your thoughts on "water cooling thoughts"

[WAJ_UK]

I've just read this http://www.overclockers.com/articles974/ and wasn't really sure what to make of it especially the following comments:

There are a couple of really silly ideas floating around out there about increasing the performance of a PC water cooling system. One has to do with so called "Swirl Technology", putting baffles or intricate channels in water blocks to increase their surface area.

I think this is a bit confusing. I've made a high swirl block it worked quite well but I'm not sure how you would encourage the water to swirl using baffles or intricate channels, any ideas?

Another bright idea is to use oversized hoses and fittings. This is supposed to somehow increase efficiency. Not so. One of the rules of fluid dynamics states that for any given diameter of pipe or hose, only a finite volume of fluid will pass through it at a fixed rate.

I really didn't understand this bit. It seems as though he is saying that larger hoses have no effect on flow and then he tells us hoses of the same diameter give us the same flow. Hmmmmmmm interesting.

Also keep in mind that water blocks are "timed". Depending on the surface area of the water block, a specific volume of fluid must pass through it at a specific rate for optimum heat transfer. The water block itself can only accommodate a set volume, it has a built in restriction - so don't waste your time. Use the correct hose diameter intended for your system.

all I can do is

Luckily the Northbridge chip set had a heat sink on it. I took it off and used a couple of plastic ties to hold the water block to the heat sink hard points already there

all I can do is cable ties For a proper engineering solution I definitely would have used sticky tape

[pHaestus]

Another bright idea is to use oversized hoses and fittings. This is supposed to somehow increase efficiency. Not so. One of the rules of fluid dynamics states that for any given diameter of pipe or hose, only a finite volume of fluid will pass through it at a fixed rate.

This is true (given a constant temperature) but not relevant to our system. When using a centrifugal pump in a cooling loop, the flow rate is not fixed but is instead a function of the total system resistance. When you decrease resistance (say by drilling out your hose barbs or running 5/8" barbs on the rad/pump/etc) then you will increase the flow rate. This is also a "law of fluid dynamics"...Darcy's Law.

Also keep in mind that water blocks are "timed". Depending on the surface area of the water block, a specific volume of fluid must pass through it at a specific rate for optimum heat transfer. The water block itself can only accommodate a set volume, it has a built in restriction - so don't waste your time. Use the correct hose diameter intended for your system.

I believe that this person completely misses the point. By removing restrictions elsewhere in the loop, one can achieve increased water velocity where it is beneficial (heat transfer is increased by turbulence which increases with water velocity).

I haven't read the article you are quoting but it sounds as if the person should have sought more relevant facts before penning thoughts to paper.

[pauldenton]

Originally Posted by WAJ_UK

I've just read this [url]
all I can do is cable ties For a proper engineering solution I definitely would have used sticky tape

actually cable ties are a lot better than you give them credit for - try opening a closed one

[Groth]

That's the sort of article that shows up on kiddie boards at the time - a writer with more money than patience, more ego than knowledge. Slap together a system with the "best name brand products you can afford' and call yourself an expert.

Is Overclockers.com's article submission standard slipping?

[pHaestus]

see Groth this is why we come across as assholes all the time

[krazy]

in reply to pauldenton:
I'll hand it to you that they "hold" pretty well, but they can't be adjusted the same way a machine screw and nut can. I would be afraid of getting one zip tie too tight and having the far side of the block not make good contact with the chip underneath.

This person definitely needs to brush up on their watercooling theory. The author seems to know just enough about fluid dynamics to be dangerous and want to criticise others' methods (that tend to work quite well).

edit:

from the article
Most chip set water blocks have a bracket which is held in place by screws.

This requires drilling holes in the motherboard - very scary. I had the drill in hand, but after looking down at all those little pathways printed on the board I just couldn't do it.

Has anyone ever seen or heard of a block manufacturer recommending drilling holes in your motherboard to mount the block?? I'm really wondering where this guy got this idea from. It should go without saying that you simply can't drill holes in motherboards. Even if there was a gap in the traces, how do you know the other side is clear? How do you make this stuff up?

[bigben2k]

Lies! Lies, and more lies, I tell you!


Any section of the loop going from 3/8" to 1/2" will improve the flow rate, albeit not in a significantly measurable way. Most blocks perform better at higher flow rates, until you hit that point of diminishing returns, but unless one is running an outrageously expensive pump, "you ain't there yet!", and if you do have one of those outrageously expensive pumps, then you probably know what you're doing anyways...

As I've stated before, 1.5 gpm is the limit at which flow rate caused a significant restriction in the tubing, at 3/8" ID. Most people achieve 1.0 gpm so either way: IT DOESN'T MATTER! but if you get up to 1.5 gpm, then be serious about your tubing, and use 1/2" ID. Besides, 1/2" fittings are so much easier to find at the local hardware store, than 3/8".

Ok, I've said my peace...

[Wildfrogman]

Heh, it seems somebody else saw the

" Most chip set water blocks have a bracket which is held in place by screws.

This requires drilling holes in the motherboard - very scary. I had the drill in hand, but after looking down at all those little pathways printed on the board I just couldn't do it."

Actually most chipset waterblocks are mounted with oh hmm bolts?...not screws. What actually is "scary" is somebody holding a power drill thinking should I or shouldnt I put holes in the motherboard..hmm where do I drill again?

[jlrii]

Some body ought to convince him to prove his point about tubing size. I propose a simple demonstration: He can try to breath through a 3ft piece of 1/4 od tubing for an hour while I get a 3' piece of 3/4 hose....Who'll still be standing.....(strain, strain, puff, gasp, thud) LOL

[Groth]

Originally Posted by pHaestus

see Groth this is why we come across as assholes all the time

I'm not too worried about seeming to be an asshole to a fellow who thinks a watercooling rig will cost "closer to five-hundred bucks" (my first setup cost <$40) or thinks that if the components aren't in the computer case "all the hoses and mechanicals are exposed" (external box?). And with jaydee mellowing, someone has to pick up the slack.

M. C. Misiolek's flow concepts remind me of a fellow I knew, whose fluid dynamics knowledge was limited to a vague recollection of an illustration of Bernoulli's equation from high school physics. You know the one, featuring an ideal fluid flowing through a pipe with a constriction. Since the flow rate is the same in the constricted and unconstricted portions, he was convinced that pipe/tubing size never mattered.

I tried to explain viscosity by challenging him to a milkshake drinking contest. I had a big straw, he had a tiny coffee straw. Eventually I gave up, and helped him plumb his new house with 3/8" supply lines. To this day, he can't figure out why it takes so long to fill the bathtub.

"A water cooling system will always follow the rules of fluid dynamics" whether you understand them or not.

[prandtl]

Originally Posted by bigben2k

Lies! Lies, and more lies, I tell you!
As I've stated before, 1.5 gpm is the limit at which flow rate caused a significant restriction in the tubing, at 3/8" ID. Most people achieve 1.0 gpm so either way: IT DOESN'T MATTER! but if you get up to 1.5 gpm, then be serious about your tubing, and use 1/2" ID. Besides, 1/2" fittings are so much easier to find at the local hardware store, than 3/8".

Ok, I've said my peace...

what exactly made you choose 1.5gpm as the "limit"? Do you have a link to the thread where you stated it first, wich, i suppose, also show the reasons behind the 1.5gpm...

[BillA]

[quote=Groth]. . . . . Eventually I gave up, and helped him plumb his new house with 3/8" supply lines. To this day, he can't figure out why it takes so long to fill the bathtub.
. . . . QUOTE]
have no idea if this is true, but it sure is funny
Thanks Groth

prandtl
Ben has a calibrated reticle in his eye, one glance and he knows all
dynamic measurements are a bit harder, he has to jiggle his eye at the same speed to freeze the frame
- practice makes perfect

[HammerSandwich]

There is only one possible explanation for this article: Overclocker's calendar is wrong, and they think it's April 1.

[krazy]

Originally Posted by Groth

"A water cooling system will always follow the rules of fluid dynamics" whether you understand them or not.

so true...

That guy really plumbed his house with 3/8"?? hahahahAHhaHAhha wow.

And you're right, HammerSandwich, It appears to be March Fools Day... or something. I'm wondering how that article made it's way up there. Don't they read through submissions at all? Anyone worth anything in this business should have caught the problems in that article if they did more than just glance at it. Not to say anything bad about whoever pusts up the articles on OC, but that one really must have slipped through the cracks.

[Jabo]

I loved this article, truly insightful piece of tekkie jurnalism muahahahah.
Good Heavens and I thought that overclockers was a respecteable site and now this....
but I loved the part with drilling MoBo, wow, almost fell from my chair....
It is a textbook example of a person using words he does not understand at all and writing about matters he's clueless about. Luckily it was short read

I bet M.C. Misiolek has his plumbed with 3/8" pipes LOL!

EDIT: I noticed title of this article and frankly he should refrain from thinking any more, it HURTS, if not him us at the very least

[Groth]

Yep. 3/8" He's always been tight with a buck, and it saved a lot of copper.

Latest thing with the same guy: sometimes the light in the bathroom flickers but mostly won't turn on at all. I jabbed the probes of my trusty multimeter into the non-GFCI socket by the sink and discovered that both hot and neutral were live. Been about two weeks, still waitin' to see he gets it fixed before the house burns down.

[BillA]

gotta give natural selection a chance

[SlaterSpeed]

Reminds me of a cirtain Apple G5 artical

[jlrii]

was waiting for him to suggest using a coule of self tapping screws to attach the block directly to th' cpu....hey I dont see any traces....

[pdf27]

Couple of other things I noticed in there...

The Southbridge chip set was a different matter - it didn't have a heat sink. I'm not sure if it even needs to be cooled, but I put a water block on it anyway.

Adding a water block to something that didn't even need a heat sink according to the manufacturer's calculations - is he trying to get the most expensive system he can? :shrug:

I'd like to try Prometia cooler unit, but hooked up to the reservoir, not the processor. Chill the fluid running through the whole system down to say, thirty-nine degrees, maybe?

Hmmm... cooler unit designed to chill down a small area to below freezing - he only uses it to cool the res before moving on to the CPU. Not only is this as thermodynamically inefficient as he can reasonably get it, but it'll probably turn a large part of his res into a solid block of ice


The fun bit is actually trying to spot correct statements in that post.... The truest one I found was this one

Some people have no idea what they're talking about: others are just plain old "liars"

[krazy]

Originally Posted by pdf27

Adding a water block to something that didn't even need a heat sink according to the manufacturer's calculations - is he trying to get the most expensive system he can? :shrug:

No wonder he quotes $500 for a good WC system...

[HammerSandwich]

Originally Posted by pdf27

Adding a water block to something that didn't even need a heat sink according to the manufacturer's calculations - is he trying to get the most expensive system he can? :shrug:

Nope - just the lowest-flow system.

[UberBlue]

Anybody have an idea what M. C. Misiolek handle in the OC forums is?

I've a few ideas but no way to confirm.


EDIT: Ran his Email addy through google and came up with this. http://www.newenglandfilm.com/guide/listing.nef?id=3745

[Cathar]

Originally Posted by UberBlue

EDIT: Ran his Email addy through google and came up with this. http://www.newenglandfilm.com/guide/listing.nef?id=3745

LOL

Years in Business: 10-20 years

I don't know about you, but a vague range of 10-20 seems pretty non-determinant to me.

"How much for this CD"
"$10-20. How would you like to pay?"

[pauldenton]

Originally Posted by Cathar

LOL



I don't know about you, but a vague range of 10-20 seems pretty non-determinant to me.

"How much for this CD"
"$10-20. How would you like to pay?"

maybe he's been in the (film) business 20 years, but for the first 10 he didn't sell anything.........