A challenge!

[pHaestus]

from the Big Players at Procooling (http://forums.procooling.com/vbb/showthread.php?goto=lastpage&t=10101)

Since there is so much hate for me and the way I do things, and since your methods are so clearly superior, this should be easy for you guys.

Your pal*

pHaestus



*and by pal I mean the guy that leaves flaming bags of poo on your doorstep and rings the bell

[Knipex]

Ph


As I have said before no one here hates you. Even the origional poster you reffer to stated that you were a decent guy.

A genuine open and frank discussion and comparision or low for and small tubing is welcomed. A continuation of personal and inter forum animosity is not.

[greenman100]

Counterproductive? Maybe.

Hilarious? YES.

[Knipex]

Fun can be had on this forum but not at the expence of any individual. We are all, after all, adults.

[pHaestus]

As a forum I propose that you draft up a statement about the superiority of low flow/small tubing. I will write an article detailing why, for a given cooling system, that using lower resistance fittings and hose (proper plumbing and bigger tubing) will always result in either the same or improved performance.

To put it simply, for a given set of parts: smaller tubing is at BEST similar in performance and at worst it performs quite a bit worse.

I am quite eager to see the math on how you can improve cooling by using more restrictive tubing and fittings.

[Knipex]

Taken from my post on your forums.


Guys

I thought we had an end to inter forum bashing.

A genuine comparison of the pro's and cons of low flow and or smaller tubing is something I would be very interested in.

But forum bashing doesn't belong on places like this or wizd. No one not even the origional poster said they hated you. As a matter of fact the origional poster even said that you were a decent guy. No one even mentioned you wife.

Can it please end now...

Also we have never stated or pushed the fact that small tubing or low flow is "better" just that it has its advantages.

All of the following quotes come directly from wizd.

Quote:
High flow has the peformance edge but only if the loop is designed carefully and the components all complement each other. Due to the nature of the components its easy to constrict your flow which can adversly affect your performance.
Quote:
avoid low flow systems with insufficient cooling as they are called kettles


Quote:
Difference in temps on an American and European watercooling systems are minimal.. At max you are looking at a degree possibly two.
Quote:
I had one too (1/2 inch), but I changed so I could harness the beauty of some of Aqua Computers "stuff". I can still overclock as well, my RAM is only PC-3500, so I get the same OC on water from either schools. Star is right, but if you like what you've got, keep it, if you want to try some German stuff, go for it. All watercooling is good, and it's up to you to choose the one you want.

The first two are taken from the infamous thread. The rest from various posts.

We have our fanboys but you also have yours.

[pHaestus]

So from those quotes I conclude that you also agree in my statements that small tubing can not result in better temperatures and that the "American" approach (your words not mine) will yield better performance as far as using larger ID tubing is concerned?

That's what I pull out of your quotes; you might wanna confer more with the resident experts here for better ammo?

[Knipex]

Nope that is exactly whay I am saying. It is my opinion only I will add I cannot speak for anyone else (all though I expect 99% of the people here will agree with me). "German style" kits will not give better temps that "American" they will give as good or close (read a degree max 2)


I have allways said that the main benifit of smaller tubing is asthetics and ease of install. I will argue the fact that my temperatures are suffering hugely as a result though. I am quite willing to run a degree hotter with the advantages I get with the smaller tubing.

Another advantage in my view is that all the equipment designed to take 8mm O/D tubing is designed to run that way so its harder for a noob to build a poorly designed system. (but that is purely my own hypothisis)

[Darv]

Seems fairly simple to me.

If you want the lowest possible temperatures from water cooling you need a high flow system. Of course it's not as simple as that because more important than the tubing is the parts that make up the loop. You will want a powerful pump, big un-restrictive heatercore, high powered fans and an effective water block.

Now that may be the best for CPU alone but when you add in a NB and GPU block you really have put a spanner in the works. There aren't very many blocks that won't cause major restrictions in this area.

Now you can go low flow and loose that degree or two but you can put a packed loop together without having to worry about the loss of flow. The pump can be less powerful and so quieter. And smaller tubing is easier to route in a case.


I'll admit I don't know everything about watercooling and the different methods (who does??) but I can put it into simple Pro's and Con's for each

High Flow Pro's

Best possible temps
Best possible overclocks


High Flow Con's

Usually requires large heatecore
Can't add HDD, NB, GPU, Mosfet blocks without lowering flow
More powerful pump needed
Can take up a lot of space
Case airflow restricted



Low Flow Pro's

Can add multiple blocks without damaging flow
Takes up less space
Less powerful pump needed


Low Flow Con's

Performance not as good



It's by all means not a complete list so please add any more thigns you can think of.


All-in-all both methods are fine. They both do exactly what they set out to do. And if you want the best you can get from watercooling then a combination of the two is the best way to go. Less restrictive blocks, ie CPU, cooled with high flow. Restrictive blocks, ie Mosfets, HDD, cooled with low flow.

[dutchcedar]

So from those quotes I conclude that you also agree in my statements that small tubing can not result in better temperatures and that the "American" approach (your words not mine) will yield better performance as far as using larger ID tubing is concerned?

That's what I pull out of your quotes; you might wanna confer more with the resident experts here for better ammo?A perspective from a resident "expert"... :rant:

I have two duallie rigs sittin' on my desk. A dual 2400MP rig that runs Innovatek stuff and an Eheim 1048 is on the left. On the right is a dual 3.06GHz Xeon rig with RBX blocks, phat tubing and an Eheim 1250 (To those of you who've seen this before, sorry, but Big Player doesn't know me from [Bill] Adam[s]). They're the rigs whose innards were shown in the "other" thread, the one that got you so irked in the first place...

http://www.pcperspective.com/images/reviews/41/systemcomparison.jpg

So you can see right off, I'm not a fanboy of AMD, Intel, little tubing or big tubing... duallies, yah, I'm a big fan. Water cooling, yah, I'm a very big fan of that too.

My point? The idea of comparing these two basic types of water cooling to determine which is "best" is futile at "best". The high flow, big tubing system will win the performance contest by a degree or two, maybe even more. Does that make it "best"? To some people, maybe. But certainly not to all.

There are distinct advantages to each and those advantages change and shift in importance depending on each end user's needs and desires. This can be argued ad nasseum without any one definitive answer that satisfies everyone, because performance is NOT the only requirement of water cooling to every one that splishes and splashes in their 'puter. If that were the REAL case, we'd all be flooding (pun:wacko: ) the evaporative cooling forums, now wouldn't we?

Water cooling PC's is akin to hot rodding, where everyone's rig is not based or judged solely on performance, but on a number of intangible things that reflects on the owner's personality, budget and ability. That's why water cooling forums are filled with threads about water dyes, visible reservoirs, better looking clamps, tubing that stays clear, attractive ways to mount radiators, etc., etc.

You seem to view it more like one would view racing on the Bonneville Salt Flats, where the only thing that matters at the end of the day is the time slip. I think this hobby is a little more elusive than that and the people involved want more than that.

Sooo... if you're in search of an answer to the question, "which can perform to cool the CPU best, fat or skinny tubing?" The answer is a resounding "fat". No more technical evaluations, formulas, bench tests, etc., are required, for me at least. I know it, see it every day on these two rigs and have seen it on others. Common sense will prove it out as well.

I think, however, that that is only one of many more complicated questions that deserve addressing in your upcoming article.

The most important question you could address is, "how much fun are you having with water cooling your rig?" I think the folks here at WizD will all give the same unanimous answer... "A whole bunch, thank you very much!!!" :blink:

Sidebar: If a thin tubing rig is performing two degrees under par, crank up the room's air conditioner... just a little tip from your Uncle Dutch... :unsure:

[Cathar]

As I've tried to explain else-where, thin vs fat tubing is not really the issue. It's a means to an end, but it's not the main issue at hand.

The issue is two-fold:

1) Pumping pressure/flow characteristics
2) Block resistance

and the characteristics of 1) and 2) affect the tubing size choice.


Low Flow Pro's

Can add multiple blocks without damaging flow
That's not exactly true. Flow rates are determined by the pump's pressure/flow characteristics vs the back-pressure that's applied. More blocks will always increase system resistance, and flow rates will always drop when more resistance is added. The issue here is the "knee" of various block's performance curves and where you're sitting relative to that "knee" as you start lowering your flow rates, although ss you start getting much below 1LPM (0.25GPM) block performance, even for many of the "low-flow" blocks, does start to drop away pretty quickly. About the only block designs I'm aware of that sustain good performance at much below 1LPM are true micro-channel designs.

To illustrate the point for high flow setups not being terribly affected by additional blocks, let's assume a DangerDen TDX block, 2m of tubing, a Swiftech MCP600, and adding/removing a Silverprop Nexus LX GPU block.

Using overclockers.com pressure drop results, we'd estimate that the pressure drop of the pump, tubing and TDX would be around 0.5mH2O @ 1gpm, and with the added GPU block would be around 0.85mH2O @ 1gpm.

Extrapolating those figures to the Swiftech MCP600's PQ curve, we'd estimate 7.0LPM without the GPU block, and around 5.8LPM with the GPU block. Performance wise at the CPU, we're talking about a 0.3-0.4C difference under full CPU load as a result of the lower flow rates.

The point being that the high-flow systems do not suffer tremendously from the addition of extra blocks, as long as those added blocks abide by the same design philosophy.

I'd do the math for a low-flow setup if I had access to the data, but that will come soon enough.

[Knipex]

Now this is what I am talking about.

Open discussion and debate.

[Darv]

Thanks Cathar. I didn't realise the difference was so low in a high flow setup.

I knew that adding blocks to a low flow system would lower the flow rate. What I meant was that it doesn't really make much of a difference to the performance.

[Starbuck3733T]

Arrrright cathar! good stuff.

What I'd love to see is some multiloop (ala-blade runner) calculations on how much worse/better the parallel combinations would be as opposed to the serial combinations.

[Cowd Tyke]

Thanks Cathar.

I'm a virtual newbie with little detailed understanding of flow, whilst your block design work is held in high esteem :) . Your explanation will make me look at PQ curves with renewed interest :blink:. At the risk of putting my foot in my mouth I'll risk a question or two:

Does that means that with a database containing

PQ curves of pumps

Pressure drop effects of adding brands of
PSU, HD, CPU, Graphic and Chipset cooler
Pressure drop effects of adding brands of radiator

the cooling characteristics of radiators with varying numbers of fans

that someone could write a "cooling simulator program" into which you enter
your various choices
plus an optional tube diameter
and your room temperature

and it could predicted

efficiency of the chosen block under the flow conditions
temperatures expected throughout the cooling system
suggestions on improvements
etc..? (What a simple life I lead)
What other effects do I need to take into account?

How far away is the cooling community from having this enough of this data available?
You mention overclockers.com pressure drop results but I've not been able to find a URL for those yet... Does this also include pressure drop from adding a HD cooler or a radiator or a PSU.
Can I, as a rough estimate, just add 0.35mH20@1gm per item added?

Cowd Tyke

[Cathar]

Crowd_Tyke, what you're asking for is somewhat possible for certain restricted scenarios. The main issue that is holding up such a thing is a lack of test data.

For example, each radiator type has it's own unique flow/performance curve, and someone would have to measure it with correct equipment and provide that curve across the range of potential flow rates, which I guess would be anything from 0.5LPM -> 12LPM (or as high as is possible given the pumping limitations of the testbed).

The same applies for each block that you intend to add to the hypothetical system. We need that flow vs performance information to be available across the range of hypothetical flow rates. It even gets a little gnarlier when you take into account that blocks can and do behave somewhat differently when the size of the heat source is changed. i.e. a T'bred B flow/performance curve will be different to a Barton performance curve, and different again to a P4 without the IHS, and different again to a P4 with the IHS intact.

i.e. to support what you want requires an absolute wealth of test data, most of which is presently being supplied for free by certain individuals with the correct equipment as they have time to do so.

However, if all one wanted to do was to roughly predict the final flow rate given

a) tubing
b) pump
c) fitting type/size
d) radiator
e) one or more blocks of known pressure drop

then yes, that could basically be done today by grabbing the eclectic set of datum that presently exists for various pumps, radiators and blocks.

The Overclockers.com results can be found here:

http://www.overclockers.com/articles373/wbsum.asp

I do have an issue with the OC.com results though. They are ranked according to a fixed flow rate, being 1GPM, however that is fairly unrealistic as in real-world setups the final flow rate is affected by the pressure-drop of the block, and for some blocks which are ranked highly (e.g. the HydroCool HS5), OC.com provides enough data to show that the HS5 would be outperformed by various other blocks once those other block's lower pressure drop is taken into account.

Sorry for the rambling post. You've just asked a heck of a question with a wide and far reaching answer that I've tried to summarise as it's now 12:15am here...

[pHaestus]

Crowd_Tyke:
You have the right idea. It's possible to measure pressure drop vs. flow rate for radiators, waterblocks, hose and fittings, reservoirs, etc and to generate P-Q curves for pumps. You can then add your resistances and solve for a final flow rate given your pump. If you also have C/W vs. flow rate data for waterblocks then you can make a rough guess at the final temperature of your hot bits too. THIS is complicated! The fan and shroud and air intake configuration will affect water temperatures. The water temperature affects viscosity of the water, and the difference between water and air temperature will affect the efficiency of the wb and the radiators.

Such a calculator is one of the things I am working on right now: generating pressure drop vs. flow rate numbers for common radiators and the wbs I have tested. Takes a while to get up to speed with the new equipment and to calibrate everything properly though. Also it's hard to work around the huge stacks of cash that are piled up everywhere "Blow" style :)

[mrplow]

Crowd_Tyke:
Also it's hard to work around the huge stacks of cash that are piled up everywhere "Blow" style :)

http://www.erostheclub.com/pimp.jpg

[Da_Rude_Baboon]

At last the opposing sides are having a reasoned debate! :wub:

Thanks Cathar and pH for your input as you've already opened my eyes to how (little) additional blocks can effect flow and performance. Heres hoping the good will continues! B)

[dutchcedar]

The debate is indeed becoming more reasonable. It will interesting to see any serious testing that comes of it.

I still get the impression that on the "other site", if you aren't overclocking your rig, you simply aren't worthy.

Despite that, I hold that the water cooling hobby covers a much larger spectrum of users than just those who want to dial up the speed of their rigs. In fact, I would wager that the majority of people who use water cooling are not only looking for an overclock. I also contend that any real serious overclocker would be using refrigeration, not just water. :blink:

[Edd]

In fact, I would wager that the majority of people who use water cooling are not only looking for an overclock.
Yeah, I primarily watercool for the silence aspect. My overclock on my 1700+ is no different on water than it was on air (though to be fair it's one of the CPUs that just WILL NOT go any higher than 1.92Ghz).

Silence (well, quiet) was my primary target, not forgetting stability when the ambient temp goes up a bit. And because it was something that I hadn't done yet. I'm sure that when I get my A64 up and running when the new waterblock gets here, I will be able to push some more performance out of it that I wouldn't have been able to do on air, but that's not my main concern.

[Knipex]

I have a feeling that you are under estimating ProCooling. I agree that some posters canot see beyond cooling performance but most see both sides.

I have a feeling that testing will go ahead IF we can agree on kits.

What i would like to see tested is.

Performance using decent fans on both systems loop with and without GPU block. (loop without waterclock will have no interest to 99% of us but it does to alot of procooling members).

Performance of both systems with quiet fans fitted. Same test as above. (I would like this to get as low as 7V Papst but I cannot see ProCooling going for this but I will ask)

Performance of both systems with extra blocks added. (tested as above)

On the results. On the CPU only I expect us to get beaten. As the overclock increases I expect the diference to get worse. The higher the temp the more they beat us.

With the GPU block included I see the results getting closer.

If I can get them to try "quiet" fans I expect us to give them a surprise.

The multiple block test isnt realy a fair one, its not suited to high flow but then again CPU only isnt really suited to low flow so fair is fair.

Any recomendations on the blocks we should submit. Remember 1048 pump.

[SlaterSpeed]

I also contend that any real serious overclocker would be using refrigeration, not just water. :blink:
If only I had the money and it was legal to build one myself :(

On the noise issue ive allways felt that a highflow setup is perfectly capible of being quiet. The only real differance between the 2 approches is pump noise and ive never found it to be a problem. i guess if your going for silence it may become an issue but ive not seen many people who can truly say they have a silent setup (Bladerunner is the only guy i know). I cant hear my pump (eheim 1250) over my psu fans on 7v so unless i watercool the psu i cant see any way to get it quieter.

IMO fans are pretty irrelivent. You can run a highflow setup fanless (or with quiet fans) if you want to and it will respond in the same way as a lowflow. One thing i have never been able to work out is if there are some rads that are less sensitive to airflow than others (In a simlar way to the differance between a low and highflow waterblock) or if lowflow has some magical advantage that means undervolting fans has less effect that on a highflow setup.

slater..

[Darv]

On the issue of noise you can make it quieter with a low flow setup than a high flow simply becasue you can cool HDD, PSU realtively easier.

The problem is different people have differnet tollerances to noise. So a system may seem quiet to one person but loud to another.


On the fan issue I've always thought that if you have the water moving slower through the rad it has more time to dissapate the heat. And with smaller channels it can go through the rad more times.

Might not work like that in real life but makes sense to me. :wacko:

[dutchcedar]

On the issue of noise you can make it quieter with a low flow setup than a high flow simply becasue you can cool HDD, PSU realtively easier.We'll see how these tests pan out, but I doubt this is the case. A high flow system uses a larger pump, but like Slater, my 1250 is virtually silent (as is my 1048). So the only "noise maker" is the radiator fan. On both my rigs, they're in shrouds to cut down their noise. I just don't see how a low flow system can be quieter... it depends on the fan and only the fan.

[Blackeagle]

Hi guys,

I think the reason the Papst fan @ 7v will be unpopuler at ProCooling is due to the fact it would not give any meaningfull air flow through a 2" thick car/truck heater core rad with it's folded and corigated fins. Just nowhere near the pressure level desired for a fan for use on a heater core rad. And a rad without good air flow is a disaster for temps.

Some questions I have for you guys regarding noise levels and how they should be considered in these tests.

1) What would you fellas consider (in dba) a low noise level for a whole rig to be? 35dba? Higher? Lower?

2) pH at ProCooling posted that he would consider it a victory for the low flow rig if it comes in within 2c of the high flow rig. So what would be a fair dba handicap for the hiigh flow rig vs the low flow one in regards to audio levels of the whole system.

I look forward to what waterblocks for the CPU & GPU your sites members pick along with what rad.....ect.
Should offer members at both sites a chance to learn something usefull.

[Blackeagle]

And I'm glad to see members of this site like the idea of the GPU block being added.:cool:

[SlaterSpeed]

On the fan issue I've always thought that if you have the water moving slower through the rad it has more time to dissapate the heat. And with smaller channels it can go through the rad more times.

Might not work like that in real life but makes sense to me. :wacko:
Yeh thats attualy the oppisite to how it works in real life:p

Rads have a performance curve just like a waterblock for both waterflowrates and airflowrates. They work better at higher flowrates due to more turbulance being created.

Your right that the more channles or vanes the rad has the better it should work but the differance between a heatercore and a traditional multipass lowflow rad is not the number of vanes, its how they are connected together. Basicly in the lowflow type of rad they are all connected in series whch makes them more restrictive to waterflow. In a hetercore the veins are connected in paralell so theres less resriction.

Im not entirely sure if a heatercore is more restrictive to airflow than traditional radiators (BillA did test it but it was a while back and i dont have a link). Basicly theres a tradeoff between the thickness of the rad (read Airflow resistance) and the surface area of the rad. That tradeoff can be made in both high and lowflow setups without effect to the rest of the setup. (thats why we are seeing blackice style rads taking over from the traditional type in europe)

gotta get to work :(

slater..

[Risky]

Rads have a performance curve just like a waterblock for both waterflowrates and airflowrates. They work better at higher flowrates due to more turbulance being created.

I take it you mean better at higher airflow rates and lower waterflow rates.

[Da_Rude_Baboon]

I interpreted it as higher water flow rates cause turbulence in the rad/heatercore which improves cooling. Heatercores require more powerful fans as they have denser fins. :unsure:

[Knipex]

Hi guys,

I think the reason the Papst fan @ 7v will be unpopuler at ProCooling is due to the fact it would not give any meaningfull air flow through a 2" thick car/truck heater core rad with it's folded and corigated fins. Just nowhere near the pressure level desired for a fan for use on a heater core rad. And a rad without good air flow is a disaster for temps..
I agree with you but to me that is a disadvantage of high flow systems.. That is why i would like to two compared.



Some questions I have for you guys regarding noise levels and how they should be considered in these tests.

1) What would you fellas consider (in dba) a low noise level for a whole rig to be? 35dba? Higher? Lower?.
Again this is a subjective question to to me 35dB is way too loud. Anything over 25dB is too loud. A 120mm Papst FGML @ 7V is the just quiet enough. (Im not joking by the way that is the loudest thing in my rig even the PSU is runng the 7V papst)



2) pH at ProCooling posted that he would consider it a victory for the low flow rig if it comes in within 2c of the high flow rig. So what would be a fair dba handicap for the hiigh flow rig vs the low flow one in regards to audio levels of the whole system.
This is one point where we will never agree. (See my answer above) That is why i would love to run the test with 7V Papst fans, I want to see a quantifiable answer to how much the high flow would suffer in comparison to the low flow.

I look forward to what waterblocks for the CPU & GPU your sites members pick along with what rad.....ect.
Should offer members at both sites a chance to learn something usefull.
As do I. If nothing else it should prove to both camps that Both systems have their place.

We'll see how these tests pan out, but I doubt this is the case. A high flow system uses a larger pump, but like Slater, my 1250 is virtually silent (as is my 1048). So the only "noise maker" is the radiator fan. On both my rigs, they're in shrouds to cut down their noise. I just don't see how a low flow system can be quieter... it depends on the fan and only the fan.
Ah here is the crux of the matter. We have allready agreed that 99% of high flow systems only cool CPU and perhaps GPU and / or Northbridge.

Radiator fans are not the only sources of noise in a computer. You have hardrive noise, PSU noise and case fan noise. We can and do put the hardrive in a watercooled soundproof enclosure (no airflow) watercool PSU's, moffsets practically everything. For most of us the only fans in our cases are 2 Papst 120mm fans @7V. (Ok I have 3 one on my PSU but thats because I cannot justify the spend on a watercooled PSU when the oneI have is so quiet)

How many Highflow systems can do that??

That is the main reason I want to test both systems with 7V Papst fans.. It is the only was we gat a true comparison of how they would behave.

[Darv]

We'll see how these tests pan out, but I doubt this is the case. A high flow system uses a larger pump, but like Slater, my 1250 is virtually silent (as is my 1048). So the only "noise maker" is the radiator fan. On both my rigs, they're in shrouds to cut down their noise. I just don't see how a low flow system can be quieter... it depends on the fan and only the fan.
What I mean is when you're cooling a HDD in a soundproof enclosure and you have eliminated the noise form the PSU fan, you end up with a quieter system.

[Etacovda]

Hi guys,

I think the reason the Papst fan @ 7v will be unpopuler at ProCooling is due to the fact it would not give any meaningfull air flow through a 2" thick car/truck heater core rad with it's folded and corigated fins. Just nowhere near the pressure level desired for a fan for use on a heater core rad. And a rad without good air flow is a disaster for temps.
I'm probably one of the only users there using a 2" thick rad, and thats because it was handy at the time; (I already have my 2 rads for 'final use' here, i just have to paint and shroud them).I had a shroud for it, and I had painted it already. Its just a simple 1x120mm fan rad with a shroud; it outperforms one of a lower fin density by water temps slightly, but its certainly not a 'standard' procooling rad. Course, i could be totally wrong... but I think the guys over at procooling will be using 1" thick max heatercores.

My final (two) rads will be a 2x120mm 1" thick rad, and a 1x120mm 1" thick rad ( 240*120 surface for the first, 150*160 for the second; both shrouded)
which means i will be able to run these at 7v or even fanless for decent enough performance; i think this is the performance, quiet part of 'performance, quiet, compact' solution, becuase i

More waterflow (faster) through a radiator is better. More turbulence is better in a rad, as more water hits the sides etc and causes more cooling. Water being in the radiator for longer doesnt make it cool better, in any shape or form... theres a good long thread over at P/c somewhere about it, im sure someone else will dig it up.

[Risky]

So if you had a pair of identical radiators, they would be more effective in series than in parallel? Would this apply regardless of design?

[Knipex]

Ok time to show my ignorance.

IMHO.

Yes two rads in series would cool the water more but more efficent well thats different.

If the first rad work well it will get the water temp close to ambient. The closer to ambient it gets the less efficent a rad becomes. Ie if water is 30 degrees above ambient you will havea big difference in the temp of the water entering and leaving the rad.

If the water is only 5 degrees above ambient the difference might be as little as a degree. You are adding extra restrictions and noise (fans) for a diminished return.

Or I could be completely wrong.

[Cathar]

For those wanting good low-fan-flow radiator performance, it may be worthwhile considering the comments being made in this thread:

http://forums.procooling.com/vbb/showthread.php?p=116301#post116301

To put it quite simply, almost everything that's on the market today is not well suited at all to work well with low-flow fans. 1"+ thick cores are too restrictive to air-flow to allow a weak fan to push/pull much air through them, and this hinders the low-flow fan's potential cooling performance.

Rather than moving to thicker cores with weak fans, we really want to be moving towards thinner cores.

There's some good information in that thread, as well as a link to a European seller (in Italy) who perhaps provides the closest thing to what low-flow fan people really want.

[Cathar]

So if you had a pair of identical radiators, they would be more effective in series than in parallel? Would this apply regardless of design?
The short answer is "it depends". There's no one easy answer. It also depends on the air-flow through the radiators too. For higher flow rates (>4lpm) it is typically fractionally better to put them in flow-series, but the maths does all work out to be fairly close. At low-flow/low-air rates, it may be a little different. I haven't crunched the math for that scenario yet.

[Knipex]

Cathar

Are you reffering to a radiator like this

http://www.wizarddesigns.co.uk/imgs/ape240.jpg

[Pug]

from the Big Players at Procooling (http://forums.procooling.com/vbb/showthread.php?goto=lastpage&t=10101)

Since there is so much hate for me and the way I do things, and since your methods are so clearly superior, this should be easy for you guys.

Your pal*

pHaestus



*and by pal I mean the guy that leaves flaming bags of poo on your doorstep and rings the bell
*Sigh* I fear the plot has been lost again slightly with this "challenge".

Show me where I said/implied that despite anything else, low flow would beat high flow in principle?
I said that my choice of European (or German, if you will) production kit would beat your choice of US production kit. I was prepared to put my money where my mouth was and posted accordingly (so I thought).

The high flow and low flow pros/cons thing would be good to quantify but performance is not everything.
I reached a hefty overclock on air but couldn't stand the noise levels of the hardware necessary to achieve that. That's why I switched to water. No loss in overclock but now I can put my PC in place of my TV and the loudest thing in this setup (http://www.wizdforums.co.uk/Pug/hexdesk.jpg) is my Xbox! (prolly a future w/c project will solve that ;))
It's fully internal, lannable and if you ask me, refined.

I unplugged it the other week to ship it 500 miles to Modfest04, put it in a box, taped it up and sent it ahead of me. When I got there, I got it out, plugged it in, turned it on and booted straight up.
End of the event, I did the same and even after sustaining damage like this (http://www.wizarddesigns.co.uk/images/rigs/hexDamage.jpg), I was still able to do the same.

Personally, I'm proud of my rig, the parts in it and the way I've built it. If there were disadvantages that precluded me recommending a particular course of action, I would do so (as I did in the "Can you have too much pump?" thread on Ars Technica when I switched out the 1046 for a 1048 and the only noticeable increase I got was on the noise level front).
If my input there prompted someone to educate me on the benefit of larger tubing to go along with the pump, I'd have entertained the idea. If I could have then adapted the system at (a worthwhile cost) to give me better performance (as in a usable increase in overclock) without too much increase in cost or noise factor, I'd have done it.
Looking back, with the wealth of experience I've gained since then (more so by practical experimentation than theoretical assumption) I find myself still using the same actual pump and blocks with my next-gen components & hardware.

Before - http://freespace.virgin.net/water.cooled/globalwinAqua3.htm

After - http://www.wizarddesigns.co.uk/images/big/allCool.jpg

I used to use 90° angles back when I had semi-rigid polyurethane tubing but now I have an improved bend radius from polyurethane polyether, I've not used so many.
The funny thing is, I've also "upgraded" the old heatercore to a continuous copper tube rad. :lol:


Back on topic of this particular thread...
I'm not quite sure what this new challenge is set to achieve, tbh.
If you want ultimate cooling, it's simple - go European.
Phase change CPU (VapoChill being the main contender but WizD may have a further ace up the sleeve here in the nearish future);) & WizD secondary and tertiary narrow-bore cooling. :p

I have a magazine in my lap (from the real industry, not the enthusiast market) with a small feature on Mikros Technologies' "high performance liquid cooled heat sink" ... Quote - "that can handle heat fluxes in excess of 1000W/cm². At a water flow rate of 1 l/min/cm², they yield a unit area thermal resistance of 0.05°C/(W/cm²) (0.0078°C/(W/in²)) with a pressure drop of 2 psi"
Hands up who knows if that's any good? :D
Why aren't the real industry using high flow?
As for small bore - it has its place. You wouldn't mod the capillary tube on a Vapo for performance improvement, would you?

Sorry for any protracted absence from me but I'll try & be back with more input when I have less on my plate and can formulate constructive input with a clear head.

...and thanks for the civility to all concerned. Conflict makes my head hurt.
I hope everyone benefits from the outcome of this. If so, it will make all the earlier grief worthwhile. :)

Cheers, Pug
Owner: WizD

PS. Knipex - imho, matched parallel rads can improve performance if set up correctly (FIFO theory may help here) but mismatched rads may be best run in serial.

PPS. pH, heard of Anter Corporation/ Anter Laboratories? Check out www.anter.com. There's some good stuff there.

[rebelbrethren]

Hmm.

We've all seen this thrashed out time and time again across forums, but i'm still lost on one aspect - why does thick tubing (ie 1/2" ID) always seem equate to a high flow system?

The system i'm currently building is based on 1/2" ID - but not because i'm shooting for performance overclocks - I'm after silence (years of listening to a 7000rpm delta screamer has left me partially deaf, I swear) and a reasonable overclock (value for money, MBFYB ). However - the whole thing will only be running a 1048 because it doesn't need more, the heater core and the 1/2" ID is partially for the asthetics.

So, "Wide Tubes", and um, "Low Flow"!

It may not be the absolute optimum combination, but the core doesn't restrict flow, and the low pump rate mean peace and quiet, and thats exactly what *I* want. Watercooling, and modding in general is about making it individual - meaning exactly whatever *you* want - so unless its a really stupid set up, I can't see what people fight over.

I can see both sides to the argument here; what loses me is why there are only two!

Me? I'm from the camp that says unless you really screw up on the design of an H2O loop, water coolings gotta be better than listening to a fan by delta.

[SlaterSpeed]

So if you had a pair of identical radiators, they would be more effective in series than in parallel? Would this apply regardless of design?
Like cather said it depends.

If they are run in parallel then the total flow resistance of the loop will be reduced allowing a higher flowrate overall and this can lead to (small) gains at the waterblock. However the flowrate through each single rad will be alot less (not exactly half tho) so this will have a negitive effect on the performance of the rad (less turbulance etc. etc.)

If you put them in series the flow resistance of the loop will increse leading to a small loss at the waterblock but the flowrate through each rad will be the same as the overall loop flowrate so you will gain (well 'not loose' relly) due to incresed turbulance etc. in the rads

Basicly the factors tend to cancel each other out and it depends on the specific PQ curves of the block and rads as to which is better. It can be calculated but you need alot of info.

(also with a highflow setup you need to consider the increce in resistance by using Y splitters may attualy cancel out the lowerd resistance of having the rad in parallel)

ps. My hcore is 50mm thick

slater..

[BillA]

Hello all,

it would seem in reading this thread that much of the difference in opinion is based (god this blue is hard on the eyes) on different, or little, data
- what is accepted as a given on procooling is unknown here

and I suspect that my old data is partly to blame
for those not famaliar with these scribblings and pages of graphs, they are collected on an old site www.thermal-management-testing.com (http://www.thermal-management-testing.com/)

many topics are addressed, but caution is advised wrt the Radiator Heat Dissipation Testing article, it contains many errors; much better is the ThermoChill article

appreciate that the motivation behind this testing was to ascertain the performance capability of wbs and rads - under any condition of use
- low flow to high flow, all conditions are covered

to those that find such reading difficult, persevere; else you will have difficulties following the explanations given

[scopEDog]

Bill I do have a question for you, if you would be so kind..

In your opinion, at what point does increasing tube size decrease efficiency of a particular setup? In other words, would we see better performance from using 3/4" tubing and accessories or would the gains just be nominal?

I've read through your testing methodology (as seen here: http://thermal-management-testing.com/methodology.htm) and all basis are covered, but I wanted to hear your opinion on the matter.

The reason i ask is because some people would find this acceptable, while others wouldn't deal with the huge tubing size. And maybe if 3/4" would ever become popular, some would say 1/2" is tiny :). So ultimately it all boils down using whats out there that suits your application IMO. But in the pursuit of science, and for the most effective waterblock; procooling pushes the envelope. Notice that i didn't say efficient waterblock, cause that would be a block with .5 gpm going through it, running 1 degree above ambient. That would be something, but as you would say Bill, theres no free lunch :)

"The initial performance tests are to determine the pressure drop across the wb in relation to the coolant flow rate. This is done with the coolant at 25.0°C at flow rates of 68, 114, 227, 341, 454, and 681lph (0.3, 0.5, 1.0, 1.5, 2.0, and 3.0gpm)." Now thats a ton of data to extrapolate...i dont envy you :)

[BillA]

"at what point does increasing tube size decrease efficiency of a particular setup?"

never, so long as the pump can clear the lines
this is the 'technical' answer, the practical considerations are quite different
note that the fluid velocity in 1/2 hose is low with the relatively small pumps we use to WC, no 'need' to upsize
- larger lines serve only to reduce the fluid velocity, which reduces the line resistance, which then yields the higher flow rate due to the reduced resistance

EDIT: not my 'opinion', these are known scientific principles - see any basic text on fluid mechanics

[Risky]

Like cather said it depends.

If they are run in parallel then the total flow resistance of the loop will be reduced allowing a higher flowrate overall and this can lead to (small) gains at the waterblock. However the flowrate through each single rad will be alot less (not exactly half tho) so this will have a negitive effect on the performance of the rad (less turbulance etc. etc.)

If you put them in series the flow resistance of the loop will increse leading to a small loss at the waterblock but the flowrate through each rad will be the same as the overall loop flowrate so you will gain (well 'not loose' relly) due to incresed turbulance etc. in the rads

Basicly the factors tend to cancel each other out and it depends on the specific PQ curves of the block and rads as to which is better. It can be calculated but you need alot of info.

Is there a "turbulance" benefit even when the rad is in the single-tube-multipass-configuration? I would think that if you were using a setup with continuous ID (hose/fittings/radtube) this would be minimised.

(In case anyone cares I'm using 12mmOD and 10mm OD tubing and 1/4 thread fittings throughout!)

[SlaterSpeed]

Is there a "turbulance" benefit even when the rad is in the single-tube-multipass-configuration? I would think that if you were using a setup with continuous ID (hose/fittings/radtube) this would be minimised.

(In case anyone cares I'm using 12mmOD and 10mm OD tubing and 1/4 thread fittings throughout!)
There will be 'some' benifit but it may be less than with a heatercore. I would say multipass rads would benifit from being in parallel because of there high resistance but then why not just use a heatercore? Are these multi pass rads perticulay 'non restrictive' to airflow? thats the only benifit i can see in using them.

slater..

[Darv]

"at what point does increasing tube size decrease efficiency of a particular setup?"

never, so long as the pump can clear the lines
this is the 'technical' answer, the practical considerations are quite different
note that the fluid velocity in 1/2 hose is low with the relatively small pumps we use to WC, no 'need' to upsize
- larger lines serve only to reduce the fluid velocity, which reduces the line resistance, which then yields the higher flow rate due to the reduced resistance

EDIT: not my 'opinion', these are known scientific principles - see any basic text on fluid mechanics

So basically you could have a setup with 1" ID tubing (and larger) and it would cool really well but you would need a monster of a pump to keep the flow high enough?

[Cathar]

Cathar

Are you reffering to a radiator like this

<snip>


No. That's nothing like what I'm talking about - almost the exact opposite in fact.

If there was a radiator I was more referring to, that would be reasonably small and compact, and offer close to as good as you'll ever get out of a single Papst 4412FGL, then imagine something like a HWLabs Black Ice Pro in thickness (not the Xtreme), but with the core area being something like 15x15 or 16x16cm in size (instead of the Pro's 12x12), with a 15 degree sloped shroud (i.e. ~3" deep for a 16x16cm core area, or ~2" deep for a 15x15cm core area), and the core should be single pass, not double pass.

[SlaterSpeed]

So basically you could have a setup with 1" ID tubing (and larger) and it would cool really well but you would need a monster of a pump to keep the flow high enough?
Nope you wont need a bigger pump flowrates will attualy go up because of the reduced line resisistance. So you could attualy get away with a smaller pump if you have bigger tubing.

Cant get away from the fact that bigger tubes = less resistance = moreflow = smaller pump needed to achive the same flowrate as with smaller tubingB)

[BillA]

Nope you wont need a bigger pump flowrates will attualy go up because of the reduced line resisistance. So you could attualy get away with a smaller pump if you have bigger tubing.

Cant get away from the fact that bigger tubes = less resistance = moreflow = smaller pump needed to achive the same flowrate as with smaller tubingB)precisely
but note the actual change in performance (CPU temp) will be small

[fillip]

When i started reading this thread, my initial thought was "here we go again..."


Thank god it hasn't gone that way.
It's nice to have a thread on such a contentious issue that doesn't disintegrate
into a slagging match. I've read a huge amount both here and at Pro-cooling about water cooling without ever actually trying it myself, so it will be most interesting to see how my kit performs when i get round to buying it.

Not that it means much, but i'm going for a 6mm ID tubing setup. Partly because i love the AC gear, partly because i'm after a totally silent system and the rest because my personal opinion is that the smaller hosing looks better than having huge 1/2" tubing...

Just my 2 pence worth

And thanks guys for an educational debate rather than a load of trash talk.
:)

[Blackeagle]

Nope you wont need a bigger pump flowrates will attualy go up because of the reduced line resisistance. So you could attualy get away with a smaller pump if you have bigger tubing.

Cant get away from the fact that bigger tubes = less resistance = moreflow = smaller pump needed to achive the same flowrate as with smaller tubingB)
As resistence drops flow will increase, I agree with that. But I doubt it will give such gains as to avoid the need of a more powerfull pump that can maintain the volocity in the lines needed to purge the lines of air.

That would seem to me to be the real practical limit on line size, along with space to route the larger lines.

I also have a 2" thick heater core for rad. I use the same one as pH, for the same reason, it's a single pass design that offers much lower pressure drop. It's also large enough for dual 120mm fans.

[SlaterSpeed]

As resistence drops flow will increase, I agree with that. But I doubt it will give such gains as to avoid the need of a more powerfull pump that can maintain the volocity in the lines needed to purge the lines of air.

That would seem to me to be the real practical limit on line size, along with space to route the larger lines.

I also have a 2" thick heater core for rad. I use the same one as pH, for the same reason, it's a single pass design that offers much lower pressure drop. It's also large enough for dual 120mm fans.
Edit for correction: Your right, its just theory. It would relly depend on specific parts. Basicly what i was trying to establish is that small tubing (not nessicaraly low flow however) attauly gives you a disadvantage on the noise issue as with bigger tubing you can use a smaller pump (quieter) and still get the same results as a bigger pump with smaller tubes.

slater..

[BillA]

TILT !

no, that conclusion is not warranted
please read your post
small tubes + small pump = small tubes + big pump

no, not at all

[SlaterSpeed]

Doh! yes your right Bill. Dont know what i was on about there, I just wrote totaly the oppisite to what was in my brain

Edited the post to correct myself :cool: