Do I need to reduce my flow?

[robertm]

Am looking at my temps on my system 44C Loaded 30C in Case my IR thermometer show my radiator at 29C.

I would assume that the radiator would get over the case temp as the air flow is from the case over the radiator. (Am assuming I better be careful :) )

Could I be getting to much flow and the water is not in the block as long as it should be?

Eheim 1048 158 G/h
5x11Inch Radiator
one 120mm 89cfm

[Christoph]

I doubt that you can have a problem from water going too quickly through the water block. It does seem counterintuitive, though. My biggest worry would be slightly faster corrosion, but any good H2OCer should know how to avoid corrosion.

[r0ckstarbob]

you can never ever ever, have too much water flowing through your waterblock.

heat transfer isn't affected by momentum or gravity.

you can however, have not-enough-time in the radiator though. if your radiator can't cool down the water in the short amount of time that it spends in there, then you'll have problems. solutions? you want to increase the time your water spends away from your waterblock and (ideally) increase the amount of time it spends in the radiator so it has time to cool down properly. here are some techniques.

slow down the rate of flow through the radiator (yes this affects your block too unfortunately) -

get a bigger radiator -

get a more efficient radiator -
(or more powerful fans)

incorporate or increase the size of your reservoir -

the last method is probably the least efficient but cheapest method. someone called it a "thermal shock". seems to be an appropriate name.

good luck!

[Christoph]

As for effective radiators, you can get a bong cooler or you can go to a auto parts store and get a couple little heater cores.

With a bong, you can get temperatures below ambient and you get to build the thing yourself. Of course, I'm biased; I'm building a bong. :D The only thing about a bong is that you can't have a weak pump, since the water will have to travel a significant vertical distance after going throuhgh your computer. How far depends on the bong. Mine's somewhere aroung a meter tall, and I'm very confident in my Eheim 1250.

[robertm]

After I did some reading, and some more thinking. I see one thing I missed some time ago. Flow doesn't mater as much as
the heat exchangers I.E. the WB and Radiator.

Water cooling with a radiator is much like a heatpipe we are moving the heat from the CPU to the Radiator.

Let me try an example.

If I don't turn my pump on until the Water in the WB was 35C Then I pump water in until the Water in the WB dropped to 25C. As long as my Radiator could drop the water from 35C to 25C in the same time as it takes the water to heat 10C then I would have a working system.

Ok its not that simple but I think it makes my point. As I think todays WBs are about as effective as they will ever be in transferring heat. The only place we have to look is the radiator.

I think I need a small refrigerator in the room for a radiator and pop! :D JK

I think am right anyway ;)

[r0ckstarbob]

heh heh heh

chuck the bong. this is what i'm in the process of building...

check it

HERE (http://216.254.0.2/~language/interface/computer.html)

:) :) :) :)

[Christoph]

That's fine for you, but one of the advantages is the SIMPLICITY and the great results you get with a relatively simple system.

[SteenkyBastage]

i agree with the "cant have too fast thru waterblock" idea, and that brought up a thought (i think i strained something).

in that scenario (slow down the water so the radiator can do it's job) could one simply have the radiator copper tubeing be larger then the tubing in the system (waterblock, tubes, etch) or to have the supply into the radiator be split into separate feeds?

both these ideas would slow the water down significantly when going thru the radiator, then when leaving, have the tubesize go back to "normal". seems to me that would allow the radiator to be more effective (water in the radiator longer).

just a brainfart... maybe i have been without sleep too long...

[Warlord2]

you can have too much flow buts its in the 1000+ range

and that because SMALL water bubbles form in the waterblock

I think I read thats somewhere on this page long time ago

[r0ckstarbob]

Originally posted by Warlord2
you can have too much flow buts its in the 1000+ range

and that because SMALL water bubbles form in the waterblock

I think I read thats somewhere on this page long time ago

don't know if thats true or not but it would make sense.

[RangerJoe34]

yea, dangerden tested a maze waterblock at 1300 psi before it exploded
you can d/l a video at http://www.dangerden.com/base_pages/explode.htm go there and get the video, quite neat

[robertm]

One point I think some poeple missed if you slow your flow in the radiator you slow it down everyplacein the system. unless you bypass water past your radiator. But I think thats not a great idea.


I think its time to sleep :) I know i need it

[Morphoius]

I really don't think you can have too much flow. the only factor is the amount of heat that your pump puts out.

[Christoph]

Although it's probably just an academic observation, at very high pressure levels, the friction from the water against the walls of the tubing/block/radiator will become a factor.

Slepp iiiis ffoor teh week.

[SteenkyBastage]

Originally posted by robertm
One point I think some poeple missed if you slow your flow in the radiator you slow it down everyplacein the system. unless you bypass water past your radiator. But I think thats not a great idea.

ahhh, but that's only if your tubeing is the same size as the rest of the sytem. if you simply got a larger (or split up) tube size for the radiator it should slow down the water speed until you downsize (combine back to one flow) it after it's left the radiator.

same idea as putting your thumb over the end of a hose, the water moves faster when it's more restricted. but when you take your thumb off (wider opening, less pressure) the water comes out slower but in a larger flow (still overall get around the same ammt of water coming out, just at different speeds).

[docJ]

Yes you can have to much flow in you block and radiator. However it is not the pressure or gpm, it is the velocity of the water in the block or radiator and the path length inside them.

[Colin]

You bet you can have too much flow. I tried an 1800 GPH Pump with a seven foot bong, my temps went way up. The specifics are in an old dual 172 watt pelt thread.

If you think you have too much flow, put a hose clamp on the output of your pump. If your temps go up, your pump is too small. If they go down, look for the sweet spot with the clamp.

[Christoph]

1800 GPH!?!?!?!?! That seems crazy, even with a 7' bong. Did you forget to do research or just get a really good deal on it?
It sounds like quite an engineering feat to get that thing to work without bursting somewhere. What's it look like? That's kinda crazy, but it'd be quite a rush having anything 7' tall cooling your computer.
Did you get it to work OK? I'd like to know how if so.

[Colin]

No, I did my research. A magdrive pump will not provide such a high PSI to blow hoses, only more head. Do a search for my dual 172 watt pelt thread to find the details.

I have a 1200 GPH impeller for the pump. It will resurface soon in a new project...

[Christoph]

Sorry if that sounded too critical. No hostility intended.
That's a lot of power. Are your temps worth the price?

[Colin]

Naw, it was not too critical. How can you put a price on low temps? :D

[r0ckstarbob]

i dunno but someone obviously decided to put the value in dollars. here are a few of our bennefactors

vantec
swiftech
thermalright
arctic silver
danger den
antec
vapochill
kryotech
panaflow
sunon
delta
koolance...




*grumble grumble*

[Al666]

I think faster the water flows better the heat transfers from the WaterBlock and longer the coolant stays in the radiator colder it gets... right? So how about using dual pumps???
Also it's a failproof setup, if one of the pumps fails...

BTW I started a new thread about this so I simply copy/paste what I posted...

I'm planning to use dual pumps in my setup...
The 145GPH submersible pump that comes with the OcersHideout aluminum reservoir, and inline Eheim 1048 158GPH...

What is the optimum placement for the second pump?
I first thought about placing it betw. the waterblocks but then it might add heat to the coolant going to the 2nd WB.
I'm thinking about placing it after the waterblocks, before the BlackIceII radiator.

Any suggestions?

[Colin]

Originally posted by Al666
.

Any suggestions?

For most watercooling setups, stick with a good 500 to 700 GPH pump and restrict the flow with a hose clamp if necessary to find the sweet spot. Dual pumps add complexity and another costly variable to work with.

[m1066ad]

RBD is on the right track. Flow speed through the radiator would be entirely dependent upon the ratio of the volumes in the water block and the radiator. I.e., if the volume of the waterblock was 250cc, and the radiator had a volume of 2500cc, water would spend 10 times as long in the radiator, as in the waterblock.
RockstarBob, it seems to me, that basically what you're trying for there, is putting most of the computer in a refrigerator, with an active dessicator. Sounds awfully elaborate, lol, but it should certainly do the trick...:D
BTW, yer either very talented with MS Paint, or you have waaaay too much time, and patience...:cool:

[Colin]

Originally posted by m1066ad
RBD is on the right track. Flow speed through the radiator would be entirely dependent upon the ratio of the volumes in the water block and the radiator. I.e., if the volume of the waterblock was 250cc, and the radiator had a volume of 2500cc, water would spend 10 times as long in the radiator, as in the waterblock.


In the real world, you have to put these two factors together. That's why you need to check the flow in your particular system and optimize it for best results.

[m1066ad]

Short answer: bigger radiator=better

[Colin]

A bigger radiator is not the universal panacea. You need to determine the optimal flow for your particular system!!!

[r0ckstarbob]

Originally posted by m1066ad
RSB is on the right track. Flow speed through the radiator would be entirely dependent upon the ratio of the volumes in the water block and the radiator. I.e., if the volume of the waterblock was 250cc, and the radiator had a volume of 2500cc, water would spend 10 times as long in the radiator, as in the waterblock.
RockstarBob, it seems to me, that basically what you're trying for there, is putting most of the computer in a refrigerator, with an active dessicator. Sounds awfully elaborate, lol, but it should certainly do the trick...:D
BTW, yer either very talented with MS Paint, or you have waaaay too much time, and patience...:cool:

dats it boss man jou gots it right on da head. isolate the components that will bennefit from the cooling apparatus and go to town.

who me? time? nah. i'm a graphic designer. i LIVE in photoshop/illustrator - MS paint is a walk in the park even if it IS like trying to split atoms with a spatula strapped to the bumper of a small sports car.

(sshhhhh... i went back in and touched it up in PhotoShop and forgot to update the site- don't tell anyone - i'm REALLY behind on updating that thing.)

elaborate? yeah... unfortunately the research thats gone into this thing is even more elaborate then it's eventual manifestation will be. it's not all for naught though... have been approached by Abit, Epox, Antec and AMD about it strangely enough. guess it's time to slap the "Patent Pending" up there too apparently. had to apply last month... for the life of me i'm not entirely certain what about it is patentable but the lawyer seems to think so...

my life is so much weirder then anyone could have possibly dreamed up.

*shakes head*

[m1066ad]

Why? Unless you're a seeker of engineering beauty, or an accounting beancounter, it boils down to the longer the water is in the radiator, the closer to ambient it will get. The waterblock has to remove the heat at least as fast as it's produced, so there are limits as to how small it can be, and how slow the water can move through it. Provided the waterblock has enough flow through it to carry the heat away, then bigger radiator=better. If the flow through the waterblock isn't enough to carry the heat away, then everything else is rather moot.

*the above was a reply to Colin, rsb beat me to the post*

Putting a computer in a refrigerator is patentable???

[r0ckstarbob]

okay boys, we're not talking about a 1000 gph pump here, we're talking about a 158 GPH. for the purposes of this, a bigger radiator WILL be more benneficial. He's got a good pump and isn't pushing too much coolant by a long shot and it's not going to be a very practical solution to reduce the pump speed in this instance, though it may provide some improvement.

yes "tuning" your system with a clamp on one of the hoses is going to be a good idea.

this jist is this.

mr man, you have an insufficient radiator to meet the efficiency of your pump. get a bigger/more efficient radiator or write off the money you spent on that very nice pump of yours and know you could have gone with a slower, cheapo RIO pump and gotten better results.

i'm afraid your options are this. upgrade or downgrade. your call.

good luck. keep us posted.

[r0ckstarbob]

Originally posted by m1066ad
Why? Unless you're a seeker of engineering beauty, or an accounting beancounter, it boils down to the longer the water is in the radiator, the closer to ambient it will get. The waterblock has to remove the heat at least as fast as it's produced, so there are limits as to how small it can be, and how slow the water can move through it. Provided the waterblock has enough flow through it to carry the heat away, then bigger radiator=better. If the flow through the waterblock isn't enough to carry the heat away, then everything else is rather moot.

*the above was a reply to Colin, rsb beat me to the post*

Putting a computer in a refrigerator is patentable???

*chuckle*

that was MY question. apparently i can patent the apparatus under some kind of stipulation in the patent law about applying known inventions in new techniques or somesuch. i think i can patent the apparatus itself, just not the individual components. i don't know. i'm letting the lawyer handle that junk. better safe then sorry. i'll know in a month or two if the patent goes through. if not, no biggie. i'm not concerned one way or the other. i'm just designing/building the thing. i pay other people to haggle over the leagalities of it (providing there are any) and CMA where i can. :D

i'm a designer. its what i do. as for the rest... bah

[Colin]

r0ckstarbob - Sorry but without knowing the details of the rest of the plumbing in the system, your comments are irresponsible. A few 90 degree bends or a bunch of small ID plastic barbs will restrict the flow to the point that the radiator or pump may not be the main issue.

BTW, in my experience, RIO pumps will result in the eventual death of your CPU. RIO pumps stink!

[r0ckstarbob]

i'm sorry. you however are apparently in posession of this knowlege then?

i'm working with what info the man has given us. my comments are not irresponsible. they are my best guess against the question posed, using the information that he has posted and referenced against my own knowlege of thermodynamic theory and practice in regards to computer cooling. my knowlege isn't perfect and not without gaps but it's pretty frickin solid and my comments not without foundation.

please take a breath and step back.

most people don't have many 90 degree angles in their tubing or barbs from what i've gathered. i can only assume that
- he doesn't either
- is running a relatively standard setup with the standard set of eventualities and 'isms' that most watercooing setups have
- standard troubleshooting techniques will apply to his watercooling aparatus.
- and that standard thermodynamic law is going to apply to his setup.

i'm not saying your wrong mr man. i'm saying that there are options to consider, yours included (if you'll note in my last comment on this).

peace brotha. its a cooling forum. not a boxing ring.

PS - i fully agree... RIO sucks bigtime. point being - a cheaper, lower GPH pump would have worked as well or better if restricting the flow is indeed the issue at hand.

[m1066ad]

Hmmm...how many gph would really be necessary, to run enough water through the waterblock to draw the heat off the processor? This is an interesting question. It would depend on the temp of the water flowing in...This is more math, than I'm up to, right at the moment...any engineers, here?

[Colin]

r0ckstarbob - Sorry I woke up thinking about this problem and realized my words were a little harsh. I really did not intend them to be. I just don't like recommending anyone spend money on new components when the current ones may not be tweaked and peaked. Please accept my apology.

While I have run into systems that don't have enough radiator capacity, more often I have found them to be shy on the pump side of the equation or full of seemingly minor things that in practice restrict the flow.

m1066ad - Take a look here (http://www.hardforums.com/showthread.php?threadid=142512) and here. (http://www.hardforums.com/showthread.php?threadid=153235) . There is also a great piece on actual flow rates by Rockhammer floating around on the Net. You will probably find a link to it in one of the threads I linked to.

[robertm]

Well I would have never started this thread had I give the problem more thought.

I need to get ride of more heat in my water. Slowing the flow might show some very slight improvements. In the end I have to remove more heat from the water at any speed it is flowing.

I need a better radiator or get more air flow across this radiator.

I have a swiftech kit when it comes down to it there pump block and radiator. Only unlike there case I have only one 120mm fan as my case is to small for any other. No 90s or other such things a few elbows as in there system.

With my system I have two choices move the radiator from the box and make it an add on unit and add a second fan.

Or I have a 2nd radiator that has about three times the water capacity of my current system do to its size and type. I am thinking very hard about using the other radiator. As it will do two things. One give me more thermal capacity as I will have more water in my system. Also more water will be in the radiator part of the system cooling. This should reduce my temps.

[BillA]

wrt flow rate:
am completing the testing of 10 radiators at flow rates from 0.4 to over 3 gpm
and can conclusively state that lower flow in a rad is always better
with an increase in flow from 0.4 to 1.0 gpm there will be a reduction of 10 to 20% in the rad's heat dissipation
(depending on the rad's design, materials, etc.)

and quite the opposite applies to wbs, more flow is always better
(at least for what is presently commercially available)
so the "sweet spot" sometimes referred to is the balancing of the compromised performance of the rad and wb

if putting together a system from scratch (as other posters have observed); a higher volumeric capacity rad is better - but this is but one of several rad parameters

look at the ratio of wb volume to rad volume (the velocity ratio will be the same)
some values:
a highend wb - 66 ml
a 5X5 in. flat rad - 72 ml
a popular heater core - 220 ml
(why don't the vendors provide this info ?)

be cool

[r0ckstarbob]

Originally posted by Colin
r0ckstarbob - Sorry I woke up thinking about this problem and realized my words were a little harsh. I really did not intend them to be. I just don't like recommending anyone spend money on new components when the current ones may not be tweaked and peaked. Please accept my apology.

While I have run into systems that don't have enough radiator capacity, more often I have found them to be shy on the pump side of the equation or full of seemingly minor things that in practice restrict the flow.

m1066ad - Take a look here (http://www.hardforums.com/showthread.php?threadid=142512) and here. (http://www.hardforums.com/showthread.php?threadid=153235) . There is also a great piece on actual flow rates by Rockhammer floating around on the Net. You will probably find a link to it in one of the threads I linked to.


's cool. no harm done.

i agree - it's all the little things about these setups that will get us and if theres a quick cheap solution, it should certainly be considered over a more expensive and complex one. in the long run, personally i've always ended up having to go the latter route eventually anyways though.

Billa, you make an excellent point about the "sweet spot" that i hadn't considered. it makes alot of sense.

learn the process, and refine the technique - and so we go...

[m1066ad]

Thanks for the links, BillA:) Now I have another forum to follow...:eek:
*Whew* That was a long thread!:cool:
However, it didn't answer the question I asked.
Just to simplify it, here, let's leave the radiator out of the equation, entirely, right now. (Stay with me, here)
Let's, oh, say we a sufficient quantity of water, at ambient temperature, that we can simply pump it past the water block, without returning it(use it to water the lawn...)
I'd like to see the curve plotted, of cpu temp/gph
I know the return will fall off, as the temperature of the cpu gets closer and closer to ambient(that is, as the gph approaches infinity, the cpu temp will approach ambient, and never *quite* get there)
We want to know where the "reasonable" range is, with a curve.(and yes, I realize 'reasonable' is subjective, here...)
I believe the best approach to the problem of cpu cooling is to start at the cpu...figure out what is necessary to cool the water block near ambient, and then design the heat exchange system(radiator/fan) to return the heated water to near ambient.
Hmm, this sounds like a positive feedback loop...but in real life, it will stabilize at a steady state, the approach to ambient at either end of the equation limited by space/cost considerations.
Of course, each end of the loop will affect the other, so theoretical curves are just that...but at least we will have something to work with, and towards.
Engineering is, after all, part art, part science, the art of using what you have, to approach the theoretical limits :D

[BillA]

this is a (mislabeled) plot of a popular wb's bp temp difference with the coolant vs. the flow rate

http://www.ecom-answers.com/wpbptemps.gif

for a 100W load from a heat die with TCs embedded in it and over the die area in the bp

the coolant temps are not so informative with a rise of 0.08^C at 3.04 gpm vs. 0.85^C at 0.42 gpm

(think for a bit on the instrumentation needed to get those temps)

be cool

[r0ckstarbob]

coo dat. lovin me some charts these days. thanks billa.