Problems with Tom’s water cooling?

[Lowlypawn]

Problems with Tom’s water cooling?

Form overclockers.com


Toms Hardware Discovers Watercooling! In a stunning example of cyber-archeology, Tom's posse unearthed the Senfu radiator, waterblock and EHEIM 1046 (my favorite) waterpump and built "our own home-grown water cooling system."<font color=red> BTW Tom: The Senfu waterblock is aluminum and the radiator is copper - you forgot to mention the galvanic corrosion problem - coloring the water red doesn't help.</font color=red>



Is this a major problem?


Thx & Cya









<font color=green>I may go to <font color=red>hell</font color=red> but at least I won't get lonely</font color=green>

 

[paulcalmond]

I thought that for 'galvanic corrosion' to occur there had to be contact between two different metals. Is this not the case ?

 

[Guest]

That is indeed correct on galvanic corrosion. here, The contact point is the water itself; it will conduct very very poorly but it will conduct. Deionized water will help, but not cure the problem. Red dye actually makes it worse, ANYTHING dissolved in the water will make it worse.

The good thing here is it takes a _long_ time, on the order of months. Any Hardware nut will probably play with his/her setup so much that they will notice it plenty early enough; I know I do. It is a very real concern tho. Just ask Dan at Dansdata.com, he has a very cool article explaining the problem indepth.

Monaco

 

[Lowlypawn]

I found the article <A HREF="http://www.dansdata.com/burning.htm" target="_new">http://www.dansdata.com/burning.htm</A>.

For those to lazy to click and read here is the good part.
<font color=green>
“Why did the aluminium leave?
Galvanic corrosion, that's why.
Take dissimilar metals, connect them electrically and put an electrolyte between them, and you make a galvanic cell. The more "anodic" (or "less noble") metal will be the negative terminal of the cell thus made, and will oxidise. You get electricity, but you slowly destroy the anode.
Aluminium is highly anodic, and the pipes in the Senfu cooling system's radiator are made of copper, which is a lot less anodic than aluminium. Tapwater's not a good electrolyte - the cleaner your tapwater is, the less conductive it'll be, generally speaking - but it's conductive enough for galvanic corrosion to happen. And for, over months, the aluminium to be eaten away and create lots and lots of nasty precipitate in the process.
Without an electrical connection between the radiator and the water block, you wouldn't expect galvanic corrosion to be an issue. No circuit, no ion transfer, no problem.
But it was.
Thanks to a kindly reader, I now know why.
The copper corrodes a little, as copper in water or air always will. This is not normally a big deal, because a thin copper oxide layer forms and protects the rest of the metal. But in the process, some copper ions go into solution and make their way around, thanks to the pump, to the aluminium water block. They precipitate onto the water block surface as teeny little metallic copper particles.
And then plain old galvanic corrosion can happen, as the copper and aluminium are in physical, and thus electrical, contact, and both immersed in the electrolyte. The dreck you end up with is mainly aluminium oxide, with its greenish colour donated by a light lacing of copper oxide. Tah-dah.
When I flame-tested some of the precipitate, by the way, this was borne out; green flame from the copper, bright sparks from the aluminium. Case closed.
Incidentally, when I was trying to track down information on this subject, I discovered that if you start searching for "galvanic corrosion" in the company of some other computer-cooling-relevant words, you're likely to find a large number of pages belonging to homebrew enthusiasts, and a smaller number of pages belonging to nuclear reactor technicians.
Neither of these groups of people want galvanic corrosion to happen in their equipment, but judging by the degree of unhappiness expressed by those who have had equipment destroyed by it, it would appear to be much more irritating when it ruins a batch of beer than when it results in radioactive liquid sodium spraying all over a control room.
But I, again, digress.
The way to avoid the galvanic corrosion problem is by using non-electrolytic coolant, or by just making sure all of the metal bits in your cooling rig are made of the same metal.”
</font color=green>

Thx & Cya


<font color=green>I may go to <font color=red>hell</font color=red> but at least I won't get lonely</font color=green>

 

[Guest]

I was just wondering about something while flicking over the pages describing the water-coolers. Would be possible to substitute the water in the cooling system for another coolant such as ethylene glycol without causing any/much problems?

If any of you are wondering, ethylene glycol is often used as a coolant in car radiators.

 

[SerArthurDayne]

Heh, I practically winced when they suggested the use of tap water. Not to mention the dissimilar metals in the system : /

I plan to go copper all the way for my water cooled rig.

LOL, when I saw the article name, I thought TH had actually MADE their own equipment and was expecting a really sweet article... but senfu? /me twitches

 

[Guest]

Just wondering, what happends (might happen) if a major leak occurs, like if one of the hoses break and and u get water in the power suply.
240 V and water doesn't get along very well ?

 

[SerArthurDayne]

If you have water spraying out of a hose all over the inside of your system, you can say goodbye to a LOT of things in that box. That is why you want to test your system for not just a measly couple hours, but more like a week. I for one am not going to stick my water rig into my box until its run leak free for two weeks straight - I'm that paranoid.

Oh yeah, and I meant to post this link about corrosion in water cooling systems -
http://www.data-detective.com/overclock/corrosion_.htm

This water wetter product seems pretty cool, anyone else know of any products like this? If I can mix metals in the system, I could probably afford to add an aluminum cooler for vid card and chipset too... but really I wanna avoid the corrosion issue at all costs.

 

[Guest]

LOL, when I saw the article name, I thought TH had actually MADE their own equipment and was expecting a really sweet article... but senfu? /me twitches

After seeing the Part 1 through 3, I'm not surprised... and for a moment I thought they really did make the blocks and all. Some of the people I know really do. Senfu... Oh, dear..

This is from the article:

physicists, please forgive this mediocre formulation!

At least the writer was a bit honest on how far his knowledge goes this time, I guess. But that doesn't stop him from making the mistake of stating resulting temperatures without details. So the water cooler got the CPU temp down to 24C... but with the water temp at how warm?

 

[Guest]

I also missed some temp readings, and a comparrison to one of the better aircoolers.

---
Engage!

 

[wapaaga]

that is why you should proablly use copper tubbing becuase it has a much lower chance of kinking or breaking

 

[Kalzone]

Copper tubing would have to be prefitted around any cards, memory bank configurations, etc. Not only would it cost more, but you'd have to have a specifically molded pieces to fit inside the box. Not much room for customizing your configuration later on.

My question: When I first read the article, along with other articles in water cooling, I thought "Why not use distilled water?". Wouldn't that have the lowest mineral count - if any - and wouldn't it therefore be less conducive to the corrosion process.? Then the thought of equilibrium and semi-permeous(sp?) membranes came to mind. Would the <i>lack</i> of minerals in the water cause it to leach minerals from the surrounding metals to compensate for its lack of saturation? Maybe this second question is way off base, but it came to mind. Now if someone that specifically deals with these processes could comment on this. (My limited exposure only deals with my environmental engineering background.)

 

[ejsmith2]

If you have a propane torch and a tube bender, it's actually pretty easy to work the tubing around your case. I run mine off the hot side, right along the case panel
You're right on the deionized water though. It's got a dielectric coefficient close to 80. But the big problem is keeping it deionized. Heat will drop that coefficient to 40 within week, if you're not filtering. Copper is soluable in water, to a small degree, and will put a few Cu (2+) atoms in there. DI water sounds like a really good solution, but in the real world, it just doesn't quite work that way.

Which is why people switch to mineral oils.

 

[Guest]

I just don't get this...
i wanna get into water coolers but it's just a bit much atm.... one thing i can't understand is why we use WATER at all, why not something like Freon (fridges) or something that actually physically COOLS instead of just, warms up and cools down... it just confuses me... anybody wanna elaborate?

<b><font color=blue>Who needs drugs when you got <font color=black><i>BasS?</i></font color=black></font color=blue></b>

 

[Guest]

>why not something like Freon (fridges) or something that
>actually physically COOLS instead of just, warms up and
>cools down... it just confuses me... anybody wanna
>elaborate?

People do use freon. Take a look at vapochill, etc.
The issue is that freon coolers use a phase change to remove heat from the cooled object. Basically, liquid freon is pumped to the object to be cooled, where it vaporizes, absorbing lots of heat in the process. But now you have to deal with that vapor with a compressor & condensor. Works great, but slightly more complicated then an aquarium pump and a bucket



In theory, there is no difference between theory and practice.
In practice, there is.

 

[wapaaga]

another problem is that you have to have alot of experonce with that stuff to make a good system and also to make it safe

plus it cost more

 

[Lowlypawn]

I wonder if anybody at THG reads this and if so I would like their comments on this.

Maybe on there next water cooling article they will address these problems with their 1st set up.

Thx & cya


<font color=green>I may go to <font color=red>hell</font color=red> but at least I won't get lonely</font color=green>

 

[Guest]

yeah very good points guys..

I do know the process of the freon means of cooling but i don't really know why it has to be so hard. I mean if you put in the time and effort it'd be easy...

anyway enough of that,
i guess you do have to be sorta experienced so i think, if i wanna get into some serious cooling i might buy myself a vapochill. I know they're like a grand in australia so... eh maybe when i get a job...

<b><font color=blue>Who needs drugs when you got <font color=black><i>BasS?</i></font color=black></font color=blue></b>

 

[Guest]

> why do we use WATER
Any serious cooling (freon, peltier) will soon bring the temperature below the point where condensation occurs, so extra measures must be taken to ensure safe operation of such a system. Watercooling setups such as Senfu's can by their very nature never reach the condensation point. I feel this is an advantage.

 

[Guest]

I have been told that condensation issues can be resolved fairly easily.

I would just once like to see my CPU temp reading be at 0

that would just be the highpoint of my computer's life

<b><font color=blue>Who needs drugs when you got <font color=black><i>BasS?</i></font color=black></font color=blue></b>

 

[peteb]

you'd be lucky to get that with a vapochill and a high end Athlon. You can keep PIIIs below 0, but it is really tough with Athlons.

-* This Space For Rent *-
email for application details

 

[Guest]

There are 2 ways of correcting this problem without searching for all copper components.

Use distilled water - it's not conductive (the minerals in water make it conductive)
Use something other than water, like a light oil such as automatic transmission fluid (about a 10W on the viscosity scale where water is somewhere near 8W) or what I would use if I go with liquid cooling is a product called Kooler Klean by International Lubericants (http://www.lubegard.com/automotive/kk.html). It's not conductive, it's a pretty blue and it's an dielectric protectant. And you can pick it up at Napa or most transmission parts stores.

Also by using a non-conductive liquid you aford yourself more security in case of leaks. There might be a mess but you can clean that up with a towel. Fry your system on a Saturday night and in most cases you mope all day Sunday because the stores won't be open until Monday.

 

[Guest]

I'm assuming you are talking about the portable refrigeration coolers that run off the cigarette lighter of your car. Coleman makes the same thing here in the US.

http://www.coleman.com/coleman/ColemanCom/prod_Detail.asp?Product_id=5615-807&CategoryID=8570&SourcePage=PROD_CATEGORY_MAIN.ASP&SubmittingPage=Prod_subcategory.asp

There is a 110 volt adapter if you are already tapped out on your power supply. Cools 40° below ambient temp.

Or if for some reason you need to heat up the unit (moved to Antarctica, AC freeked out, ect) you can reverse the polarity and heat the unit up to 100°!

 

[Guest]

One other thing I was comptemplating. A means of indicating pump failure. 2 methods I can see are pressure switch and pump voltage.

By monitoring the pump as though it were a fan you could rig up a connector to the MoBo and use the CPU fan speed warning in the BIOS.

Using a pressure switch would allow you to hook an alarm up in case of failure.

Both have their pros and cons; CPU fan method could be used to perform an automated shutdown, but actual flow failure wouldn't be reconized.

Pressure switch method requires someone to be there to shut down.

Just some random thoughts.

 

[killall]

a copper block will also be a lot more efficient wouldnt it... oooh lets try silver... are you considering the fact that aluminium is already coater in an oxide layer?

you do not strengthen the weak by weakening the strong