Fast and slow pump water cooling idea

[Owenator]

After reading Bill Adams article on "Radiator Heat Dissipation Testing", I was struck with an idea. Why not use a higher flow rate pump to supply the waterblock and a lower flow rate pump to supply the radiator with a common reservoir for both. This way the block gets the high flow rate it needs but the rad gets a lower one to optimize its heat rejection? The common reservoir I was thinking of using would be pretty big to be able to hold my pondmaster 250 and a littler pump but ti should be a interesting test. What do you think?

O

[SavageHenry]

I do something similar with my bong, except that I supply my waterblock with the low-flowrate pump and my bong with a high-flowrate pump. The high pump (a Rio 2100 -- rate ~800 gpm at 0 head) is useful because I have an inline drinking water filter filter on the tower to filter my water, but it drops a lot of pressure, so a smaller pump wouldn't be able to pump to the top of the tower.

I tried using the high-flowrate pump with my waterblock, but it produces too much pressure and I had a lot of problems with leaks in my system.

Using a radiator system as you describe sounds doable. Make sure that your flowrates aren't too mismatched, though, or you may end up cycling your coolant to the waterblock much faster than you can get the head dissipated through the radiator. Ultimately your cooling is going to be limited by how well you get the heat out of the system.

[combatmedic]

I was discussing this with another member, and here's my thoughts.
You'd be best of with two connected reservoirs. One would supply water to the water block, and one would supply water to the radiator. The water from the radiator would go to the water block's res and vice-versa.
This way you aren't mixing hot water with cold water.
You might also want to link the two reservoirs together at the top as kindof a overflow protection incase one of your pumps is moveing too much water.

Just my .02

[Al666]

I like this idea and going to do it very soon...

but this idea was originally posted by Unaclocker of procooling.com some time ago... He even posted a diagram of his idea @ PRO/Forums... I'm sure he'll send you the direct link 'cause I don't have time to find it for you right now... sorry about that... Hey Una listening ;0)

Also you can search HardOCP forum for a thread that I started about a month ago called RAIP (Redundant Array Of Independent Pumps)...

Keep it cool,

AL

[Owenator]

Originally posted by combatmedic
I was discussing this with another member, and here's my thoughts.
You'd be best of with two connected reservoirs. One would supply water to the water block, and one would supply water to the radiator. The water from the radiator would go to the water block's res and vice-versa.
This way you aren't mixing hot water with cold water.
You might also want to link the two reservoirs together at the top as kindof a overflow protection incase one of your pumps is moveing too much water.

Just my .02

combatmedic,

Thanks for the input!

For what I'm thinkng of this won't work because I want to have one pump flow at 250 GPH and the other at 90GPH. If I cross pipe them the 250 will suck the 90 output reservoir dry pretty fast. I suppose I could have the 250 filled reservoir overflow into the other but seems to be the same as just one?

attached is a pic of my idea (top) and what I think you are saying (bottom).

One reservoir is not neccessarily a problem because the water will be cooled by the rad more effectively and the block will be cooled by the rad cooled water. The idea is to let the rad work at it's best flow rate so it can cool the water that is being heated by the block. I will probably stick with inline pumps to keep their heat out of the reservoir.

O

[Owenator]

Originally posted by Al666
I like this idea and going to do it very soon...

but this idea was originally posted by Unaclocker of procooling.com some time ago... He even posted a diagram of his idea @ PRO/Forums... I'm sure he'll send you the direct link 'cause I don't have time to find it for you right now... sorry about that... Hey Una listening ;0)

Also you can search HardOCP forum for a thread that I started about a month ago called RAIP (Redundant Array Of Independent Pumps)...

Keep it cool,

AL

Al666,

Thanks, I'll check it out!

O

[Dissolved]

i was thinking of posting an idea just liek this last night.. i guess i got beat to it... its a great idea.. altho im useing a basic cpu/respump/rad setup.. but im useing a 300gph pump.. when i come acorss some money im gonna buy a 200gph pump and test the 2 and see if the lower gph will better my temps.. seeing mine are pretty hight for being water cooled..

[Holst]

Another simpler method would be to have a radiator bypass tube with a ajustable valve in it.

That way you could reduce the flow through the radiator without the need for a aditional pump.

Savagehenry has a very good point here

Ultimately your cooling is going to be limited by how well you get the heat out of the system.

If reducing the flow through the radiator helps cooling then your system may be of use, although you probably wont see a huge improovement.

[sciss]

Anybody ever tried a parallel radiator setup?
Then, theorically, a 180 gph flow will be split to 90 gph per radiator.

[Al666]

Owenator when you use different rated pumps like described in your diagrams (in your case a 250GPH and a 90GPH), they work @ the speed of the slowest one, so you'll end up getting 90GPH.

Take a look @ this
POSTED BY UnaClocker @HardOCP forums (http://www.hardforum.com/showthread.php?threadid=250233&highlight=Redundant) :

"Better idea... Since you have 2 pumps, have the slow pump pump water through your radiator and back to the res, and the other pump pump through the CPU and back to the res.. 2 seperate loops, 1 reservoir.. That way the water spends more time cooling in the radiator, and less time heating on the CPU.. ;)"

[fatshlink]

i brought this up about a month ago, and i decided bad idea. the radiator can only remove so much heat, and the processor isent gona change how much heat its removing so it will be verry close to the same

[Owenator]

Originally posted by Al666
Owenator when you use different rated pumps like described in your diagrams (in your case a 250GPH and a 90GPH), they work @ the speed of the slowest one, so you'll end up getting 90GPH.

Take a look @ this
POSTED BY UnaClocker @HardOCP forums (http://www.hardforum.com/showthread.php?threadid=250233&highlight=Redundant) :

"Better idea... Since you have 2 pumps, have the slow pump pump water through your radiator and back to the res, and the other pump pump through the CPU and back to the res.. 2 seperate loops, 1 reservoir.. That way the water spends more time cooling in the radiator, and less time heating on the CPU.. ;)"

Al666,

That's what I was trying to show in the first pic. Two pumps with a common reservoir. Each pump would draw water from the reservoir and then discharge back into it. The idea being that the rad would get a slower flow pump and the water block the higher flow pump. The water would then mix in the res. I may do this this weekend if I have the time. I also am working on an air chiller project also.

O

[UnaClocker]

I'd have tested this theory months ago if I had a spare pump laying around. As for the guy that thinks his radiator only removes so much water.. Check out BillA's article.. Some of those radiators have a huge performance increase when they're in their "happy spot".. I'd guess waterblocks are the same way, tho BillA has already said, from his testing so far, waterblocks love lots of flow.
As for the bypass idea, it might work, I'd trust it more with a flow meter hooked up on the radiator output, so you can make sure the flow rate through the radiator stays the same, and possibly even ring an alarm if it goes too low or high.. :) (read: pain in the arse...)

[SteenkyBastage]

i may be way off, maybe i dont understand the bypas idea.

but if i'm thinking right, wont that just mean that whatever is bypassed wont go thru the radiator, thus meaning that water doesn't get cooled?

that would raise the temps of the reservoir getting the hot water. i cant imagine it'd be able to compensate for the water that does go thru the radiator getting a little cooler.

just my thoughts

::edit::

also, it seems to me like the "single res, dual pump" idea would actually warm the reservoir water up. since heat transfer works depending on the difference in temps between the substances, wont this actually decrease performance?

ex: the hot water from cpu goes to reservoir, so does cooled water from radiator. they mix, making the reservoir temp somewhere inbetween.

the warmish water then goes to the waterblock, in effect not being able to cool it as well because it isn't ambient temp water.

the warmish water also goes to the radiator, which doesn't dissipate as much heat because the water isn't hot, it's only warm, and thus cant lose much temp.

i'm not a thermal scientist, this is just how it seems like it would work to me, i could be way off...

what seems like the right thing to do would be to simply put two radiators in parallel, as was suggested earlier. that would double the time the water is in the radiator (due to slower waterflow) as well as double the ammount of water in the radiators at one time.

[UnaClocker]

I see where your coming from, and I agree, that is a possible problem.
Here's what I see happening, plan a, your running the water through the radiator too fast, the water is 1 degree hotter than it would have been if it had gone through at the right speed.
Ok, so instead of doing that, you go with the other idea. 2 pumps, fast and slow.. Now, one pump is slowly taking the water through the radiator, it's losing it's maximum amount per pass. Let's say, for argument's sake, it's 1c cooler on the return from the radiator now, than it was before. Now, I'd probably try to setup the reservoir so that the return from the radiator was near the intake for the block, and the return from the block was near the intake for the radiator. Hmm.. Ok, I'm getting myself lost here. Fact is, this theory needs to be tested. I can see it going either way.
Yes, 2 radiators is great, but alot of people don't have the space, let alone money to run 2 radiators. Much better to get every penny's worth of performance out of your current radiator. :)

[Diggrr]

A couple of thoughts:
The system with two reserviors wouldn't gain anything over the single res system. That 250gph pump will eventually empty it's res to the point that it's actually feeding itself throught the overflow from the other tank. (if the're water/air tight) If it's not that tight, then it'll overflow...bad.

The idea with the single res, two pump system isn't bad. One way to allow you to hit the sweet spot of the rad.

I kind of like the flow control bypass idea the best though. My res can't take the mods needed for two pumps. It's about the same thing as the single res, but less power consumtion/cost. It's just using the tubing as the res.
Besides, I have a flow control valve in the junk drawer. (Aero equipment)
Looks like another trip to the hardware tomorrow. I've got no fittings.

By the way, to work right the flow control needs to be between the output of radiator and the intake side of the pump. This should put the bypass water back into the radiator loop.

[DodgeViper]

I for the life of me cannot understand how you use a water block and chill water below ambient temps within your home or office. The system is sealed so it’s not like an evaporator cooler. If you have a radiator with 2-120mm fans blowing onto the radiator, you cannot cool the water below the ambient temps of the environment in which the computer is operating. Can someone please explain?

[Diggrr]

I don't think anyone here's below ambient on watercooling alone. We're just trying to narrow the margin.
Right now, I'm at 32C cpu/26C case/25C amb. Most would say that's pretty good, but I'm a true addict and won't stop. If it hadn't snowed today, I probably would have dug a hole and buried some copper tubing and gotten rid of the radiator altogether.

[DodgeViper]

My problem with the water blocks is the expense and not really gaining over just moving air with proper fan location. What other steps are you taking inline with the water block to drop the temps? Are you using some type of additive to the water?

At work we have cooling towers that work in relation with chillers. The tower work on evaporation and pre chills the water before entering the chillers. This chilled water is then pumped throughout the property to fan coil units that have coils with blowers blowing through the coils producing cool air.

I have a small water cooler with a compressor, evaporator, and condenser. Water goes in and comes out cold. Now if I install a reservoir, pump and pump the water through a radiator mounted on the back of my tower with a single 120mm fan I now have chilled water supplying cool air into my case. Now condensation may be a problem, but if I only drop the cooled air 10 degrees below the ambient temp I should not have any problems. I can lower my DEW point using a product called DRIERITE.


DRIERITE desiccants are made from the naturally occurring mineral, gypsum (Calcium Sulfate), and are available in several varieties designed to solve all types of drying problems. DRIERITE is an all purpose drying agent for the efficient and rapid drying of air, industrial gases, refrigerants, organic liquids and solids. It is used to maintain a dry atmosphere in storage spaces, vaults, commercial packages, and other enclosures; and for the protection of hygroscopic materials or materials subject to mildew, corrosion, rust, or other deterioration caused by high humidity.

Forms of Drierite
The granular forms of DRIERITE are available in sizes ranging from powder to 4 mesh ( 1 /4 inch granules) weighing approximately 65 pounds per cubic foot, (60 pounds per cubic foot for 200 mesh powder.) The powder is very well suited for removing the last traces of moisture from organic liquids and as an additive in polymer formulations. The most popular granule sizes are 200 mesh, 20-40 mesh, 10-20 mesh, 8 mesh, 6 mesh, and 4 mesh granules. (“Mesh” means the number of granules per inch.)

Efficiency
The moisture remaining in gases dried with DRIERITE at 25° - 30° C is 0.005 mg./liter, (data verified by the National Bureau of Standards). Air is dried to a dew point of -100° F. Organic liquids are dried to the low ppm range. There is little change in efficiency at temperatures up to 100° C.Drierite DESICCANTS

Capacity
For the drying of liquids, DRIERITE instantly absorbs 6.6 weight percent water by chemical action, creating the hemihydrate of calcium sulfate. For the drying of gases, DRIERITE has a water capacity of 10 to 14 weight percent. This increased capacity is due to combined chemical and capillary action. DRIERITE granules have 38 percent pore space volume, which accounts for the additional capacity. Above 6.6 percent, the capacity varies inversely with temperature and directly with pressure and the partial pressure of water vapor.

[Diggrr]

DodgeViper,
I'm using about 10cc of antifreeze to prevent corrosion, and a very small amount of dishsoap to kill surface tension. Besides that I'm not really using anything special.
I built my own copper/lexan waterblock partly to test a theory, and party for the economy over store bought. For $12 bucks (not including shipping) I get enough copper to make 3 blocks. And it's working well...only 7C over ambient using ambient air to cool it.
I use a radiator very similar to 'big momma', it too sheds 4500btu/h, the t-bird can't overpower that.

Your air chiller sounds like a great idea, you could add a flow control between the two tubes to the radiator and possibly control the temp of the air you're blowing. Very creative idea though. Drierite may not be needed if you can get control over the air temp going in.

I'm about to hang a copper coil out the window at my desk and bypass the radiator for the winter. I'll have to trade the antifreeze for alcohol due the the test results in the stickies at the top of the forum. The window at my desk goes to the porch and not outside, the temps should stay above 0F. But I'll have to insulate the waterlines/block as if using a peltier or chiller.

[DodgeViper]

Currently my problem is my wife keeps the house to warm. The small water-cooling has a control that will allow me to set the water temps. Like you mentioned above by placing a couple of flow valves inline I could control the water entering and exiting the radiator. I will need to locate a small radiator or make one. Years ago when I was involved in off road racing Bilstien made these small round fins that would slide over 3/8” copper tubing. They were made for cooling the oil in shocks. I am going to look for a small transmission cooler.