Sign in with
Sign up | Sign in
Your question

Liquid cooling & capacity

Last response: in Overclocking
Share
September 25, 2006 12:01:31 PM

Everytime I see a liquid cooling system (mine included), I always wonder why people don't use larger capacity reservoirs. It would seem more logical that the more thermally conductive fluid you have, the less your system as to work to cool it since it takes more heat to raise the over all temperature of the fluid. I've run zalmans (the big tall passive heat ones) fine (with the exception of the pumps dying, but thats another story). It took much longer to heat all the fluid in the reservoir. Seems to me that it might be better if you have the room to maybe hold 1/2 a gallon vs a liter or less. Is this sound logic?
September 25, 2006 12:46:27 PM

That seems very logical, but with all the internal/external accessories that the modern cases have, space is more important for future upgrades, unless it's an external reservoir, they may not have adequate room for the bigger resevoir. I have been considering upgrading to the larger resevoir as well.
September 25, 2006 1:46:26 PM

Just rambling here and I am not a physicist so please feel free to challenge my assertions.

I think the more coolant you have, the better you are able to cool, at least I am hoping so as I just upgraded my rad. to a larger 2x120 model. Wusy is absolutely correct in that eventually you will reach equilibrium, regardless of total fluid, based on the ambient air. However, I think Synthetickiller's has a point in that the more liquid there is to absorb the heat the easier it is to cool your system.

Why. Let's take it to extremes. Which system would cool better, one using 5 ozs of water or one using 1 gallon of water? Obviously the 1 gallon model, right? Why? 5 ozs of water has only so much thermal capacity. Now, given enough cooling capability (Wuzy's cold packs in a bucket) you can cool that 5ozs of water quickly enough for it to pick up more heat and deliver it outside the loop. Is that practical for most users though? No, most want a relatively quiet, fairly automatic cooling solution.

So what do you do in that case? Well, if you increase the total mass of heat absorbing media (be it copper or water) then you can use a less radical cooling mechanism. Instead of a bucket of cold packs you can use a couple of slow moving fans because you now have more time to cool the media as its greater mass takes longer to heat. Of course there is always a point where you will pass equilibrium and start getting that temperature creep. Also, you are still dependent on ambient air to cool everything and as it heats up so will the temp of your cooling media.

So, we have two ways to cool, lot's of media with a slower cooling mechanism or less media with a more radical cooling mechanism, or something in between. For most folks more media with a quiet cooling mechanism is preferable. Always a balancing act and environment is a factor in all cases but I'd be interested in your opinions (I've only ben watercooling for 7 months or so and would like to hear the more experienced voices)
Related resources
September 25, 2006 2:57:13 PM

It is not the amout of liquid so much as the amount of surface area able to dissipate heat. 5 gallons of water in a small radiator will probably not work as well as 1 cup with a very large radiator.

Since larger radiators mean more volume, (All else being equal) then you will, by default, get more liquid.

It comes down to a large enough radiator will always allow you to cool the water down to within a few degrees of ambient. About the only way to do that with water is to have a continuous supply like a tap. (We all know not to use tap water in a water cooler right?? :)  )

Both radiator size and resevoir size is limited by the case for internal systems. Externals can go much bigger if you want to pay for them. Still there comes a time when the returns on cooling don't match the investment.

I use an external system on mine. I use the external because I wanted to dump the bulk of the heat outside an enclosure where my computer resides and the external one actually cools the CPU very well. Unfortunately, it cools only the CPU.

All systems today compromise between cost/size/cooling. I've always wanted to combine an active cooler like a peltier or something with my water cooler. Let the water cool the CPU but have something very cold cool the water down. :)  Keep the water about 70F (21C) to keep consendation down and let it rip.
September 25, 2006 3:00:42 PM

Good point Gneisenau. Of course, if you use the swimming pool as a reservoir then you can probably live without fans :D  That would be interesting, tryingto keep the kids away from those tubes at the bottom of the deep end.
September 25, 2006 4:05:19 PM

Additional liquid and increased surface area can both help to cool down your system. Case in point. In '97 (could have been '96 I can't remember) I built a DIY water cooling system from the spec's given on tomshardware. I had no radiator or fan. Instead I had an open system where water would flow down in the air like a waterfall into essentially a bucket. I was using probably 3-4 liters of water. the sheer amount of liquid and surface area, combined w/ the water passing over air while falling, dissipated the heat. FYI, I would not recommend this system since it allowed for contaminants to enter the water an eventually clog the cpu block.
September 25, 2006 5:55:37 PM

Well, it's not enough to just have more liquid but a combination of factors combined with the radiator and pump as well as the king and placement of the waterblocks.

For some reason, people get an idea that if coolant is moving slowly through a waterblock, it gives the coolant enough time to collect more heat for transfer. I keep reading about this from time to time but it just doesn't work that way. Slower moving coolant allows for a buildup of heat in the waterblock as it is moved from the source slower. Having a higher flow rate guarantees that cooler coolant will come in contact with a heat source faster.

While relikemethods mention of that home made cooling solution could prove effective strictly based on its performance, it is impractical for several reasons, one of which he mentioned. Like Gneisenau, I to use an external radiator setup (2 dual 120mm rads) to dump my heat. The radiator is the key to removing the heat from the loop and there are several factors that can influence that. The choice of fans as compared to what type of radiator (dual or single pass rads - low or high cfm fans). If you use low cfm fans then you'd better use single pass rads (like the Black Ice X-flow or Thermochill P series - sorry wusy, I am tired and can't remember the exact model) because they won't function efficiently if you use them with dual pass rads (like the Black Ice Xtreme rads).
September 25, 2006 11:20:14 PM

redlikemethodz, you did have a radiator. It wasn't some piece of metal with fins, it was the waterfall. If you would have been sucking water out of and returning to a jug of water, it wouldn't have worked nearly as well.
September 26, 2006 3:24:49 PM

Quote:
Well, it's not enough to just have more liquid but a combination of factors combined with the radiator and pump as well as the king and placement of the waterblocks.

For some reason, people get an idea that if coolant is moving slowly through a waterblock, it gives the coolant enough time to collect more heat for transfer. I keep reading about this from time to time but it just doesn't work that way. Slower moving coolant allows for a buildup of heat in the waterblock as it is moved from the source slower. Having a higher flow rate guarantees that cooler coolant will come in contact with a heat source faster.

While relikemethods mention of that home made cooling solution could prove effective strictly based on its performance, it is impractical for several reasons, one of which he mentioned. Like Gneisenau, I to use an external radiator setup (2 dual 120mm rads) to dump my heat. The radiator is the key to removing the heat from the loop and there are several factors that can influence that. The choice of fans as compared to what type of radiator (dual or single pass rads - low or high cfm fans). If you use low cfm fans then you'd better use single pass rads (like the Black Ice X-flow or Thermochill P series - sorry wusy, I am tired and can't remember the exact model) because they won't function efficiently if you use them with dual pass rads (like the Black Ice Xtreme rads).


I was refered here from another post, and this maybe an old thread but I find it interesting... More flow and pressure is better, too a point... The problem is that a waterblock could be designed for lowflow systems, and could prove to be very effective at cooling these systems... Remember we deal with water which takes allot of energy to heat up water... In fact it has one of the highest Calories per degree change of any material in the world.

Then we deal with a very small surface area, so even at low flow rates water moves very fast accross the heat source. Low flowrates and high pressure are not as restrictive to smaller passages.

No one, including myself, will disagree with your statement that higher flow and pressure is better. The biggest benifit is probably the added turbulance of the water that comes with these high flow and pressure rates.

The problem is that there are very few good benchmarking provided. How many waterblocks reviews have you read that compair multiple pressures and flowrates with each waterblock available?

Take into account the added heat higher flow/pressure pumps exert, and things become very difficult to optimize. Unfortunately this can be VERY expensive without the proper resources.

Now, More water will take longer to heat up... and the larger surface area will increase cooling (as stated). But if you put that water in a plastic container, you wont get the same benifit as you would lower amounts of water through a larger radiator.

Then again, a big container might take 24 hours of straight constant usage to max out its overall temp.

I just cant see myself having a 20 gallon bucket of water next to my PC.

hehe...
September 26, 2006 3:52:09 PM

Your points are well taken.

However, sometimes getting into the physics of things just really complicates what should really be an easy decision. As for reviews and such, there is such a thing as just being too thorough. You know, half the joy of watercooling is just putting the loop together - after doing that, then you decide how far into efficiency you want to carry it,

Nothing wrong with testing one pump, one reservoir, one radiator and several water blocks. Or, you could test one reservoir, one radiator, one waterblock and several pumps. Better yet, how about one pump, one reservoir, one water block and several reservoirs. How about one reservoir, one pump, one water block and several radiators. If you have time you could test several reservoirs, multiple pumps, a water block and a radiator. But, if that doesn't satisfy you then test several reservoirs, one pump, a few water blocks and a radiator OR several radiators, one pump, one water block and a few reservoirs.

Or

Just take in what your buddies tell you, get what you need and enjoy your creation.

This was a message from the Emergency Broadcast System. This was only a test. If there had been an actual emergency, someone would have come along and slapped yer nutz. Please leave your trays in their upright and locked position. You may now walk about the cabin.

I haven't slept in 36 hours because of my job (law enforcement - investigations) so I am NOT responsible for any of this (at least it shouldn't be held against me).
September 27, 2006 12:24:38 AM

Im having trouble with astablishing a base line... I think I got to remove my waterblock and run it with the stock cooler for a bit... see how things are...

At least then I would have some idea of how things are... If I knew that my systems wasnt working up to par, I would do something about it... but I dont know that... and from all Ive read, it seems to be working good... though the flow is REAL slow...

For me to change pumps, I would have to buy a new resivor... my pump is a submerged pump, and it just doesnt have the power... It also runs off my powersupply and it will definately add heat to the system, so I know that a new pump will help me...

You are right about over calculating everything, but when you dont want to spend thousands on parts, its real important to be smart with them...

Whats going to end up happening is that I buy the Eheim 1048
http://www.eheim.com/universal.htm

and then a resivor, and a new waterblock... then all I need is a 1/2 inch replacement barb for the intake, and then 1/2 inch barbs for my radiator...

Then probably I will still not be happy so I will add more powerfull fans to my 120mm radiator, and I will add a switch so that I dont have to listen to them when Im watching a movie or working on documents...

Mike
Mike
September 27, 2006 12:39:42 AM

If you're running 1/4ID you'll need to change tubing also. If you are running 3/8 ID then you could opt for the MCP350 or MCP355 (although, on this pump, the sound is recognizeable). If you go to 1/2 ID I'd go for the DD Laing 5 (MCP655). I think that you have sufficient interest in water cooling to know that it's somehting you want to get into (judging by the comments you've made and the lengths you've researched things). That being said, you probably also recognize that, sooner or later, you will have to make these upgrades because the fact that your cooling loop is not functioning like you want is going to gnaw at you like a dog gnaws on a bone - until you do something about it.
September 27, 2006 3:38:06 AM

follow phreejack's analogy here my friend! although i'm not really sure what that means... maybe something to do with getting to the marrow?!?

anyways, the volume of a reservoir makes no difference in a system that runs longer than it takes for the fluid to reach the system's point of "equilibrium"

capacity is not important, only the components which absorb, pump and dissipate the heat...

you have to look at your components... how much heat can the chips generate, how much heat can the waterblocks transfer to the coolant, and how much heat can the radiator dissipate...

you only need enough coolant to fill the loop... a reservoir isn't even necessary, but it sure makes filling a system a hell of a lot easier
September 27, 2006 4:03:54 AM

The reason for all that nonsense above is that sometimes you can overthink some things and take the fun out of.....having fun.

I have a blast when I design a cooling loop, collect the components, test it, work out the kinks and enjoy the results.

When I give people advice, I do so from the standpoint of trying to help them build the most efficient cooling solution within whatever budget they have set - the lowest flow resistant components, the best position to arrange the components in, etc. However, people are different and just because I might say that one water block is better than another in terms of flow characteristics doesn't mean that they have to floow that advice. I mean, I've got a Danger Den Koosah waterblock - which affects the water flow in a cooling loop alot because of all the bends in it - but, I like it and I still use it in another system.
September 27, 2006 4:16:33 AM

Why DDLang over Eheim?

Ive heard good things about Eheim, and nothing about DDLang...

I will look at the products you mentioned...

I am more interested in head then in pressure... Just talk to my Girlfriend...

I got limitted space, and I am going to upgrade too 1/2" pipe... Im using 8mm now... I use a DD ATI Radeon Block that is big, and it came with 1/2" barbs... I think its the kooash...

I dont worry much about resistance as long as I got the pressure to push past it... My radiator was also suppose to be 1/2" barbs, so I am going to get my hands on some.

That leaves the pump and resevor...

I was thinking of using a double bay 3.25 inch bay resevor, but they are made out of plastic, and I dont like this... I have around 17cm TOTAL from powersupply too the HD Drive bays, and the default width of the coolermaster. I will be fitting the pump in this spot...

I got to see these pumps and make sure they fit... Im going to look locally for a pet supply store that carries them...

Frozen CPU isnt far from me... but Im not happy with their service.

They sell Hydor... and a few other products... but I dont want to buy a pump from company like dangerden... I would rather buy a hobbiest pump...

Mike
September 27, 2006 4:43:23 AM

If it's head pressure you want then the best pump on the market is the new MCP355 - 20+ ft of head.
September 27, 2006 6:14:25 AM

Hey, check out my reservoir. It's just another tubing hooked up in the loop using t-connector.



With this kind of reservoir I think it has less restriction compare to loop with regular reservoir. No it doesn't have much extra fluid on it but even with my pump (MCP655) at the lowest setting 1 of 5, I'm still getting good cooling. My 840 runs 3.920 @ 1.45vcore and idles 35c and full load of 40c. Sure having a huge reservoir can increase cooling performance if it's cooled down with a good radiator. I see some review of that Zalman reserator water cooling kit with external passive cooling reservoir that hold lots of water, it's not better than a cooling kit with 1 120mm rad. So that said I think the size of the reservoir has little impact on cooling the fluid core temp.
September 27, 2006 12:52:15 PM

I love that Resivor Idea... I think... Wait, I know, thats what Im going to do... Ill buy a big fat 1" Tube, and run it up, then put a fill cap on it, like you did...

What a great Idea...

Maybe even run a Copper Pipe accross the side of my Radiator and out the back where the top PSU goes, but that is instead blocked by the radiator.

My case comes with 2 spots for a Powersupply: One on the top and one on the bottom... This will save allot of money, and If Im smart, I can make it so that the copper will help cool the water... a Bit...

Ill use a Clear PVC Pipe On top for the fill line and on the bottom to connect to the regular hose... Due to gravity I can even run the copper pipe allong the bottom of the case:) 

Thanks so much, this is a very inexpensive and very usefull idea, that will solve so many problems at once...

As someone once told me... the biggest pipe is the size of a small car! I dont like 12v pumps that run off the powersuppy. I would much rather fabricate another 120v outlet (like the one on all powersupplies) that would then go into a regular female 3 or 2 pin connector (depending on size, pump need etc).

I like the idea of not running off the powersupply because like every PSU, the more current you pull from it, the shorter the lifespan becomes... Anyone know where I can get an AC version of this pump?

Thanks for the great ideas... I am going to research these pumps...
Mike


Edit: Swifttechs pump Looks amazing... except for the 2 AMPS it pulls from the powersupply... YIKES!
September 27, 2006 2:30:26 PM

It's a great reservoir but bleeding out the air in the first place when filling up the fluid was time consuming. Took me about 30 minutes to fill it up and another hour and a half to bleed out the remaining air pockets. About after a week there were no air pockets left in the loop.

I used this one for I can't find any reservior that fits my cooling loop and besides it's pretty expensive to them plastic containers. Some of them goes for over $50! With this one I have very low restriction and so I am able to set my pump to the lowest setting. I'm using Aspire 600 watts psu and so far it's holding up this power hungry pump.

With this set up you don't have to worry about head pressure as the pump is pushing and pulling the water at the same time. I like this setup very much and few out there have the same as mine. :D 
September 27, 2006 4:07:47 PM

It is a great idea... I just will use a very large pipe or use something else... I dont have room for much, and thoose plastic containers specially made for this are large, and they are expensive... Air bubles could be a problem... I guess if you put a air compressor on the fill end, you could increase the water pressure on the resivor, and thus eliminate this problem... small amounts of pressure should be ok... you may even be able to just blow it in with your mouth...

Like you said, this isnt an inline resevor, and I doubt the inlines provide much restriction. I could also see this as reducing the capacity, allowing the water temp to fluctuate faster between performance boosts and drops.

...Which is both good and bad...

But we must remember, water has a GREAT ability to absorb heat, and disapate it... and it takes allot of power to warm up water... Oil for example heats up MUCH faster... so does alchool and other compounds... In fact almost every aditive you use, will decrease this ability. I would be hard pressed to find a better liquid that requires so much energy to warm and cool the substance a single degree. Also heat rises... so hot water should float to the top...
September 27, 2006 4:18:26 PM

I using the Innovatek non-electrical fluid for my loop. Yeah, it's slightly lower performance than water but it protects my loop as well as my pc. :) 
September 27, 2006 5:19:53 PM

Quote:
anyways, the volume of a reservoir makes no difference in a system that runs longer than it takes for the fluid to reach the system's point of "equilibrium"

capacity is not important, only the components which absorb, pump and dissipate the heat...


Just a reminder, Nowhere is it written in stone that the only thing that disappates heat is a standard looking radiator.

Your reservoir will also exchange heat with the ambient.

John
September 27, 2006 8:17:02 PM

Quote:
Additional liquid and increased surface area can both help to cool down your system. Case in point. In '97 (could have been '96 I can't remember) I built a DIY water cooling system from the spec's given on tomshardware. I had no radiator or fan. Instead I had an open system where water would flow down in the air like a waterfall into essentially a bucket. I was using probably 3-4 liters of water. the sheer amount of liquid and surface area, combined w/ the water passing over air while falling, dissipated the heat. FYI, I would not recommend this system since it allowed for contaminants to enter the water an eventually clog the cpu block.


I like this idea. I was thinking of using one of the high performance pumps (the one that gets 317 gpm, i can't remember the name although its in this thread) and build a foutain in an acrylic case. I'd have 1 or 2 fans (with filters) blowing air across the waterfall. This would take a while to make of course.

I also though of building a tiered foutain since there's a lot of surface area to it. Just ideas. Nothing set in stone.

You have have brought up a lot of good points. Maybe I'll just put a kiddie pool in my room and have that as the reservoir. :lol: 
September 27, 2006 10:41:04 PM

The problem I have is that a I will probably go with a passive resivor... The issue with a resivor that isnt part of the cycle is that it doesnt allow for easy fill, as you noted...

My problem with the resivor I have seen is that they are A) too big to fit with a pump (though I havent seen the size of the pump mentioned above) and that if I cant get a regular resivor, then I would have to get the double bay

Typhoon Dual 3.5" Bay Reservoir - CLEAR

Which is made out of plastic... This makes me sad, because its essentially like wrapping water in plastic, which wouldnt help cool the system...

If they made these systems with a metal, the heat of the water will be transfered partially to the case... which in turn would cool the system...

It is also too bad that they dont make a 5" resivor that would fit behind (or infront) of a 5" fan... This resivor could be like a radiator, that would be built to remove bubles... Throw a fan in front of it, and you got a VERY low restrictive radiator/resivor...

Too do this you would have a series of verticle pipes with a horizontal pipe on the front and a pipe on the bottom...

Mike
September 27, 2006 10:59:08 PM

Beware getting metal reservoir. Make sure it's copper, well I haven't seen one but if it's aluminum see to it that the aluminum surface on the inside are painted or anodized. Copper and Aluminum in the same loop of a water cooling system is not good as they can cause each other to corrode over time and could ruin components like pump, blocks and radiator. I know metal transfer heat better that acrylic but there's a reason why they use plastic, since a reservoir that can be fitted inside the case might move around and the aluminum reservoir could cause shortage on circuits. That and it's cheaper as some could go well to $100 for single unit.
September 29, 2006 3:11:08 AM

how well does that set up with the half inch tube as the resovior actually work? i might try it
September 29, 2006 3:58:24 AM

Quote:
It is also too bad that they dont make a 5" resivor that would fit behind (or infront) of a 5" fan... This resivor could be like a radiator, that would be built to remove bubles... Throw a fan in front of it, and you got a VERY low restrictive radiator/resivor...


If you're going to go so far as to have another fan in the system, why not just for for a larger rad in the first place? You're not going to get huge heat transfer by blowing air over a reservoir. You could make one with heat sink fins soldered to it and do a little better. I've seen pics of a guy that made his case out of copper tube. It had enough surface area to transfer heat well. So if you want a reservoir to work as a radiator, go for high surface area to volume ratio. Take that far enough and you'll have a reservadiator...
!