The impact of tubing sizes

[Turtle 1]

Yeah, I get kind of itchy whenever I come over to XS as there are lots of trolls here as well. It's okay because they do the same dog and pony show, right?

Look TN there are trolls everywhere in all forums . Their are differant kinds. You have the grammer spelling trolls / You have the Intel vs. AMD trolls. ATI vs. NV trolls.

You are infact a small tube troll. But you go one step further by pushing AC products like you own the company.

Some consider myself an Intel troll . I don't however and in Jan. of 08 will find out if I was a troll or had very good insight. Fact is I was banned from both the news section here at XS and banned @ [H] for what has now turned out for telling the trueth. I had my facts pretty much inorder. Yet was punished for it. Its ok tho the world isn't a fair place as we all know.

I read the small tube post @ [H] and your disrespect for cathar was just shamefull. Cathars figures do infact work . However in real world there are variables that when applied in real world terms might go against the math.
That however doesn't matter as its impossiable for cathar to figure how each and every system will work because there are so many . But if given the specs for each system his numbers would win the day. Cathar has forgotten more about flow and pressure in a closed loop system than you will ever know.

I will tell you this straight out. I don't care what happens. If you were my neighbor I would have walked over and B-I-T-C-H slapped you for your disrespect. Cathar has done more for the PC water cooling community than anyone.

I personnally believe your a guy with female genes. Get over it .

[MotF Bane]

Look TN there are trolls everywhere in all forums . Their are differant kinds. You have the grammer spelling trolls / You have the Intel vs. AMD trolls. ATI vs. NV trolls.

You are infact a small tube troll. But you go one step further by pushing AC products like you own the company.

Some consider myself an Intel troll . I don't however and in Jan. of 08 will find out if I was a troll or had very good insight. Fact is I was banned from both the news section here at XS and banned @ [H] for what has now turned out for telling the trueth. I had my facts pretty much inorder. Yet was punished for it. Its ok tho the world isn't a fair place as we all know.

I read the small tube post @ [H] and your disrespect for cathar was just shamefull. Cathars figures do infact work . However in real world there are variables that when applied in real world terms might go against the math.
That however doesn't matter as its impossiable for cathar to figure how each and every system will work because there are so many . But if given the specs for each system his numbers would win the day. Cathar has forgotten more about flow and pressure in a closed loop system than you will ever know.

I will tell you this straight out. I don't care what happens. If you were my neighbor I would have walked over and B-I-T-C-H slapped you for your disrespect. Cathar has done more for the PC water cooling community than anyone.

I personnally believe your a guy with female genes. Get over it .

Minor issue Turtle... Vapor already closed off the Aquaduct thread, told people enough with the flaming/trolling, and somebody banned TN, so he/she won't even see your lecture. Good lecture though.

[ranker]

Look TN there are trolls everywhere in all forums . Their are differant kinds. You have the grammer spelling trolls / You have the Intel vs. AMD trolls. ATI vs. NV trolls.

You are infact a small tube troll. But you go one step further by pushing AC products like you own the company.

Some consider myself an Intel troll . I don't however and in Jan. of 08 will find out if I was a troll or had very good insight. Fact is I was banned from both the news section here at XS and banned @ [H] for what has now turned out for telling the trueth. I had my facts pretty much inorder. Yet was punished for it. Its ok tho the world isn't a fair place as we all know.

I read the small tube post @ [H] and your disrespect for cathar was just shamefull. Cathars figures do infact work . However in real world there are variables that when applied in real world terms might go against the math.
That however doesn't matter as its impossiable for cathar to figure how each and every system will work because there are so many . But if given the specs for each system his numbers would win the day. Cathar has forgotten more about flow and pressure in a closed loop system than you will ever know.

I will tell you this straight out. I don't care what happens. If you were my neighbor I would have walked over and B-I-T-C-H slapped you for your disrespect. Cathar has done more for the PC water cooling community than anyone.

I personnally believe your a guy with female genes. Get over it .

Minor issue Turtle... Vapor already closed off the Aquaduct thread, told people enough with the flaming/trolling, and somebody banned TN, so he/she won't even see your lecture. Good lecture though.

Amen to both. She took trolling to a whole new level.

[RickCain]

Why is this thread becoming a TN obsession thread? Can we get back on topic, maybe?

I for one am going to give a try with 10mm press to fits and see what kind of temp differences I get, if any. How are other users of this forum going to use the information from the OP?

[migueld]

Why the hell was top nurse banned? I found her contributions to be educational and constructive. If anyone did the trolling is the people at XS who can't tolerate anything unknown to them.

Some time ago I would get flamed and trolled about my use of 3/8. Some people at XS deserve a long overdue ban. Not top nurse.

[Turtle 1]

Back on topic . I have always preferred small tube to large ones. . That said when I say small tube I mean 3/8 = GPU cpu' Smaller yet NB/sb and regulators 5/16 / 1/4 . harddrives 1/4 memory 1/4 .

The thing I believe many people are missing here is cathar simply stated that because of higher head(pressure) more flow is now available to smaller tubes. In other words flow threw 3/8 today is compareable to flow a few years back to 1/2 tubes.
I debated with the large tube crowd over small tubes being just as good as large back than and got spanked. Now the small tube guys are coming out of the wood work saying they were right all along which is false .

I got into with the small tube guys also .Back than . OH he contredicts himself. Not.
I have always been a multi loop type guy . Flow is the Key word here. Today we have the pumps to push flow threw those small tubes.
Now I have read that to much flow stuff and I won;t touch on that right now but that can get very interseting. When I go on that one I will have the numbers to back it up.
But in a normal today set up there is no such thing as to much flow.

Pumps in parallel Vs. series we shall get into that a bit later in the thread. But again it going to prove very intersting because the math just doesn't lie.

Bottom line is this Cathar new info is based on todays pumps in very simple single loop systems. The small tube guys can spin it anyway they want this did not apply 4 years ago as many are saying.
Now some of the german stuff did work fairly good in a single block low flow system that is true. But add another block and things went to hell in a hurry.
I feel good about what cathar new info. Not all the small tube guys were wrong back than . there were guys like myself that new what to do to get the performance out of small tubes. It was pretty simple really more loops and more money. And the use of pumps that others wouldn't use.

As for you Migueld . I have read a few of your post. I find you to be a Bob Hope type. TN was trolling bigtime. Not only that he attacked the messenger. Your right their are other trolls here. But most are long time members that helped build this fine forum . Not all debates are troll based.
Some would say I flamed you by calling ya Bob hope type not true I just find you amusing and someone I wouldn't take seriously . Bob Hope was a good man tho.

[RickCain]

I debated with the large tube crowd over small tubes being just as good as large back than and got spanked. Now the small tube guys are coming out of the wood work saying they were right all along which is false .

Except the ones that were using small tubing with DDC level pumps to begin with Funny part is that configuration was met with ridicule too.

[zhangmaster12]

wow, nice lecture! it was great!

[Martinm210]

FYI

I created this chart that may be of use if you are looking at a comparison of pressure drop vs flow rate between water blocks and tubing. It's theoretical, but I get results that are within about 10% using either Darcy's or Hazen Williams equations. As Cathar already pointed out in the OP, tubing is relatively insignificant.



If you compare the same pressure drop vs. flow rate of something like a D-tek fuzion it's really obvious. At 2 gallons per minute a D-tek fuzion is about equivalent to 36 feet of 1/2" tubing, the difference is huge.

Hope this is a useful tool, I plan to do some real world pressure drop checks on it soon to confirm it's correct.

[aussie-revhead]

G'day Cathar , nice work there mate , very informative . I had felt this was the case after many people got nice clocking results with Corsair kits and even Gigabyte nub-level kits (while they lasted)

[Cathar]

G'day Cathar , nice work there mate , very informative . I had felt this was the case after many people got nice clocking results with Corsair kits and even Gigabyte nub-level kits (while they lasted)

Cheers. That's a different kettle of fish right there though. You're talking about "The impact of flow rates on over-clocking". The differences here are far smaller than many think. While you might get a 1C difference from 1/4" tubing, the difference in peak overclock would be unlikely to be more than 1MHz on the FSB (i.e. ~10-20MHz or so, and may even be no difference).

[Berg]

Hi all, I would like to share some of my experiences with different types of tubing and fittings. For the record, I am co-owner of the watercooling shop www.vcore.dk - located in Denmark. We started back in 2003 with the White Water, when 1/2" ID still ruled the earth supreme (except in Germany).

First some comments regarding the tubing and fittings discussed here:
Cathar’s excellent graphs show much the same as some local tests have shown here in DK: That 8mm ID is a sweet spot with the right barbs, 3/8” ID is plenty for high performance and 1/2” is overkill for 99,99% of users.

And precisely 8mm ID hose has for years been the most common in European watercooling – but in 2 different versions:

Version 1: So called “10mm Push-IN” (push-fit) fittings with 10x8mm stiff PU hose.
(This has also been used by Asetek for their well-known, and now outdated, Waterchill systems, although they used a special hose with only 6,5mm ID)
Version 2: So called “10/8mm Push-ON” (compression) with 10x8mm soft PVC hose.
Pros: Both are easy to fit, secure, gives decent flow, and looks good and clean.
Cons: PU Hose is very stiff and does not bend well at all. It is often necessary to use 90 degree angles to get around the sharp bends. The thinwalled 10/8mm PVC hose is not stiff at all, but rather it almost collapses under its own weight after a while.
The Push-ON fittings has only about 6.85mm ID, which reduces flow some.

To get around the problems with the stiff PU hose, some people instead use a soft 10mm OD PVC hose in the Push-IN fittings. This is possible to do, but NOT recommended, as they will easily leak, especially after a few months - as others have also stated in this thread.

However using a 10x7mm ID (10mm OD) soft PVC hose with Push-ON (compression fittings), works very well, and is to me the perfect combo for many things. Only drawback is that 7mm ID is a tad small for super high performance. But that is easily fixed in a system with parallel loops, by using 3/8” ID with hi-flow barbs for the critical parts between reservoir, pump and radiator, and the 10x7mm for connections to and from the waterblocks.

Regarding Push-IN fittings for 3/8” tubing: (aka Push-Fit aka Quick-Fit)
Cathar: I can certainly see that 3/8” ID with Push-IN fittings are a very sweet spot theoretically. But unless I have misunderstood (and then I humbly apologize), you need a stiff PU or PE type tubing with 1/2” OD for this, and a PU hose with sidewalls that thin is completely unsuited for the tight bends and radia in a PC-case. It would be necessary to use a lot of 90 degree angels I would think (as with the common 10/8mm PU) – which would completely ruin the good flow.
And if you instead use a soft PVC hose, then you have problems with leaking over time and evaporation through sidewalls (plus it pulls out of the fitting very easily!).
So despite its theoretical merits, I don’t feel large diameter Push-fits are suited for practical watercooling.
- What is your experience with it? Which hose have you used?

And the positive things about the Push-IN system:
The smaller the diameter, the better it works really! Aqua Computer uses 8x6mm PU hose (they call it “Plug&Cool”), and this hose is actually quite nice. Handles bends well enough, although 90 degree angles will still be needed sometimes – but that is true of most tubing.
And when we get down to the very cute 6x4mm PU, as used by www.watercool.de for their “MICRO” concept, it works even better. Perfect in a seperate loop for DDR RAM watercooling and similar stuff.


And some of my views and thoughts on the topic:
Our webshop is basically a very time consuming hobby, so I have had ample opportunity to try a lot of different watercooling equipment and setups. We sell everything from 4mm ID (!) hose to 12,7mm (1/2") tubing, and use all types of fittings: Regular barbs, Push-In (aka push-fit) and Push-ON (aka compression) - often combined in the same system using multiple loops, and varying tubing size according to the heatsource.

Generally my personal objective with watercooling is not to obtain the lowest possible temperatures (contrary to many in this forum I realize), but instead to make the quietest possible systems, with watercooling on most all components: CPU, NB, SB, GPU, Mosfet voltage regulators, HDD and RAM.
Systems with this many waterblocks are "multi-parallel" using a distributor manifold and perhaps 3-5 seperate loops.

What you need to keep in mind, is that in Europe it is quite common to also have watercooling on HDD and perhaps RAM and even voltage regulators around CPU. So even if the tubing ID is only 7mm, the actual area for water to flow can be perhaps 2-4 greater when using multiple loops, and then the flow/backpressure situation suddenly looks quite different. It would be very interesting to make simulations or tests of flow and backpressure in that type system...

Also when using watercooling in this way, you automatically start using smaller tubing, else it is simply not possible to fit all the hose in the case. I usually use a combination of 10mm ID (or 3/8”) and hi-flow barbs with 10mm ID, mixed with 10x7mm (7mm ID) Clearflex on Push-On (compression) fittings.
The Push-ON fittings we use are actually made for 10x8mm hose, but as the thin sidewalls make this type very prone to bending/collapsing, we started using 10/7 mm PVC instead. As mentioned above, this has proven to be an excellent combination to work with, as it provides a very tight bending radius, together with acceptable flow, and very secure and nice looking fittings. The ID of these fittings (with G1/4 thread) are 6,85 mm - so it almost 7 mm straight through: http://www.vcore.dk/shop/1-4-10-470p.html
The compression nuts can be a bit hard to fit as the 10x7mm tubing is 1mm oversized for the fittings, but tools are not needed, and it sits very secure.
I am wondering if this combo would somewhere right between the two types of 8mm (yellow and green curves) on Cathar’s graph – or if it perhaps is about the same as using 8mm with 6mm barbs (yellow line) ?

We have been trying to find the same type Push-ON fittings to use with Clearflex or Tygon 3/8 x 5/8 (9,6 x 15,9mm) which has a good inner diameter for high flow, and also excellent bending radius due to the thick sidewalls (which most people here probably know all about). Ideally it should be a bit oversized on the inner diameter to keep a full 3/8” ID through, so it should be around 15/11mm or 16/11mm maybe.
But apparently that doesn’t exist – so should anyone know of such a fitting, I would love to hear about it, as I think that would be very near perfect! This would give the same flow that Cathar gets with 3/8” and “quick-fit”, but instead with a good hose with tight bending radius.

Sorry for the long post!

[Snyxxx]

Berg,

Excellent information. Thanks for posting.

P.S. I also watercool for silence with a tad overclocking as well.

[fgw]

berg,

missing a word on the 11x8mm PVC tubes. could this be a good compromise having the advantages of both worlds:

• 8mm ID
• 1.5mm wall size, which gives a bit more stability
• cheap and more flexible than PUR

what's your opinion on this one?

i use this tube size for about 6 month now and had no problems with leaking so far. although i don't use any sort of quick-fit or push-fit fittings. using the standard thread type fittings here.

anyway, interesting post!

[Marci]

And if you instead use a soft PVC hose, then you have problems with leaking over time and evaporation through sidewalls (plus it pulls out of the fitting very easily!).
So despite its theoretical merits, I don’t feel large diameter Push-fits are suited for practical watercooling.
- What is your experience with it? Which hose have you used?

Are you using the Tube-End inserts?? http://www.johnguest.com/part_spec.asp?s=TSM_S1
Generally solves the leaking / pulling out issue, especially when coupled with locking collars - http://www.johnguest.com/part_spec.asp?s=CM18_S1

[Mord]

Hi there, nice work, Cathar!

I've been recently conducting some simulations regarding tubing ID, connections, bends, etc. Though I am not into fluid dynamics. I am using the PipeFlow Expert Software. There are apps for unit conversion and pipe estimation (ID, flow, head, max length) - pipeflow wizard.
I posted it here a few hours ago. Since then I've been reading this thread (dunno how I missed it till now ).
I only provide some results (theoretical) - I will not argue with you, Sir! It only shows flow rate vs. tubing ID. I cannot comment on heat dissipation, absorption, C/W - only flow rates. It's still WIP.

Simulation run with no connectors, barbs and so.


And with different pumps @ 0.95m tubing length:

And some calculations for the tubing only (no blocks, nothing):


Also can would you share what kind of software/simulator are you using for your estimations/calculations/plotting.

Thank you, Sir, for your labor.

[RickCain]

I have recently switched out my 6mm tubing with 3/8" to see what the temp differences are. Right now I would say that my idle temps are the same but my load (via TAT) as about 5c lower. I need to do a couple of remounts to see if the data changes but it appears though my load temps are better by a pretty good margin.

[Cathar]

I have recently switched out my 6mm tubing with 3/8" to see what the temp differences are. Right now I would say that my idle temps are the same but my load (via TAT) as about 5c lower. I need to do a couple of remounts to see if the data changes but it appears though my load temps are better by a pretty good margin.

A gentle reminder for when doing comparitive tests to conduct the tests under the same conditions for each scenario. If you've mounted the 3/8" with a fresh mount, make sure you do the same for the 6mm scenario too. Thermal joints can dry out, or you may simply have had a bad mount before.

So many times in my testing, I've put something new on, and thought "Fantastic!", and then gone back and tried the old thing on a fresh mount, only to find that it's now pretty "better" too.

[Turtle 1]

Berg . I like your approach very much. We have been working in the same area. I have gone away from the simple manifold system your talking about to complete flow control system . But your correct people would be very surprised at the results that a dual manifold system with shutoff valves provide.

[Berg]

berg,

missing a word on the 11x8mm PVC tubes. could this be a good compromise having the advantages of both worlds:

• 8mm ID
• 1.5mm wall size, which gives a bit more stability
• cheap and more flexible than PUR

what's your opinion on this one?

Yes I decided to leave the 11x8mm (and others) for later, otherwise I fear the first post would have been too long - I can ramble on about hoses and fittings all day really

But you are right: 11x8mm PVC, with 1,5mm sidewalls are a good compromise also. This hose started seeing use, as a result of the problems I mentioned with the standard 10x8, which has only 1mm walls, and collapses much to easily. There is a Tygon 3603 which is 11,2x8mm and that has some succes over here, but is too expensive I think, and you need special 11mm compression nuts for the fittings - you can't just use standard cheap 10/8mm Push-ON fittings like you can with 10x7mm tubing (which is available in a very nice Clearflex version).
We have also been selling (and using) this 11/8mm combo, but personally I still prefer the 10/7mm for 3 reasons:
1. Fittings and hose are cheaper.
2. The hose has a smaller bend radius.
3. The flow advantages of the slightly larger 8mm ID are probably marginal at best, as the fittings used are still only about 6,9mm ID (as also shown by Cathar's graphs) - where as with the 7mm ID, there is hardly any ID flow restriction going through the fittings.

BUT 11x8mm PVC is sure much better than 10x8 PU or PVC - so not a bad choice at all.

As the hose ID gets larger, the sidewall thickness also needs to increase to still have a good bend radius, so something like a 12x8 with a Push-On fitting with ID of 8mm would be really great. (and guess what diameter Zalmans blue hose is.. yes 12x8mm exactly. Sometimes the big companies do get it right!) Only problem is that sourcing the fittings and hose for this size is not easy. I have only been able to find suitable fittings in plastic, and we really prefer nickel plated brass (with o-ring sealing).

[Berg]

Marci >> Yes, with tube end inserts, 10x8 pvc could be used in Push-IN fittings, but the hose is still crap due to the 1mm thin sidewalls.
I have never seen tube end inserts for 7mm ID, but maybe they are around somewhere. I wonder if they wouldn't cause some flow restriction themselves though - which would make them a bit pointless.

Turtle 1: >> Yes I noticed you said something about multiple loops earlier - which is really the only way to go when you have more than 2 or 3 waterblocks if you ask me. Solves pretty much all flow problems.
When you say "complete flow control system" do you mean you have valves for each loop? - or something different? - have any pics?
I started using seperate ball valves for each loop in my own systems a few years ago, so I could control and balance flow for each block as I wanted. I use little anodized aluminum manifolds with 1/4" thread, where you can put in a row of ball valves if you like: http://www.vcore.dk/shop/pro-mini-ventil-1-975p.html
(have not been able to find a similar manifold in copper..) - usually I will have a 3/8" ID feed into the end of the manifold, and then ideally an identical unit after the waterblocks to join the small ID hoses again - and meet in a 3/8".
This type of solution is not really necessary for most people though, but is nice for the "control freak" - and really nice if you are often replace a waterblock or fiddle with the system, as you can shutdown and isolate any one block without having to drain the entire system!

[Turtle 1]

As I already said I like the way you think.

We are working with these now both the W098/ $798 and the W099/ $1,124. Yep thats expensive but we really don't care.
http://www.emiplastics.com/water_regulators2.htm

These are available also but their not what we want . They are way cheaper tho.

http://www.emiplastics.com/water_regulators.htm

As for manifolds I like stainless steel . You can get those here or make them as we do .

http://www.plastixs.com/burg01/b01_03.html

look around on this site they got some nice stuff .

[Berg]

lol - yes now that is what I call a water flow regulator!
I like it a lot Turtle
- but definitely stuff in a whole other price bracket than my simple manifold with ball valves! Wouldn't mind one of those brass flow regulators to play with though

[Cathar]

Great post, Berg. Good info.

I've personally tried:
8mm ID over barbs
9.6mm (3/8") over barbs
3/8"ID | 1/2" OD into push-fits
7/16" ID over 1/2" barbs
1/2" over 1/2" barbs

I agree, and estimate as you stated above, that 8id/12od mm sizing into push-fits would be the near ideal in terms of wall strength, flexibility, radius, and minimal flow-resistance impact in an in-case full system. With larger ID's than that, flow rates don't get impacted by a great deal. If 9/13mm was available, that'd be what I'd prefer to run with. The closest to that is 9.6/12.7, but wall thickness is too thin.

Like yourself though, there's a difference between deciding what would be best, and finding the tubing/fittings to go with it. I'm still in the process of hunting down my "ideal" parts.

[LogAn'sRun]

As the hose ID gets larger, the sidewall thickness also needs to increase to still have a good bend radius, so something like a 12x8 with a Push-On fitting with ID of 8mm would be really great. (and guess what diameter Zalmans blue hose is.. yes 12x8mm exactly. Sometimes the big companies do get it right!) Only problem is that sourcing the fittings and hose for this size is not easy. I have only been able to find suitable fittings in plastic, and we really prefer nickel plated brass (with o-ring sealing).

McMaster has Tygon 3603 in those measurements with 2mm walls, which is almost what Cathar wished for wall thickness somewhere else in this thread (if I remember right, he wanted 25mm). Anyway, have you tried Legris fittings? Don't they have some push-ONs that would work with that? What's the disadvantage of plastic vs metal fittings?