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

water pumps?

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

junk30

Registered
Joined
Dec 23, 2000
I am going to get water cooling for my 1.2 amd cpu, but I was wondering what pump to go with how much gph do I really need to cool it down I have picked my radiator and water block from danger den but need a water pump I just want a in line one that's all if you know of good ones let me know
 
hmm, colin gets the best results using a dangerden setup and danner 700 pump. get at least 300gph.
 
dunno260 (Jun 02, 2001 10:11 p.m.):
hmm, colin gets the best results using a dangerden setup and danner 700 pump. get at least 300gph.

But I aint that far off at all with my setup either. I just need to secure my cube to the intake hole on my case but I need to mod it and I'm lazy right now. I got the one YS-Tech 131cfm 120mm fan sucking air out of the radiator but I want to get another to push air through it at the same time for maximum cooling in the radiator. I read just about 4-6C over his idle and loads.
 
are the mag drive pumps any good? and what is the gph limit where effectiveness starts dropping off? i can get a 350gph or for $5 more a 500gph...

Justin
 
I would recommend something way slower, like 150gph. You have to give the water a chance to cool down and if it''s gonna pass through that radiator way too fast its not gonna have enough time to transfer the heat. Something like the Rio 180 is good enough.
 
Ok I don't have anything to add to this accept to say that overclockedpc is completely wrong.

How water cooling works 101.

Lets say at 100gph the water pics up 10 heat units (this is all gestimation) and if gets to the radiator and gives off 10 heat units

Now at 500gph the water only pics up 2 heat units and gives off 2 at the radiator. but since the water is moving 5 times faster its still transferring 10 heat units in the same amount of time.

So why do you care if its 100 or 500? Well if the water is only carrying 2 heat units at a time then your water temperature is 2 heat units higher then ambient not 10.

This whole "got to give the water time to cool down" Is totally wrong because your only looking at one side of it. Sure the water doesn't loose 10 heat units at a faster speed like it would at a slower speed but it doesn't pick up 10 either.
 
BlakeN (Jun 04, 2001 10:22 a.m.):
Ok I don't have anything to add to this accept to say that overclockedpc is completely wrong.

How water cooling works 101.

Lets say at 100gph the water pics up 10 heat units (this is all gestimation) and if gets to the radiator and gives off 10 heat units

Now at 500gph the water only pics up 2 heat units and gives off 2 at the radiator. but since the water is moving 5 times faster its still transferring 10 heat units in the same amount of time.

So why do you care if its 100 or 500? Well if the water is only carrying 2 heat units at a time then your water temperature is 2 heat units higher then ambient not 10.

This whole "got to give the water time to cool down" Is totally wrong because your only looking at one side of it. Sure the water doesn't loose 10 heat units at a faster speed like it would at a slower speed but it doesn't pick up 10 either.
I understand what you're saying but in my experience with my camaro, I took the thermostat out to improve the waterflow cause I thought it was running too hot and the temps went up! I put a hose clamp on the radiator hose and tightened it down to constrict the hose and slow the water down and guess what? The temps dropped. I slowed the water flow through the radiator to allow it to cool down before going back to the engine to get more heat. I know a car is not a computer but the water cooling setup is exactly the same.
 
I cant find the article but there is one that debates the exact same point you just made as soon as I can find it Ill post it.
 
BlakeN (Jun 04, 2001 10:22 a.m.):
Ok I don't have anything to add to this accept to say that overclockedpc is completely wrong.

How water cooling works 101.

Lets say at 100gph the water pics up 10 heat units (this is all gestimation) and if gets to the radiator and gives off 10 heat units

Now at 500gph the water only pics up 2 heat units and gives off 2 at the radiator. but since the water is moving 5 times faster its still transferring 10 heat units in the same amount of time.

So why do you care if its 100 or 500? Well if the water is only carrying 2 heat units at a time then your water temperature is 2 heat units higher then ambient not 10.

This whole "got to give the water time to cool down" Is totally wrong because your only looking at one side of it. Sure the water doesn't loose 10 heat units at a faster speed like it would at a slower speed but it doesn't pick up 10 either.

You're basically right. However, as the water passes through the waterblock, it's going to heat up. As it heats up, the efficiency at which heat is transferred from the block to the water drops, which is why you want it moving faster through the waterblock. Unfortunately in a closed system, the water can't move fast through the waterblock, and slowly through the radiator. If that happened, eventually you'd have a backup somewhere in the system. For the most part, faster=better, but obviously extremes are no good. I mean, if you really wanted to you could go out and buy a 1200gph pump, but it wouldn't be worth it.
 
Maybe don't assume linear relationships between the two rates of heat transfer at each the end of the system. 1.)rad/air 2.)block/cpu ie. a warm block may still absorb heat well as it gets warmer but a warm rad maybe looses efficiency ....or vice versa. It maybe a small factor but could be determining in this puzzle.
 
I have found that the Dangerdan Maze 2 blocks add a bit of resistance to the waterflow. There is a big differance in the flow rate using my Eheim 1048 (160gph) with an old cross drilled block as compared to the Maze 2. To add to that the cooling cube has a lot of passages and this also cuts the flow down. You need a good pump with setup to get a decent flow. I would recomend nothing less then 300gph. I use a Eheim 1060, 600gph in my sub-zero water cooled system along with a Maze 2. The 1048 could not do the job.
 
wow I never realized how cheap the Eheim 1060 was I guess I just assumed a pump that could do 500 GPH would be much more expensive.
 
Some good links regarding flowrate here and here.

More flow is better, but there is a feasible limit. Go any higher than that, and you are just pushing your luck.

P.S. I have a Rule24 Bilge pump that does 360GPH at 12VDC. Machined some fittings on a lathe so the outport is 1/2" instead of 3/4" O.D., and it's inline instead of submersible.
 
great thread but unfortunately both speed flow theories make sense to me so now what the hell do i do?!#%$
lol
 
lol ....exactly why I invented the non-linear theory. Both made sense so it was a case of "That does Not compute". Considering other variables allowed a good nights rest. Whew that was a close one.
 
I'll make this real simple. I have found the best pump for the Maze 2 and a Cooling Cube is a Danner 7. Restricting the flow causes temps to go up. The Danner 1800 I tried was way too much. Please stay away from Rio pumps, the anemic 180 will give you very poor results with the Maze and Cube.
 
the problem with most "high flow " pumps , is they rely on their large fittings to acheive this flow rate , but at a very low pressure . if you have a 200gph pump that puts out 10 psi from a 3/8 fitting , it will prolly push water through the system FASTER than a 500gph pump that puts out 1/2psi from a 5/8" fitting , because the of the static back pressure from all the stuff (tubing -block-rad ). They will all work as long as they will pump high enough to get a siphon going .but just reading the GPH rating on the pump without considering pressure can be very deceiving
 
I use a maxijet 1200 that works great. It's 295GPH and was highly recommended by the fish shop I bought it from. It was around $30.
 
Back