10mm vs 12mm ?

[*klonk*]

I've searched and searched but I can't seem to find any articles or posts that handle this subject thoroughly, so maybe someone could point me in the right direction.

How does going from 12mm tubes to 10mm tubes afftect the cooling abillity of a system?

The reason for this question is that 10mm tubes are easiere to work with and I have a radiator that has 10mm in/outlets.

[Etacovda]

Generally it'll lower flowrates, which effects performance. How much that effects the system overall will depend on the pump, radiator, and blocks used. To effectively answer the question, you'll have to post the other parts of your system for consideration.

[*klonk*]

it's not so much the performance on the system-to-be that im concerned about, Im just looking for some general information on tubesize and how it -generally- affects system performance.

now really... didn't they do an article here on Procooling.com on that very thing? - can't find it no matter what I search on.

[Long Haired Git]

Perhaps they did, but that was back in the day of high flow waterblocks.
These days people run highly restrictive waterblocks, and hence the need for 1/2" tubing is reduced rather significantly.
I'm working some numbers here.......ding - finished.
3 ft of 1/2" tubing will flow 300 GPH (choking on dumb empirical units), which calcs to be 0.6m of head per meter @ 16.7 LPM (inc coversion to sane metric).
10 ft of 1/2" tubing will flow 12.24 LPM, which calcs to be approx 0.17m of head.
It stands to reason that as the flow rate approaches zero, the head will approach zero.
Now I've forgotten the maths to generate an equation for a graph, but some ballpark guesses and sketching has 1m of 1/2" tubing generating 0.02m of head @ 3 LPM.
So, even if 3/8" tubing was five times as restrictive as that, it would be a massive 0.1m of head.

[nexxo]

In short, there is little difference between 1/2"ID and 3/8"ID tubing.

[*klonk*]

awright! thx for the info, that was what i was guessing at - less important than I was let to believe

[Long Haired Git]

When blocks were but channels in solid copper, and used sheer volume of water to promote turbulence for heat transfer, the move to 12mm was justified.
Times have changed.

[redleader]

Depends on the system. With a swiftech block, 1/2 isn't going to make a difference.

With a dangerden block, the difference will be more significant.

[Cathar]

When considering the tubing size, one must also consider the level of restriction of the fittings that the hose attach to, when deciding how much restriction it all offers.

One important aspect of tubing size is not so much the performance, but the level of restriction that is affecting the pump inlet. Centrifugal pumps don't much like having to suck water into them, and excessive restriction at the pump inlet can greatly shorten the life-span of the pump. It should be noted that some specific centrifgual pump designs are more susceptible to this sort of failure than others. Regardless of what your final tubing size choice is for the loop, the pump's inlet should not be of a tubing size that is less than the pump is designed for, and the path of water into the pump inlet should be as unrestricted as possible.

Restricting the outlet of the pump generally has little ill-effect though other than lowering flow rates.

Rather than blindly saying that 1/2" ID is always superior, it makes more sense to have a balanced approach to the issue. If we instead say that in a "well balanced system" that the tubing plus fittings should not account for more than 20% of the total system's resistance, then this is perhaps a better approach to take when evaluating the significance of a tubing size choice.

Doing so presents the uncomfortable issue for many that this needs to then be balanced with the pump and block choice.

Up to about 18 months ago almost all pumps that you could buy were quite low pressure and high flow by the standards of some of the pumps available today. Also there were a number of waterblocks about that relied on very high (again, by today's standards) flow rates to achieve close to their peak levels of performance. This combination of pumps and blocks almost necessitated the use of 1/2" ID tubing, and to this day this belief is still largely followed.

A large amount of research has found that flow rates in the 1-1.5gpm (4-6LPM) range is however quite desirable for various radiators, as well as for many of the high-end waterblocks on the market. Flow rates much above 6LPM are largely unneeded any more. Because of this, 7/16" ID tubing is perhaps the best balanced tubing size to use if one were picky about performance, and 3/8" or 10mm ID is also perfectly acceptable if targetting 4-5LPM (1-1.25gpm) flow rates.

8mm ID tubing even with a modern pump largely limits oneself to 1gpm or less for achievable flow rates in a full system. This is getting towards the start of the "knee" below which many common blocks and/or radiators will start to noticably lose cooling performance.

In rough summary:

8mm ID = borderline acceptable balance of tubing size vs performance
3/8" ID or 10mm ID = good balance of performance with a manageable tubing size
7/16" ID = excellent performance with a still somewhat easily manageable tubing size
1/2" ID = pure overkill for most any modern waterblock/pump scenario with unwieldly large and difficult to route tubing

There are however certain exceptions to that summary, for which there is no substitute for doing a bit of good research with respect the to any specific setup that you have planned.

[TbirdX]

Ok, so lets give it some real world meaning...

MCP600
BIX 120.3
Swiffy 6002
DD Maze 4
Bayres

That should be a relatively common sort of set up.

10mm or 12mm???

[*klonk*]

Thx Cathar, you should get the thread edited and stickied - good info here.

[Risky]

Cathar, great post there. I'm sure it will be linked to a lot in the future.

That said my rig is on 7mm ID, though with a 10mm ID from Res to pump. Mostly based on where I happended to start really. If starting from scratch I might go for the 1/2" OD hose, though its harder to get the fittings over here

[nexxo]

Originally Posted by Cathar

When considering the tubing size, one must also consider the level of restriction of the fittings that the hose attach to, when deciding how much restriction it all offers.

One important aspect of tubing size is not so much the performance, but the level of restriction that is affecting the pump inlet. Centrifugal pumps don't much like having to suck water into them, and excessive restriction at the pump inlet can greatly shorten the life-span of the pump. It should be noted that some specific centrifgual pump designs are more susceptible to this sort of failure than others. Regardless of what your final tubing size choice is for the loop, the pump's inlet should not be of a tubing size that is less than the pump is designed for, and the path of water into the pump inlet should be as unrestricted as possible.

Restricting the outlet of the pump generally has little ill-effect though other than lowering flow rates.

Rather than blindly saying that 1/2" ID is always superior, it makes more sense to have a balanced approach to the issue. If we instead say that in a "well balanced system" that the tubing plus fittings should not account for more than 20% of the total system's resistance, then this is perhaps a better approach to take when evaluating the significance of a tubing size choice.

...Because of this, 7/16" ID tubing is perhaps the best balanced tubing size to use if one were picky about performance, and 3/8" or 10mm ID is also perfectly acceptable if targetting 4-5LPM (1-1.25gpm) flow rates.

...In rough summary:

8mm ID = borderline acceptable balance of tubing size vs performance
3/8" ID or 10mm ID = good balance of performance with a manageable tubing size
7/16" ID = excellent performance with a still somewhat easily manageable tubing size
1/2" ID = pure overkill for most any modern waterblock/pump scenario with unwieldly large and difficult to route tubing

There are however certain exceptions to that summary, for which there is no substitute for doing a bit of good research with respect the to any specific setup that you have planned.

Nice to know I got it right then.



(excuse the messy cabling, it's still a work in progress.)

I adopted 3/8" ID tubing because the barbs of my blocks were 3/8" ID (1/2" OD). I made sure all the other barbs in the system were at that ID. The pump inlet however has an unrestricted 3/4" ID tube straight down from the res (through a 3/4" barb, of course). The whole circuit was designed to be as short as possible --hence the unconventional setup of putting the GPU block before the CPU blocks.

The result? A watercooled dual Opteron 250 with a Radeon 9800Pro in a single loop, cooled on a single BIX with a low-flow Panaflo (67.9CFM) fan, running at CPU1=43C, CPU2=40C and system=26C with room temperature at 18C (PC temps are measured by motherboard, with all the inaccuracies that involves, of course, but still, not bad).

[Long Haired Git]

Check out graph 18 at BillA's rad test article:
http://thermal-management-testing.com/ThermoChill.htm