPUs with parallel flow, which helps by splitting the flow into each GPU inlet and bringing it back at the outlet. Serial flow is generally more restrictive in a single segment compared to parallel setups within the same segment.
I’m not a fluid dynamics expert, so much of this technical advice should come from someone more experienced. Here’s the link for reference:
https://www.overclock.net/forum/61-water...allel.html
This topic has been debated extensively, and the decision mainly comes down to personal preference and flow speed.
I’ll always go parallel for dual GPU blocks, though I’ve thought about running parallel my CPU and GPU blocks just to compare results.
I would anticipate the temperature to eventually become more balanced. The variation between flow and return reflects the amount that has been removed. If these two are equal, it likely means either a) very little heat was lost (rads too small or insufficient? b) heated water is returning before it can cool down by the rads (pump operating too quickly). The water only needs to move slowly through a rad to function properly. A type of circulator would be more effective, assisting in even heating across all rads. Nevertheless, heated water can naturally rise and fall without a circulator. It uses thermosyphons because of the density difference between hot water in the flow and cold water in the return. The issue is that it tends to expel heat more effectively through the highest rads. Assisting it helps all rads warm up uniformly. By the time the water exits a rad, it should feel noticeably cooler than when it entered. If not, the rad isn’t doing its job and you’re pushing hot water back out. I think I might be overlooking something important here.