Overclock.net banner

Special loop setup. Will it work.

627 views 22 replies 10 participants last post by  EMUracing 
#1 ·
Hi all,

I have a short question. I'm planning my first custom water loop. But I want to know if my design will work. I know that a normal loop can go two ways. The one most common is.

Code:

Code:
Pump -> blocks -> reservoir -> radiator -> pump
However my idea due to design will be.

Code:

Code:
Pump -> reservoir -> radiator -> 2nd reservoir -> blocks -> pump
Does this work. The pump will be in the bottom of the case, lowest point in the loop. The two reservoirs will be the top of the loop. The bottom of the reservoirs will be just above the top level of the radiator (360mm).

I hope anyone can help me with the question if the will work. I know that thanks to a couple of liquid laws there will be some flow in the radiator. But will the be enough to cool a intel 4920k and a nvidia gtx970?

Best regards,
Xerohumoris
 
See less See more
#2 ·
The most common setup is:

> Pump > Rad > Block(s) > Res >

Having the block(s) right after the rad assures that the coolest water enters the blocks, though the water temperature anywhere in the loop varies little - less than 1°c, typically 0.5 °c

Your setup will work, but you will have difficulty controlling the water levels in the reservoirs.

Watercooling loops are "closed Loops" and gravity has no effect on water flow. Other than making bleeding the air out easier, there is no performance difference no matter what the relative heights at which the components are.
 
#3 ·
I know there will be a difference in liquid level of the reservoirs. But a can't find the flow diagrams of any EKWB radiators any where so I can calculate the difference. If I know this I can calculate the height difference I need to get them equal under load.
 
#4 ·
Quote:
Originally Posted by Xerohumoris View Post

so I can calculate the difference. If I know this I can calculate the height difference I need to get them equal under load.
What "difference" do you want to calculate?
 
#5 ·
As long as you go Res > Pump, the rest doesn't really matter.

Why do you have two reservoirs? Is it aesthetics?
 
#6 ·
Quote:
Originally Posted by billbartuska View Post

What "difference" do you want to calculate?
I will push the water in to one of the reservoirs. So where will be a high pressure. Then it will flow to the radiator and the next reservoir. Due to the resistance in the loop and the radiator the pressure in the second radiator will be lower. There will be some air in the reservoirs. Due to the pressure difference under load the liquid levels in the reservoirs will be different.

So to make it short I want to calculate the pressure drop that the radiator causes in the loop. This way I can compensate in the filling stage.
Quote:
Originally Posted by haha216 View Post

As long as you go Res > Pump, the rest doesn't really matter.
Not a option. If I want to do this I have to machine a other reservoir bottom, with a bypass. At this point that is not a option.
 
#7 ·
With the pump off and res caps open (for filling) gravity will control the level since it's not a "closed" system.
With the pump on and the caps closed pressure will control the level.
 
#8 ·
I will make it a closed system. This is the best thing to do for the loop.

But due to the resistance in the radiator the level in the two reservoirs will be different. To calculate the pressure, and therefor liquid level, difference in the two reservoirs I need to know the resistance of the radiator. I can calculate the resistance in the rest of the loop.

I'm thinking about making a air connection between the tow reservoirs. This way the pressure in both will always be equal. But there will, even in this case, be a difference in liquid level in the two reservoirs. This has all to do with the resistance in the radiator and loop.
 
#9 ·
Because the loop will be a closed system, the volumetric rate of flow throughout the loop will be constant, such that the rate of flow into the res will be equal to the rate of flow out of the res, so the levels of the res themselves is not dependent on loop order, but how you fill the loop.So it doesn't really matter if one res is before a rad and the other after.

However, the reason why it is often suggested that you run a res before a pump is so that you can ensure, especially during the filling process, that your pump will not run dry and destroy itself. Is there a particular reason you cannot reverse the flow order?
 
#10 ·
I have a bit of a small case that has no real room for water cooling. To make it all fit I have to mod the case and have some external parts. So I was planning the radiator and the reservoir on to of the case. To make this look nice I'm planning to put two reservoirs vertically on top of the case with the radiator behind them. If I had the option to machine a insert for the reservoir bottom to make it a bypass I would.

That the pump will run dry won't be a problem I think. The two reservoirs will be on the top and all the other parts will be above the pomp.

p.s. No a other case is not a option. Don't have space for a bigger one and I need a inverted case to have the window face me.
 
#11 ·
There is only one hard and fast rule of loops.

Res > Pump, and make sure the head of coolant in the res is higher than the pump.

Unless you like replacing pumps, I suggest you follow it.
 
#12 ·
Well, technically, you don't even need a reservoir - can just use a fill port. Obviously makes things a little trickier, but i think its more suited to your situation.
 
#13 ·
I prefer to run Res- Pump- Block- GPU Block- Rad-(Res)

It just made things easy for me doing this route, however I want to believe that it expels the heat right after the blocks-- but the water is generally moving so fast around the loop it shouldn't really matter.

I'll reiterate what everyone is saying though, it doesn't *really* matter as long as your pump is fed, too much air can burn up a pump very fast.
 
#14 ·
Feeding the pump was never a concern of mine. Like I explained the pump will be in the bottom of the system. They only thing about the pomp I have to figure out is the direction it will pump in. But this will be depending on the exact orientation of the pump, not going to cross tubing.

The only thing I was wondering if I would still have enough pressure to have flow in the radiator when I don't have a pump pushing or pulling from it. But apparently this is not a issue. When I start building I'll upload pictures in a build log.
 
#15 ·
Quote:
Originally Posted by Xerohumoris View Post

Feeding the pump was never a concern of mine. Like I explained the pump will be in the bottom of the system. They only thing about the pomp I have to figure out is the direction it will pump in. But this will be depending on the exact orientation of the pump, not going to cross tubing.
1) Asks for advice
2) Doesn't like it, so ignores it
3) Burns out pump.

Some people just have to learn the hard way I suppose....
 
#17 ·
Quote:
Originally Posted by Wihglah View Post

1) Asks for advice
2) Doesn't like it, so ignores it
3) Burns out pump.

Some people just have to learn the hard way I suppose....
I know that if you don't do it the right way you can and will burnout the pump. But If I read the command on my question I get contradicting advice.
1) pressure in the loop will always be equal. I know for sure this is not true. But this would mean that pump placement is not a issue ones the loop is filled. Remember my pump will be at the bottom so ones the loop is filled correctly it will always have a constant flow of incoming water. According to the idea of equal pressure.
2) The pump has to be run after the reservoir to not have run out of water.
Quote:
Originally Posted by Jidonsu View Post

I don't see why you don't take your plan for flow and just run it in reverse so the pump gets fed by the reservoir instead of the other way around. Nothing has to change except switching the hoses on inlet and outlet on the pump itself.
This really is going to depend on the orientation of the pump. I know it will only fit one way. But I probably don't have the room to get the pipes of in the correct direction to make that fit. But If I can I will have a reservoir be the input of the pump.
 
#18 ·
Quote:
Originally Posted by Xerohumoris View Post

I know that if you don't do it the right way you can and will burnout the pump. But If I read the command on my question I get contradicting advice.
1) pressure in the loop will always be equal. I know for sure this is not true. But this would mean that pump placement is not a issue ones the loop is filled. Remember my pump will be at the bottom so ones the loop is filled correctly it will always have a constant flow of incoming water. According to the idea of equal pressure.
2) The pump has to be run after the reservoir to not have run out of water.
This really is going to depend on the orientation of the pump. I know it will only fit one way. But I probably don't have the room to get the pipes of in the correct direction to make that fit. But If I can I will have a reservoir be the input of the pump.
Im a noob but here we go. Pressure in the loop WILL always be equal and its a fact. If you have a very restrictive gpu block for example it will cause back pressure. That back pressure will cause your pressure to slow (aka youll have LESS pressure).

The reason you always run res-pump or a combo is because, if the pressure becomes too great and slows down after the gpu/cpu block and thats whats feeding the pump youll get nothing but air back into the pump, making bleeding virtually impossible, youll get the pump running dry and kill it within minutes.

I hope you do understand that anything running at 2000+rpm without any lubrication can and WILL get hot very VERY quickly.
 
#19 ·
Pressure and flow are not the same thing. Yes the flow in the loop will be the same in all components. However do the restrictive nature of ALL components in a loop, even the way tubing is run, there will be a pressure, not a flow, drop.

If you would use two different sizes of piping in your build, lets say a 10mm ID and 12mm ID. The flow in both have to be the same. But the pressure in the 10mm has to be higher then in the 12mm to have the same flow.

Pressure depends on flow (equal at every point in the loop) and the service area. Depending on the configuration of a loop pressure will change.

The pressure before a restrictive part will there for be higher then after. Specially if you have something that can be compressed, aka air, in your loop. And most of the water loops will have some air in them.

But to get back to my main question. How bad is the restriction of a radiator. According to the answers I got very little.

A small video off YT to explain it a bit better:
 
#20 ·
Not sure what case you have, but you definitely can't call it small if you can fit a 360 mm radiator in there and two reservoirs. My Fractal Core 1000 is borderline small, but not mITX small where you need SFF power supplies and cases designed with water-cooling in mind. I can fit a 240 mm RAD in there with cutting, maybe a 120 mm with some more cutting.

You really only need one reservoir to act as a fill point for the entire loop and to help get rid of air bubbles. If you do use a reservoir, you ideally want it right before the pump so that you can fill the reservoir and feed the pump from there so the pump never runs dry. Putting the reservoir above the pump allows gravity to help with filling your loop and air bubble removal as well.

If you don't have the reservoir right before the pump and you have a radiator or block after it, then the pump will have to push enough fluid in to the reservoir to pressurize it along with the small bit of air then you have to make sure the entire reservoir is air-tight and don't make the mistake of opening it up while the loop is running.

If the reservoir is right before the pump though, it's the last point before the pump and more or less just a collection spot for the water/coolant before feeding in to the pump, you won't have as much pressure in the reservoir this way and you can typically open the fill port no problem to let air bubbles out, or add more fluid while the system is running.

Now you don't need to run any reservoirs either, a single T-fitting before your pump up to a fill port will work, but you will need to control the pump at its lowest speed and in short bursts to get the loop filled without causing damage to the pump. It also helps to fill your RAD with water/coolant first if you have it installed in a way that puts the fittings at the top, this saves a good amount of filling later on.
 
#21 ·
Quote:
Originally Posted by Xerohumoris View Post

Pressure and flow are not the same thing. Yes the flow in the loop will be the same in all components. However do the restrictive nature of ALL components in a loop, even the way tubing is run, there will be a pressure, not a flow, drop.

If you would use two different sizes of piping in your build, lets say a 10mm ID and 12mm ID. The flow in both have to be the same. But the pressure in the 10mm has to be higher then in the 12mm to have the same flow.

Pressure depends on flow (equal at every point in the loop) and the service area. Depending on the configuration of a loop pressure will change.

The pressure before a restrictive part will there for be higher then after. Specially if you have something that can be compressed, aka air, in your loop. And most of the water loops will have some air in them.

But to get back to my main question. How bad is the restriction of a radiator. According to the answers I got very little.

A small video off YT to explain it a bit better:
Haha sorry to fix this but NO!!!!!!!!. 1. Loops will not have some air in them, theyll actually have NO air. Air is the poorest of poor thermal conductors. 2. Air can be compressed as you stated but water cannot, it never will and never can (auto tech for 10yrs i can tell you what happens when you try to compress water). 3. Theres a reason 99% of builders use a reservoir, and theres a reason why auto techs who know what there doing ADD a reservoir when doing any type of cooling fix/repair/refilling. It looks like this and works like a miracle. Its technically a reservoir for your coolant system (and no the bottle with coolant in your car is not a reservoir but an overflow bottle)



If someone gives you advice take it, as theyve all been there and know what they're doing. You could always buy that funnel and rig it to work in your system provided your radiator has a free port (aka alphacool and so on)
 
#22 ·
Quote:
Originally Posted by axipher View Post

Not sure what case you have, but you definitely can't call it small if you can fit a 360 mm radiator in there and two reservoirs. My Fractal Core 1000 is borderline small, but not mITX small where you need SFF power supplies and cases designed with water-cooling in mind. I can fit a 240 mm RAD in there with cutting, maybe a 120 mm with some more cutting.
Yes it is small. That's why I have the two reservoirs and the radiator outside on top of the case. The case itself doesn't have a spot for any radiators. None of the fan mounts have additional space for a radiator.

Now for the offtopic part.
Quote:
Originally Posted by bluej511 View Post

Haha sorry to fix this but NO!!!!!!!!. 1. Loops will not have some air in them, theyll actually have NO air. Air is the poorest of poor thermal conductors. 2. Air can be compressed as you stated but water cannot, it never will and never can (auto tech for 10yrs i can tell you what happens when you try to compress water). 3. Theres a reason 99% of builders use a reservoir, and theres a reason why auto techs who know what there doing ADD a reservoir when doing any type of cooling fix/repair/refilling. It looks like this and works like a miracle. Its technically a reservoir for your coolant system (and no the bottle with coolant in your car is not a reservoir but an overflow bottle)



If someone gives you advice take it, as theyve all been there and know what they're doing. You could always buy that funnel and rig it to work in your system provided your radiator has a free port (aka alphacool and so on)
1. Yes most custom loops have some air. In most cases this is in the top of the reservoir where you want it to be. But it still can be compressed.
2. I have a bit more then just a high school science background. So I got this part.
3. That is the only reason I use a reservoir. As a space to store the excess coolant. But I will use the tip of the funnel. I was still looking for a tool to fill my loop. No just to think of what to put in to it.

I do take all normal advice in consideration. But the space I have to work with is limited so I don't know if I can do it all. And the part where I needed advice on I got the least, aka the pressure drop due to the resistance of the components in the loop.
 
#23 ·
What case are you using? Other than aesthetics, I don't really see the point of having reservoirs after the pump. There is no need to store excess coolant, that is not the purpose of the reservoir in a pc cooling application. The reservoir serves to fill and remove air from the loop.

I would personally forgo using a reservoir than place one after the pump. It just doesn't make much sense to me. My next build I am thinking about not using a reservoir, and just using a fill port on a t-line to pump inlet. It would be a very small build as well.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top