The "What's your Water-cooling System's Flow Philosophy" Poll

What's your Water-cooling System's Flow Philosophy?

  • Total voters
    138
Top Nurse said:
So I am curious as to why you guys don't use push-fits? Seems you could really get some flow gains by actually having a true 1/2" ID. Also makes me realize that a 1/2" ID tube may only be seeing say 3/8" ID...so a 3/8" ID setup with push-fits may be a better way to go in that scenario?
Click to expand...

The main reason is lack of space. 1/2" ID tubing pretty much implies 3/4" OD, and can you imagine trying to squeeze a 3/4" push-fitting onto most waterblocks? This is especially an issue with middle-in blocks. Where does one then locate the outlet fitting?

Maybe less of an issue with modern CPU's with widely spaced mounting posts, but with the older Socket A and Socket 370 (Pentium 3) mountings it was seriously an issue.

The scale of the issue isn't terribly serious though unless trying to achieve >10LPM flow rates at which point then the barb fitting transition resistance does start to become somewhat significant for many pumps. Still, even at this point, 3 transition points worth of resistance still only equates to about the same level of resistance as the 1/2" ID tubing itself is offering, so it's not a total flow killer.

One of the issues with push-fit fittings though is that they are (obviously) OD dependent, and this can limit the range of tubing choices. If we stuck 3/4" push-fittings on every-block, then people would be forced to use 3/4" OD tubing, and couldn't use anything smaller if they wanted to.

Finally, with more modern waterblock development, actual flow rates of >8LPM generally are totally unnecessary, unlike 3 years ago where >10LPM flow rates were desirable, nowadays that's more like a far more reasonable 6-8LPM actual flow rates that makes for a top-end performing high-flow setup, and at 6-8LPM, the barb transition resistance isn't heavily significant at all, especially in comparison to the waterblocks.
 
Have you noticed that the poll graph is starting to resemble a standard bell curve?

Okay lets assume I were to change from a 6mm ID tube to an 8mm ID tube on my Cuplex XT, Aquastream, Aquatube, and BI Pro loop. I get with the push-fits a pretty much unadulterated 6mm flow pretty much all the way through the system. I just measured the 8mm x 1/8 BSPP compression fitting I have here and its ID is about 5.5mm. So how does increasing the tube size to 8mm give me better flow?

I also wonder how much of this change in ID in other tube sizes is chasing a red herring? Almost all blocks made have to make 90 degree turns just to get the water in the block and out again, but I hear constantly that using 90 degree elbows are a bad idea. But it also seems to me that the commonly used plastic 90 degree fittings for bigger bores are not very well hydrodynamically manufactured. Like when I look at Legris push-fit 90 degree elbows I see a very nice transition from one direction to the other. Also the distance from the end of one tube to the beginning of the next tube is only about 12-14mm.
 
Cathar said:
The main reason is lack of space. 1/2" ID tubing pretty much implies 3/4" OD, and can you imagine trying to squeeze a 3/4" push-fitting onto most waterblocks? This is especially an issue with middle-in blocks. Where does one then locate the outlet fitting?

The scale of the issue isn't terribly serious though unless trying to achieve >10LPM flow rates at which point then the barb fitting transition resistance does start to become somewhat significant for many pumps. Still, even at this point, 3 transition points worth of resistance still only equates to about the same level of resistance as the 1/2" ID tubing itself is offering, so it's not a total flow killer.
Click to expand...

So what your saying is that there really isn't that much of a difference in actual tubing size between a system using a 3/8" push-fit design and a 1/2" barb or compression fitting design? Is that why the people using barb or compression fittings get penalized 1 point on the poll questionaire? I think I'm starting to get a headache, but from what you are saying, is that all other things being equal, you should get just as good results with the 3/8" push-fit design as you would the 1/2" barb correct?

I have been thinking I might ask for the next Swan-Ganz catheter setup that goes out of date and do a little investigating on my own of intra-line pressure and flow readings with the use of the ICU monitors. We can do very accurate flow calculations based on temp difference between cool water injected into a line and the time it takes to go from injection point to the thermistor. I can also look at the pressure and see what effect it has when you introduce one or more 90 degree elbows into the loop. Essentially I would be putting a very small catheter inside the water loop. Just got to figure out how to get it in. ;)
 
Happy Hopping said:
Aqua is consider as ?? moderate flow I presume?
Click to expand...

Well that depends a lot on how you configure your system, what pump, what tube size, what type of radiator, what fitting type, how many blocks and what type of design they are. I cheated because I have Black Ice rads in my setup and I have the Aquastream pump wound up to a little better than Eheim 1048 levels. Depending on your loop it could be, according to Cathar's poll, anywhere from low flow to moderate high flow. Real eye opener, huh?
 
Cathar said:
...the AquaStream is an over-volted Eheim 1046 that behaves somewhat like an Eheim 1048.
Click to expand...

While your synopsis of the Aquastream is correct in function it is not correct in design type. The Aquastream pump is actually an Alternating Current (AC) pump. Yes it does run on Direct Current (DC), but the controller converts the 12 VDC to 9VAC before it actually gets to the pump. The reason it can get such a wide range of flow is that the controller changes the Hz (or cycles per second) of the AC voltage from something like 45 Hz to a maximum of 74Hz IIRC. Changing the HZ rate on an AC motor changes the RPM of the motor. So you still get long life and varied flow at the same time.

This was one of the main reasons I bought this pump as German engineering + high tech design = winner IMHO. :) That and the Aquaero allows you to connect the Aquastream controller to the software interface called the Aquasuite. I think you would really like that interface as it gives a lot of specialized control over the watercooling system and also gives you a lot of data back as to what is actually going on inside your loop. I also think it is possible to utilize the Aquaero with just about any size tube you might want to use with just a few hardware mods to the fittings and some minor adjustments to the calibration sequences.
 
Top Nurse said:
So what your saying is that there really isn't that much of a difference in actual tubing size between a system using a 3/8" push-fit design and a 1/2" barb or compression fitting design? Is that why the people using barb or compression fittings get penalized 1 point on the poll questionaire? I think I'm starting to get a headache, but from what you are saying, is that all other things being equal, you should get just as good results with the 3/8" push-fit design as you would the 1/2" barb correct?
Click to expand...

3/8" + push-fit = +2 points
1/2" + barbs = +1 point

That's about right in the scheme of things. Higher pressure drop (lower achievable flow rates) with 3/8" + push-fit as with 1/2" + barbs, but not dramatically so. I mean think about it, if the barbs are 3/8" ID, then you've taken 1/2" the whole way until you reach that 3/8" bit at the end, whereas with 3/8" + push-fit it's still 3/8" for the entireloop.
 
Top Nurse said:
While your synopsis of the Aquastream is correct in function it is not correct in design type. The Aquastream pump is actually an Alternating Current (AC) pump. Yes it does run on Direct Current (DC), but the controller converts the 12 VDC to 9VAC before it actually gets to the pump. The reason it can get such a wide range of flow is that the controller changes the Hz (or cycles per second) of the AC voltage from something like 45 Hz to a maximum of 74Hz IIRC. Changing the HZ rate on an AC motor changes the RPM of the motor. So you still get long life and varied flow at the same time.
Click to expand...

Yeah, I knew how it worked, I was just relaying it in a more commonly understood fashion, but thanks for the explanation anyway. It's the same reason as why Iwaki pumps work better in North America than in Europe (60Hz vs 50Hz).
 
Cathar said:
3/8" + push-fit = +2 points
1/2" + barbs = +1 point

That's about right in the scheme of things. Higher pressure drop (lower achievable flow rates) with 3/8" + push-fit as with 1/2" + barbs, but not dramatically so. I mean think about it, if the barbs are 3/8" ID, then you've taken 1/2" the whole way until you reach that 3/8" bit at the end, whereas with 3/8" + push-fit it's still 3/8" for the entireloop.
Click to expand...

Okay that is where I don't understand. If you have a 1/2" tube and drop it down to say 3/8" at multiple places in a loop you shouldn't see any real additional flow gains as it is probably increasing pressure at that point and decreasing flow, right?
 
Cathar said:
Yeah, I knew how it worked, I was just relaying it in a more commonly understood fashion, but thanks for the explanation anyway. It's the same reason as why Iwaki pumps work better in North America than in Europe (60Hz vs 50Hz).
Click to expand...

I thought as much on your explanation, :) but a lot of people seem quite ignorant of how those 12 Vdc Eheim pumps really work so I thought I would make sure.
 
Top Nurse said:
Okay that is where I don't understand. If you have a 1/2" tube and drop it down to say 3/8" at multiple places in a loop you shouldn't see any real additional flow gains as it is probably increasing pressure at that point and decreasing flow, right?
Click to expand...

Total flow resistance is a sum of effects.

Here's a nice pressure-drop calculator => http://www.pressure-drop.com/

Let's assume 2m of tubing. For the 1/2" scenario, let's assume 4 places of barb based reduction, being the inlet of the pump, the inlet of the waterblock, inlet of the reservoir, and the inlet of the radiator. Let's assume a high flow rate of 9LPM, and a 26C water temp.

2m of 3/8" tubing the whole way = 1.89mH2O of pressure drop due to the tubing alone

as opposed to

2m of 1/2" tubing = 0.42mH2O of pressure drop
1/2" ID to 3/8" ID sudden contraction = 0.21mH2O of pressure drop per contraction
Total pressure drop = 0.42 + 4 x 0.21 = 1.26mH2O

Need to break it down into its total of pressure drop effects.

If you had 6 or more points of restriction, then the 3/8" push-fit would be a better solution.
 
i just love it that erasmus, plywood and top nurse have basically very similar "flow philosophies"..... wonders if cathar built this poll specificlaly for those and whoever else usually in the arguement.

i aint got my kit yet but i already knwo what im gettin so lemme see how my "philosophy" stacks up when i already knwo it means "budget"

so "budget = ? "
 
ok voted wheeee

cuz i want the swiftech h20 kit thingy:
6pts -- pump mcp350 [92gph / 350lph]
2pts -- pump mcp350 [13ft / 4m]
2pts -- 3/8" ID tubing / 1/2" OD
4pts -- want the storm CPU block with the kit + GFX wb + either chipset or ram blocks.
4pts -- want quick disconnect
3pts -- radiator with continuous tubes? not sure abotu this one.
____
21pts

hmm but i miscounted before and voted 17pts (not including the quick disconnecting poiints)
 
el rolio said:
i just love it that erasmus, plywood and top nurse have basically very similar "flow philosophies"..... wonders if cathar built this poll specificlaly for those and whoever else usually in the arguement.
Click to expand...

ROFL, but I doubt that it was designed that way. But it does really point out that we are all probably tripping over each others use of language that has no common meaning. I have found this thread to be quite inclusive rather than the typical BIB/SIB thread that tends to be exclusively a flamefest. :eek:
 
Cathar said:
These fans are pure gold as far as near-silent water-cooling radiator fans go. If I were you I'd collect a whole bunch from customers and use them.

At 12v they aren't exactly silent - just quietish, but as you get below 10v they really quieten up nicely while still offering good air-flow.

These are the sorts of fans that I model my near-silent radiator cooling designs on.
Click to expand...

didja say those nexus fans are? and where do you get that lil thing you plug in to slow it down to like 10v or 7v or whatnot?
 
el rolio said:
i just love it that erasmus, plywood and top nurse have basically very similar "flow philosophies"..... wonders if cathar built this poll specificlaly for those and whoever else usually in the arguement.
Click to expand...


Lol. I don't argue just to argue. Hmm, well sometimes. :D But it is usually when incorrect thinking or info is being presented as fact.

What I like about this poll is it helps people to see what kind of system they "really" have. And as mentioned by TN, it is taking on a bell pattern, with most users falling into the modderate categories.
 
and where do you get that lil thing you plug in to slow it down to like 10v or 7v or whatnot?
Click to expand...

you use a series resistor in line to the fan + wire. Voltage = Current x Resistance.
 
Well you could get one of these, but the shipping is going to kill you :( So look at how it is made and go to the electronic store, get a resistor, and solder it yourself :)
 
A simple fan controler works very well and they are pretty cheap these days. Something like the Sunbeam Tech Rheobus

DSC01887.JPG
 
@bio-hazard: good lord that thing is wide. it really didn't show as much in you previously posted pics.

however, the sunbeam rheobus is a good option for multi-fan controlling, and the zalman fanmate 1 or 2 is a good option for something that does not take up a full bay, and controls a fan or two.
 
They also have a software programable controler card coming out that looks very interesting.................. :D

Sunbeam Theta series is a PC software based fan controller which enables many user-defined features by its PC software.
Theta supports up to 8 independent channels of fan (max 10W per channel). With the PC software of theta, users can tune
every individual fan settings into automatic or manual modes according to their need. The settings can be saved/loaded in
PC, which enable users keep all PC fans to run at the most optical way in different situations with just a click of mouse.

Users can also schedule up those settings by time! Theta introduces Sunbeam Advanced Auto Control Mode, which user
can set an upper and lower bound temperature, regulating your fan running 50%-100% speed linearly between these 2
temperatures.



tp101_multi-dunction_controller_r1_c1.jpg


http://sunbeamtech.com/PRODUCTS/Theta_TP101_Multi-Function_Controller/theta_tp101_multi-_function_controller.htm
 
Don't have any idea yet............... :eek: I'm still waiting on more info and a review sample............... :D
 
Top Nurse said:
Well you could get one of these, but the shipping is going to kill you :( So look at how it is made and go to the electronic store, get a resistor, and solder it yourself :)
Click to expand...


ahhh thank you TN, thats all i was askin. im not gonna make it myself i'll jsut wait till i need to buy supn from there and gang uup the shippin. no reason for me to be solderin or learning electronics to make my fan go slower. heh
 
not 100% sure where I stand, I have the laguna statuary pump #3 shown in this review here: http://www.procooling.com/articles/html/project__pro_lan_-_page2.php specs are in the pdf they link to. here's the flow rates:
flow3zp.jpg


I have 2 heatercores in series, one is a bix the other is a chev heatercore. 2 waterblocks all running on 1/2" tubing with barbs.

I don't have much flow left when the water goes back into the res it just kinda trickles in as opposed to pouring in which it used to do before i added the second heatercore. I was going to take that out of the loop this weekend and test to see which gave me better results, not sure if i will or not though.

your thoughts?

edit: if you go to the end of the posted review, turns out my system is laid out just like his. i have a res on the back of my case feeding the pump though (in same spot as his), then it goes to the bix in the same spot, then my gpu block, then i have the other heatercore way at the top of my case then down to the cpu block then out to the res which is in line with the cpu block.
 
So why did adding a heater core cause such a decrease in flow? I thought heater cores were supposed to be like extremely high flowing parts? :confused:
 
I think that your problem is that your pump doesn't have much head at all. Going by what the chart you posted says, you only have 5.58 feet of head at the pump, add to that all the extras that's required to add a second radiator, you may very well be over taxing the pump.............. :eek: It could also be that there is something blocking your heater core............. :D
 
What with all the high vs low going on I thought it might be a nice diversion to roll this thread out again. It seems to pull people here together rather than what has been happening lately.
 
score of 5 guess its high flow for me. This could be a really great forum if some would put away their elitist know it all attitudes. I check thru the forum daily but post very little because its like jumping in a pissing contest just not worth it nobody ever wins. I am far from a noob and have been OCing for many years so some may know me from other forums. And I love to learn new things everyday and usually do. Cather has the knowledge and experience,but his posts are informative and knowledge based without the accompanying attitude to go with it. Many of us could learn from cather more than just WCing facts. Heres wishing everyone a great day and to a great forum.
 
I got an 8. seems about right i guess. the Koolance block is prolly slowing things downa bit but it works awesome (see sig for temps). Can't wait to get a dual core under it.
 
I don't think the Koolance block is that terribly restrictive compared to impingment designs. It's not as freeflowing as some of the maze type designs but still not a big flow killer.
 
madmat said:
I don't think the Koolance block is that terribly restrictive compared to impingment designs. It's not as freeflowing as some of the maze type designs but still not a big flow killer.
Click to expand...
i wouldnt call it a "flow killer" but having done some pretty extensive testing on it i know its a tad less restrictive than the storm and a hell of a lot more restrictive than the apogee. (common knowledge i know but it was part of the job). the thing i kind of find funny is how things are coming around full circle. between the storm and the apogee i would say the CPU-30X is alot closer to the apogee in design yet is waaaaaaay older than both. i get some kick ass temps too. these blocks really stretch their legs on an IHS with a good high head pump. *hint, hint, Koolance ive been saying this for a year now!*
btw i am not saying the apogee is superior to the storm just newer and supposedly designed for dual core
 
Probably the only thing I think missing from the equations here is the different style blocks. Some are more restrictive than others and I would imagine they would affect the overall rating. My case: CPU: AquaXtreme MP-05 SP LE - high restriction impingement block, and GPU: AquaXtreme MP-1 - high restriction block also. As opposed to a Maze 4 or non-impingement style blocks. Cathar, if you are still reading this thread now and then, what's your take?
 
You must log in or register to reply here.
Back
Top