The impact of tubing sizes

[Waterlogged]

LogAn'sRun, do you mean Push-INs? A quick look through the Legris online catalog doesn't show any soft line compression (Push-ONs) fittings.

I have Legris Push-INs in the rig I'm posting this on. The thing with Legris (and most other manufacturers) is, they put a stupid hex for an Allen wrench inside. On the fittings I have in use now (G1/4'x12mm w 12x8 Ether based Polyurethane tubing) they use a 7mm hex, I think this is what Catar is trying to avoid . IIRC, it drops to 6mm for 10mm O.D. tube fittings.

With the Tygon, the material itself isn't strong enough to resist the forces placed on it by the fittings. You would need to add a thin brass, copper, or delrin(?) insert to support the tube from the inside.

FWIW, I have some Delrin fittings on order from Mcmaster that don't use a internal hex that look like they "might" be from John Guest. I also have some more of that polyurethane coming and I also order some tubing that hopefully will work out even better.

[LogAn'sRun]

yupp, my bad. It was push-ins that I meant. But wouldn't 2mm wall be the real deal? Or is that tygon in itself is just too soft (at least the 3603)?

[Waterlogged]

Yeah, even at 2mm 3603 (@ 55A durometer) is likely a bit soft to use without some sort of internal support. If you get a good seal right off the bat (without support), over time (could be as short as a few hours or as long as a few years[it really is Russian Roulette at it's best ]) the tygon starts to give in to the force of the gripper ring and o-ring inside the fitting (if not the suction side of the pump) and will start to leak.

The softest durometer Mcmaster lists for clear tubing rated for use with "instant" fittings is exactly what I bought, Ether based Polyurethane (@ 95A durometer) (in fact, it's the only clear tubing listed). Now, the softest durometer they list is 60A, but it's EPDM rubber tubing and is black (it's also not rated for pressure).



I've included a durometer chart for anyone wanting to do a rough comparison of hardness. It's from Mcmaster but I'm hosting it.

[Berg]

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?

Yes a Tygon 3603 with 8mm ID and 12mm OD, would be just about perfect as such. The problem lies in finding the matching Push-On (compression) fittings - which I have not been able too. As Waterlogged also said, Tygon 3603 (and all other nice soft PCV based tubing) is to flexible for use in Push-IN fittings - like Legris. The Legris are great, except you need a stiff PU type tubing to have a secure connection, and even then you may have trouble in some negative pressure situations as described earlier in the thread.
That is why Push-ON type fittings are a better solution - especially for larger diameter and softer tubing.

Plastic vs. brass: Plastic is just fine for Push-IN, but for Push-On brass is better due to the thread strength and holding power of the compression nut.
- also they look nice and shiny

[davidzo]

I read a large part of this thread and i am not totally convinced with everything said.

Way back in the day, when many w/c setups were 3/8", and only some were 1/2", we had the super-hot AMD T'bird CPU's, and poorly designed open-flow waterblocks with little internal furniture that demanded that as much flow be rammed through them as possible to perform. Further, most pumps that were available were like the Eheim 1250. High-flow, low-pressure pumps that dumped a fair amount of heat, and radiators taken from cars that also demanded flow rates in excess of 10lpm to get past their performance knee.

Are you sure it is not the other way round?
With open flow blocks having laminar flow characteristics, the impact of flowrate should be considerable less than with pin structures like the apogee series. The 1/2" tubing and high flowrate pumps back then dind't make any sense because the benefit was even smaller than what we have now.

[QUOTE=Cathar;2247245]
i.e. 1/2" ID made sense then. The benefit still wasn't huge over 3/8", but it was noticeable. Tests back then showed that the move from 3/8" to 1/2" meant anything from a 0.2-1.0C improvement, depending on various factors. A 1.0C improvement was enough for most people to make the jump, and so 1/2" tubing got its following. I also feel responsible in part for the jump to the 1/2" band-wagon, since I conducted a fair few tests back then to justify it.
[/QUOTE=Cathar]
I remember it when i argued about that with you over at procooling.com.
I thought back then my 10/13mm PVC whas all you need, while 8/10 or 8/6 are only little worse. In fact the 10/13 tube had good flow characteristics but it was a very poor solution in therms of wall thickness and it was too stiff compared to todays preferred masterkleer or tygon.

[QUOTE=Cathar;2247245]
Fast-forwards to today, and we have well designed middling restriction blocks which are more flow agnostic. We have well designed radiators that are more flow-agnostic (both in terms of fan power and liquid flow-rate). We have pumps that are near optimal for PC water-cooling that strike a good balance between pressure, peak-flow, noise, and heat. In short, everything has, quite rightfully, been pegged back from old-school high-flow excess, and tempered with a more balanced approach.
[/QUOTE=Cathar]
I agree with you, that the Apogee GTX is a middling restriction block. But this doesn't apply to aquaextreme blocks, nexxos series, heatkiller, storm, etc. which cripple the flowrate a lot more to get better turbulences, less laminar boundary layers and better performance through these. The apogee is a good example for a well designed block which also performs good in lowflow scenarios and offers a huge surface area well suited for todays increasing core sizes. Fuzion is probablyquite similar until it gets a jet injection plate. But there are still other blocks on the market which scale better and often also cool better at least on limited die sizes. especially nexxos xp (rev2, not 1) is extremely well on small dies when you are a lucky one and got a unit whith a good flat base. but on big dies it is only average, while nexxos xp highflow is a total joke in my eyes.

The one thing remaining is tubing size. It's also the thing that seems to cause massive grief. Years back when running tests with the Cascade/Storm designs, I realised that flow rates didn't have to be that high. When I started factoring in pump heat coupled with radiator performance, we learned that there is such a thing as "too much pump". There really is a happy middle ground.

Nice to see you writing such things. when i remember back to the procooling days (during cascade period, shortly before the first storm), i was talking about combining lowflow aspects like quick connects and turbulence designs with highflow techniques like loop planning, no elbows, big tube diameter...
i was dissed back then, i don't remember from whom, but it wasn't exactly nice. It was before the nexxostest on procooling, which came close to the storm g4 in therms of performance even with a concave base and while being only rev1. probably most people were thinking that i was a lowflow freak at procooling and at german kaltmacher.de most people thought of me as a highflow fanatic. both of them werent exactly nice to me. it isn't easy t sit between chairs, but finally i see something what one could probably describe as the globalisation effect on watercooling.

I saw that people loved their low-restriction 1/2" tubing, but people also complained about how big it was. I agreed. 3/4" OD tubing is rather large. People didn't want to give up the idea of the benefits of 1/2" ID tubing, so that's when I investigated and came up with the idea for 7/16"ID|5/8"OD tubing. I ran the tests and the maths. No net difference. The amount of people who adopted the smaller tubing showed just how many people were unhappy with huge tubing. People want small tubing, but that don't want to lose the performance benefits of larger tubing.

I love the Masterkleer 7/16" and i love Nalgene 180 7/16" (like tygon r3603 shore 55). It is much less stiff than any common lowflow tubing and i really love to work with it. I didn't knew that you were it who came up with that, but i ordered it after reading about it on XS only about a year ago.

Also, I'm just sick of small-bore/big-bore bickering. I've always been a "middle-ground is best and cut-the-crap" kind of guy. Been running tests and for some time now I've even considered converting to 8mm ID tubing with quick-fittings because the differences are so small on modern hardware.

I guess the "middle-ground ist best and cut-the-crap" side of you didn't show often back then. I'm glad it does show now.



Just back to the current topic:

Pushin fittings have great flow characteristics. In fact they are better than similar sized conventional barbs. cathar explained that well.
But the problem lies in the Details. Pushin fittings work with a sealing o-ring which seals on the outside. In order to have a realiable seal, one has to consider following rules:
- Normal tubings are inner-tolerated. for pushin you need outer tolrated tubing. Almost everytime the tolerance is so low that it works, but this is important for the industry to make clear for what purpose and how the tubing is manufactured.
- Tubing diameter must stay the same even when bending the tube. this is only given with stiff tubes like PUR, PU, PUN, Teflon, and metal tubes. Standard industrial PVC tubing has only shore a 85° which isn't enough.
- The biggest caveat is that you can't combine tubing like tygon with pushin fittings. Tygon/Masterkleer/clearflex/nalgene are all too soft for use with pushin fittings regardless of wall thickness.
- Pushin fittings are only available in limited diameters. maximum for G1/4" is 10mm, for G3/8" is 16mm. Full metal nickel plated fittings are often only available in smaller sizes such as G3/8" 14mm.
- Fitting G3/8" is not possible on some blocks for your motherboard or videocard. most blocks come with 1/4" threading and there is no space for more on some of these blocks.

Often it works anywas, but there is no manufacturer with a warranty for selfmade solutions or materials which aren't commonly used in the industry. For example i udes G3/8" 12mm pushins on a whitewater clone i dremeled some years ago (it was my first waterblock ) and fitted 13mm PVC tubing on it.

There us a trick to overcome some of these shortcomes, butit does come with its own downsides.
You can use thin stainless steel tubing with the outer diameter matching the inner radius of your tubing and push it into the tube before pushing the tube into the quick connect. this will make tygon and masterkleer and other soft pvc tubes stiffer and suitable for the use with quick connects.
but when usind thin metal tube inside your tubing, isn't that the same as using a very optimized barb?


I came back to barbs, because they are really practical to use and more realible when it comes to longtime sealing. they do not need an additional o-ring and i can see through the transparent tube when it leaks. i have already had original legris which leaked after four years of usage. nickel plated full brass wuick connetcs are better, but they still have the o-ring. barbs are safer on the longtherm.

I prefer Dangerden highflow barbs and 7/16" Nalgene tubing. It fits me perfect. and i do get some more degrees out of that than a well on low-flow performing GTX like in your test.
I will continue to use big and soft tubes and i will also continue the trend to make blocks with higher restriction built to extract more performance with bigger flowrates. There must be a solution to make Blocks with better scaling than Fuzion and Apogee GTX. Both blocks are perfect for the most watercooling fans, but not for agressive highflow use in my opinion. such better scaling blocks do perform worse in some scenarios, but if optimized correctly they also offer better performance. since this is xtremesystems, some people will probably also prefer these kinds of blocks.

Since i have holidays now, i will make some blocks and maybe test some structures. Maybe i can post a few pictures and numbers of what i mean with next generation of agressive blocks in a few weeks. sty tuned

David from Hamburg (Germany)

[CyberDruid]

Thanks for a great deal of useful information Cathar and all you other Water Heads.

I have been following liquid cooling for many years now and have to say that I too went big in response to the original gear.

Like so many things tubing diameter is an application driven choice. As Cathar pointed out in this thread a small nuclear reactor like what I run for Folding needs to be looked at a little differently than say my P4 specimen.

I have a question about a particular pump that is never mentioned in this thread.

I tried out the D-tek Db-1 pump in lieu of my favored MCP 655 or 50Z DC-12 choices that I usual revert to and was well pleased with all of it's characteristics in a low restriction loop.


Is this pump just a rebranded pump that you are already familiar with and have plug ins for?

BTW I find my Storm Waterblocks still outperform the FuZion, Apogee, Apogee Gt and Apogee GTX on my s 478 and s939 applications...but without doubt the FuZion works better on the s775.

I was not at all impressed with the GTX--and I think you might want to actually get a hold of a FuZion for your own data needs as it is not just a GTX in a different package.

Thanks again for a great thread on a hotly debated topic!

CyberDruid

[Marci]

I tried out the D-tek Db-1 pump in lieu of my favored MCP 655 or 50Z DC-12 choices that I usual revert to and was well pleased with all of it's characteristics in a low restriction loop.

Is this pump just a rebranded pump that you are already familiar with and have plug ins for?

Nope - it's a new pump that no-one has done any indepth testing on as of yet, that hasn't had a chance to establish itself fully within the scene, hence little attention is paid to it at the moment.

[Liquid3D]

This is probably the most useful thread I've read about water-cooling components. I've been discouraged with 1/2" barbs and results where some tubing will outright fail, especially with the provided plastic tensionsers compared to 10mm OD / 8mm ID hoses. And then how 10mm/8mm perform whether they're mated to O-rings (push-in) and tighten-down where there's a mini-barb upon which the outer metal seal tightens down the tube. The latter seems to be the best simply because it dosn't involve finding the "correct" tube, although that's the point isn't it? Finding the orrect tube only makes your system better.

It's amazing how often improper tubes (and size doesn't matter heree that's prima facie) outer strength vs. inner strength (or rather material characteristics) are supplied with supposedly matched "kits."

It's really up to the consumer (if they don't want to see leaks down the road) to find the proper tubes. This thread is the best I've seen.

Thank you - Bookmarked

[HESmelaugh]

Great article. I'm glad it was stickied, since I haven't been in these forums for long.

[HESmelaugh]

It's not a sticky, you Necromanced it...

True, it's not a sticky itself. But, fortunately, it's linked in one of the stickies. :-)

[Eternalightwith]

Isn't it obvious?

[HESmelaugh]

Erm... I just wanted to show some gratitude for the work the original poster and many of the following posters put into it. I didn't realize there was something wrong with that.

[Xilikon]

Nothing wrong, this thread is still very useful, hence why it's linked to a sticky

[jonny_ftm]

Thanks for letting me discovering it, I missed this one

[Cornelious0_0]

VERY nice thread Cathar...just what I needed to see...as I'm in the planning stages of a completely internal water loop in my Shuttle SP35P2 and was seriously considering 6mm tubing with compression fittings...good to know that I wouldn't be losing out any great deal.