Liquid Cooling Myths

[zanzabar]

d5 is more reliable or stronger than the ddc with top

[DarthBeavis]

tubing size has importance over 3/8ths ID. <--- myth

I wish my wife did not believe that . . .

[OC Maximus]

I wish my wife did not believe that . . .

your on a roll today.

[Waterlogged]

I've been doing some reading, going back through info published in the last 10 years, and got to thinking about the "Myths" or common misconceptions people have about water cooling. I thought it would be interesting to see what the group here thinks are the most common myths/misunderstandings still prevalent today.

I'm thinking about "principles" here, not specific products, so please don't mention brand or vendor names. Thanks - it will be interesting to see what the group thinks.

Could be a little difficult for some of the myths seeing the "myths" are actually about what the specific product is like and what it can do.

Oh, and I'm not talking about any non conductive fluids..

[mav2000]

Non Conductive liquids....enuf said

[OC Maximus]

lol

[voigts]

I've got a few myths, some already mentioned, that I'll summarize in one shot:

1. Thermaltake = watercooling
2. non-conductive liquids will make sure that you don't short anything out if you have a leak
3. a larger reservoir will result in better temps (we are not talking about a 50 gallon drum buried in the ground here)
4. gpm on a pump is what we should look for, hence any pond pump will work
5. the more money you throw at a watercooling loop, the better it will be.
6. water needs to go slower in a loop to have more time in the radiators to shed heat
7. there is some magic clear tubing (that uses plasticizers) that will forever stay clear
8. the bigger the pump, the better the loop performance
9. the more fancy Bitspower fittings you can cram into a loop, the better it is
10. the more expensive your watercooling loop is, the better your potential overclocks will be
11. Darthbeavis ever had a hope of being the kind of guy who could pick up chicks in the first place

[CrazyNutz]

You can water cool your wife down after she finds out how much money you spend on water cooling parts.

[Church]

Myth that LC is omnipotent & all-powerful while ignoring need for adequate rad size and expecting uber performance (common to factory made kits/sets with 1x120 rad).

Shazza: can you put myths scattered in many thread posts in OP for them to be easily readable and to lessen repeated cases?

[Martinm210]

This one comes up now and then:

Pump before CPU block because high pressure improves thermal performance of the CPU block.

We know that pressure changes throughout the loop and flow rate remains constant, so this would imply that high pressure would somehow change the thermal properties of water at pressures as low as 10PSI. Unfortunately water is for all practical purposes particularly at the low pressure ranges considered in-compressible. So if you can't compress the water and increase it's density, you can't improve water's thermal properties with pump pressure. I suppose if you ran a system full of air bubbles it would have some impact, but otherwise...myth. As far as loop order goes, if you want to split fractions of a degree, the best place to put the CPU block is immediately after the radiator where the water is it's coldest. For practical purposes, it really doesn't matter because most systems will see less than a degree variance in water temperature throughout the loop.

[shazza]

[QUOTE=Waterlogged;4967248]
Could be a little difficult for some of the myths seeing the "myths" are actually about what the specific product is like and what it can do.

Oh, and I'm not talking about any non conductive fluids.. [/QUOTE


I get what you're saying, Waterlogged. I just wanted to avoid getting into the whole negative vendor bashing regime we've seen so much of. I've deleted one such post, left a couple others ...

@ Churchy - I've mentioned before that it would be useful to update the Info Guides, maybe do a FAQ as well - which would include debunking of Myths. I just need to find the energy (and fortitude) to get it done.

Thanks Voigts and Martin210. Voigts - I understand what you're saying about Thermaltake, but it does seem they are trying to make some new products which may be better. No experience with their stuff, so can't say for sure.

[SoMBrA]

If you view Bundymania's watercooling parts pictures before going to sleep, you will pee on your bed :>

+1 to what Voigts said!

@Shazza great post idea, deserves a sticky with Q&A format... will help a lot noobies like me

[Levesque]

Full cover blocks for GPUs are better then universal GPUs waterblocks + RAM heatsinks.

[NaeKuh]

Myth that LC is omnipotent & all-powerful while ignoring need for adequate rad size and expecting uber performance (common to factory made kits/sets with 1x120 rad).

this changes as you introduce this tho:

This one comes up now and then:

Pump before CPU block because high pressure improves thermal performance of the CPU block.

We know that pressure changes throughout the loop and flow rate remains constant, so this would imply that high pressure would somehow change the thermal properties of water at pressures as low as 10PSI. Unfortunately water is for all practical purposes particularly at the low pressure ranges considered in-compressible. So if you can't compress the water and increase it's density, you can't improve water's thermal properties with pump pressure. I suppose if you ran a system full of air bubbles it would have some impact, but otherwise...myth. As far as loop order goes, if you want to split fractions of a degree, the best place to put the CPU block is immediately after the radiator where the water is it's coldest. For practical purposes, it really doesn't matter because most systems will see less than a degree variance in water temperature throughout the loop.

lulz.. answered this already:

CPU b4 GPU... can be possible.. depends on how many gpu's u have, as having 2 can increase the water temp b4 it hits cpu while in loop... your playing with holding capacity of H2O here... Anytime u go over 350W of heat collected... your going to be playing with holding capacity.

Full cover blocks for GPUs are better then universal GPUs waterblocks + RAM heatsinks.

^ this is not a myth.... u can only get water channels @ VRM's with FC blocks.
There is no VRM block yet, unless i missed something.

Watercooling VRM's is very benificial, and more so since we saw how fast they can blow out after skinnee pop'd one on his bed during his GTX's tests.

[zeropluszero]

I've deleted one such post, left a couple others ...

killjoy.

[warriorpoet]

How about:

Water cooling can never be quieter than air? I have a system here that kills that myth dead.

Water cooling can not be economical. While it is certainly more expensive than stock air cooling, it can be, when longevity of a carefully planned and constructed loop is taken into account, be nearly equal to top-performance air-coolers.

When considering cooling alternatives, Watercooling's risk of leakage and potential total-system destruction must be taken into account. Seriously. This one's still alive.

[Water Hippo]

Cooper can not look as nice as nickel plated copper. :P I noticed over the years a few people say how ugly copper is, but I must admit I find the raw material alluring, I can't be the only one. (I know this sounds wrong but it was the best way I could put it. )

[NaeKuh]

How about:

Water cooling can never be quieter than air? I have a system here that kills that myth dead.

And if i was to move the pump and radiator into another run.
Have the tubing extend from room A into computer B, and move the heat out from the PC into the Other room.

It would be on PAR with a passive system, yet hold better cooling pontential because H2O has almost 10x the thermal conductivity as AIR does.


The key thing in anything science is to work with the numbers and then apply them in real life.
And if the numbers work out... then the theory is a working theory.

But dont get me wrong... because numbers work out, does not mean it will work out in Real Life.
I and a few other pioneers have learned this first hand... sometimes the numbers look TOO GOOD to be true.

[Raul-7]

How about:

Water cooling can never be quieter than air? I have a system here that kills that myth dead.

Me too, I just moved the radiators two-floors down into my garden. Gets me super cool temperatures at night as well.

[bds71]

a couple i have seen (and might be guilty of thinking myself):

1. temps in the loop change from component to component (20C drop after rad, 20C increase after CPU, etc) <---- EPIC FAIL loop maintains constant temp within .5C throughout loop

2. **all** prefab WC units/kits suck <---- no so much: some do, but others are actually pretty decent performance per dollar (especially the kits that come with double/triple rad)....and once the prefab units start coming with at least a double radiator they might be pretty good as well!!

3. i've heard the one about the bigger res makes better temps, but it's an older myth <---- FAIL it might take a little longer to max out water temps, but they will still reach the same temp as with a smaller loop - this is only true if you don't let the loop reach max temps (take reading too early)

[NaeKuh]

a couple i have seen (and might be guilty of thinking myself):

1. temps in the loop change from component to component (20C drop after rad, 20C increase after CPU, etc) <---- EPIC FAIL loop maintains constant temp within .5C throughout loop

3. i've heard the one about the bigger res makes better temps, but it's an older myth <---- FAIL it might take a little longer to max out water temps, but they will still reach the same temp as with a smaller loop - this is only true if you don't let the loop reach max temps (take reading too early)

no its not constant... its thermodynamics and you got it wrong also.

The value your mentioning in A is the holding potential of H2O.
What that means is the value in which it takes X amount of heat @ Y galons / minute, to make it go up 1C.
This value is noted at somewhere around 300-350W @ 1-1.5gpm flow.
As you decrease flow, you decrease the holding potiental of water and vice versa.

The reason why we say .75gpm - 1.5gpm is the sweet spot is due to the holding potential number. Rougly 275W-350W b4 water goes up 1C.
And yes its been shown many times that the difference between a .5gpm system vs a .75-1.5gpm can be quite substantial.

Its a science of fine tuning part A to carry just enough for Part B, so Part C can have just the right amount of both.
This is what we call efficiency.


Bigger res translates into longer equalibirum time...
Meaning it will take more to see bigger change.
In a closed loop system, yes, it wont matter much at the finish line.

[Marci]

O_o

Heh - Things never change really, do they...
What can be said has been said. Read the archives. Laws of Physics don't change. The original 'myths' article (http://www.overclockers.com/watercooling-myths-exposed/), and all the emp data here, on OCAU and at ProC always did cover it ALL for those who took the time to read.

[wondergod]

Putting more than 2 rice size drops of thermal paste is the reason for your temp that is 10-20 degrees higher than it should.

I see this all the time when someone is like my CPU is 10-20 degrees higher than it should be, first thing someone says is too much thermal paste. If you mount the cpu block correctly it should squeeze out the excess anyways.

[bds71]

temps in the loop change from component to component (20C drop after rad, 20C increase after CPU, etc) <---- EPIC FAIL loop maintains constant temp within .5C throughout loop

no its not constant... its thermodynamics and you got it wrong also.

i didn't mean the overall loop temp doesn't fluctuate - of course it can increase upwards of 20C under heavy load. i simply meant that **within** the loop (ie: between components) the temps are within .5C. so there **might** be a slight temp difference between the radiator outlet/inlet, but it won't be more than about .5C ---> not a 5-20C change (from rad-in to rad-out) as the MYTH states.

[Computurd]

OMG...
This ONLY applies to open loop my friend.

In a closed loop, you got 0 net movement.. because what is pulling up is being canceled by whats going down so you initially got 0 movement. The Kinetic Value in the system is negated by the potential value.
Once heat is added, you got convection (adding energy)... warm water wanting to go to the highest point while cold dropping.

explain please?
your saying that there is no movement?