Started thinking about WC again, quick question!
- Thread starter Ibanez81
- Start date
SieB
New member
8 if it's just a CPU and 10 if it's CPU and GPU.
For a CPU loop
2 for the CPU block, 2 for the res, two for the pump and 2 for the rad.
For a GPU + CPU loop
Same as CPU loop but + 2 for the GPU.
You just have to think about which way your tubing will go and if a strait, 45 degree or a 90 degree angled one would be best.
Well I cant do anything more to it till probably Tuesday.
I need a 90Degree fitting, Not so much need it would just look better. I also Bought a pump top for the hell of it, The stock pump i find is a pain in the arse to get hose on to So went for a Pump Top and stick some Bitspower fittings on.
I went onto scan and bought an ax750 too for the hell of it lol xD
Everything is flushed out anyway, I did The Rad seperate And gave it a good shake for a while full of liquid, Nothing bad came out, clear as it was when it went in.
Just have to fill it up now which is easier said than done lol
I need a 90Degree fitting, Not so much need it would just look better. I also Bought a pump top for the hell of it, The stock pump i find is a pain in the arse to get hose on to So went for a Pump Top and stick some Bitspower fittings on.
I went onto scan and bought an ax750 too for the hell of it lol xD
Everything is flushed out anyway, I did The Rad seperate And gave it a good shake for a while full of liquid, Nothing bad came out, clear as it was when it went in.
Just have to fill it up now which is easier said than done lol
S_I_N
New member
I dont disagree or anything but all in all the fluid moves quickly enough thru that loop order isnt totally important as long as its neat. I'd wager if ya put temp sensors throughput the loop you'd not see any significant variations in temps. I doubt pumps add alot to the temp issue since alot of manu's put them in the actual rez.. But most pumps are designed to push fluid not pull it. this is given by head pressure which is the amount of lift (or push) a pump has given the amount of tubing fittings bends and blocks in the setup. Basically I'm saying a pump sucks up the fluid and pisses it out lol.
as long as there is fluid in front of the outlet it will also pull it given there is fluid (or trapped air) behind it pushing
SieB mate I'm not saying ur wrong but from ur pic it looks like you have the flow (arrows) backwards. Pretty sure that top barb on pump is inlet and the side one is out.
as long as there is fluid in front of the outlet it will also pull it given there is fluid (or trapped air) behind it pushing
SieB mate I'm not saying ur wrong but from ur pic it looks like you have the flow (arrows) backwards. Pretty sure that top barb on pump is inlet and the side one is out.
My res is the lowest part of my loop, Its On the Side of the HDD bays, and the Pump Sits on top of the HDD bay, inside where The Opticals go, and it pushes water Into The rad just a few inches above it, The hose from the pump to the rad is literally about 3-4nch long. As long as the loop is full it should pump round right?
Master&Puppet
New member
Personally when I design a loop I care more about gravity aiding the pump than anything else.
The pressure is going to be greatest from immediately after the pump so it makes sense to me to get the coolant as high as it needs to go straight away and let gravity help it through the rads and blocks then trickle back into the res & pump. Even better if the res and pump is a combo.
Botched a diagram
Otherwise my priorities are radiator placement:
1: Rad before pump/res
2: Rad after pump
3: Rad between cpu and gpu if on single loop.
If I have a dual loop system then I cross feed the returns to keep the whole system equal in temps yet gaining the bonus lph and pressure of 2 pumps.
Just my thoughts.
M&P
The pressure is going to be greatest from immediately after the pump so it makes sense to me to get the coolant as high as it needs to go straight away and let gravity help it through the rads and blocks then trickle back into the res & pump. Even better if the res and pump is a combo.
Botched a diagram
Otherwise my priorities are radiator placement:
1: Rad before pump/res
2: Rad after pump
3: Rad between cpu and gpu if on single loop.
If I have a dual loop system then I cross feed the returns to keep the whole system equal in temps yet gaining the bonus lph and pressure of 2 pumps.
Just my thoughts.
M&P
Master&Puppet
New member
Yea that looks like what I'd end up doing in your situation. Sieb is clued up on this stuff though so it has made me wonder about res placement (well I use res/pump combos so I avoid this issue!) but I have no facts to sway it either way.
SieB
New member
S_I_N and a few others are clued up and know their stuff too though.
I think it's the same as with a lot things and that it's the case of personal preference because really loop order doesn't really make much difference, once the loop is up and running and the coolant is flowing it's the same no matter what. As I said before though, I think it makes more sense to have the coolant pass through all the hot parts and then out through the cool parts, but thats just my opinion.
I don't think there is any real right or wrong just as long as you are not doing something out of the ordinary.
I think it's the same as with a lot things and that it's the case of personal preference because really loop order doesn't really make much difference, once the loop is up and running and the coolant is flowing it's the same no matter what. As I said before though, I think it makes more sense to have the coolant pass through all the hot parts and then out through the cool parts, but thats just my opinion.
I don't think there is any real right or wrong just as long as you are not doing something out of the ordinary.
Master&Puppet
New member
You are 100% right there. If you take a look at something like the laing DDC or D5 then they are quoted as having a maximum operating temperature of 60C - if you don't have a radiator between the components and your pump then start digging an early grave.
My loop does that i guess right? Going through the CPU & GPU and also the Pump Before Then being cooled again. But as master said, all the hot water will be going through the pump :/
Should have it all running Tuesday as i think that's when my 3 specialtech orders will turn up lol Dont ask why i have 3 separate >.<
Should have it all running Tuesday as i think that's when my 3 specialtech orders will turn up lol Dont ask why i have 3 separate >.<
Master&Puppet
New member
Yea sorry, this thread seems to have been giving out some confusing advice which is rare on this site and I know that some of that has come from me. I realise that I completely missed a point when I looked at your first diagram.
1. I believe that 1 radiator in a system should be placed immediately before the pump/res because obviously the coolant will be hotter before it goes through the radiator.
2. Beyond that, my personal opinion is that the pump should be placed at the lowest point in the loop and should pump straight out to the to the highest point because the leg work is done when the flow pressure is at its highest leaving gravity to aid it through the most restrictive part of the system (the blocks). However if you have a powerful enough pump this does not matter and should not be considered at the expense of point 1.
You are obviously restricted by the nature of the case ie you have to put the radiator in the roof. So the radiator has to be the last point in the loop and therefore the pump cannot pump to the highest point becuase it has to flow in the other direction. That is fine since you are thinking about the ddc 18W which has a monster head pressure and would be my choice of pump too.
So yea, your last diagram looks good. Sorry for messing up in the posts above!
1. I believe that 1 radiator in a system should be placed immediately before the pump/res because obviously the coolant will be hotter before it goes through the radiator.
2. Beyond that, my personal opinion is that the pump should be placed at the lowest point in the loop and should pump straight out to the to the highest point because the leg work is done when the flow pressure is at its highest leaving gravity to aid it through the most restrictive part of the system (the blocks). However if you have a powerful enough pump this does not matter and should not be considered at the expense of point 1.
You are obviously restricted by the nature of the case ie you have to put the radiator in the roof. So the radiator has to be the last point in the loop and therefore the pump cannot pump to the highest point becuase it has to flow in the other direction. That is fine since you are thinking about the ddc 18W which has a monster head pressure and would be my choice of pump too.
So yea, your last diagram looks good. Sorry for messing up in the posts above!
S_I_N
New member
sorry but y'all are being abit daft here. Gravity, cooling fluid before pump all hogwash imho. I've been doing this shite for over a decade and never before have I heard gravity lmao. ITS A FARKIN PUMP (pumps create pressure) gravity has nothing to do with it. now explain to me why one would water cool a cpu and make the coolest part going into the pump before hitting the cpu???? are we not trying to cool the cpu not the pump??? overall as i have said before the temp differences in the loop between parts (blocks, pump, rads) is negligible. Your coolin a cpu and gpu for heaven sake you want the rad before those. loop should be rez>pump>rad>block(s)>rez. The fluid moves to quickly to be worried about hot fluid hitting the pump. Which is utter horseshite And operating temps for a pump lmao those are external temps mate not internal. Basically one wouldnt be water cooling with 60c ambient temps anyways. they'd be using some kind of chiller but thats another topic lol.
overall though a well designed loop will as i said before have very similar temps at every stage of the loop. The only thing that will make a difference is the voltages used to oc the cpu or gpu. Water can only remove so much heat no matter how big a rad you use. The more volts you use the less effectively the heat is removed by the blocks water and rads. This same principal is true for air cooling. The more heat the harder it will be to cool.
overall though a well designed loop will as i said before have very similar temps at every stage of the loop. The only thing that will make a difference is the voltages used to oc the cpu or gpu. Water can only remove so much heat no matter how big a rad you use. The more volts you use the less effectively the heat is removed by the blocks water and rads. This same principal is true for air cooling. The more heat the harder it will be to cool.
Master&Puppet
New member
Lol S_I_N this is how I came to my conclusion.
Pressure will differ across the system. Take a human body for example. The arteries which the heart pumps into are larger and stronger, to absorb the pressure, than the veins which return the blood to the heart. In fact the veins in the legs have non return valves because the pressure can be that low. It's exactly the same in this situation - the pressure immediately after the pump will be greater than when it returns to the pump. Now in a smaller system, like pcs, the difference will be far less noticable I expect, but why make a pump work harder than it has to?
If that's an external pump temp then I've learnt something, but it does say here under special features "maximum system temperature 60C".... In any case it still makes sense, conservatively, to cool the cpu before the gpu right? Because the gpu can take more heat and the coolant will be hotter after the cpu. I get the point about not really caring about the coolant temp since it's the temp of the components we are worried about - it's not like we are stressing the water lol.
I admit that I might be being very picky here but that's the way I do it, am I the only one which this makes sense to?
To be honest it probably doesn't make much difference I guess.
Pressure will differ across the system. Take a human body for example. The arteries which the heart pumps into are larger and stronger, to absorb the pressure, than the veins which return the blood to the heart. In fact the veins in the legs have non return valves because the pressure can be that low. It's exactly the same in this situation - the pressure immediately after the pump will be greater than when it returns to the pump. Now in a smaller system, like pcs, the difference will be far less noticable I expect, but why make a pump work harder than it has to?
If that's an external pump temp then I've learnt something, but it does say here under special features "maximum system temperature 60C".... In any case it still makes sense, conservatively, to cool the cpu before the gpu right? Because the gpu can take more heat and the coolant will be hotter after the cpu. I get the point about not really caring about the coolant temp since it's the temp of the components we are worried about - it's not like we are stressing the water lol.
I admit that I might be being very picky here but that's the way I do it, am I the only one which this makes sense to?
To be honest it probably doesn't make much difference I guess.
S_I_N
New member
M&P I apologize if I came off snooty. I was just laughing my arse off when i read alot of the posts and well my fingers went to banging on the keys faster than my mind caught up to it. Had a very ruff day with work ( got alot worse in the afternoon after I posted that) anywho not an excuse for it but yeah there are many ways to run a loop best is the tidy way. But its not always possible to make tidy but one can only do the best they can.
