A challenge!

[Les]

Originally Posted by unregistered

this is all as obvious as the tail on a donkey

Yes, but it is the length of the tail
WAJ's numbers look about right for 1m of tube

Larger effect of couse for 2m, roughly :

For the Eheim1048 would give increase in C/W of ~ 0.02 for reduction in tubing ID from 12.5mm to 6mm

[Cathar]

Originally Posted by unregistered

and your edit is why low flow systems with BIG tubing won't work (though simple numbers would suggest its optimum)
we seem to be a total of 2 who understand this

hmmm - won't work? Surely it is dependent upon whether or not the waterblock is using the fitting/barb as a middle-in impingement velocity acceleration device, as opposed to waterblocks that implement their own acceleration devices.

[BillA]

no no Cathar
pH was referring to the fluid velocity necessary to purge 'big hose' of bubbles
i.e. you can't use 1/2" hose with a very small pump
now you can put the system through gyrations to clear the lines, then what happens over time ?

[pHaestus]

Les: That's probably closer to right. but how many effective pipe length are added by all those 90 barrel connectors that seem to go hand in hand with the little tubing? Enough that probably a "high flow" system is closer to 1m tubing and a "german engineered low flow" one is closer to 3m. And that's assuming the same parts are used for both.

[Cathar]

Originally Posted by unregistered

no no Cathar
pH was referring to the fluid velocity necessary to purge 'big hose' of bubbles
i.e. you can't use 1/2" hose with a very small pump
now you can put the system through gyrations to clear the lines, then what happens over time ?

I must confess to being up too early - I read Les's earlier posts about turbulent water flow through tubings and assumed that there was an implication about improved performance through using smaller tubing sizes, ala as an explanation for the differences between the MCW6000/6002. I guess that teaches me not to open my mouth without clicking every link first.

[BillA]

coffee mate, coffee

[AngryAlpaca]

Quite true that I offer nothing, but whining (and not comprehending) is what I do. If we use those numbers for low flow, then we'll prove conclusively that anything under 1.2LPM for 12mm ID tubing is a poor performer, but will that help the argument for either side, or will it get people screaming that Procooling is biased (yet again)?

Maybe we should measure typical flow in what we find to be the "typical" system in both a German and an American system and then observe how it works with their radiators and blocks?

Do we know numbers for bends? I've been interested in these for a long time, and it would make a large difference (moreso in smaller systems, right?) to all of these numbers... What are the figures for 8mm or 10mm (much more common than 6mm in systems I see)?

[BillA]

I posted bend resistance here long ago
'our' problem is that 'we' have no memory, and won't search !

EDIT
not only you AA, this is the same thing that IMOG described,
noobs in -> net clue down

[Cathar]

Grab "SF Pressure Drop 5.0" from here. Very handy free tool.

[LPorc]

Originally Posted by pHaestus

Les: That's probably closer to right. but how many effective pipe length are added by all those 90 barrel connectors that seem to go hand in hand with the little tubing? Enough that probably a "high flow" system is closer to 1m tubing and a "german engineered low flow" one is closer to 3m. And that's assuming the same parts are used for both.

pHaestus, you indirectly raise a point of consideration. How is the typical small tube system plumbed? I am guessing that many of these multi-block CPU/GPU/NB/HD/PSU small tube systems are not plumbed in series as is typical of larger tubes. I can even see some of what appears as gratuitous use of fittings on the surface being intentional to add restriction to one loop to tune the flow between parallel loops.

Anyone familar with the "German" plumbing arrangments? I would love to hear how they are routing flow for these totally W/C systems with little tubes and small pumps.

It is obvious that the "German" approach has been to maximize what they can do within limitations, while the "U.S." approach has been to refuse to accept any limitations. It is finesse v.s. brute force. Finesse will prevail up to a point, then it is overwhelmed. I think a more interesting question would be if the typical big tube system has 3-4x the flow, why isn't it doing better more than a degree or three?

I think looking for the differences in methods is likely to be more productive than cross-accusations of tube envy and tube compensation.

[Myth]

I'm afraid that the real difference in German and US style systems originate from the fact that almost all comparisons are made with grave errors.. the test of waterblocks on WCP for an example.. And furthermore WHEN people agree that US style can give more performance, yet another funny thing happens.. people somehow always end up saying well it can only give an improvement of a degree og two..

[Cathar]

Originally Posted by LPorc

I think a more interesting question would be if the typical big tube system has 3-4x the flow, why isn't it doing better more than a degree or three?

Got results to back that up?

Wasn't that the whole point of this debate?

Can a ~1.5LPM system offer performance within 1-3C of a >6LPM system that's designed to run at those flow rates? 3C is fairly significant though.

[nigelyuen]

are you guys comparing big tube vs small tube or high flow vs low flow with same tube?

i think it also depends on what block you test with
if you use a US block, the difference can be big, but if you use a german block, maybe not as much.

[Cathar]

Originally Posted by nigelyuen

are you guys comparing big tube vs small tube or high flow vs low flow with same tube?

Really it's about the flow-rate, not so much the tubing size.

However, the lower the flow rate, then the smaller that the tubing size can be and not play a significant role in the system-wide resistance. i.e. the lower your flow is, the smaller ID tubing that you can use.

If we set an arbitrary limit of 0.5mH2O Pressure-Drop of the tubing in the system, then the following flow-rates are where 2m of that tubing size would offer 0.5mH2O of resistance. The following calculations assume a 0.05mm "roughness" in the tubing's inner surface.

4mm ID => 0.5 LPM
5mm ID => 0.8 LPM
6mm ID => 1.4 LPM
7mm ID => 2.1 LPM
8mm ID => 3.0 LPM
9mm ID => 4.2 LPM
10mm ID => 5.5 LPM
11mm ID => 7.2 LPM
1/2" ID => 10.6 LPM

Originally Posted by nigelyuen

i think it also depends on what block you test with
if you use a US block, the difference can be big, but if you use a german block, maybe not as much.

Actually that's a presumption I don't think anyone should form until some blocks are first tested. Have any of us seen such characteristics in "German" blocks? Is it a characteristic of all "German" blocks? Could some "German" blocks have the same performance characteristics as "US" blocks, but because they are so hindered by small tubing and low flow rates that such block designs have been discarded along the way thanks to test-beds like WCP, and never given a chance to shine?

[alexwai]

Originally Posted by Cathar

Actually that's a presumption I don't think anyone should form until some blocks are first tested. Have any of us seen such characteristics in "German" blocks? Is it a characteristic of all "German" blocks? Could some "German" blocks have the same performance characteristics as "US" blocks, but because they are so hindered by small tubing and low flow rates that such block designs have been discarded along the way thanks to test-beds like WCP, and never given a chance to shine?

Cathar

That is the main reason I am hunger to see the test result of those high-rank "German" Blocks done by pH.

[LPorc]

Originally Posted by Cathar

Got results to back that up?

Wasn't that the whole point of this debate?

Apologies, please prefix with "If, as some have claimed in this thread and others about the differences between the 'German' and 'U.S.' approaches to water cooling".

As for the debate, it is a little one-sided to be a debate, isn't it? Both sides have done different optimizations for different reasons, and instead of comparing notes and seeing what we can borrow from each other's approach we're busy doing nerdy little chest thumps.

It's not like both sides don't have working systems, so I see little need for the fervor. When I go looking for a new automobile, I don't arrange head-to-head challenges between economy cars and muscle cars.

[pHaestus]

I don't think that many people argue that the Japanese are such great engineers that the economy cars can hang with the brutish American muscle car at the racetrack though. That is the very argument I have seen elsewhere:

Originally Posted by wizd

Difference in temps on an American and European watercooling systems are minimal.. At max you are looking at a degree possibly two.

Originally Posted by knipex

Looks better. Performs as well.

Its engineering V's brute force.

Now to ME, that is not based on any sort of accurate test date. What is being proposed really in essence is that a German wb will perform better at 1LPM than the Cascade-class block does at 8LPM. That is a VERY bold statement; performance of wbs drops DRAMATICALLY at 0.25GPM; even for thicker bp blocks.

[Les]

Worth noting Bill's radiator data, which indicates that the lower the air-flow the greater loss in performance as coolant-flow decreases::-
http://thermal-management-testing.com/Thermo9.gif
http://thermal-management-testing.com/Thermo15.gif

[pHaestus]

Sure it does. It makes it all a rather ridiculous argument really; you can surely design a wb for "good" performance at lower flow rates (usually at the expense of continued increase at high flow rates) but everyone is using pretty typical radiators that take a big hit at lower flow rates.

[Les]

The main point was making is :-
The bigger hit is for low fannage(quiet?) systems.

[pHaestus]

I would hazard a guess that the combination of 6mm tubing and low speed Papst fans is very common

[BillA]

worth noting is that some rads perform much better at low air flows than others;
never exceed what more air flow can deliver, but can easily out perform a 'better' rad that has insufficient air flow

and who is to know ?
one rad mfgr does, but apparently chooses not to publicize

this is why good data on component performance is so useful
and lacking such it is quite easy to understand the mis-direction provided by poorly designed systems testing

[knipex]

Originally Posted by LPorc

pHaestus, you indirectly raise a point of consideration. How is the typical small tube system plumbed? I am guessing that many of these multi-block CPU/GPU/NB/HD/PSU small tube systems are not plumbed in series as is typical of larger tubes. I can even see some of what appears as gratuitous use of fittings on the surface being intentional to add restriction to one loop to tune the flow between parallel loops.

Well I cannot speak for everyone (I know a number of people who cool dual CPU's in paralell) but in general (read my situation) its all in series. I also have no right angle connectors (unless you include the one attached to my GPU block).

It goes pump - rad - cpu - Nb - GPU - HDD - res.

The mofset cooler used to be in series also untill I changed my board and removed it.

I can post a pic if anyone wants...

Also most low flow / small tbning users will admit to a sacrafice in temps but will argue its only a degree or 2 (as I myself have done many times). What I get in return is neat routing, a completely (or damn near) watercooled near silent system and a good looking set up. To me its well worth the sacrafice.

[Les]

Originally Posted by unregistered

worth noting is that some rads perform much better at low air flows than others;
never exceed what more air flow can deliver, but can easily out perform a 'better' rad that has insufficient air flow

and who is to know ?
one rad mfgr does, but apparently chooses not to publicize

this is why good data on component performance is so useful
and lacking such it is quite easy to understand the mis-direction provided by poorly designed systems testing

Which takes me back to radiator these design thoughts.
Relevant?

[knipex]

Originally Posted by pHaestus

I don't think that many people argue that the Japanese are such great engineers that the economy cars can hang with the brutish American muscle car at the racetrack though. That is the very argument I have seen elsewhere:


Now to ME, that is not based on any sort of accurate test date. What is being proposed really in essence is that a German wb will perform better at 1LPM than the Cascade-class block does at 8LPM. That is a VERY bold statement; performance of wbs drops DRAMATICALLY at 0.25GPM; even for thicker bp blocks.

Sorry PH.

I can see how thats the way you take it but (as I posted both quotes) I was thinking more along the lines of British or European Sports (particularly earier ones) car V's American Hotrod.(of the same era)

On straight line performance is no contest.
Take into account looks, handleing, size etc and the answer becomes less clear.

BUT when the too work together you get the AC Cobra ..

But you are seeing this allready in watercooling. Look at the Kit Pug suggested......

By the way if anyone has a right hand drive cobra they want to get rid of let me know..


And thats another argument I want to hear discussed. Right hand drive cars have to be so much better than left hand drive..... You can hold the steering wheel in teh right hand when changing gears....... :shrug: