How big is too big?

[Korndog]

Oh come on.
What about fluid dynamics and turbulance as considered in a closed flow system?
They have NO effect whatsoever?

you can consider turbulance when you're talking about speeds of something like a rocket.
The mechanical energy by the pump gives off a lot more heat then turbulance in the common setup, and even that is ignored. Heat generated by electrical resistance in the pump motor is really the only energy worthy enough to be included. Even as a "closed" system, which it isn't. Now a car engine, u can consider that to be closed system, when engineers design they, they consider that there is NO heat dissapation because so much heat is generated. automotive thermodynamics are insane, and i like to keep away from them, lol.

 

[clocker2]

So turbulance and manging flow is not even a minor consideration when it comes to the design/construction of a waterblock, radiator or pump?
Only size and throughput matter.

Interesting.

 

[Korndog]

So turbulance and manging flow is not even a minor consideration when it comes to the design/construction of a waterblock, radiator or pump?
Only size and throughput matter.

Interesting.

i think i misunderstood you the first time. i thought u mean turbulance causing heat.
turbulance has a great effect on heat transfer. In cathar's whitewater for example, the thin slit causes pressure which accelerates the flow of water going towards the channels, cooling the cpu. Although its less flow, its at a higher velocity.
here i quote myself from an old discussion about flow

i think you're missing a few things. its not about quantity of water.. water can store a lot of heat so even a little bit of flow is good enough, but the problem with it is that its a bad heat conductor of heat. Now with moving water, or whats called convection, the transfer rate of heat improves greatly. The faster and harder water hits your copper block, the faster it conducts. When i mean stronger, i mean pressure. Try blowing on your arm with your mouth wide open, then try it with your lips almost closed like you're whistling. It gets colder and it gets harder to blow.

 

[SureFoot]

clocker2, a thermostat is a regulating device. Thank you for your compliments but i very well know what it's meant for. A car engine runs better when hot. The thermostat, and restricting plates, are used to make it run hotter. The thermostat only fixes a chosen temperature (according to the spring rate, or if it's electrical, to a resistor / program / whatever). Anyway this is borderline offtopic.

fafnir please don't quote pseudo science, and logic exercices. "Proving that bumblebees cant fly" is inherently flawed by the lack of data and the reliance on flaky hypothesis. And offtopic.

Here we discuss heat transfer in a closed watercooling loop. It can only follow the 3 rules of Thermodynamics, and of course heat transfer laws.
Facts:
* turbulence increases heat transfer. The bigger the Reynolds number is, the bigger the surface area per second is presented to the interface.
* velocity increases heat transfer. Think surface area of renewed interface per second.
* cavitation in a conventional watercooling loop won't happen before long. It happens at about 14m of head, of which we are FAR far away.
* scientific laws NEVER contradict reality (quantum field applications nonwithstanding, and even for that we're finding new laws that work..), because the day they contradict reality they are no longer scientific laws... Cars are today designed only with equations. Rockets are sent with incredible precision only thanks to equations. Please don't mix engineering problems with underlying science. The newtonian laws of gravity are shown to be real, and universal (again, quantic / submolecular fields excluded..). Predictions from Einstein, which were mere equations, are regularly confirmed by experience. clocker2, you may have car mechanics experience, but a lack of scientific knowledge makes you build false assumptions about science and how it works.
Back on our tracks: with a constant heat input, a bigger flow in a watercooling loop is always better. But of course the heat input FROM THE PUMP alone will increase more than the gains earned by higher flow / heat transfer.

If my observed results show that higher flow does not necessarily translate to better temps, then no number of charts, reputable experts or science books will convince me otherwise.

And what about real life measurements ? It's not an impossible lab experience, since Bill Adams and Phaestus made it already. Negating the obvious *can* be called stubborness you know...

i would prefer the same performance from a smaller pump...

A pump which offers the same performance with a smaller size is a good thing, provided it produces less heat.

 

[fafnir]

SureFoot, you've just said the thing about five times over as well almost as it could be said, and i don't think you need more scientific proof to swing the jury, they're already convinced and sold on what you've said

its just that i don't think you can convince him no matter what you say or do or no matter how much evidence you bring out

thats what i was getting at, i know that lower flow can't be better, but i just don't think you're going to be able to convince him, thats all

 

[matttheniceguy]

This thread is to long, it should have ended about 30 posts ago, with the conclusion that higher flow rates are better, as long as pump heat is not a factor.

First of, I have seen the law Q=M*C*deltaT a bunch, and although this law is correct, it is used out of context. Is states that heat transfer Q = Mass * heat constant * change in temperature of the mass. This law could be used to measure the heat transfer from the waterblock to the water, but it doesn't tell us anything about how the flow rate relates to the heat transfer. (the delta T refers to the change in temperature of the water)

The thermodynamic law that we should be paying attention to here is this:

Q=h*A*(deltaT) where Q is heat transfer, h is the convection heat transfer coefficent, A is area, and delt T is the temperature difference between the waterblock and water.

So for more heat transfer we want: higher h, higher A, higher delta T.

To get a high A, we use fins, pins, weird shapes, bigger blocks, anything to help.

h, the convection heat transfer coefficent, is related to many factors and is usually found experimantally. It is influenced by the renolds number, a higher reneolds number will mean a higher heat transfer coefficent due to increased turbulence, decreased boundary layer thickness, and some other effects. h is also influenced by the surface conditions of the materials, the Nusselt number, and some other effects. All you really need to know, is that a higher flow rate will result in h being higher. If you don't believe this, read "heat transfer" by J.P. Holman. I know this makes a poor argument for a forum, but I don't think anyone would appriciate me going through several years of university physics here.

Finally, we want a higher delta T. The bigger the difference in temperature between the block and the water, the more heat will be transfered. (this is the whole idea behind chilled water coolers and phase change coolers) When water enters the block, it is warmed up by the block lowering the delta T between the block and the water and reducing the heat transfer. This means we want the water to be in the block for the shortest period possable befor being replaced by fresh water. A higher flow rate will mean a higher delta T over the block, and will raise the heat transfer.

This explanation has been made refering to the waterblock, but is just as true for the radiator.

A HIGHER FLOW RATE WILL INCREASE THE HEAT TRANSFER BETWEEN THE WATERBLOCK AND WATER, AND WILL RESULT IN LOWER CPU TEMPERATURES!!

 

[SureFoot]

Yes, thanx for the correction. Both formulas are telling basically the same thing through, as the 'm' can be seen in a flow as the renewal of 'cooling mass', and C is water specific heat. To put a real thing behind 'm' with a waterflow one needs this 'convection coefficient'. After a quick play with the formulas you can jump from one to another I vaguely remember a physics teacher doing so... even breaking up the 'h' part, with nightmare-ish formulas... well those are bad memories (when i was young i didn't like thermodynamics at all..)

A HIGHER FLOW RATE WILL INCREASE THE HEAT TRANSFER BETWEEN THE WATERBLOCK AND WATER, AND WILL RESULT IN LOWER CPU TEMPERATURES!!

I might add, with a constant pump heat input.

 

[clocker2]

A HIGHER FLOW RATE WILL INCREASE THE HEAT TRANSFER BETWEEN THE WATERBLOCK AND WATER, AND WILL RESULT IN LOWER CPU TEMPERATURES!!

OK.
I concede.
There's no fighting the immutable laws of science.

Especially when written in caps.

 

[Sjaak]

dude, this thread almost started looking like a kiddy fight

once again i will have to refer to my dad he is a science / physics teacher, and he confirmed that a higher flowrate will lead to more heat transfer, regardless of the kind of material.

 

[SureFoot]

clocker2: the fact is it's about *this world* (like in, "not a parallel dimension") and how it's working.

dude, this thread almost started looking like a kiddy fight

once again i will have to refer to my dad he is a science / physics teacher, and he confirmed that a higher flowrate will lead to more heat transfer, regardless of the kind of material.

Of course A close friend who's composite materials engineer (and specialist in fluid dynamics..) confirmed, too. Clocker2 do you want to speak to him live ? I've got his phone number. He's not at all in watercooling (although in the future.. hmm) and doesn't even know about these forums, so i guess he's neutral.

 

[clocker2]

SureFoot,
Thanks for the offer, but really not necessary.

Out of curiousity, what kind of temps does your system run?
I am assuming that is a maxflow setup.
Just want to see what I have to look forward to.

 

[SureFoot]

(offtopic)
According to my OC'ed TBird 1000@1500, with 200MHz FSB (400MHz DDR), i'm running 42°C at peak load, with a 27°C ambient, and a 34°C mobo temp. That's with the PAPST, which is quite flimsy in the airflow department, but strong enough since i cannot overclock any further. I tried another fan, more powerful, and i fall to 37-38°C max load in about the same conditions. Yes my setup is optimized for flow as the CPU loop has very little backpressure (full 1/2", etc). The high sensibility to airflow variations confirms this. I've got a good selection of 120mm fans at home now So i can compare directly....
Note that the TBird is rated between 95W and 105W at these overclocking levels. I'm satisfied with the overclock, and the one and only fan in my system is very quiet (no fan in PSU, no fan nowhere, and HDDs are in silent enclosures)
Flow is excellent through the system, and going to a parallel setup (CPU // GPU+NB) helped a lot. Now it's too much for my small airtrap, as there's so much turbulence the bubbles are recycled...

I'm ordering a XP Mobile 2600+ now, it's rated as cooler (45W !!!) than the TBird, so temps may drop quite a bit in the CPU department, even overvolted and overclocked to the max, but as i'm putting a 6800 as well the GPU will add more heat Might rethink my flow distribution, and give more to the GPU now... Still waiting for my new case too (lian li V2000) so i can put my new monster heatercore (not in my sig yet, as it's not mounted, it's the WA2 one)..
(/offtopic)

 

[clocker2]

Thanks.

 

[gbaz]

Well thanks for all the replys, most ive ever gotten on anything But i do think that my original question is still unanswered.... I know that the higher the flow rate the better, but at what point is it no longer worth it? A 1000gph pump requires over 100w while a 370ghp only needs 18w. Now this does not mean that they put that much heat into the system, but no doubt some will be introduced. Heat transfer is not liener, so going 2X the flow will not result in 2X the heat removed, but where in water cooling a PC is the "best" trade off? Rad size and system restrictions would play a big part, but in general all water cooled setups should act the same. Mabey ill just go cheap grab a AV1300 for $16 and go from there.

 

[Etacovda]

lol, this thread is a f**king giggle.

Surefoot, after reading that you thought your thermostat was electrical 'because there were 2 wires sticking out of it' and INSISTING that it was so, everything you have said makes you looking like a bloody moron. Sure, youre right about flow etc, but dont spout off about things you obviously dont know about.

Clocker, obviously you know this now, but more = better, to a point, that point being watts generated by the pump. Theres not many pumps around that will accelerate the water to the point that friction on the tubing becomes an issue - the ones that are put out ridiculous heat and arent suitable anyway.



WATERCOOLING PUMPS ARE HEAD PRESSURE DEPENDANT. Damn, this needs a sticky, with just that sentence in it.

To the original poster - what waterblock(s), rad(s) etc are you using? If its a single two pass rad (ie BIX, most heatercores) and an average new block, an eheim 1048 (1.4m head, 600lph 15ish watts) will do you about 1.0gpm(ish) -1.5gpm(ish) through your loop, which is enough.

A 1250 (2.0m, 1200lph, 30ish watts) will do around 1.75-2gpm. If you plan on more than one block, then get a 1250-like pump. A single, and a 1048 like pump is good.

My rule of thumb is to look at the eheim pump ratings, and then use wattage as a factor.

Stick above 1.4m head and 600lpm, and you'll be right. If you can get a pump that does about 2m head for under 30watts, get that.

http://www.procooling.com/reviews/html/swiftech_mcw6000-a_review_-_5_.php
http://www.procooling.com/reviews/assets/images/mcw6000_comparo.gif

If you look at that gif above (courtesy of Phaetus @ procooling) you'll see that at around 1.5gpm for most blocks, things start to level out a bit, with diminishing returns.

Its all restriction dependant really. If 1 degree matters to you, get the more powerful pump. If not, get the less powerful one. Just look at the head pressure, and the pressurehead/flowrate chart.

Stay under 35-40w, I reckon

 

[SureFoot]

Surefoot, after reading that you thought your thermostat was electrical 'because there were 2 wires sticking out of it' and INSISTING that it was so, everything you have said makes you looking like a bloody moron. Sure, youre right about flow etc, but dont spout off about things you obviously dont know about.

Please cut off the insults. English is *not* my native language. To me "thermostat" meant that electrical device i got on the side of my engine. Obviously in my language the term differs. No need to call people names.

back on topic: 100W is definitely overkill for a single PC pump. The heat input of that pump will be more than significant compared to what waterblocks and radiators can deliver. If you read those "C/W vs head loss" graphs on 'good) waterblock tests, you want to be in the "knee" of the curve, when efficiency is 90% of the maximum. You'll see you can get it with a reasonable pump. Most pumps under 30W are up to the task for a single PC.

 

[clocker2]

more = better, to a point, that point being watts generated by the pump

Believe me, I have no desire to reenter the mudwallow of the past few days.
The "thermostat" debacle is totally irrelevant to the main discussion and shouldn't reflect on the validity of SureFoot's scientific assertions.
That said, I still have a question.

My original post posited that the coolant needed a given amount of time in the heat exchanger ( radiator) to effectively dissipate it's heat load, thus "more" was only better up to a certain point. Beyond that, "more" was actually worse.
Much discussion of my statement ensued, the upshot being that the Laws of Thermodynamics conclusively prove that I am a fool.

My statement was predicated on the observation that removing the thermostat completely from an automotive cooling system ( which effectively increases flow, as the thermostat is a restrictor) will ultimately lead to higher temps.

I have already conceded that my explanation for this phenomonon violates at least one of the Laws.

Today I visited a local race shop that I occasionally frequent and explained the gist of this thread ("More flow=better heat transfer...ALWAYS!").

There was unhesitant, 100% disagreement from the ( highly talented and very experienced) staff.
Much salty language ensued.
To a man, they echoed my original theory about the length of time in the rad, etc.
I courteously explained that the Laws of Thermodynamics insisted that this was, in fact true.
More salty language.

I admit a less than intimate knowledge of these Laws ( I was urged several times to read them, and in fact may do so), but I'm relatively certain that there is not an exemption written into them "excepting race cars/bikes".
Most of the vehicles in question operate with mechanical water pumps (there are exceptions, but not many), so "pump wattage" is not a factor.

@SureFoot, matttheniceguy and now,Advocate...
Would any of you care to venture an explanation for this seeming contadiction of immutable Law?
If "MORE FLOW=BETTER HEAT TRANSFER...ALWAYS" is true, what are we "car guys"( Fafnir's grouping) observing?

I have already suggested that perhaps other physical laws were in play here, but this was roundly dismissed.

Clearly, we are missing something.

 

[greenman100]

Believe me, I have no desire to reenter the mudwallow of the past few days.
The "thermostat" debacle is totally irrelevant to the main discussion and shouldn't reflect on the validity of SureFoot's scientific assertions.
That said, I still have a question.

My original post posited that the coolant needed a given amount of time in the heat exchanger ( radiator) to effectively dissipate it's heat load, thus "more" was only better up to a certain point. Beyond that, "more" was actually worse.
Much discussion of my statement ensued, the upshot being that the Laws of Thermodynamics conclusively prove that I am a fool.

My statement was predicated on the observation that removing the thermostat completely from an automotive cooling system ( which effectively increases flow, as the thermostat is a restrictor) will ultimately lead to higher temps.

I have already conceded that my explanation for this phenomonon violates at least one of the Laws.

Today I visited a local race shop that I occasionally frequent and explained the gist of this thread ("More flow=better heat transfer...ALWAYS!").

There was unhesitant, 100% disagreement from the ( highly talented and very experienced) staff.
Much salty language ensued.
To a man, they echoed my original theory about the length of time in the rad, etc.
I courteously explained that the Laws of Thermodynamics insisted that this was, in fact true.
More salty language.

I admit a less than intimate knowledge of these Laws ( I was urged several times to read them, and in fact may do so), but I'm relatively certain that there is not an exemption written into them "excepting race cars/bikes".
Most of the vehicles in question operate with mechanical water pumps (there are exceptions, but not many), so "pump wattage" is not a factor.

@SureFoot, matttheniceguy and now,Advocate...
Would any of you care to venture an explanation for this seeming contadiction of immutable Law?
If "MORE FLOW=BETTER HEAT TRANSFER...ALWAYS" is true, what are we "car guys"( Fafnir's grouping) observing?

I have already suggested that perhaps other physical laws were in play here, but this was roundly dismissed.

Clearly, we are missing something.

Tell me, then, how would one calculate this magical point at which heat transfer becomes worse?

 

[sandman001]

Damn, it's just a computer. And I thought I went overboard.

 

[Attak82]

One of you is going to put one of these 1000+ gph pumps in your computer and it will burst and there will be water all over your rig and I will luagh