• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Question about a Dual Pump setup

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

IFMU

The Xtreme Senior Nobody
Joined
Jun 21, 2001
Ok, I have an idea for a new styled waterblock. Ive made a very lame display model, mostly so I can get it more physical to check out and test some. So, due to the way its made it will need to have an external line for the heater core. i.e. 2 pumps would be needed.
What I was wondering was what would the effects be to have a res, with dual pumps (1 for CPU cooling, the 2nd for the heater core) be?
I know the water would mix and circulate some of the warmer water back into the CPU pump. Which I can fix some of that issue with the way the res would be laid out internally. So that I am not to worried about.
Would it be better to have the dual pumps different?
Have the pump that is hooked to the heater core be a low GPH pump, and the one to the CPU be slightly higher.

The reason behind that is I was thinking with the lower GPH pump pushing water through the heater core it would have more time to cool water.
Then the higher, would be slightly faster and be able to put the water into the CPU in and out faster.

Thoughts? Did I make myself understandable? LoL

Once I start working on the CPU Block Ill get up some images, but this "model" is just way to .... generic to show. LoL
 
i dont get the RAD part....... but the rest i understand......
Do you want to have both pumps pump to ONE raditor?
I would just use 2..... and use 2 resivors~ or one big one divided and sealed into two~
 
Hrm... Ok, 1 Res, 2 pumps inside.
1 pump goes to the Heater Core, the 2nd goes to the CPU.
 
a basic picture may help some.

so you're having a master res, from which water is pumped to the cpu and the heater core (which is in parallel with the cpu) ?

I'm thinking you'd want high gph pumps on each. I think the higher the flow the better the cooling. Just put a good fan on the heater core.


edit: wow, two posts while I was typing my one....
 
Last edited:
OHHH i get it!!! One pump moves the water threw the CPU loop back to the resivor (side with the heatercore pump) and the second pump cirulates water threw the heatercore and returns back to the pump side with the cpu pump, providing cooled water by using a lower GPH pump to better cool the water!
 
Figures, I explain like crap.. heh.. Ok, well heres a pic.
Res.JPG

Link
www.members.cox.net/ifmu/Res.JPG
 
The Spyder said:
OHHH i get it!!! One pump moves the water threw the CPU loop back to the resivor (side with the heatercore pump) and the second pump cirulates water threw the heatercore and returns back to the pump side with the cpu pump, providing cooled water by using a lower GPH pump to better cool the water!
YES! LoL
Yea, thats basiclly what I mean.
Would that cool the water better with a lower GPH pump? So it has more time inside the Rad?
Then keep a standard/faster GPH pump for the CPU.
 
WC.JPG

i was bored :) with this split resivor, you will get better cooling. Hot goes cold, while supplying low and high GPH :-D good luck!
 
Yea thats basiclly it. The split would be a little more difficult due to the way Im laying out the WB, but thats the basic Idea.
Small drawback to having fully 2 physical different res's. If there were to be 2 different speed pumps, the faster would overload the slowers water holding capabilites.
There would have to be an opening between the 2 so the extra being pushed into that one, would have some place to escape.
So, would this help keep the water cooler?
The slower pump on the heater core?
 
If it we're up to me, I'd hook both pumps up in series and make a block that could make use of the extra head generated, that, will give you better temps, not this imho.
 
Volenti said:
If it we're up to me, I'd hook both pumps up in series and make a block that could make use of the extra head generated, that, will give you better temps, not this imho.
The only problem with that is that the block idea I have in mind, this is not an option. It would have to have 2 pumps and 1 would have to be set for the radiator because the other would be used on the block itself and nothing else would be able to be on the block.
 
The pumps will have to be the SAME with a split res. Otherwise, one will be "filled" or pressuried by the faster one.

Either way, I would use the same pump to ensure the least amount of mixing of hot/cold water. If you slow down the cold too much, you will recycle the hot water ... which is not good

IMHO...
 
I don't know what exactly do you want, but I agree with Volenti two pumps in series is the only good solution, and basicaly that is drawn on spyders picture. You don't need the res, you can just have two inline pumps. With open res you can't have differen't type of pumps because more powerfull will flood second res, only if you can calculate all the preasure drops in firstline and in second line and then choose the pumps. But like I said before this is exactly the same as having two pumps in series, except they are not one after another but there is waterblock in between.
And your first idea with the same res for two pumps is bad because warmer and colder water would mix up and you have no control what kind of water will go into cpu waterblock pump.
 
I'd say give it a try. There's been talk about the sweetspot for radiators being slower than for waterblocks, but I don't think anyone's actually done it.

You're right about them being in a common res. If they were split, they'd be a series pump, as the bigger one would push through the smaller one...as long as the lid was watertight. If it weren't then it'd just be wet. :eek:

Instead of trying to size two different pumps, why not just use a tubing clamp to slow the flow through the radiator's loop. It's adjustable to optimize it, and reversible without making yet another purchase. A ball valve would also work, just try to find one with the inside hole's (the hole through the ball) diameter that matches your tubing.

That looks like a good setup there, Frag...
 
why not come off the block to a Y splitter, then convert to a much larger size hose to two different heatercores, then have both of them dump back into a res. or just back to another Y connection and convert it to your smaller hose size then pump then block?
 
Ok, doing series is NOT an option. The layout is the way It HAS TO BE. Period.
See image below.
Red is the heater core
Yellow is the res
blue is the block
WBcorereslayout.JPG

Gotta go, explain more later.
 
The Spyder said:
WC.JPG

i was bored :) with this split resivor, you will get better cooling. Hot goes cold, while supplying low and high GPH :-D good luck!

I thought about the two pump setup a while ago. I did not complete it becuese i dont think it will work.

Take a pint of water at 70F and a pint at 75F combine them and you get, well i dont know. n e ways what i am trying to say is that there is more volume of warm water vs. cooled water thus the water temp will rise MUCH higher than you might want.

does that explian enuff? it is a great ideal, but flawed.

but if you are nutz you can have a 500gph pump running the CPU loop and two 300~400 gph pumps running though two heatercores. but yet 300~400 might be too much still. If you use 5/8" tubing for the rad loops volume is present but speed is a bit slower. (i think)
 
Back