Mixing tube sizes in the same loop? - Awesome diagram inside

[|Oc|REvO_HsArC]

Hi there
I'm about to do some changes, again, in which the title is my current concern, but first let me explain the situation :)
Its my first "commercial" WC... My last one was ghetto and lasted 4 years (hurray, probably the time I was absent from overclocking, and XS).

I saw it beat in disbelief by the tunic tower 2 months ago when the ghetto died.
So I built a new rig.
I had 2 revisions in the last week. One without the hd4870 in the loop, and one with her on it with a second radiator!
Terrible results in both revisions.

first revision, The material and loop:
- radiator black ice pro (240... 4 years old the poor bastard),
- Pump/reservoir innovatek 600L 12v
- the cheap ALPHACOOL NEXXXOS XP LIGHT REV2 cpu wb.
- 8/10 tubes
- 2x silenx 47cfm 120mm fan (11db!)
pump -> Cpu -> Radiator -> pump
water in system: about 350ml
Result: E8400 @ 4ghz 1.37v 71º ! ouch... Good cpu, bad WC. The radiator is pumping ALOT of heat!


Second revision, The material and loop:
All the same, plus:
- hd4870 EK wb
- a second radiator do deal with her, the Black Ice GTS-Lite 240
pump -> radiator -> cpu -> radiator -> Hd4870 -> pump
water in system: 500ml
Result: E8400 @ 3800mhz 1.3v 70º! ouch..... lost 200mhz. The radiators are pumping ALOT of heat Hd4870 45º idle, 65º load.







I figured... must be the pump and the Low amount of water in the system, right? 600Lph and 500ml... at the end of the loop the height is merely 1cm! The stream its like a flaccid peins.
So to my last revision, I already bought the following:
- A real men pump. the MCP 655
- 1/2" tubes and fittings all around
- 350ml reservoir

But my old blacice Pro 240, doesn't have replaceable fittings... Its only size is half of the 1/2 tube. But I wan't to keep it! My new thin 240 GTS radiator probably isn't enough to handle my 4ghz cpu and hd4870...

And this is what I was planning (beware the mona lisa):
http://i38.tinypic.com/3586hax.jpg

-will the thin tubes in the latter part of the loop affect my performance?
-Am I better off with a single radiator with the entire system 1/2?
-How come my first revisions failed??? was it really the pump?

thanks!

[brinox]

there are several parts of the loop where the velocity of the coolant will vary, and sometimes vary greatly. in the case of your system, simply adding smaller diameter tubes will make the coolant flow through that part of the system more quickly, and in turn will add a very slight amount of friction to the overall system, but all-in-all there is nothing wrong with doing so.

[kinghong1970]

awesome diagram there Picasso!

why would you want to keep a 4 year old radiator and risk new equipment when you can pick up a good and effective radiator such as the Swiftech MCR320 at a good price?

i'd be more comfortable with a new radiator and 1/2" tubings...

well, my un-educated guess would be the coolant will flow as fast as it does in your thinnest tubes... will it be enough?

[|Oc|REvO_HsArC]

Well, I am putting my hopes that what is wrong with my current revision (nº2) is mainly the slow pump (600lph) and perhaps the very restrictive system (2 rads). 15º worse than a 50$ air cooler is not OK.

I guess what I was trying to know with those questions is: will the mcp655 heat up and blow up because of those 2 thinner tubes ?

[brinox]

will the mcp655 heat up and blow up because of those 2 thinner tubes ?

no, this is what i was trying to explain. your system is generally going to have the same flow-rate (volumewise) at any point in the system, once fully bled. when the coolant is flowing through the smaller tubes, sure it WILL be moving faster, but there will be less of it moving through any point in that diameter size of tubing, as opposed to the larger diameter tubing.

the major downside (and not really a large impact at all) to using smaller tubes is it creates a slightly larger amount of friction than the larger tube size.

[|Oc|REvO_HsArC]

Ok so I have this Wc setup up and running

http://i38.tinypic.com/3586hax.jpg
http://i35.tinypic.com/207m4gi.jpg

Objective is to be dead silent when not playing games, yet fast. And silent when playing games, yet very fast.
I got all I wanted.

I'm using the cheap aquacomputer wb, that doesn't have 1/4 fittings so I couldn't use 1/2 tubes on it.
I already ordered D-tek fuzion rev2, but right now these are the results:

running orthos E8400 @ 4ghz 1.38v with hd4870 in the loop, my temps are:
ambient 25ºc
cpu 65ºc
GPU 40º
Radiators temp ~ 31ºc
Water temp 40ºc

idle temps are:
cpu 42º
water temp 36.6º

So, 15º between ambient and water with the fans fullthrottle (but still silent). How much difference you guys have?

[Eddie3dfx]

Ok so I have this Wc setup up and running

http://i38.tinypic.com/3586hax.jpg
http://i35.tinypic.com/207m4gi.jpg

Objective is to be dead silent when not playing games, yet fast. And silent when playing games, yet very fast.
I got all I wanted.

I'm using the cheap aquacomputer wb, that doesn't have 1/4 fittings so I couldn't use 1/2 tubes on it.
I already ordered D-tek fuzion rev2, but right now these are the results:

running orthos E8400 @ 4ghz 1.38v with hd4870 in the loop, my temps are:
ambient 25ºc
cpu 65ºc
GPU 40º
Radiators temp ~ 31ºc
Water temp 40ºc

idle temps are:
cpu 42º
water temp 36.6º

So, 15º between ambient and water with the fans fullthrottle (but still silent). How much difference you guys have?


Fyi, aqua computer, innovatek, and other european and older blocks can use bitspower new G1/8 to G1/4 adaptor which will let you use 1/2" barbs!!!!!
performance pcs sells them and they are nice

[|Oc|REvO_HsArC]

I bet they would come handy if I didn't already ordered the dtek :)

but those adaptors kind of defeat the purpose of 1/2 barbs... the entrance to the wb is so tight as it is with 1/8. No point in using those I suppose

[NaeKuh]

the two systems are completely different.

Also mixing tubing brings your entire system down to the smallest tubing.

So if you had all 1/2 and you ran a small section of 3/8th your entire system would perform at 3/8.

[RADCOM]

Is it significant though? May I refer my honorable gentleman to the link below. BTW any houses for sale in Grinchville :)
http://www.over-clock.com/articles.php?action=show&id=23&perpage=1&pagenum=4

[IanY]

Grinchville residences are free. Applications are always welcome. No guarantees of acceptance though lol

[duckm4n]

you should think about reversing the airflow over your rad as you are currently cooling the radiators with warm internal air

[skinnee]

Grinchville residences are free. Applications are always welcome. No guarantees of acceptance though lol

NaeKuh still hasn't been down to drink beer...IanY, hell he knows the combo on the beer fridge. ;)

[|Oc|REvO_HsArC]

you should think about reversing the airflow over your rad as you are currently cooling the radiators with warm internal air

thats true.
But that way I would be dumping 30ºc air into the case, heating up everything (hd4870, ram, chipset etc)

[brinox]

the two systems are completely different.

Also mixing tubing brings your entire system down to the smallest tubing.

So if you had all 1/2 and you ran a small section of 3/8th your entire system would perform at 3/8.

this makes no sense whatsoever. care to backup such a claim? i am willing to admit that i am completely wrong, but not without some data. until then, i am completely stumbled on this and cant go quietly into the night. sorry :(

the addition of 3/8" tubing into a majority 1/2" does indeed add restriction, as previously said, but the claim of said addition making the entire loop drop in performance to that of the smallest tube diameter however short of distance is ludicrous! again the flowrate will decrease when using any length of 3/8" tubing, thats simple physics, but the system will not drop to the level of a total 3/8" system.

the momentum that all the coolant in the 1'2" tubing carries is going to force the coolant flowing through the 3/8" lengths at a faster pace, and once it enters back into the 1/2" size tubing it will slow down back to its original velocity.

[hotdun]

this makes no sense whatsoever. care to backup such a claim? i am willing to admit that i am completely wrong, but not without some data. until then, i am completely stumbled on this and cant go quietly into the night.....

NaeKuh is right.....think about the flow.... like a network bottleneck

[systemviper]

but NaeKuh doesn't say it just will effect flow rates, which is all it will do, he says
mixing tubing brings your entire system down to the smallest tubing
but when you have 1/2 inch flowing thru a block, how does that perform like a block with 3/8 tubing?


NaeKuh is right.....think about the flow.... like a network bottleneck

I don't get that either. All that gets effected is different flow rated at different points, and lots depends on the pump.

[ThrustSSC]

NaeKuh is right.....think about the flow.... like a network bottleneck

I don't think fluid dynamics resembles network transmission at this part, since the flow/data transmission in network is fixed.
I'm with brinox, When water will travel from the 1/2" tubing to 3/8" they will accelerate.
The reason I think that way is because that a closed loop maintain the same flow in each point, there for if water will remaine at a constant speed at different tubing diameter the flow will vary from 3/8" to 1/2" tubing.

I hope you understood what I wrote, I'm not a native English speaker but I try my best :up:.

[hotdun]

I don't think fluid dynamics resembles network transmission at this part....

I see what you're saying, that's a good point. Basically, if a block restricts the flow to below that of say 3/8" tubing, then what's the point of even using 1/2" tubing on that loop? Something along those lines, right?

[gillbot]

I don't think fluid dynamics resembles network transmission at this part, since the flow/data transmission in network is fixed.
I'm with brinox, When water will travel from the 1/2" tubing to 3/8" they will accelerate.
The reason I think that way is because that a closed loop maintain the same flow in each point, there for if water will remaine at a constant speed at different tubing diameter the flow will vary from 3/8" to 1/2" tubing.

I hope you understood what I wrote, I'm not a native English speaker but I try my best :up:.

The fluid going from 3/8 to 1/2 won't accelerate. It will only move as fast as the smallest diameter will allow it. If the fluid would accelerate going from 3/8 to 1/2, it would outrun the fluid still caught in the 3/8 line. :ROTF:

[ThrustSSC]

The fluid going from 3/8 to 1/2 won't accelerate. It will only move as fast as the smallest diameter will allow it. If the fluid would accelerate going from 3/8 to 1/2, it would outrun the fluid still caught in the 3/8 line. :ROTF:

Fluid travaling from 3/8" tubing to 1/2" tubing will decelerate.
Fluid traveling from 1/2" tubing to 3/8" tubing will accelerate.

[NaeKuh]

The fluid going from 3/8 to 1/2 won't accelerate. It will only move as fast as the smallest diameter will allow it. If the fluid would accelerate going from 3/8 to 1/2, it would outrun the fluid still caught in the 3/8 line. :ROTF:

Guys Gillbot looks like a noob but he he isnt.

He coaches me when it comes to pumps. So im vouching for his credibility.

So listen to him. He knows his pumps. Cuz of him, he's giving me a hard ass time in setting my RD-30 properly.

And its bugging the crap out of him cuz im trying to mickey mouse it. :rofl:


Fluid travaling from 3/8" tubing to 1/2" tubing will decelerate.
Fluid traveling from 1/2" tubing to 3/8" tubing will accelerate.

and did you forget the first concept of closed loops?

I believe it was oh yeah:

Flow is uniform inside a closed loop! <- remembering this?

So your flow is dependant on the most restrictive bottle kneck. IE smaller tubing, which brings your entire system = to the smaller tubing.
AS gillbot said, how is your 3/8 section going to outrun your 1/2 section if the loop is closed?

You gonna try passing it like you do on a freeway when a granny gets in the carpool lane? :rofl: <sarcasm dont take offense to it>

[ThrustSSC]

Yes flow rate will remain uniform at a close loop, thats why I sayd that water must accelerate while traveling from 1/2" to 3/8" ID tubing. If the speed will remain the same than the flow will not be uniform.

Flow rate and flow speed is not the same.

[NaeKuh]

Yes flow rate will remain uniform at a close loop, thats why I sayd that watter must accelerate while traveling from 1/2" to 3/8" ID tubing. If the speed will remain the same than the flow will not be uniform.

Flow rate and flow speed is not the same.

ahhhh okey i mis understood your statement.

Sorry about that! :up:

And yes your correct, because the same volume of water must pass though a smaller space, and the only means of doing it is by pushing it faster. This is the principles of injectors on plates. :P


Basically my original statement still stands. You change a section of a smaller tubing, your system effectively operates at that smaller tubing.

[gillbot]

I know flow rate and flow speed are not the same BUT the pressure drop at 3/8 will increase flow speed but reduce flow rate overall. The highest restriction in your loop determines flow rate.

[Koen]

The highest restriction in your loop determines flow rate.

No it doesnt. If you were right, guys with nozzled fuzion should be running 3,5mm ID tubing. A part of 3/8" tubing in a 1/2" loop will reduce the flow but not to a level of full 3/8" loop.

[gillbot]

No it doesnt. If you were right, guys with nozzled fuzion should be running 3,5mm ID tubing. A part of 3/8" tubing in a 1/2" loop will reduce the flow but not to a level of full 3/8" loop.

So by your logic, I can put a 3/8 piece in with 1/2 and it will flow like 1/2?

The highest restriction WILL determine your overall flow rate. If it did not, your flow rate wouldn't change if you removed your nozzled fuzion. We all know pulling the fuzion WILL increase your flow rate because there is no longer a restriction to flow.

I also never said 1 piece of 3/8 would reduce the loop to the same level of an entire 3/8 loop.

[NaeKuh]

No it doesnt. If you were right, guys with nozzled fuzion should be running 3,5mm ID tubing. A part of 3/8" tubing in a 1/2" loop will reduce the flow but not to a level of full 3/8" loop.

*sigh*

pressure is not uniform, flow is uniform.

I think when gillbot said, the 3/8ths can never overtake the 1/2, thats about as simple as it can get.

[Koen]

So by your logic, I can put a 3/8 piece in with 1/2 and it will flow like 1/2?

The highest restriction WILL determine your overall flow rate. If it did not, your flow rate wouldn't change if you removed your nozzled fuzion. We all know pulling the fuzion WILL increase your flow rate because there is no longer a restriction to flow.

I also never said 1 piece of 3/8 would reduce the loop to the same level of an entire 3/8 loop.

Ah sorry, by "determines" i thought you meant "brings down to its level" :up:
And of course...

So by your logic, I can put a 3/8 piece in with 1/2 and it will flow like 1/2?
...I didn't have anything like that in mind, just misunderstood your statement.

[systemviper]

So if what i am hearing is correct
the statement
So if you had all 1/2 and you ran a small section of 3/8th your entire system would perform at 3/8.
is incorrect.

It will impact the flow of the system decreasing it to make up for the added pressure of the 3/8th reduction,
It wouldn't make the entire loop "perform at 3/8" but it would reduce the flow from full 1/2 to a lower flow rate, but not the flow rate of a 3/8th loop

[hotdun]

So if what i am hearing is correct
the statement
is incorrect.

It will impact the flow of the system decreasing it to make up for the added pressure of the 3/8th reduction,
It wouldn't make the entire loop "perform at 3/8" but it would reduce the flow from full 1/2 to a lower flow rate, but not the flow rate of a 3/8th loop

Good question.....

[ThrustSSC]

I don't understand the question, adding a 3/8" section into a 1/2" loop just increase the resistance compare to 1/2".
The flow rate of 1' of 1/2" ID tubing + 9 feet of 3/8" ID tubing is smaller than the flow rate of 1' 3/8" + 9' of 1/2" ID tubing

[brinox]

I don't understand the question, adding a 3/8" section into a 1/2" loop just increase the resistance compare to 1/2".
The flow rate of 1' of 1/2" ID tubing + 9 feet of 3/8" ID tubing is smaller than the flow rate of 1' 3/8" + 9' of 1/2" ID tubing

thats the point im trying to get across. would you agree that 1' of 3/8" tubing and 9' of 1/2" tubing is not equal to 10' of 3/8" tubing in terms of flow resistance/pressure drop and overall volumetric flow rate?



Basically my original statement still stands. You change a section of a smaller tubing, your system effectively operates at that smaller tubing.

operates how? or with respect to which measurement?



I know flow rate and flow speed are not the same BUT the pressure drop at 3/8 will increase flow speed but reduce flow rate overall. The highest restriction in your loop determines flow rate.

restriction in our case is a measurement of diameter (or cross section) and length. but dont let that fool you. the smallest length of a large change in diameter/cross section is still pretty substantial. look at martin's testing with the fuzion's nozzles. conversely, a long length of a minutely small change in diameter/cross section does have added restriction, but by no means approaches a large change in diameter/cross section of any length.

what we need to take away from this is the idea of change in diameter/cross section. thats most important when all else is left constant.

[hotdun]

There's gotta be some kinda formula for this kinda thing no?. On Martin's Liquid Lab website, he states that:

"A system flow rate is determined by the intersecting point of the pump and system pressure drop curves...."

So for this thread's discussion, the longer the length of the restriction, the higher the pump's pressure drop.

Check out his flow rate estimator here (http://www.martinsliquidlab.com/MartinsFlowRateEstimator.html).

[Martinm210]

FYI,
You can plug in various lengths of tubing in the flow rate estimator if you want of different sizes.

Nothing wrong with running mixed tubing sizes, I would just watch out for the smaller fittings more than anything. Smaller fitting are quite a bit more restrictive than 1/2" fittings and that is not figured into the estimator, everything in there is 1/2" fitting based. The resulting flow rate is a result of the "Accumulation" of restriction and pressure drop of the system as a whole.

Adding one 6" piece of 3/8" tubing is NOT as restrictive as running the whole system with 3/8" tubing. It's as simple as the pump adding a pressure curve or series of pumps, and the system resisting that as a whole with a system curve that consists of adding up every little restriction in the system. What get's confusing is how much more significant some restrictions are. Tubing size generally is very low in restriction compared to nozzled waterblocks, so it's easily perceived as the block controlling...but that's just a matter of scale. Everything adds restriction and flow rate is a result of all those restrictions added together.

With that, I doubt the resulting flow rate is bad enough to really be a big factor. It depends on the waterblock, but most of todays micropin types are very good performers with lower flow rates.

My 2C

Awesome sketch though, that's cool!

[|Oc|REvO_HsArC]

thanks guys :)

right now I'm on one radiator alone and 1/2" all across the loop.

My water is 4ºc hotter.
As I expected, a second radiator pays off even if its tubing is only 8/10 (as the mosa lisa shows)

[hotdun]

Martin has just "kicked some knowledge" at us. I was waiting for that...:)

[fusionmkx]

Guys, it's very simple, it's just fluid dynamics. Yes, The mass flow rate is the mass that passes a given point per unit time. The flow rates at any two points must be equal, as long as no fluid is being added or taken away. This is only true with laminar flow, (let's assume the loop is laminar and not turbulent), mass flow rate = dM / dT (mass / time), when mass = density (p) x vol (v). This gives us the equation of continuity: dM1 / dt = dM2 / dt. Therefore. p1v1A1 = p2v2A2 where A is area. If the density doesn’t change – typical for liquids – this simplifies to A1V1 = A2V2. Where the pipe is wider, the flow is slower.

Hence in 3/8" tubing the flow is faster, in 1/2" tubing, the flow is slower, all this can be calculated based on the flowrate of the pump.