Water Flow Characteristics

[XyBeRWaReZ]

Does anyone have any info or links to sites that talk about water flow characteristics? I'm interested on how water is affected when you introduce obsticales and it's efficiency. I can't seem to really find any info, or I'm just not searching hard enough.

[bobo5195]

try putting all your questions in one thread and then they might get answered

[bobkoure]

Yup, I'm afraid you're not searching hard enough.
Try civil engineering or plumbing for actual in-pipe flow characteristics (the CE stuff is IMHO more interesting as it'll go into why and the plumbing stuff is typically just formulas).
Nope, don't have any links handy. Google is your friend.

[bigben2k]

Google for ELSEVIER Referex. I can't say any more.

[XyBeRWaReZ]

Originally Posted by bigben2k

Google for ELSEVIER Referex. I can't say any more.

DAMN!!! I've seached this info on Wikipedia and read up some about it. Is there any type of software and not just reports on the matter? It'd be a lot easier for me to understand if I could see it in action using software then trying to picture it all in my head. Also isn't there some report that has just laymen terms on how water flows and turbulence?

I feel like my mind will blow up if I keep reading those types of reports, hehe.

BAAAAAAAAAAAMMMMMMMMMMM!!!!!!!!!! Too late

[bobo5195]

CFD (computational fluid dynamics) software really needs a engineering degree to know how to use it. Modeling turbulence directly needs a fast computer and alot of time. For youre average waterblock it might take 24 hours and a supercomputer to work out fluid flow and turbulence. Can use turbulence modeling and lamina flow of course but even then the software is complex.

If you want to learn more get a book on fluid dynamics from the library. I always stand by my recommendation of this book for water block design. Probably need a book on fluid mechanics as well as it is light in that area.
http://www.amazon.com/gp/product/047...lance&n=283155

White is quite good but there is a huge range of text books when it comes to fluid mechanics.
http://highered.mcgraw-hill.com/sites/0072402172/

To do the analysis to the standard that i think that you want needs an engineering degree + maybe a pHd. After that you might come back and edit your question because it is too hard.

[XyBeRWaReZ]

Originally Posted by bobo5195

CFD (computational fluid dynamics) software really needs a engineering degree to know how to use it. Modeling turbulence directly needs a fast computer and alot of time. For youre average waterblock it might take 24 hours and a supercomputer to work out fluid flow and turbulence. Can use turbulence modeling and lamina flow of course but even then the software is complex.

If you want to learn more get a book on fluid dynamics from the library. I always stand by my recommendation of this book for water block design. Probably need a book on fluid mechanics as well as it is light in that area.
http://www.amazon.com/gp/product/047...lance&n=283155

White is quite good but there is a huge range of text books when it comes to fluid mechanics.
http://highered.mcgraw-hill.com/sites/0072402172/

To do the analysis to the standard that i think that you want needs an engineering degree + maybe a pHd. After that you might come back and edit your question because it is too hard.

Thanks, duely noted! I think I'll stop asking about it now and just try the trial and error method.

[bobo5195]

you can do abit but trust me its not simple.

Best thing to do is a thought experiment. Imagine how the flow goes in your head, avoid sharp edges and expansions/ contractions. More surface area is a good thing, more velocity near a surface is a good thing. From that you get tightly spaced fins. you start thinking the fins should be very thin, but wait heat can't get to the top then. So the fins have to be a good size etc etc.

If there was a magic bullet that was easy for these things someone would of tried it a long long time ago.

[XyBeRWaReZ]

Originally Posted by bobo5195

you can do abit but trust me its not simple.

Best thing to do is a thought experiment. Imagine how the flow goes in your head, avoid sharp edges and expansions/ contractions. More surface area is a good thing, more velocity near a surface is a good thing. From that you get tightly spaced fins. you start thinking the fins should be very thin, but wait heat can't get to the top then. So the fins have to be a good size etc etc.

If there was a magic bullet that was easy for these things someone would of tried it a long long time ago.

Very true. But I would appreciate it if I could get some tips on water flow, from what everyone else knows and have experimented with.
As I undertand it, when desigining a waterblock you want water to go through the block and carry as much heat off as possible. But the only thing is tying to figure how slow the water has to go through in order to pick up as much heat along the way.
That's where I guess designing fins and what not inside of the blocks. With those fins, it should slow down the water enough to pickup the head but be able to flow easily enough out afterwards.
Actually wouldn't a design of fin that's thicker in the middle and is sharp at both ends work pretty well? Water would enter pretty easily at the beginning because of the sharp edge and then move along to the middle where hopefully a lot of the heat has passed through from the CPU, and then exiting the back end of the fin where it's thin again.
Please let me know what you guys think of that? It seems pretty logical, but I'm wondering if water would move too quickly through it and not pick up enough heat, causing it to be somewhat inefficient.

[bobo5195]

tapered fin idea may work, lower pressure drop for one. Bit hard to make though.

More water speed is good. Don't worry about it going through to fast for the moment. Worry about the amount of wb touching the fluid.

[bigben2k]

Bobo's on the right track; you have to visualize in your head, the flow. Think about where the speed changes (slow for large openings, fast for small ones). Go over some existing designs for this exercise (save the Storm/Cascade block for last).

[Long Haired Git]

Originally Posted by XyBeRWaReZ

But the only thing is tying to figure how slow the water has to go through in order to pick up as much heat along the way.

Slow water is bad. Faster water is more water and as there is water behind it that is cooler the more water the better.

The only issue with higher water velocity is that it creates more resistance which adds to the back pressure which reduces flow rate. The balance between "head" and velocity is always, errm, a balance.

Originally Posted by XyBeRWaReZ

Actually wouldn't a design of fin that's thicker in the middle and is sharp at both ends work pretty well?

Swiftech think the Apogee works very well....

[XyBeRWaReZ]

Well I'm thinking of using 2 Via Aqua 1300 to solve the water flow problem.

It should right? One is good, but two would be better? I'm going to put it in a push-pull system. One to push at the beginning of the system, and the second half way through pulling, just to make sure that it can handle anything thrown at it.

My system will consist of many waterblocks.

1. CPU
2. GPU
3. Ram
4. Hard Drive
5. Custom waterblock for 8 pelts, in a 2x4 setup
6. Northbridge

So making sure enough water pressure will be a most serious consideration,

[qet77]

These pumps are of the centrifugal variety and therefore they do not have the suction you are thinking of that would benifit a system if it were placed after the water blocks. A positive displacement pump would be necessary to get the full effect of placing a pump after the waterblocks, typical positive displacement pump are of the roots, screw, rotor, and parastaltic variety (I am sure there are others as well). These types of pumps will displace a given volume for each rotation as opposed to providing a volume based on the inlet pressure and volume, and the rpm of the spinning compressor (This means that if you plug a centrifugal pump nothing would really happen aside from excessive load on the motor, but if you blocked a positive displacement pump something would break as you are stopping a component in a system with mechanical connections).

I am not an engineer so my understanding of this may be somewhat limited but I feel the above is a reasonable explaination of the concept.

[XyBeRWaReZ]

Originally Posted by qet77

These pumps are of the centrifugal variety and therefore they do not have the suction you are thinking of that would benifit a system if it were placed after the water blocks. A positive displacement pump would be necessary to get the full effect of placing a pump after the waterblocks, typical positive displacement pump are of the roots, screw, rotor, and parastaltic variety (I am sure there are others as well). These types of pumps will displace a given volume for each rotation as opposed to providing a volume based on the inlet pressure and volume, and the rpm of the spinning compressor (This means that if you plug a centrifugal pump nothing would really happen aside from excessive load on the motor, but if you blocked a positive displacement pump something would break as you are stopping a component in a system with mechanical connections).

I am not an engineer so my understanding of this may be somewhat limited but I feel the above is a reasonable explaination of the concept.

Damn, that kinda killed my hopes for a two pump setup. What brand of pump could I use in a push pull setup, or should I just stick to one?

[bobkoure]

You can run pumps in series. If they're the same model, then you're essentially adding pressure at the same flow rate.
But why not just start out with a pump that has the characteristics you're looking for?

Originally Posted by LHG

Swiftech think the Apogee works very well....

But does anyone else? From the pictures it appears to be a 5000/5002 sink (diamond pins in diagonal array) with different top and hold-down.
I realize that just a minor change in pin dimension, alignment to flow, and deck height can make a big difference - so what did they do? And is it really that much better than, say, the 6000? It's touted as having less flow resistance, but that could just be a higher deck height (like the resistance difference between the 5002 and 5000). Just wondering aloud...

[XyBeRWaReZ]

Originally Posted by bobkoure

You can run pumps in series. If they're the same model, then you're essentially adding pressure at the same flow rate.
But why not just start out with a pump that has the characteristics you're looking for?

If it's possible then I'll be running two Via Aqua 1300 pumps. Using two pumps in series won't be a problem? I'm assuming that you mean one pump hooked right after the other? What type of flow rate change might I get from hooking it up in that way?

[bobkoure]

In theory, you get no flow rate change, but a pressure increase (assuming a zero restriction loop).
Remember that some pumps are sealed based on the assumption that there'd be low pressure at the center of the rotor - and that the second pump now doesn't have this. (oft discussed issue with the first verson c-systems pumps, but may be an issue with others)

[XyBeRWaReZ]

Originally Posted by bobkoure

In theory, you get no flow rate change, but a pressure increase (assuming a zero restriction loop).
Remember that some pumps are sealed based on the assumption that there'd be low pressure at the center of the rotor - and that the second pump now doesn't have this. (oft discussed issue with the first verson c-systems pumps, but may be an issue with others)

Well with a water cooling setup, there will never be zero restriction right? I'm not sure how much restriction my system would have using six different different waterblocks. Would one Via Aqua 1300 be enough in my system or should I use two in seris like suggested?

[BGP Spook]

Yes, there is no such thing as zero-restriction. (unless it is standing still...even then on the molecular level H2O does vibrate and "bounce around" so I guess even then there is no such thing as zero-restriction...) Reality is never as simple as it should be.


I am currently using one Via Aqua 1300, so I have some experience with it. I am not, however, privy to it's *real* performance characteristics.

I have about 8' of 1/2" ID tubing, one CPU block (6002-A), and a reservoir in the loop.
The impression I get is that I chose the right pump for the job. If you were to put as much restriction as you are talking about in a loop then I doubt even two could maintain acceptable flow rates. They are not that powerful, the biggest selling point for me was actually the price.

You could try it of course.

You could also not put so many blocks in the loop. Most consider RAM, hard drive, and northbridge blocks to be unnecessary. Some 10k rpm hard drives, the NF4 NB, and high voltage(3.2v ?) memory aside you probably will never know the difference and even then it is debatable sometimes. It may actually make them hotter overall, depending on the total heat absorbed and dumped by the system.

I do not want to rain on your parade, by all means if you think you want/need that many blocks in you loop then do it. Regardless I would probably not use two Via Aquas in series.

[XyBeRWaReZ]

Originally Posted by BGP Spook

Yes, there is no such thing as zero-restriction. (unless it is standing still...even then on the molecular level H2O does vibrate and "bounce around" so I guess even then there is no such thing as zero-restriction...) Reality is never as simple as it should be.


I am currently using one Via Aqua 1300, so I have some experience with it. I am not, however, privy to it's *real* performance characteristics.

I have about 8' of 1/2" ID tubing, one CPU block (6002-A), and a reservoir in the loop.
The impression I get is that I chose the right pump for the job. If you were to put as much restriction as you are talking about in a loop then I doubt even two could maintain acceptable flow rates. They are not that powerful, the biggest selling point for me was actually the price.

You could try it of course.

You could also not put so many blocks in the loop. Most consider RAM, hard drive, and northbridge blocks to be unnecessary. Some 10k rpm hard drives, the NF4 NB, and high voltage(3.2v ?) memory aside you probably will never know the difference and even then it is debatable sometimes. It may actually make them hotter overall, depending on the total heat absorbed and dumped by the system.

I do not want to rain on your parade, by all means if you think you want/need that many blocks in you loop then do it. Regardless I would probably not use two Via Aquas in series.

I really wouldn't use the waterblock for everything else if I wasn't planning on overclocking a lot. I'll be doing a voltage and bios mod to to an eVGA Geforce 6800GS video card and pushing that to it's extreme limits, so I will need as much cooling as possible. Also my CPU is an AMD XP 1600+ and I'll be pushing that way over 2ghz, and it's really hot as it is now. So in reality, waterblocks on pretty much everything will be required, at least I think so, and being in Hawaii where the weather is pretty hot doesn't help it at all.
If using 2 Via Aquas not work well with my setup, would you have any other pump recommendations? I need to make sure enough water is going through everything to effectivly cool it.

[Long Haired Git]

Adding a second pump improves flow.
Whether the extra heat from the pump offsets the benefits of the extra flow is another question.

Here's "a" PQ graph for the 1300: http://www.wcaquatics.com/cgi-bin/it...aqua_pump.html

I've not got approximator to where it needs to be, but the closest pump to it appears to be the Laing D4 which can push the same LPM but can handle more head, and the MCP350 which can get the same LPM but not as much head.

Anyway, I've used a MCR320QP rad and have not factored any tubing (coz its broken) and have used 2 x Apogee blocks , 2 x MCW60 blocks and 2 x MCW30 blocks. The Laing D4 gets over 4 LPM which is certainly enough, whilst the MCP350 gets over 2.5 LPM which should be enough.
Averaging suggests the 1300 would get around 3 LPM which is enough.
I'd almost be certain any extra flow from a second pump would be offset by its heat and not make a noticeable gain (or something like half a degree).
Far better to put a second radiator in for that much heat...

[XyBeRWaReZ]

Originally Posted by Long Haired Git

Adding a second pump improves flow.
Whether the extra heat from the pump offsets the benefits of the extra flow is another question.

Here's "a" PQ graph for the 1300: http://www.wcaquatics.com/cgi-bin/it...aqua_pump.html

I've not got approximator to where it needs to be, but the closest pump to it appears to be the Laing D4 which can push the same LPM but can handle more head, and the MCP350 which can get the same LPM but not as much head.

Anyway, I've used a MCR320QP rad and have not factored any tubing (coz its broken) and have used 2 x Apogee blocks , 2 x MCW60 blocks and 2 x MCW30 blocks. The Laing D4 gets over 4 LPM which is certainly enough, whilst the MCP350 gets over 2.5 LPM which should be enough.
Averaging suggests the 1300 would get around 3 LPM which is enough.
I'd almost be certain any extra flow from a second pump would be offset by its heat and not make a noticeable gain (or something like half a degree).
Far better to put a second radiator in for that much heat...

Thanks for this info. This will help me out. So I'll only use one pump then. Maybe I might try two pumps out one day anyway as the 1300's are pretty cheap. It'll cost me only an additional $20. Could always use a backup pump if anything should go wrong anyway.

[Long Haired Git]

Yep, not sure what the PQ graph would be for a non-spinning pump, but centrifugal pumps do allow fluid to pass when they are not powered, so two pumps in series do allow a degree of "backup".

[Butcher]

I'd suggest a pump with more head capacity than the via aqua for that many blocks. Or even two pumps with high head maybe.