Flow doesnt change thru a loop,you can only flow as much as your narrowest restriction.
Presuming "flow" refers to volumetric flow rate (as in liters per second), then this is right. If I recall correctly this is basically what the
continuity equation says, for the academically interested.
"Water should remain in radiator as long as possible to cool down": That's a
misconception. Consider this:
- Radiator one is completely serial. One continuous thin tube. This results in maximum flow velocity and extremely high restriction (such radiators are not used in practice because they would require a very strong pump which would dump much more heat into the loop which would require more radiators and so on and so forth...:huh
.
- Radiator two is completely parallel. The water flow is split into many tubes and flows through the radiator slowly (velocity, not flow rate).
The time the water remains in radiator one is exactly the same as the time it remains in radiator two. As B Negative said, this is due to
constant volumetric flow rate in the entire loop and has to do with the continuity equation. If you want the water to stay in the radiator for a longer time, the absolutely only thing you can do is reduce the flow rate in your entire loop.
Flow velocity however does of course not have to be constant (again, this is reflected in the continuity equation, among other things). A wider tube in a closed loop has the
same volumetric flow rate as any other area of the loop (so long as it is connected in series and not in parallel of course), however it will have a lower
flow velocity than a tighter tube in the same serial loop.
Whether or not water really should pass through a radiator slowly (as in: low velocity, not low flow rate) and through a block quickly (again, velocity, not flow rate) can be inferred from a few basics:
- Heat transfer usually functions better in turbulent environments (explaining why would make this post even longer, but it's true).
- Turbulence in a water cooling loop is mostly the result of a fluid flowing along a non-moving surface (the inner side of a tube or a block etc.) and the friction that is caused by that movement.
- Higher flow velocities result in more turbulence (due to higher friction).
- As mentioned, tighter channels result in higher flow velocities.
- Higher surface areas improve heat transfer.
This is why most modern CPU blocks use either a pin matrix or very narrow channels (increase surface area, increase friction) preceded by a jet plate (increase flow velocity).
The same could however be said for a radiator: Turbulent flow along with higher flow velocity should increase heat transfer from the water to the tubes. And indeed some radiators have structures inside their tubes to increase turbulence (I think the Aquacomputer AMS belong to that group if I'm not mistaken).
The trouble is,
creating turbulence costs energy. This is for example why airplanes, ships and cars are designed to minimize turbulence. Because that energy costs fuel. Or in our case: You need a stronger pump (a lot stronger). A stronger pump however dumps more heat into the loop, which starts the never-ending circle of escalation mentioned above.
This is why radiators are built at least with a partially parallel configuration in order lower
flow velocity and therefore turbulence. This reduces flow resistance. Having the same turbulence as in a CPU block would be awesome for heat transfer, but cannot be sustained by any pump without dumping an excessive amount of additional heat into the system.
As everything else in engineering, it's a balancing act
. There is no silver bullet.
PS: It's really late here, and it's a really long post. I hope there aren't any logical fallacies buried in here, but if there are, feel free to point them out (with supporting arguments of course
). I'll review the post again once I've gotten some sleep.
.
