Question for water chillers

[000]

I'm curious as to why my water/antifreeze solution is so much colder than my cpu is, and also why when my water/antifreeze solution gets colder my cpu doesn't. For example I started up my cooler and let my water get down to -14c and then started my computer up. In windows at idle my processor was at about -5c (fluxuates alot for some reason).

After a few hours my water temperature had gotten down to -20c but my cpu temperature was still about the same (-5c fluxuating between -2c & -7c)

Anyhow I was curious as to why there's such a big difference between my water temperature and my cpu temperature.

For those who want to ponder I'll give a description of my system:

dehumidifier with evaporator submersed in 4 gallon plastic ice chest full of water/antifreeze.

Plastic Ice chest contained in a 10 gallon styrofoam cooler with about 2 inches of spray polyurethane foam in between the 2 coolers.

700 gph danner pump. Inlet side left bare, outlet side connected to 1/2" tygon 3603 tubing insulated with 1/2" rubatex tube insulation.

5 ft of tubing to a swiftech mcw462-uh waterblock which has been insulated with rubatex insulation tape completely to the motherboard (the tape comes off the water block and seals to the motherboard).

5 ft of tubing from the waterblock back to the reservoir.

The pump is on one side of my evaporator and the outlet tube from the waterblock is on the other side of the evaporator in the opposite corner.

I think this is pretty standard as far as water chillers go, but I guess the temperature difference may also be pretty standard also?

Any Ideas?

Thanks,

-Sidney

[airspirit]

This is easy. Antifreeze sucks the donkeys nutz for heat transfer. Use alcohol (ethanol in this case, you have no need for methanol at those temps, I don't think) and Water Wetter to drop your temps dramatically. That is why you need to use less antifreeze in your car in the summer ... it doesn't transfer heat anywhere near as good as water. Get rid of the antifreeze and get yourself some better coolant.

Also, check that your block is lapped good, that you're not using too much goop (minimum is best) and that it is seated square. Also, I hear that that Swiftech isn't the best for cooling ... I could be wrong on that, but the interior geometry of the block doesn't look that promising.

Overall, your coolant might be what is hurting you. If the fluid can't absorb the heat from the core, the core won't cool as much, right?

[000]

Hmm well I've tried using alcohol for a little bit before, but since my evaporator is copper tubes with a bunch of aluminum fins, and my block is copper, I get the battery effect pretty easily.

And that experiment was with 1 gallon of of antifreeze 1 bottle of 91% alcohol and the rest water, so the antifreeze wasn't helping enough with the battery effect there.

But you may be right about the heat transfer. One problem though is why didn't the cpu get colder with colder water, even with poor heat transfer you'd think it'd still get colder right?

I did lap the block real good (although for some reason they say not to), so that should be ok. And I checked the mounting of it very carefully, had the motherboard completely out of the case and looked at it from all the sides, with a flashlight to check for gaps.

One thing I was thinking is that at lower speeds my processor was much closer to my water temps. So what this seems to say to me is that the heat isn't getting carried away fast enough. Although this is kinda weird since I have a 700 gph pump, I would think it's getting a pretty strong flow going through it but I think I'll check and see to make sure tonight.

[airspirit]

The block may be the culprit, then. Those blocks have a big internal tank with an inlet and outlet. They aren't the most efficient buggers in the world. Do you have another block you can try? (Yes, the better lapped the better the connection ... just use next to no goop or you'll have WORSE connection ... you want to only see the sparkle on the metal if lapped perfectly).

As for the battery effect, I don't believe antifreeze will help. Water wetter might. I'm not to up to speed on that, though. Anybody else have any ideas?

[000]

Ok I've figured out what it was. It's my pump, my pump is hardly pumping this stuff at all. I took a look at how fast it's coming out of the outlet and it's just a steady trickle (prolly 50 gph or something). Now I know when the coolant was warmer it was coming out very strong.

So does anyone know why the pump is going so slow? I imagine the liquid is a bit thicker but I don't think it's really that much thicker (it's still very much a liquid). Does the cold affect how the pump works somehow?

[WebMasta33]

If you attach each part to your case with a copper wire wrapped around it, it'll ground all the parts out. No more battery effect. (I think)

[TigerOats]

Your problem is caused by the viscosity of the coolant. As the coolant temps drop the viscosity increases. At about -10C the coolant viscosity is increasing exponentially and hence the reason for the temps leveling off.

[bigben2k]

Originally posted by TigerOats
Your problem is caused by the viscosity of the coolant. As the coolant temps drop the viscosity increases. At about -10C the coolant viscosity is increasing exponentially and hence the reason for the temps leveling off.

and that's about the best assesment one can put out, right on!

You might consider the windshield wiper fluid: although it still has water, the methanol in it actually behaves very nicely at lower temps.

Hey, it's a buck or two, give it a shot! Save your old coolant, in case you decide to go back to it (or dispose of it safely, please!) and flush clean in between.

[airspirit]

Grounding your parts to the case will not help. The potential between the different metals is what is causing the battery effect, not anything to do with the unit's power supply. There really is nothing that you can do about the battery effect in the fluid except to use fluids that are electrically inert, or additives that lower the conductive potential of the coolant. Sorry.

If you're having severe viscosity problems, try a commercial surfecant if you can get your hands on one. I have no idea where to get one, though .... It makes water super slippery (drop a vial into a mall fountain and watch the fun begin as the water starts spraying everywhere ... this was once a common practical joke ... haven't heard of it happening in a while). If you can't find any, try water wetter. It doesn't work as well, but you'll notice a difference.

Rather than trying wiper fluid, try the stuff in pure plus the additive. The water, even when mixed with alcohol, will still lower the overall viscosity of the fluid. If you're going below freezing, use it straight with a touch of water wetter.

[000]

Well I decided to try a bigger pump, I found an article for pump engineers (didn't realize there was such a thing until recently) that was saying how to calculate the correct centrifugal pump size for a certain viscosity. Anyway I decided to get a danner 12 to see if that helps or hurts.

I decided to try this since I'm pretty leary about messing up my dehumidifier (would hate for it to corrode and then start leaking), since It would be very hard to replace it for the same amount of money spent ($76 for a $250 dehumidifier is hard to find).

I also tried watering the solution down a little more to try to help things, but it didn't help much.

Would methanol possibly corrode things? If it won't, I may try that if the pump doesn't fix it.

[bigben2k]

Well, the methanol won't create corrosion, but it won't inhibit galvanic corrosion either, so if your rig is all copper, then it's good, if not, then you're got to figure out which additive to add. ("Water Wetter" or "Hyperlube" is probably good).

On the plus side, the windshield wiper fluid will prevent algae growth.

[bigben2k]

Originally posted by 000
Well I decided to try a bigger pump, I found an article for pump engineers (didn't realize there was such a thing until recently) that was saying how to calculate the correct centrifugal pump size for a certain viscosity. Anyway I decided to get a danner 12 to see if that helps or hurts.

I decided to try this since I'm pretty leary about messing up my dehumidifier (would hate for it to corrode and then start leaking), since It would be very hard to replace it for the same amount of money spent ($76 for a $250 dehumidifier is hard to find).

I also tried watering the solution down a little more to try to help things, but it didn't help much.

Would methanol possibly corrode things? If it won't, I may try that if the pump doesn't fix it.

Adding water would only increase the problem, because it's the fluid that's freezing up on you.

Do you have a link to that article?

and by the way, galvanic corrosion isn't something that should be of great concern, unless you plan to run the rig for 20 years, without an anti corrosion agent.

[000]

Here's the link:
http://www.pdhonline.org/courses/m102/m102.htm

So more water will make it worse? Would more antifreeze make it better?

Well if galvanic corrosion shouldn't affect me for awhile then I have another question.

I've run the setup with just water (when I was testing stuff) and I got alot of white gunk on the copper pipes of the evaporator and also the inside of my water block developed a rough layer of stuff on the inside, now this was only for a day or so, and it was a pain trying to get this stuff off of things. Any idea what it was? I thought it was galvanic corrosion but maybe not?

Also it did a similiar thing when I tried a mixture of a bottle of 91% rubbing alcohol 1.5 gallons antifreeze and tap water, except it was slightly different with this solution it was more fizzing of stuff around the evaporator.

Was that galvanic corrosion or something else?

[bigben2k]

That was calcium, from the tap water. Use distilled water instead. You can pick it up pretty much anywhere.

Yes, more anti-freeze will help prevent your coolant from freezing

Thank you very much for the link, I was looking for something just like it, just last week!

[nexxo]

With all this stuff, keep in mind that heat transfer from CPU to block is not 100%; more like 20% I read somewhere. You will never get your CPU as cold as the block/coolant no matter what you do.

[TigerOats]

I use 50% ethylene glycol based antifreeze and water and my coolant effect is about 50 % worse than pure water. Was getting a 10C differential with plain water and 15C with the mix. There is an increasing diff due to increasing viscosity as the temp drops. At -20C its about 17C. If you are using a submerged pump it may prove counter productive to use a larger pump. The actual increased output my be off set by the increased heat of the pump. I replaced my Maxijet 750 with the 1000 and my temps were 2C higher due to the fact that the 1000 is 14W and the 750 is 7W.

[gooFy]

why not use something like this. don't know where you can buy it in the US, but try asking.

-
gooFy

[000]

Hmm well I can't use Freezium since it says don't use it with aluminium

[000]

Although thanks for the freezium link, cuz it made me think to look for other heat transfer liquids, since I didn't really realize that they do similiar things that we are doing commercially. So I found Enviro-Kool, which looks promising www.enviro-kool.com although I dunno how you buy it (I emailed them to see). Hopefully it's not expensive and I don't have to buy a 55 gallon drum of the stuff.

[000]

bigben2k,

I think I may try the windshield wiper fluid route. (since I imagine the specialty stuff is gonna be expensive)

Is there a certain windshield wiper fluid that I should look for? Any certain place to find it it (autozone, walmart, etc?)

Also should I use a mixture of it and water or just straight windshield wiper fluid?

Thanks
-Sidney

[000]

I went looking for the windshield wiper fluid and didn't have any luck. Unless ya'll have a specific place for the winter version then I might not be able find it, since there's not much need for cold formula anything in the summer in TX.

What about pure methanol/ethanol, any ideas where I might find that?

Thanks,

-Sidney

[bigben2k]

Originally posted by 000
bigben2k,

I think I may try the windshield wiper fluid route. (since I imagine the specialty stuff is gonna be expensive)

Is there a certain windshield wiper fluid that I should look for? Any certain place to find it it (autozone, walmart, etc?)

Also should I use a mixture of it and water or just straight windshield wiper fluid?

Thanks
-Sidney

Well, I went to the dollar store yesterday, and found "Crown" brand wiper fluid. It's labelled as "67-56-1" methanol, with a mix ratio that escapes me. It also said that it would keep bugs off, but it didn't say if there was any additive.

The wife and kids were with me, so I didn't pick it up, but I know where to look now...

[bigben2k]

No, don't add water to it.

Been looking at this enviro-kool stuff. The company also makes Thermogreen (www.thermogreen.com, of course!). No idea where I'd get some though, but it says that it performs better than all Glycols. It still won't match water though...

This stuff seems to be made by a company called Clearwater (www.cwichem.com). They have an office in Houston... they have a phone number in Houston... hehehe They make products for oil drilling (drilling fluids). I work for an oil drilling company... hehehe

Gotta go!

[000]

I didn't think of looking in dollar stores, I'll have to go by some tommorrow.

Wow, that's pretty lucky. I didn't notice that they were in houston. Wonder if oil companies use heat transfer liquid (I'm guessing I'll find out fairly soon )

[bigben2k]

Also found therminol (therminol.com)

There's a whole sleuth of products out there that perform better than the glycols, and bonus, some are environmentally friendly, and non-toxic!

Also found www.nu-chem.com/heat.htm, a company that makes heat transfer fluid for optician labs. anti-corrosive, and self cleaning!

This is my favorite: http://www.anl.gov/OPA/logos19-2/nanofluids01.htm Looking for an additive? Try copper!

Here's a bit of theory http://www.process-cooling.com/CDA/A...4,7424,00.html