Two pumps into one HC...

Can O' Beans · 08-21-2002, 02:10 AM

Well, I think I'm gonna go all-out-insane & do a dual pump/dual loop system.

Loop1 will be a 500gph Danner feeding a homemade high-flow water block. Pump -> Block -> Rad

Loop2 will be either 250 or 500gph Danner(probably the 500) feeding GF4 -> Chipset -> RAM

-> PSU -> Flow sensor/indicator -> HD -> Rad.

_
Basically: (ignore periods for spacing)

Pump1 -> Loop1 -\................/-> Pump1
...........................>Radiator<
Pump2 -> Loop2 -/................\-> Pump2

90% sure there will be a back-lit plexi waterwindow/res in there too, will sit behind my Dragon cutout..., Two systems could dump into the res then into rad. I might put the GF4 into the first system loop though...

Rad is an Escort Heatercore, about 6"x8"x2" core or so, shrouded with fans on both sides.

I'll be getting the new Soyo P4X400 DRAGON Ultra P4 board with a 1.6A P4 or 1.8A P4 at the least, OC'ed as far as it will go.

Can anyone tell me if this HC will handle all this?

BrianH · 08-21-2002, 06:37 AM

I have also built a 2 loop system. Check it out here
http://forums.procooling.com/vbb/sho...&threadid=4046

It was be fun to build. Will you be using one rad or 2. Using only one may be a problems???

8-Ball · 08-21-2002, 07:33 AM

As far as I can tell (don't quote me on this), but they're still using the same water which will be mixing in the heater core. There fore the water temp will stabilise at a uniform temperature.

So what's the advantage?

FoolOnTheHill · 08-21-2002, 08:25 AM

It's not a bad concept, especially to save space, but I can see some potential issues already. One issue would be trying to keep the water flow from both loops fast, and not have them interfere with each other at the rad. Also, wouldn't this setup push water through the rad even faster? If so, this isn't good, because you want a lower flow rate through the rad.

Just thinking out loud...:shrug:

Can O' Beans · 08-21-2002, 10:57 PM

I'm aware the temps would equalize, as I'm gonna be pulling the water from the same res to feed the pumps.

1. - It's a test of a few ideas.
2. - I just feel like going a bit overboard

3. - I wanna see how much crap I can fit into a self contained system and have it actually fit into the case in a somewhat still maintenanceable configuration.

It would cause the water to flow through the radiator even faster, BUT the water going thru each loop would be going a lot faster, meaning it wouldn't be picking up as much heat per unit of water so it doesn't have to stay in the rad as long.

FoolOnTheHill · 08-22-2002, 12:23 AM

I say go for it, at the very worst, you get average tems, and you either get a single loop, or another heater core. Be sure to share pics and results!

jamicon · 08-23-2002, 07:32 AM

im planning on doing the same thing in my case.

the difference being, im going to get a good pump for the CPU, and i got a really crap flow pump for the GPU.

for those who dont fully understand what im on about, its like this:

8-Ball · 08-23-2002, 07:43 AM

Surely it would be better to have on good pump and have say 1/2" tubing going through the cpu block and 3/8" everywhere else. Or maybe eve 1/4". That way you'd still have more flow through the cpu block, and there would be no problem with the powerful pump trying to push water backwards through the weak pump!

8-Ball · 08-23-2002, 07:46 AM

just out of curiosity, is RESERVOIR spelled differently in every english speaking country in the world?:shrug:

jamicon · 08-23-2002, 08:07 AM

oops, spelt reservior wrong in the pic, oh well , its late here

|PuNiSh3R| · 08-23-2002, 01:29 PM

Actually, Foolonthehill

A slow flow rate through the rad proves nothing..

If you have a slow flow rate through the radiator you will have a slow flow through the block.. over all you want a fast flow rate.. There is always water going through the radiator.. so water is always being cooled..

Now a system that WOULD kick ass would be having 1 res... having a fast pump send water through your whole system.. then having another small pump in the reservoir pump water slowly through a radiator.. then back to the reservoir.. see what I mean?

That would work great.. because you'd have a slow flow through the radiator.. yet still have the fast flow through your system..

Skulemate · 08-23-2002, 01:34 PM

But why would you want slow flow through your reservior? They're heat exchangers, just like water blocks, and they like high water velocities just as much.

In that setup you propose all of your water would mix in the reservoir... I don't think it would work too well. I bet that the water temps would be quite a bit worse than a conventional setup... this setup would also hurt the heat exchange efficiency in the rad, and the block as well. No good.

8-Ball · 08-23-2002, 01:58 PM

Punisher, I've heard a lot of people suggest that idea. However, it is flawed, as you are still not moving any more heat.

The water flowing fast through the block will pick up a small amount of energy per unit volume. Conversley the water flowing through the radiator will give out a large amount of energy per unit volume.

So that makes it better right?

Wrong, because you are moving many more unit volumes of water through the block, each picking up a small amount of energy, and it adds up.

I think what people are trying to envisagewhen they say fast through the block and slow through the rad is,

a. cold water into the block to maximise the temp difference between core/block and coolant.
b. lots of time in the rad so as to get rid of all the heat.

The best way to do this is to have either a wide radiator so it moves slowly through the rad, or a long radiator, so it moves at the same speed just in there for longer. Both of these mean having a bigger radiator. That's the only way of having each unit volume of water spend more time in the radiator for each unit of energy it picks up from the block.

8-Ball · 08-23-2002, 02:10 PM

Jamicon, I'm gonna try and answer your question now

First Statement

Would you agree that if there are two opposing pumps in a loop, that water would still flow in the direction of the weaker one.

Second Statement

Thus, the weaker pump is merely acting as an obstacle to flow, much as a waterblock or radiator might.

Make sense so far?

Now look at the two pictures below. The one on the left is the one you have drawn.

Would you agree that they are equivalent?

8-Ball · 08-23-2002, 02:17 PM

Now look at the right hand image.

Surely the water will split at the junction above the cpu block, by an amount representative of the resistances of the two routes, having established that the small weak pump acts as an obstacle to flow.

Although Heatercores have low flow resistance compared to other radiators, they still have a good flow resistance.

As such, water will either flow the wrong way through the small pump or the flow will be very slow. Either way, this will most likley stall the smaller pump or cause it to fail, at which point you would have a lot of water flowing through the right hand circuit which will never go through the Heater core.

Does all of that make sense, or have I wasted my time.

mashie · 08-23-2002, 04:10 PM

8-Ball, BMP images are not designed for internet, please use GIF or JPG...

jamicon · 08-23-2002, 07:52 PM

Thanks 8-Ball

i understand what your getting at now, and i see the problem.

guess i will just have to build a single loop then, as i got no more room (let alone money) for another radiator.

Can O' Beans · 08-24-2002, 05:04 AM

Well, I revised my system.

2x 500GPH Danner mag drives.

Res -> Pump1 -\....................../ ->Loop1 -> Res
.........................>Heatercore<
Res -> Pump2 -/......................\-> Loop2 -> Res

Heatercore will be modified for larger input & output. Pumps will be equally pulling from res, and dumping right into the heatercore. The output side of the HC will have a Y on it, splitting to each loop, then both loops would dump into the rad.

This would equalize any flow/pressure issues between the pumps.

I don't have the 2nd pump yet, so one question:

Has anyone tested whether 2 pumps are better in series or in parallel? And would one be better than the other in certain low/high flow rate, etc??

Skulemate · 08-24-2002, 08:48 AM

Well, considering the characteristics of most mag drive pumps that we're using (high flow, low head) I'd say that a series setup would be the most efficient.

bigben2k · 08-24-2002, 10:39 AM

Quote:

Originally posted by Can O' Beans
I don't have the 2nd pump yet, so one question:

Has anyone tested whether 2 pumps are better in series or in parallel? And would one be better than the other in certain low/high flow rate, etc??

Yeah, we went through that in another thread: some guy who used to be in the military kindly reminded us that centrifugal pumps work better in series. The 2nd pump acts as a booster pump.

check it out The Navy man is Airspirit.

Can O' Beans · 08-25-2002, 12:38 AM

Thanks for the info.

I knew about the basic pros & cons, but didn't know the extent of each...

MadDogMe · 08-26-2002, 09:05 AM

Seems kinda pointless without two rads :shrug: .

I was thinking of two pumps but a weak pump pumping water from a resevoir into a rad and back to the resivoir, and an Eheim 1250 pumping from the res through the CPU block and back to the res'. The thing is I think it'll only work if I get a certain balance. otherwise the CPU will heat the water faster than the rad is cooling it.
Best to have two weak pumps that match the flow of the Eheim?.and two rads?. even then it'd be difficult cause I'd want actual flow rate of both single & double pumps, not theoretical. Well the two weak pumps having more flow than the 1250 would'nt matter, it'd be a boon!.

Sugestions?. Criticisms?. Bad~brain thinking corrections?

...

8-Ball · 08-26-2002, 04:53 PM

maddog,

I tried to explain the exact same set up a few posts up, as you may have noticed if you had read the thread.

Punisher originally put the idea forward.

I'm not saying I'm definitely right but it seems to make sense.

To repeat, you can optimise the flow ratefor the components and the coolant. beyond that you need a better components. The, way you set the components up makes little difference.

8-ball

08-21-2002, 02:10 AM	#1
Can O' Beans Cooling Savant Join Date: Aug 2002 Location: Pa - USA Posts: 264	Two pumps into one HC... Well, I think I'm gonna go all-out-insane & do a dual pump/dual loop system. Loop1 will be a 500gph Danner feeding a homemade high-flow water block. Pump -> Block -> Rad Loop2 will be either 250 or 500gph Danner(probably the 500) feeding GF4 -> Chipset -> RAM -> PSU -> Flow sensor/indicator -> HD -> Rad. _ Basically: (ignore periods for spacing) Pump1 -> Loop1 -\................/-> Pump1 ...........................>Radiator< Pump2 -> Loop2 -/................\-> Pump2 90% sure there will be a back-lit plexi waterwindow/res in there too, will sit behind my Dragon cutout..., Two systems could dump into the res then into rad. I might put the GF4 into the first system loop though... Rad is an Escort Heatercore, about 6"x8"x2" core or so, shrouded with fans on both sides. I'll be getting the new Soyo P4X400 DRAGON Ultra P4 board with a 1.6A P4 or 1.8A P4 at the least, OC'ed as far as it will go. Can anyone tell me if this HC will handle all this?

08-23-2002, 01:34 PM	#12
Skulemate Cooling Savant Join Date: Jul 2002 Location: Toronto, Canada Posts: 381	But why would you want slow flow through your reservior? They're heat exchangers, just like water blocks, and they like high water velocities just as much. In that setup you propose all of your water would mix in the reservoir... I don't think it would work too well. I bet that the water temps would be quite a bit worse than a conventional setup... this setup would also hurt the heat exchange efficiency in the rad, and the block as well. No good. __________________ Michael E. Robbins M.A.Sc. Candidate, University of Toronto 12.1 GHz of AMD's finest (17.7 GHz total) crunching proudly for the AMDMB.com Killer Frogs SETI BOINC: Dual Opteron 246s (Iwill DK8N) \| XP2800+ (Shuttle SN41G2) \| 3x XP2400+ (ASUS A7N266-vm) SETI BOINC: 2x P4 2.8E (ASUS P4R800-vm) \| Crunching 24/7

08-23-2002, 04:10 PM	#16
mashie Cooling Savant Join Date: Dec 2001 Location: Reading, UK Posts: 149	8-Ball, BMP images are not designed for internet, please use GIF or JPG... __________________ mashie

08-24-2002, 08:48 AM	#19
Skulemate Cooling Savant Join Date: Jul 2002 Location: Toronto, Canada Posts: 381	Well, considering the characteristics of most mag drive pumps that we're using (high flow, low head) I'd say that a series setup would be the most efficient. __________________ Michael E. Robbins M.A.Sc. Candidate, University of Toronto 12.1 GHz of AMD's finest (17.7 GHz total) crunching proudly for the AMDMB.com Killer Frogs SETI BOINC: Dual Opteron 246s (Iwill DK8N) \| XP2800+ (Shuttle SN41G2) \| 3x XP2400+ (ASUS A7N266-vm) SETI BOINC: 2x P4 2.8E (ASUS P4R800-vm) \| Crunching 24/7

08-21-2002, 06:37 AM	#2
BrianH Cooling Savant Join Date: May 2002 Location: Ottawa, Ont, Canada Posts: 115	I have also built a 2 loop system. Check it out here http://forums.procooling.com/vbb/sho...&threadid=4046 It was be fun to build. Will you be using one rad or 2. Using only one may be a problems???

08-21-2002, 07:33 AM	#3
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	As far as I can tell (don't quote me on this), but they're still using the same water which will be mixing in the heater core. There fore the water temp will stabilise at a uniform temperature. So what's the advantage?

08-21-2002, 08:25 AM	#4
FoolOnTheHill Cooling Neophyte Join Date: Aug 2001 Posts: 30	It's not a bad concept, especially to save space, but I can see some potential issues already. One issue would be trying to keep the water flow from both loops fast, and not have them interfere with each other at the rad. Also, wouldn't this setup push water through the rad even faster? If so, this isn't good, because you want a lower flow rate through the rad. Just thinking out loud...:shrug:

08-21-2002, 10:57 PM	#5
Can O' Beans Cooling Savant Join Date: Aug 2002 Location: Pa - USA Posts: 264	I'm aware the temps would equalize, as I'm gonna be pulling the water from the same res to feed the pumps. 1. - It's a test of a few ideas. 2. - I just feel like going a bit overboard 3. - I wanna see how much crap I can fit into a self contained system and have it actually fit into the case in a somewhat still maintenanceable configuration. It would cause the water to flow through the radiator even faster, BUT the water going thru each loop would be going a lot faster, meaning it wouldn't be picking up as much heat per unit of water so it doesn't have to stay in the rad as long.

08-22-2002, 12:23 AM	#6
FoolOnTheHill Cooling Neophyte Join Date: Aug 2001 Posts: 30	I say go for it, at the very worst, you get average tems, and you either get a single loop, or another heater core. Be sure to share pics and results!

08-23-2002, 07:32 AM	#7
jamicon Cooling Neophyte Join Date: Aug 2002 Location: Australia Posts: 44	im planning on doing the same thing in my case. the difference being, im going to get a good pump for the CPU, and i got a really crap flow pump for the GPU. for those who dont fully understand what im on about, its like this:

08-23-2002, 07:43 AM	#8
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	Surely it would be better to have on good pump and have say 1/2" tubing going through the cpu block and 3/8" everywhere else. Or maybe eve 1/4". That way you'd still have more flow through the cpu block, and there would be no problem with the powerful pump trying to push water backwards through the weak pump!

08-23-2002, 07:46 AM	#9
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	just out of curiosity, is RESERVOIR spelled differently in every english speaking country in the world?:shrug:

08-23-2002, 08:07 AM	#10
jamicon Cooling Neophyte Join Date: Aug 2002 Location: Australia Posts: 44	oops, spelt reservior wrong in the pic, oh well , its late here

08-23-2002, 01:29 PM	#11
\|PuNiSh3R\| Foo's Been Banned Join Date: May 2002 Location: Pennsyltucky Posts: 255	Actually, Foolonthehill A slow flow rate through the rad proves nothing.. If you have a slow flow rate through the radiator you will have a slow flow through the block.. over all you want a fast flow rate.. There is always water going through the radiator.. so water is always being cooled.. Now a system that WOULD kick ass would be having 1 res... having a fast pump send water through your whole system.. then having another small pump in the reservoir pump water slowly through a radiator.. then back to the reservoir.. see what I mean? That would work great.. because you'd have a slow flow through the radiator.. yet still have the fast flow through your system..

08-23-2002, 01:58 PM	#13
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	Punisher, I've heard a lot of people suggest that idea. However, it is flawed, as you are still not moving any more heat. The water flowing fast through the block will pick up a small amount of energy per unit volume. Conversley the water flowing through the radiator will give out a large amount of energy per unit volume. So that makes it better right? Wrong, because you are moving many more unit volumes of water through the block, each picking up a small amount of energy, and it adds up. I think what people are trying to envisagewhen they say fast through the block and slow through the rad is, a. cold water into the block to maximise the temp difference between core/block and coolant. b. lots of time in the rad so as to get rid of all the heat. The best way to do this is to have either a wide radiator so it moves slowly through the rad, or a long radiator, so it moves at the same speed just in there for longer. Both of these mean having a bigger radiator. That's the only way of having each unit volume of water spend more time in the radiator for each unit of energy it picks up from the block.

08-23-2002, 02:17 PM	#15
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	Now look at the right hand image. Surely the water will split at the junction above the cpu block, by an amount representative of the resistances of the two routes, having established that the small weak pump acts as an obstacle to flow. Although Heatercores have low flow resistance compared to other radiators, they still have a good flow resistance. As such, water will either flow the wrong way through the small pump or the flow will be very slow. Either way, this will most likley stall the smaller pump or cause it to fail, at which point you would have a lot of water flowing through the right hand circuit which will never go through the Heater core. Does all of that make sense, or have I wasted my time.

08-23-2002, 07:52 PM	#17
jamicon Cooling Neophyte Join Date: Aug 2002 Location: Australia Posts: 44	Thanks 8-Ball i understand what your getting at now, and i see the problem. guess i will just have to build a single loop then, as i got no more room (let alone money) for another radiator.

08-24-2002, 05:04 AM	#18
Can O' Beans Cooling Savant Join Date: Aug 2002 Location: Pa - USA Posts: 264	Well, I revised my system. 2x 500GPH Danner mag drives. Res -> Pump1 -\....................../ ->Loop1 -> Res .........................>Heatercore< Res -> Pump2 -/......................\-> Loop2 -> Res Heatercore will be modified for larger input & output. Pumps will be equally pulling from res, and dumping right into the heatercore. The output side of the HC will have a Y on it, splitting to each loop, then both loops would dump into the rad. This would equalize any flow/pressure issues between the pumps. I don't have the 2nd pump yet, so one question: Has anyone tested whether 2 pumps are better in series or in parallel? And would one be better than the other in certain low/high flow rate, etc??

08-25-2002, 12:38 AM	#21
Can O' Beans Cooling Savant Join Date: Aug 2002 Location: Pa - USA Posts: 264	Thanks for the info. I knew about the basic pros & cons, but didn't know the extent of each...

08-26-2002, 09:05 AM	#22
MadDogMe Thermophile Join Date: Aug 2002 Location: Just shut up ;) ... Posts: 1,068	Seems kinda pointless without two rads :shrug: . I was thinking of two pumps but a weak pump pumping water from a resevoir into a rad and back to the resivoir, and an Eheim 1250 pumping from the res through the CPU block and back to the res'. The thing is I think it'll only work if I get a certain balance. otherwise the CPU will heat the water faster than the rad is cooling it. Best to have two weak pumps that match the flow of the Eheim?.and two rads?. even then it'd be difficult cause I'd want actual flow rate of both single & double pumps, not theoretical. Well the two weak pumps having more flow than the 1250 would'nt matter, it'd be a boon!. Sugestions?. Criticisms?. Bad~brain thinking corrections? ...

08-26-2002, 04:53 PM	#23
8-Ball Cooling Savant Join Date: Feb 2002 Location: Oxford University, UK Posts: 452	maddog, I tried to explain the exact same set up a few posts up, as you may have noticed if you had read the thread. Punisher originally put the idea forward. I'm not saying I'm definitely right but it seems to make sense. To repeat, you can optimise the flow ratefor the components and the coolant. beyond that you need a better components. The, way you set the components up makes little difference. 8-ball

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