View Full Version : How small can we go?
Trying to nail down my post on reservoirs got me to thinking. We are always talking about the biggest baddest block, rad, pump, etc. A lot of the folks who come into the forum ask what should I get, what is the best, etc.
We seem to lean towards bigger is better. Lets face it, 120.3 sized radiators aren't exactly small.
So I got to thinking. Has anyone done any expermients with smaller components, smaller radiators, smaller pumps, etc. to find setups that are both compact, fitting easily into mose cases WITHOUT modifiaction, and also provide better or equal cooling power to air?
I've been looking all over, but all I find are reviews, and or statistics that relate to the whole "bigger is better" theory. (except when it comes to tubing length :D )
Anyone done any research into miniturizing? Or know of any?
alinosa
03-15-05, 01:14 PM
well, in europe, they usually use smaller fitting and hoses. ie ~1/4 inch so they must know something we don't... or are ignoring.
joelkyr
03-15-05, 01:20 PM
yah i think i saw 1 post in this forum regarding this matter, as i remember it was posted by someone from europe, i think he posted a question about why americans loved big block, loud pump, etc.....comparing the WC hardware they use in europe and stating that european WC are much better. I didnt get to follow the whole discussion though.
I'm just curious because I do have friends who are looking to jump onto the WC bandwagon, but eithe rlack the tools, knowhow, or desire to Mod their cases to the point of slapping in big ole rads.
Now I have a BIM2 and it actually performs surprisingly well on a single block loop, but the question is, can a small rad liek the BIM2 which does fit easily in most cases, or an ever smaller rad, be used to gain performance at least equal to if not better than the most agressive aircooling?
Perhaps this will be an experiment over the summer. I got a hunch a big cylidrical Rad, a DDC pump, (insert jet impegnent WB here), and a small rad could be a great performer with the right combonation of tub size, and water volume.
t4runnr
03-15-05, 01:33 PM
yah i think i saw 1 post in this forum regarding this matter, as i remember it was posted by someone from europe, i think he posted a question about why americans loved big block, loud pump, etc.....comparing the WC hardware they use in europe and stating that european WC are much better. I didnt get to follow the whole discussion though.
I think most europeans wonder why americans like anything big. Big engines and big cars, etc. :p
I have 2 watercooled systems. One system is using a BIM2. Its effective but temps are not as low as I'd like them to be while using a Barton 2500 OC'd to 3200. Of course that could just be like that base on how I got the fans and radiator placed in the case. The case could use more work to get better airflow for the radiator (which I intend to do sooner or later). And that system is only cooling the cpu. It has a Hydor L30 (I think its an L30 anyway), Asetek waterchill block (not Antarctica) and the fans are 2 TT SF2's running full speed (ack!). I think once I cut the intake grates (very restrictive) out of the front of the case, things will be much much better and I could turn the fans down.
Yeah, I'm sure you could drop the temps with a better WB, a pump with more head pressure, and less restriction on the airflow over the Rad.
I think I'm going to have to try a 1/4" setup using a BIM2, DDC, and maybe a TDX with a #5 plate and a Big Res. See what I can accomplish, assuming the Mrs. lets me lol.
Albuquerque
03-15-05, 01:37 PM
When people talk about "normal" air cooling, I get a picture in my head of the stock cooler, stock fan, making a medium amount of noise and keeping your CPU temperatures in the high 40's to low 50's centigrade under full load -- GPU temps being a bit higher under the same full load.
You could more than certainly use a very small radiator, smaller blocks, smaller tubing and a smaller pump to obtain the same level of cooling with almost no noise. But what fun is spending ~$100 on watercooling equipment if it's "only" going to get you the same temperatures as air cooling?
I didn't spend the money to get the same cooling as air but quieter. If that was to be my goal, I would be sitting over on the silentpc forums instead :) (no offense meant to the SilentPC folks who visit here too) I wanted watercooling so I could reduce my temperatures by a large amount, to further increase and/or stabilize my overclock. That means I'm not going to use 1/4" diameter hoses, a 90lph pump with 11" of head and an 80mm radiator. Because if I did, while I'm sure I'd be fine at stock speeds, I likely wouldn't be stable with my X800Vivo unlocked / voltmodded / overclocked and my 3ghz processor overvolted and overclocked too.
Go big, or go to Europe ;) j/k
well, the europeans usually use low-flow pumps w/ small dia. tubing, plus restrictive blocks
with a low-flow pump, the water moves slower through the rad. so that it can dissipate more heat (to a certain point) - the smaller dia. tubing is probably to keep up the pressure in the loop, so that the restrictive blocks can function properly
but for now, i'll stick to my gigantic american cases, radiators, pumps, tubing, and waterblocks :p
t4runnr
03-15-05, 01:40 PM
See what I can accomplish, assuming the Mrs. lets me lol.
My work on the PCs are covert operations. Wifey doesn't even know that watercooling even exists. hehehehehe. If she did, I don't think she'd appreciate the extra costs.
Those components were leftovers from a friend who had them laying around so I bought them off of him for like $50. I just had to buy the 3/8" tubing and the hose clamps. It was my first attempt at watercooling so I used my slower pc as the guinea pig. So now my 3 year old son has a slightly overclocked and watercooled pc to play his reader rabbit games. :p
Well thats exactly my question. What paremeters need to be changed to use such small components and to cool at least as efectively as the most agressive air cooling.
Obviously we are onto somethign when it comes to pressure, or else JI type blocks wouldn't be the norm these days. Can the same principles be applied to the radiator to more effectively transfer heat? Can an ideal coolant volume be reached? Can the design of a reservoir produce more effective cooling?
Can it all be done on a small scale?
Albuquerque
03-15-05, 01:47 PM
the water moves slower through the rad. so that it can dissipate more heat (to a certain point)
Not to be picky, but this is a common misconception in both the PC watercooling and automotive worlds. Slower movement of the water does not mean "more time to transfer heat", as thermal energy transfer does not have a specific "speed" like you're suggesting. The greater the temperature difference between two materials, the faster energy will transfer.
to cool, you need surface area, this you cannot change.
i guess you could spend your money on some small parts, and do ok, but why not just buy a 120mm heatsink?
Well thats the point, can we achive something small, and effective, that is at least as good as sticking a delta on an XP-120?
Sure we need surface area to cool, but we are also using a different medium of heat transfer, water instead of air.
but in the end the heat still has to get to the air. its just kindof detatching the heatsink fins(radiator) from the heatsink base (waterblock). in order to cool as well as the xp-120 at similar or better noise levels, you will need surface area of the xp-120 in your radiator and perfect efficiency with your waterblock.
Hey_Its_Cole
03-15-05, 02:33 PM
See, with my watercooling system i just put in i cant get my prime temps below 50-52c. I have a swiftech 6002, aqua 50z pump, a maze gpu block, cooled by a chevette core with one 120mm sunon 90cfm fan. Idle temp usually stableizes at around 34c-36c. I am on a 3500 winchester overclocked to 2.53 ghz at 1.44 volts, on a epox 9nda3+. These temps seem high but the only thing i have to compare it to is my xp90 on a msi k8n neo 2 which was about 30c idle, 48-52c load. Any ideas? Or could my new motherboard just be reading temps high. My gpu stays at 45c under load.
I think your Mobo is reading high. Either that or you don't have a good seat on your waterblock.
But back to topic.
Hrmm, how to best describe this. The water is able to "absorb" more heat than the same volume of air, so in reality, the same surface area's aren't needed. You don't to expelle as much heat back into the air, as the water is able to "absorb" more of it.
If that makes any sense. I beleive there is a post in the stickies section that goes into this.
if you were drawing from an infinite lake, then you would be correct, the water absorbs the heat and then its gone.
in our closed loops, the water merely acts as the connectoin between the block and the radiator. its just a good way to transfer the heat, and then it is dissapated somewhere else, instead of directly on the cpu. the radiator where its dissapated acts the same as the cpu heatsink, transfering the heat to the air.
you do expend the exact same heat into the air as you would with air cooling
I'l agree to disagree a little bit :D
But regardless, question is, we know that the radiator is essentially just a medium to transfer the heat from the water to the air. We have seen how Jet impigent designes on water blocks can increase turbelence and the ability of the water to absorb more of the heat. Wouldn't the same idea apply in reverse. Couldn't we create massive turbelence in the radiator to increase the waters conact with the radiator and thus its ability to transfer more heat to it? Just one idea.
The other, can we design a reservoir that can increase the cooling performance of the loop?
Can it all be done on a small scale?
Im curious what everyone thinks. Obviously the more surface area we have to work with the more heat we can disapate. But in the quest to make things function better, trying to make them smaller is a great way to find ways to boost performance of the full sized version.
i cant see how you disagree. where does this heat that the water absorbs go?
Over time it will go back out into the air. Just not as rapidly as it would on air cooling since its able to basorb more heat and hold onto it. That is my understanding anyway.
I've been known to be wrong, just ask the wife :p
But regardless, anyone have any ideas on how to miniturize?
nikhsub1
03-15-05, 03:20 PM
Couldn't we create massive turbelence in the radiator to increase the waters conact with the radiator and thus its ability to transfer more heat to it? Just one idea.
This is not just an idea, it is a reality. Many tranny coolers use turbulators inside the rad, to do exactly what your thought is. Problem is, most are aluminum. IMO, the smallest I would want to go that I believe would whoop any air cooling would be something like a swifech kit... 6000 block, MCP350, 3/8" ID tubing and the BIP with their rad box. It doesnt get much easier or smaller. The BIM's are just too small to effeciently handle anything more than say a GPU. Then again, the BIM could probably handle a Prescott and a GPU IF the system were run at stock speeds. I would bet the farm that good aircooling would kill that setup though.
Probably, but even the BIM2? Its rated higher than a BIP for cooling potential, and its dual 80mm fan configuration is much more case friendly.
Hrmm, we need to get some all copper rads with tubulators then.
nikhsub1
03-15-05, 04:02 PM
Probably, but even the BIM2? Its rated higher than a BIP for cooling potential, and its dual 80mm fan configuration is much more case friendly.
Hrmm, we need to get some all copper rads with tubulators then.
I missed the part about the BIM2, I assumed (wrongly) BIM. Anyway, the BIM2 is problematic IMO, it is WAY TOO THICK! That sucker is 46mm thick, compare that to the BIP which is 26mm thick. You would need some serious 80mm fannage to properly outfit a BIM2, perhaps 4 80CFM deltas would do, the noise would be horrible. Sticking 2 or even 4 quiet fans on there wont do the rad justice. I stick with the notion that the BIP with a quiet fan is a better solution.
Well actually the thickness of the BIM2 is slightly misleading. On the backside of the BIM2 is a built in shroud. The depth of the rad does not extend across the entire depth of the housing. I haven't measure the depth, but my best guess would be aproaching 2cm. Giving the rad an actual dpeth of roughly 26mm. Again this is my best guess. I've got two fans rated at 32CFM running on it and you would probably be surprised at the performance.
I just checked Danger den on the dimensions, and they are a bit misleading.
It says it has an overall dimension of 160 mm x 73.4mm which seems to indicat ethe hieght and width of the actual core.
and then says Compact 193 x 82 x 46 mm which seems to be the over all size of the housing.
It makes no mention of the depth of the core. Im guessing its roughly equal to that of the DIP so around 25mm.
vBulletin® v3.8.7, Copyright ©2000-2012, vBulletin Solutions, Inc.