Single pump or dual pumps?

niner · 01-30-2002, 07:19 AM

Hi all,

First post in this forum.

I'm putting togather another WC computer. I'm going to need a lot of water flow in this setup. My question: is it better to have a single big pump, 500gph Eheim 1060, or a couple of smaller pump, 300gph Eheim 1025, configured as push-pull? The setup will be in-line with 1/2" connection throughout.

Thanks.

gmat · 01-30-2002, 07:45 AM

Using 2 pumps you may have a better output pressure, and less tear on the pump mechanics.
However the solution is more expensive and of course you need the real estate for 2 pumps instead of just 1....
All in all it depends on your setup. How many loops ? How many waterblocks ?

niner · 01-30-2002, 08:36 AM

It will be a single loop, a single waterblock with a good size Mocal radiator. Water is used to cool a dual peliters setup.

mashie · 01-30-2002, 08:36 AM

Hmm, sounds like an interesting setup that requires a 500gph pump... please give us some information about the remaining system.

mashie · 01-30-2002, 08:39 AM

Quote:

Originally posted by niner
It will be a single loop, a single waterblock with a good size Mocal radiator. Water is used to cool a dual peliters setup.

A single Eheim 1250 should be enough since you are running 1/2"

niner · 01-30-2002, 09:21 AM

I tried a single Eheim 1250 in this setup, can push less than 40gph. The Mocal radiator needs at least 90gph to reach its maximum efficience. Of course, waterblock just loves higher flow rate. That's why I think I need a really powerfull pump, or two.

Butcher · 01-30-2002, 09:57 AM

40 gph on a 1250? damn I get 75 gph on my 1048 you should really considere different block / rad not pump...

niner · 01-30-2002, 10:28 AM

Butcher,

Do you measure the flow rate in the case or out of the case? I put the radiator on the top of a full tower case, which is almost 2' higher than the pump at the bottom. 2' of height along would cut the flow rate in half.

gmat · 01-30-2002, 10:40 AM

Yep, make sure your whole water circuit is not too high. Where is the lowest point and where is the highest point ? Add elbows and fittings, waterblocks and turns to this height.
Those small mag-drive pumps aren't designed to raise the water - they produce a real small backpressure.

mashie · 01-30-2002, 03:29 PM

As long as you are running a closed loop it doesn't matter at all how high it is since it is exactly the same pressure caused by gravity on both sides of the pump. The water is _circulated_ not lifted. Some myths just don't want to die...

DigitalChaos · 01-30-2002, 03:39 PM

Quote:

Originally posted by mashie
As long as you are running a closed loop it doesn't matter at all how high it is since it is exactly the same pressure caused by gravity on both sides of the pump. The water is _circulated_ not lifted. Some myths just don't want to die...

exactly

you can try this: get like... 20 feet of tubing and make a loop with your pump... lay the tubing on the ground with the pump (0ft lift) then lift up the tubing to give several feet lift... there will be NO change in flow rate. (maybe a very non measureable ammount due to the slight elasticity of water.. and the slight bulge on the tubing from gravity.

niner · 01-30-2002, 03:52 PM

Well, my setup is not truely close-loop. It has a reservoir with the pump being external. May be that's my problem? Do you think it's a good idea to remove the reservior? I had a close-loop setup before, and it's a real pain to fill and remove air.

mashie · 01-30-2002, 04:12 PM

Make it inline, just put a T in a good location so it will bleed of your system and also gives you a place to fill it.

My inline system was free from air in a couple of hours.

Brad · 01-30-2002, 08:58 PM

It might be an idea to have one pump before the cpu block and one after, and see how your temps go compared to just one pump

futRtrubL · 01-30-2002, 11:15 PM

As long as you aren't pumping air around the system it makes no difference whether there is a res or not, water going up = water going down still.

Edward

Quote:

Originally posted by niner
Well, my setup is not truely close-loop. It has a reservoir with the pump being external. May be that's my problem? Do you think it's a good idea to remove the reservior? I had a close-loop setup before, and it's a real pain to fill and remove air.

niner · 01-31-2002, 06:45 AM

At the moment, I have single OASE Aquarius 8 in my system. It works out pretty good. The CPU idle at 32c, system at 22c. The Athlon is running very stable at 1896MHz (12x158) with 2.15 volt. I want to go overkill and see how far I can push it. I just ordered another Aquarius 8 to give it a try. I'll report back if it does anything to the temp and flow rate.

Right now my system contains:
Athlon XP 1900+ at 1896MHz
Innovatek waterblock
ASII, waiting for ASIII to arrive
1/2" Si tube
Abit KR7A RAID
PC Power & Cool 425ATX
OASE Aquarius 8
Mocal 14"x6"x2" radiator
Lian_li PC70 case

In the future, I would like to try dual-peltier setup.

gmat · 01-31-2002, 10:53 AM

Digitalchaos: a myth ???

Did u try the very experience you described ?
Take a pump with an indicated water height of 5ft for example.
Put a hose on the output and the pump in its own water (or the input in a large basin). Make sure the water comes down to its origin.
Then raise the hose manually up to 5ft. No flow. Try it. It's not important where the hose finishes, for example let it lay down in the basin while you raise its middle part.
Closed-loop or not, it's not the problem.
The only thing that counts is "total water height". If you exceed that height the pump "cavitates". Cavitation is well known by house owners who want to bring water from their basement to their 2nd floor, it produces a 'klang' in pipes when you open/close a tap. To raise water further you need extra pumps at intermediate heights, even with straight pipes !
Moreover elbows and fittings *do* add height to the circuit.
I learned those things during engineering school (fluid mechanics courses), and i tried this experience at home, just to make sure....

So ppl who put their mega-tower on their desk, and their radiator somwhere on the floor on in their basement, need to consider seriously this total "water height".
[edit] oops i forgot:
* gravity still affects water in a closed-loop. If not, i WANT your anti-grav system !
* there -is- a slight overpressure at the outlet and an underpressure at the inlet. This pressure differential is weak on little magdrives we use in watecooling, but still exists. Joe's wonderful pump produces a very high output pressure, with such a pump you'd better check your fittings against leaks...
* You can see this pressure differential if your res is transparent, and placed near the outlet or near the inlet.

Butcher · 01-31-2002, 02:35 PM

Quote:

Originally posted by niner
Butcher,

Do you measure the flow rate in the case or out of the case? I put the radiator on the top of a full tower case, which is almost 2' higher than the pump at the bottom. 2' of height along would cut the flow rate in half.

It was measured with the system installed. I have about 12-15 inches of total lift on my system as I run external rad / pump.

gmat · 01-31-2002, 02:54 PM

Whats the height of the Lian Li PC70 ? With the pump at bottom and rad over the top that may be already a pretty good height. If you have 1/2" tubing all the way it's good though, it won't add much lift.
If you can try to put one pump at bottom and one pump at top.
Or each pump at opposite parts of the circuit, then to compare this with both pumps next to each other...
Results may be very interesting

futRtrubL · 01-31-2002, 03:26 PM

That is a bad example as you are using an open system. Not only that your example uses enough hight that a high presure pump must be used, high pressure enough that the pressure at the inlet side of the pump blade is reduced so much that water can boil a room temp (the cause of cavitation, unless you talk boats in which case it can mean air ataching itself to the prop).
The use of the word hieght is confusing to those who don't work with pumps etc. I will use hieght to mean distance vertically, what elbows etc. do is add resistance to flow.
Lets do an even easier demonstration, one that almost everyone has done before. All you need is a bucket and a hose, put one end of the hose in the bucket, take the hose over the rim of the bucket or higher (this is your water height) no take the other end below the first end and suck on it. Water will flow very easily, your water height is not making it difficult. In this case it will be dificult to start as we are raising the water at first with none going down. This time submerge the entire hose in the bucket so that there is no air left in it. Block one end and bring that over the lip down below the first end, it will flow without help, no pump. Distance down helps while distance up hinders, they will both cancel in a closed loop system. The only thing that will affect it is hose length and the resistance it creates. If you found that you got different resistances with hight make sure that you aren't bending the hose any more than it was on the ground.
Nobody said there wasn't a pressure differential at inlet and outlet.
People who have there comp in their office and there rad in the basement need to think about hose length.

Edward

GMAT wrote:

Cavitation is well known by house owners who want to bring water from their basement to their 2nd floor, it produces a 'klang' in pipes when you open/close a tap. To raise water further you need extra pumps at intermediate heights, even with straight pipes !
Moreover elbows and fittings *do* add height to the circuit.
I learned those things during engineering school (fluid mechanics courses), and i tried this experience at home, just to make sure....

So ppl who put their mega-tower on their desk, and their radiator somwhere on the floor on in their basement, need to consider seriously this total "water height".
[edit] oops i forgot:
* gravity still affects water in a closed-loop. If not, i WANT your anti-grav system !
* there -is- a slight overpressure at the outlet and an underpressure at the inlet. This pressure differential is weak on little magdrives we use in watecooling, but still exists. Joe's wonderful pump produces a very high output pressure, with such a pump you'd better check your fittings against leaks...
* You can see this pressure differential if your res is transparent, and placed near the outlet or near the inlet. [/b][/quote]

gmat · 01-31-2002, 04:49 PM

Flow resistance = water height. Strictly. I insist. (its not the place for a physics course so i wont argue over this)
I know, powerful pumps out there can bear like 11m or even 13m of water.
But not tiny weeny aquarium pumps we usually put in our comps. And there, having a 1m high tower case counts. No matter how you put it you've got 1m of water over your pump outlet (if it's at the bottom). Specially if you have 1/2" or wider all around. And the pressure of this water column will bring down your pump performance. Even in a sealed-closed loop. The pump has to start up and get things moving... And "centrifuge" pumps are not designed to withstand even moderate pressures.

That's why some ppl like Joe buy such powerful pumps.
They're less concerned by water height than by leak prevention... Those pumps dont produce more flow (on paper...) but have a great pressure differential.
My initial point was, putting 2 pumps in the circuit would help reduce that differential and therefore raise flow.

Oops i forgot. Air traps, resevoirs - most ppl have those. And ususally at the highest point. How do they count ? Open loop? Closed loop ?

mashie · 01-31-2002, 06:27 PM

Quote:

Originally posted by gmat
Flow resistance = water height. Strictly. I insist. (its not the place for a physics course so i wont argue over this)
I know, powerful pumps out there can bear like 11m or even 13m of water.
But not tiny weeny aquarium pumps we usually put in our comps. And there, having a 1m high tower case counts. No matter how you put it you've got 1m of water over your pump outlet (if it's at the bottom). Specially if you have 1/2" or wider all around. And the pressure of this water column will bring down your pump performance. Even in a sealed-closed loop. The pump has to start up and get things moving... And "centrifuge" pumps are not designed to withstand even moderate pressures.

That's why some ppl like Joe buy such powerful pumps.
They're less concerned by water height than by leak prevention... Those pumps dont produce more flow (on paper...) but have a great pressure differential.
My initial point was, putting 2 pumps in the circuit would help reduce that differential and therefore raise flow.

Oops i forgot. Air traps, resevoirs - most ppl have those. And ususally at the highest point. How do they count ? Open loop? Closed loop ?

gmat, time for a little picture...

Black square = pump

In the left example you are actually lifting the water to another level since it won't flow back and that will require a strong pump.

In the right example you see that there are no lifting involved at all since it is exactly the same pressure on both inlet and outlet of the pump created by gravity. The work the pump has to do is to move the water from left to right (which means it will go up on one side and down on the other).

The diameter of the tube is of absolutely no importance for creating water pressure it's only the height that decides = a 1" and a 1' pipe full of water creates exactly the same pressure per square inch as long as they have the same height.

Regarding reservoirs and so at the top, as long as the water goes back to the pump again is it considered a closed loop (right example)

The reason most people buy oversized pumps are to compensate for smaller things

Now please tell me which part you don't understand.

Brad · 01-31-2002, 08:17 PM

Quote:

Originally posted by mashie

The reason most people buy oversized pumps are to compensate for smaller things

Now please tell me which part you don't understand.

The only thing I don't understand is the smaller thing part. Are you talking about a slower cpu?

futRtrubL · 01-31-2002, 09:27 PM

Couldn't have said it better myself mashie.

Edward

gmat · 02-01-2002, 02:18 AM

Well, i didnt talk about the left drawing. What is this basin placed at top ????
Just let the water flow all the way back to the pump. With or without a hose, your choice.
With a weak pump the effect is noticeable enough.
With a stronger pump add elbows, waterblocks and so on, and the effect will still be noticeable.

01-30-2002, 07:19 AM	#1
niner Cooling Neophyte Join Date: Jan 2002 Location: Boston, MA Posts: 7	Single pump or dual pumps? Hi all, First post in this forum. I'm putting togather another WC computer. I'm going to need a lot of water flow in this setup. My question: is it better to have a single big pump, 500gph Eheim 1060, or a couple of smaller pump, 300gph Eheim 1025, configured as push-pull? The setup will be in-line with 1/2" connection throughout. Thanks.

01-30-2002, 08:36 AM	#4
mashie Cooling Savant Join Date: Dec 2001 Location: Reading, UK Posts: 149	Hmm, sounds like an interesting setup that requires a 500gph pump... please give us some information about the remaining system. __________________ mashie

01-30-2002, 09:57 AM	#7
Butcher Thermophile Join Date: May 2001 Location: UK Posts: 1,064	40 gph on a 1250? damn I get 75 gph on my 1048 you should really considere different block / rad not pump... __________________ Once upon a time, in a land far far away...

01-30-2002, 03:29 PM	#10
mashie Cooling Savant Join Date: Dec 2001 Location: Reading, UK Posts: 149	As long as you are running a closed loop it doesn't matter at all how high it is since it is exactly the same pressure caused by gravity on both sides of the pump. The water is _circulated_ not lifted. Some myths just don't want to die... __________________ mashie

01-30-2002, 04:12 PM	#13
mashie Cooling Savant Join Date: Dec 2001 Location: Reading, UK Posts: 149	Make it inline, just put a T in a good location so it will bleed of your system and also gives you a place to fill it. My inline system was free from air in a couple of hours. __________________ mashie

01-30-2002, 07:45 AM	#2
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Using 2 pumps you may have a better output pressure, and less tear on the pump mechanics. However the solution is more expensive and of course you need the real estate for 2 pumps instead of just 1.... All in all it depends on your setup. How many loops ? How many waterblocks ?

01-30-2002, 09:21 AM	#6
niner Cooling Neophyte Join Date: Jan 2002 Location: Boston, MA Posts: 7	I tried a single Eheim 1250 in this setup, can push less than 40gph. The Mocal radiator needs at least 90gph to reach its maximum efficience. Of course, waterblock just loves higher flow rate. That's why I think I need a really powerfull pump, or two.

01-30-2002, 10:28 AM	#8
niner Cooling Neophyte Join Date: Jan 2002 Location: Boston, MA Posts: 7	Butcher, Do you measure the flow rate in the case or out of the case? I put the radiator on the top of a full tower case, which is almost 2' higher than the pump at the bottom. 2' of height along would cut the flow rate in half.

01-30-2002, 10:40 AM	#9
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Yep, make sure your whole water circuit is not too high. Where is the lowest point and where is the highest point ? Add elbows and fittings, waterblocks and turns to this height. Those small mag-drive pumps aren't designed to raise the water - they produce a real small backpressure.

01-30-2002, 03:52 PM	#12
niner Cooling Neophyte Join Date: Jan 2002 Location: Boston, MA Posts: 7	Well, my setup is not truely close-loop. It has a reservoir with the pump being external. May be that's my problem? Do you think it's a good idea to remove the reservior? I had a close-loop setup before, and it's a real pain to fill and remove air.

01-30-2002, 08:58 PM	#14
Brad Thermophile Join Date: Oct 2001 Location: Nuu Zeeelin Posts: 3,175	It might be an idea to have one pump before the cpu block and one after, and see how your temps go compared to just one pump __________________ 2x P3 1100's at 1400, Abit VP6, 2x Corsair 256mb PC150 sticks, 20gb 'cuda ATA-III, 2x 40gb 'cuda ATA-IV in raid 0. 20" Trinitron. No fans 2x 2400+ at 2288mhz (16.0 x 143), Iwill MPX2, 2x Kingmax PC-3200 256mb sticks, 4x 20gb 60gxp in Raid 5 on a Promise SX6000. Asus Ti4200 320/630. Cooled by Water

01-31-2002, 06:45 AM	#16
niner Cooling Neophyte Join Date: Jan 2002 Location: Boston, MA Posts: 7	At the moment, I have single OASE Aquarius 8 in my system. It works out pretty good. The CPU idle at 32c, system at 22c. The Athlon is running very stable at 1896MHz (12x158) with 2.15 volt. I want to go overkill and see how far I can push it. I just ordered another Aquarius 8 to give it a try. I'll report back if it does anything to the temp and flow rate. Right now my system contains: Athlon XP 1900+ at 1896MHz Innovatek waterblock ASII, waiting for ASIII to arrive 1/2" Si tube Abit KR7A RAID PC Power & Cool 425ATX OASE Aquarius 8 Mocal 14"x6"x2" radiator Lian_li PC70 case In the future, I would like to try dual-peltier setup.

01-31-2002, 10:53 AM	#17
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Digitalchaos: a myth ??? Did u try the very experience you described ? Take a pump with an indicated water height of 5ft for example. Put a hose on the output and the pump in its own water (or the input in a large basin). Make sure the water comes down to its origin. Then raise the hose manually up to 5ft. No flow. Try it. It's not important where the hose finishes, for example let it lay down in the basin while you raise its middle part. Closed-loop or not, it's not the problem. The only thing that counts is "total water height". If you exceed that height the pump "cavitates". Cavitation is well known by house owners who want to bring water from their basement to their 2nd floor, it produces a 'klang' in pipes when you open/close a tap. To raise water further you need extra pumps at intermediate heights, even with straight pipes ! Moreover elbows and fittings do add height to the circuit. I learned those things during engineering school (fluid mechanics courses), and i tried this experience at home, just to make sure.... So ppl who put their mega-tower on their desk, and their radiator somwhere on the floor on in their basement, need to consider seriously this total "water height". [edit] oops i forgot: * gravity still affects water in a closed-loop. If not, i WANT your anti-grav system ! * there -is- a slight overpressure at the outlet and an underpressure at the inlet. This pressure differential is weak on little magdrives we use in watecooling, but still exists. Joe's wonderful pump produces a very high output pressure, with such a pump you'd better check your fittings against leaks... * You can see this pressure differential if your res is transparent, and placed near the outlet or near the inlet. Last edited by gmat; 01-31-2002 at 11:08 AM.

01-31-2002, 02:54 PM	#19
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Whats the height of the Lian Li PC70 ? With the pump at bottom and rad over the top that may be already a pretty good height. If you have 1/2" tubing all the way it's good though, it won't add much lift. If you can try to put one pump at bottom and one pump at top. Or each pump at opposite parts of the circuit, then to compare this with both pumps next to each other... Results may be very interesting

01-31-2002, 03:26 PM	#20
futRtrubL Cooling Savant Join Date: Jul 2001 Location: Kingston, Jamaica Posts: 204	That is a bad example as you are using an open system. Not only that your example uses enough hight that a high presure pump must be used, high pressure enough that the pressure at the inlet side of the pump blade is reduced so much that water can boil a room temp (the cause of cavitation, unless you talk boats in which case it can mean air ataching itself to the prop). The use of the word hieght is confusing to those who don't work with pumps etc. I will use hieght to mean distance vertically, what elbows etc. do is add resistance to flow. Lets do an even easier demonstration, one that almost everyone has done before. All you need is a bucket and a hose, put one end of the hose in the bucket, take the hose over the rim of the bucket or higher (this is your water height) no take the other end below the first end and suck on it. Water will flow very easily, your water height is not making it difficult. In this case it will be dificult to start as we are raising the water at first with none going down. This time submerge the entire hose in the bucket so that there is no air left in it. Block one end and bring that over the lip down below the first end, it will flow without help, no pump. Distance down helps while distance up hinders, they will both cancel in a closed loop system. The only thing that will affect it is hose length and the resistance it creates. If you found that you got different resistances with hight make sure that you aren't bending the hose any more than it was on the ground. Nobody said there wasn't a pressure differential at inlet and outlet. People who have there comp in their office and there rad in the basement need to think about hose length. Edward GMAT wrote: Cavitation is well known by house owners who want to bring water from their basement to their 2nd floor, it produces a 'klang' in pipes when you open/close a tap. To raise water further you need extra pumps at intermediate heights, even with straight pipes ! Moreover elbows and fittings do add height to the circuit. I learned those things during engineering school (fluid mechanics courses), and i tried this experience at home, just to make sure.... So ppl who put their mega-tower on their desk, and their radiator somwhere on the floor on in their basement, need to consider seriously this total "water height". [edit] oops i forgot: * gravity still affects water in a closed-loop. If not, i WANT your anti-grav system ! * there -is- a slight overpressure at the outlet and an underpressure at the inlet. This pressure differential is weak on little magdrives we use in watecooling, but still exists. Joe's wonderful pump produces a very high output pressure, with such a pump you'd better check your fittings against leaks... * You can see this pressure differential if your res is transparent, and placed near the outlet or near the inlet. [/b][/quote]

01-31-2002, 04:49 PM	#21
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Flow resistance = water height. Strictly. I insist. (its not the place for a physics course so i wont argue over this) I know, powerful pumps out there can bear like 11m or even 13m of water. But not tiny weeny aquarium pumps we usually put in our comps. And there, having a 1m high tower case counts. No matter how you put it you've got 1m of water over your pump outlet (if it's at the bottom). Specially if you have 1/2" or wider all around. And the pressure of this water column will bring down your pump performance. Even in a sealed-closed loop. The pump has to start up and get things moving... And "centrifuge" pumps are not designed to withstand even moderate pressures. That's why some ppl like Joe buy such powerful pumps. They're less concerned by water height than by leak prevention... Those pumps dont produce more flow (on paper...) but have a great pressure differential. My initial point was, putting 2 pumps in the circuit would help reduce that differential and therefore raise flow. Oops i forgot. Air traps, resevoirs - most ppl have those. And ususally at the highest point. How do they count ? Open loop? Closed loop ? Last edited by gmat; 01-31-2002 at 05:22 PM.

01-31-2002, 09:27 PM	#24
futRtrubL Cooling Savant Join Date: Jul 2001 Location: Kingston, Jamaica Posts: 204	Couldn't have said it better myself mashie. Edward

02-01-2002, 02:18 AM	#25
gmat Thermophile Join Date: Mar 2001 Location: France Posts: 1,221	Well, i didnt talk about the left drawing. What is this basin placed at top ???? Just let the water flow all the way back to the pump. With or without a hose, your choice. With a weak pump the effect is noticeable enough. With a stronger pump add elbows, waterblocks and so on, and the effect will still be noticeable.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)