• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Are CPU dies waterproof?

C

cdabc123

Supreme [H]ardness
2FA
Joined
Jun 21, 2016
Messages
5,107
If I was to delid my CPU and cover the caps with epoxy could I theoretically have tap water running directly over the die?
 
They are waterproof technically, but there's a couple problems here.

One, your water isn't PURE, meaning all those impurities are going to cause two things...One, electrical shorts/grounds. Two, erosion of the silicon. Three, you could also have water leak UNDER the die (between the die and PCB, since that's not a water-tight joint) All three of which will very shortly result in a dead chip.

So this would be a very bad idea for a lot of reasons, and would have almost zero benefits. If the temps are that much of a problem, de-lid and relid the chip to reduce the IHS-to-die distance and replace the TIM with one of the liquid metal types.
 
People have done direct contact liquid cooling, but it's been a while since I've seen that. The trick is of course getting a reliable seal with the cpu package. Maybe use a lesser cpu / motherboard to test with first ...
 
Bandalo is the chip itself effected by a conductive liquid? I figured the layers of silicon would do enough to protect the actual die. Would a fluid like mineral oil work any better?

Basically this will be a upgrade to my thermoelectric rig. The water temps are as close as possible to the condensation point so I'm trying to find the best way to cool the chip with that. I have tried a normal block with little success and I currently have the Tec and thermister mounted directly on the heat spreader. The CPU is a 2500k at 4.7 GHz and currently my temps max at 50 under full load.
 
I've seen mineral oil builds, seems to work just fine if you don't mind the labor intensive set up and disgusting oily components if you ever need to do maintenance. Oldschool Cray supercomputers used to be cooled using immersion cooling https://en.wikipedia.org/wiki/Fluorinert

The issue, as Bandalo stated, is that water is hard to keep pure. It reacts to and acts a solvent to so much junk even just in the air that unless you're working in a incredibly controlled environment it's going to get contaminated and lead to a end to your experiment.
 
cdabc123 said:
Bandalo is the chip itself effected by a conductive liquid? I figured the layers of silicon would do enough to protect the actual die. Would a fluid like mineral oil work any better?

Basically this will be a upgrade to my thermoelectric rig. The water temps are as close as possible to the condensation point so I'm trying to find the best way to cool the chip with that. I have tried a normal block with little success and I currently have the Tec and thermister mounted directly on the heat spreader. The CPU is a 2500k at 4.7 GHz and currently my temps max at 50 under full load.
Click to expand...

There's a thin layer on the top that's mostly water-tight, but it's going to wear down if it's getting a constant stream of tap water or dyed coolant or almost anything really. Mineral oil might work, but it's harder to work with as others have mentioned. The conductivity of the liquid and the abrasive qualities of the liquid are the key problem. Once the coolant gets dirty, it's going to cause problems.

I think you're looking at a LOT of effort, cost, and risk of damage to save a couple degrees C over a well-mounted IHS. If it's a science experiment or a competition and you have money to burn, go nuts with it. But for any other purpose at all, it's probably not going to be worth it.
 
I used to have a similar idea in my head.

Delidding not necessary.
Get a copper tube and place it vertically on top of the CPU.
Next, use solder or silicone to create a seal.
Once you have that, put some mercury in the tube and seal the top so it won't poision you as a side effect.

Basically it'd be a vertical heatpipe with the mercury constantly heating up and thus hovering up and above the IHS, where the heat can be dispersed by a rad, small fan, TEC or all of the above.

I may or may not have been high at the time this occured to me. :)
 
A company called Kryotech....not sure if they're still around....used to make a set up with the original Athlons that was basically just a compressor refrigerator that chilled a little insulated compartment around the CPU. It was the first commercially available 1ghz machine. I always wanted one, but the 100 pound monstrosity was like $4000 so a bit out of my teenager price range....or my old fart price range now either for that matter.
 
Bandalo said:
There's a thin layer on the top that's mostly water-tight, but it's going to wear down if it's getting a constant stream of tap water or dyed coolant or almost anything really. Mineral oil might work, but it's harder to work with as others have mentioned. The conductivity of the liquid and the abrasive qualities of the liquid are the key problem. Once the coolant gets dirty, it's going to cause problems.

I think you're looking at a LOT of effort, cost, and risk of damage to save a couple degrees C over a well-mounted IHS. If it's a science experiment or a competition and you have money to burn, go nuts with it. But for any other purpose at all, it's probably not going to be worth it.
Click to expand...
any idea on how quick the chip would degrade? Would tap water be better as it would be less acidic?
 
cdabc123 said:
any idea on how quick the chip would degrade? Would tap water be better as it would be less acidic?
Click to expand...

I don't think that's something you can really know. And it's not acid necessarily that's your only corroding product....tap water has all kinds of potential junk in it you wouldn't want touching sensitive equipment. Just the chlorine in tap water would be enough to consider it fairly poor choice for use in this scenario, let alone whatever other junk your local water folks either put in or don't take out.

If you were to toy around with this, I'd use distilled water only. It'll cost you a buck a gallon at the store and dramatically increase your chances of very briefly making this work :)

Also....try it on something old you don't care about breaking!
 
michalrz said:
I used to have a similar idea in my head.

Delidding not necessary.
Get a copper tube and place it vertically on top of the CPU.
Next, use solder or silicone to create a seal.
Once you have that, put some mercury in the tube and seal the top so it won't poision you as a side effect.

Basically it'd be a vertical heatpipe with the mercury constantly heating up and thus hovering up and above the IHS, where the heat can be dispersed by a rad, small fan, TEC or all of the above.

I may or may not have been high at the time this occured to me. :)
Click to expand...

There's a lot of reasons why this wouldn't work well. Mercury isn't the greatest conductor and it's a bit too viscous at CPU temps. You're not going to get a lot of "flow" from the thermal driving head in your design, so basically you'd be relying on conduction. So it'd work better to just put a solid copper slug on top of the CPU and make some fins on the top end.

You'd get far better performance than either option with a $30 off-the-shelf heat pipe heatsink.
 
Bandalo said:
There's a lot of reasons why this wouldn't work well. Mercury isn't the greatest conductor and it's a bit too viscous at CPU temps. You're not going to get a lot of "flow" from the thermal driving head in your design, so basically you'd be relying on conduction. So it'd work better to just put a solid copper slug on top of the CPU and make some fins on the top end.

You'd get far better performance than either option with a $30 off-the-shelf heat pipe heatsink.
Click to expand...

I really appreciate this. Thank you. I'll probably try something similar anyway. But for now I can sleep soundly knowing I'm not wasting time!
 
Ok well I'm probably going to order some fittings tonight I have a old 920 I wouldn't mind trying it on however it has a soldered heat spreader so deliding will be a pain.
 
lol i just killed my 920 deliding it :p pushed abit to hard on a razor and heard a small crack (this was before heating it) guess i have a new keychain :) im probrably going to try with a dual core xeon now.
 
and i killed one of the dual core xeons. deliding these cpu's is hard :( i think my razor is a tiny bit too thick so when i push it in it puts too much stress on the die (i should have used the xacto knife instead). if anyone has a cpu they would like to donate (preferably not soldered on, i have a mobo for socket 1366, 1155, 1150) feel free to send it my way (ill pay shipping here and i can give it back once i mod it :p
 
Last edited:
bluesynk said:
I have a 920 floating loose in a drawer I never did anything with.
Click to expand...

I would be down to giving it another chance on one. I think I may be able to use either a thin peice of fishing line or floss to break the seal before I heat it up but these 1366 chpis are a pita.
 
yess!!!!! i managed to sucsessfully delid one of the xeons (a thinner blade did the trick) i will start a new thread with the building and testing of the project soon.
 
cdabc123 said:
If I was to delid my CPU and cover the caps with epoxy could I theoretically have tap water running directly over the die?
Click to expand...

It won't work well. Theoretically direct water cooling is the best, however the lack of surface area kills it. It also gets worse with these puny sized die which further reduces the surface area.

I still have my DT Direct somewhere...

d548f8d3_IMG_1946.jpeg
 
thesmokingman im too far committed now :p i have no idea if it will work or not and i have no idea (if it does work) that it will work well. im also concidering carving fins in the leftover solder but that seams kinda stupid. I also just lapped the top of the cpu for the fuck of it so it will be pretty and match the copper fittings.
 
well i got one prototype done today with a xeon and turns out cpu dies ARE waterproof. ill put out a post soon for the build/results and give it a try with a i7 920.
 
Just in case you dont already know. (Or it wasnt mentioned already)

Perfectly pure water does not conduct electricity. Making it a little more safe.

Once the tiniest bit of salt, minerals, sugar, or anything that dissolves into a polar solvent gets in the water then it does start to conduct.

I would be very careful putting tap water near the insides of a computer you are not ready to trash.
 
Elf_Boy said:
Just in case you dont already know. (Or it wasnt mentioned already)

Perfectly pure water does not conduct electricity. Making it a little more safe.

Once the tiniest bit of salt, minerals, sugar, or anything that dissolves into a polar solvent gets in the water then it does start to conduct.

I would be very careful putting tap water near the insides of a computer you are not ready to trash.
Click to expand...

i think im goint to run tap through this build. When i post the build results you can see the only part really exposed is the very top of the die everything else is coated in epoxy. my main worry with using distilled is its tendency to wear away some of the metal and as this is basicly a jet directly on the die i would rather minimize that effect the most i can.
 
btw i have already tested the cpu (minimally) with tap and it isnt dead yet.
 
cdabc123 said:
i think im goint to run tap through this build. When i post the build results you can see the only part really exposed is the very top of the die everything else is coated in epoxy. my main worry with using distilled is its tendency to wear away some of the metal and as this is basicly a jet directly on the die i would rather minimize that effect the most i can.
Click to expand...

I am kinda interested how this goes. Please understand I get curious about things and the best cure is to ask questions :)

I am not brave (rich) enough to do water cooling. I would stress about leaks or forgetting to fill the tank.

Do you have a plan to deal with the residue tap water will leave?

If you boil away a pot of tap water how much crap does it leave sticking to the inside of your pot?

I dont know your area, most tap water while more or less "safe" (unless your in Flint) has a fair amount of solids dissolved in it.

Does the water loop get hot enough that even small amounts of the tap water will boil or even sublimate? Will the water be recirculating or going down the drain?

I am not saying dont do it, that's your business. I am curious however.
 
Elf_Boy said:
I am kinda interested how this goes. Please understand I get curious about things and the best cure is to ask questions :)

I am not brave (rich) enough to do water cooling. I would stress about leaks or forgetting to fill the tank.

Do you have a plan to deal with the residue tap water will leave?

If you boil away a pot of tap water how much crap does it leave sticking to the inside of your pot?

I dont know your area, most tap water while more or less "safe" (unless your in Flint) has a fair amount of solids dissolved in it.

Does the water loop get hot enough that even small amounts of the tap water will boil or even sublimate? Will the water be recirculating or going down the drain?

I am not saying dont do it, that's your business. I am curious however.
Click to expand...
so i ran my last water build with tap soley out of the sake of lazyness (it was actually a pretty cool build) and never noticed too much of a issue. as for the tap in my area i have never noticed much of a residue when boiling water. and i will mostlikly put in a water chiller if this works decent so i cant imagine the water would ever get above 50c. and im defenently going to do a post on this as soon as i get the cpu and mobo mounted in a actual loop.
 
You must log in or register to reply here.
Back
Top