[SOLVED] 1 radiator 2 components, help pls

mantinziza

Distinguished
Dec 10, 2012
28
3
18,535
I’m currently trying to figure a cooling loop for my system. I have a great radiator its too big actually so it works great, but I’d like to try to avoid getting another due to it’s size and my space limitations. So I was trying think of a way.

First of I plan on getting 2 pumps like this, I already have 1 and it’s works great.

https://www.ebay.com/itm/SC-300T-12...iquid-Cooling-System-/401443006468?nav=SEARCH


Now I found another reservoir that has 3 holes altogether.

https://www.ebay.com/itm/200ml-Wate...m=142515810971&_trksid=p2047936.c100373.m3226

So basically my loop would be,

I’m going to start at the 3 hole reservoir to a 1/2 hole output and tube (very short less than 4 inches) , too the radiator (which is sized 1/2 as well, input and output ,holes and tubes), too a 1/2 inch input on a Y split, which will split into two 3/8 outputs, these 2 outputs/hoses go too the inputs at the 2 pumps, separately.

(the pumps intake is both a g1/4 to a 3/8) which goes straight into a tank that the pump sits in, AKA in the link you can see this), 1 pump will go through a 3/8 tube to my cpu block, then back to 1 hole in the original 3 hole reservoir,

the other pump will goe through a 5/16 tube (I think, I still need to measure accurately but it’s either a 5/16 or 7/16 Inner diameter hole) in my mofset,

(my board is an old asrock z77 oc formula if that helps),

Which will then go back to the original 3 hole resevior, but I plan on using 1/4 ID tube at this part, or at least 1 step bigger coming out of the mofset, than going in.

Considering the pumps sit in a reservoir and they draw through a larger area than they pump into individually.
It seems like it would work to me, but I would like some input from the pros before hand, thank you all
 
Solution
Ever put a pan on the stove that's full of water? Takes forever to boil. Water has an amazing ability to absorb energy, consequently it takes a huge amount of energy to even raise that pans water 1°C. The cpu and mosfets are nowhere near what the stove can dish out, that's in the neighborhood of 1500-2000w. It'd take hours at full loads to get a noticeable increase in coolant temp on a loop. That's what the radiator is for, balance. It takes that absorbed energy and dissipates it.

So no, it doesn't matter where in the loop the components are, but most will follow an easy path around the case. They put the rad on top or in front. The pump on the floor. So from pump it'll go up to the gpu, right to the cpu in/out of the rad to drop into...

Karadjgne

Titan
Ambassador
You only need 1 loop. Continuous.

Res/pump to cpu to mosfets to rad to res/pump. And even that is interchangeable, you can stick the radiator anywhere. If the pump can handle the flow restriction you could have 10 components in a row, including a gpu, second rad, flow meters, ram coolers etc. Once you actually fill the loop/res it's all the same, total volume has changed (there's a little more coolant due to the added piping and mosfets) but what's going through the pump at any given moment is the same.

The only calculations you'd need are the total wattage being cooled, just to see if the rad is sufficient, and the amount of restriction on the loop.
 

mantinziza

Distinguished
Dec 10, 2012
28
3
18,535
You only need 1 loop. Continuous.

Res/pump to cpu to mosfets to rad to res/pump. And even that is interchangeable, you can stick the radiator anywhere. If the pump can handle the flow restriction you could have 10 components in a row, including a gpu, second rad, flow meters, ram coolers etc. Once you actually fill the loop/res it's all the same, total volume has changed (there's a little more coolant due to the added piping and mosfets) but what's going through the pump at any given moment is the same.

The only calculations you'd need are the total wattage being cooled, just to see if the rad is sufficient, and the amount of restriction on the loop.
The hotter coolant coming out of my cpu won’t effect the cooling efficiency of my mosfet? I like the idea of what your telling me but if I run it like that wouldn’t it make sense to go mosfet to the CPU since the mosfet runs cooler naturally than the CPU does?

And the fact that the mosfet tubes are the smallest part of the entire tubing system, wouldn’t that effect the order as well? I need to max my cooling out as much as possible.

Does everything really even out that much due to the coolant flow that order doesn’t matter?

Ty for the response
 

Karadjgne

Titan
Ambassador
Ever put a pan on the stove that's full of water? Takes forever to boil. Water has an amazing ability to absorb energy, consequently it takes a huge amount of energy to even raise that pans water 1°C. The cpu and mosfets are nowhere near what the stove can dish out, that's in the neighborhood of 1500-2000w. It'd take hours at full loads to get a noticeable increase in coolant temp on a loop. That's what the radiator is for, balance. It takes that absorbed energy and dissipates it.

So no, it doesn't matter where in the loop the components are, but most will follow an easy path around the case. They put the rad on top or in front. The pump on the floor. So from pump it'll go up to the gpu, right to the cpu in/out of the rad to drop into the reservoir which drops into the pump. One nice circle, little or no crossed piping or hard angles. Flow direction (other than res to pump) doesn't matter, you could run it opposite, as long as inlet/outlet designations are correct.

Mosfets constantly run 80ish+°C when the cpu is under loads, can even go upwards of 95°C on occasions, nobody with half a brain allows the cpu to get that hot for any length of time. So mosfets are almost always hotter running than the cpu.

Almost all fittings used in loops are the same G1/4. So tubing ID isn't going to have as great an impact as the amount of fittings in the loop. That tubing would have to be seriously small ID to place enough restriction on the loop to warrant a dedicated pump.
 
Solution