Burn

-----Theory------


Let us take sub-zero cooling to the average phase-change setup....Say, -40C.

We all know that frost obviously forms at that temperature, right? No single TEC is capable to reach that temperature, so this one's for the phase change/DICE/LN2 guys.

So we have frost forming on the various components of the motherboard, because water condenses on the board because it's below the ambient temp. It turns to frost, because we're chilling to -40C. So, we run the setup for a while, anyhere from an hour to several days, I will have my TEC rig up for multiple weeks at a time if the Windows Gods permit it. It doesn't really matter for how long, just that frost is forming.

So now we shut down the computer. Phase change/DICE/LN2 no longer is cooling the motherboard, so it warms back up. Remember the frost we formed? No longer at -40C, it too begins to warm up. Bang, it defrosts and we're left with several tiny puddles of water. Here's where we get damage, these puddles dripping onto our various capacitors, video card(s) and motheboard components.

^That is where the damage occurs- Not while we're benching, but after we've stopped cooling everything down.

Why doesn't it apply to TEC users? Mainly because we don't get cold enough. We might get down to freezing or under about 10 deg. C during the day-to-day, but we just don't get cold enough to get frost forming from what I have read and observed if insulated properly. Insulation provides a good enough barrier that we don't need to worry about frost forming because insulation provides a good enough barrier from the -20C TEC module and coldplate to the air inside our cases. Not so much with phase change, DICE and LN2. As the temperature differential increases (From -20 to -40 or -50), you need to use more insulation to provide that barrier.
------------------

Does it make sense at all? Or am I spouting foolishness?

__________________
My CNC Mill Project
We cannot change the cards we are dealt, just how we play the hand.
--Randy Pausch

DuckieHo

I don't know. DICE and LN2 cooling are short term methods. The user is pretty much their standing over it the whole time to monitor. I don't think anyone careful enough to do custom benchmark cooling would be so careless to allow dripping water when they are done.

Phase change is day-to-day cooling. The heads are well insulated. Again, if water condesation was a day-to-day issue... it would obviously be noted. "Hmmm... why do all phase change mobo burn out after a month?" This is not the case though. If it was, it would be addressed.

Being careful (hopefully) and extreme overclockers, I would think they would take extra precaution on insulating.

Kopi

I believe waxshop coats his motherboard and parts in some weird goopy, non conductive stuff

__________________

kidwolf909

dielectric grease*

__________________

Burn

DuckieHo, exactly. What I'm saying is when such insulation is not "good" enough, doesn't isolate the outside air good enough, is when frost and such form.

__________________
My CNC Mill Project
We cannot change the cards we are dealt, just how we play the hand.
--Randy Pausch

Kopi

If that was true, than how do some phase cooled machines last for very long times?

__________________

pauldovi

I strongly disagree.

1. Water condensation will form at all temperatures below ambient assuming the air is not dry.
2. Frost does still conduct electricity.
3. Dielectric grease, insulation, and heater pads will ensure that the motherboard and other components stay warm (weird isn't it?)

__________________

MADMAX22

well burn the thing is that the frost will form on the coldest parts butyou have that transition area that is warmer thent he coldest part so thats where you get the frost or condensation that is actually droplets of water and it can happen very quickly, you can fry a mobo in a matter of minutes of not protected properly (not that ive done that or anything )

but also its not jsut the condensation whichis a short term problem but the long term problem is the corrosion of having the water on the soldered parts of your mobo and the littel metal parts, it will corrode and very quickly might i ad

__________________
E6400@3856E4300@4158E6600@4510,Opti148@3102,opti170@2918,Xeon3220(B3)@3430(490fsb)

Burn

What I was postulating was that, if the provided insulation was not sufficient, that such a situation might ensue.

__________________
My CNC Mill Project
We cannot change the cards we are dealt, just how we play the hand.
--Randy Pausch

pauldovi

I postulate, that, if you don't seal the tubing onto the waterblock then you might have water leak all over the processor. This is the fundamental reason why no one should use watercooling.

__________________