Can we submerse radiators?

[cvaldes]

Would that depend on the size of the container. If my container was the size of a swimming pool, there no way i need a fan. ..OK a little OTT but can you see where I'm going with this. If the container is large enough, the CPU/GPU will have a hard time getting pass a certain temperature.

Of course, but at what point does it become impractical? The cooling pond needs to adequately dissipate during a heat spell during the summer.

If you have a 100W TDP CPU and a 300W TDP GPU, is a standard bathtub enough to dissipate the heat? I'm not sure but I do know that it's not practical.

With stagnant water, you'd have to treat it with chemicals (like a swimming pool) to prevent algae growth and deter pests (like mosquitos). And then you'd probably have to protect it to prevent pets and children from drinking biocide treated water. If the cooling pool is deep enough, you'd probably need to set up safeguards for drowning. Would you like to check the cooling pond's chemistry (pH, chlorine, etc.) weekly as is done for swimming pools? The list goes on and on.

The original premise is flawed and poorly thought out.

 

[maxfly]

Eliminate the rad and use copper tubing. Bury it coiled, deep enough under ground and it will naturally cool the water in the loop.
Its been done so googling might find a how to.

 

[delshay]

Of course, but at what point does it become impractical? The cooling pond needs to adequately dissipate during a heat spell during the summer.

If you have a 100W TDP CPU and a 300W TDP GPU, is a standard bathtub enough to dissipate the heat? I'm not sure but I do know that it's not practical.

With stagnant water, you'd have to treat it with chemicals (like a swimming pool) to prevent algae growth and deter pests (like mosquitos). And then you'd probably have to protect it to prevent pets and children from drinking biocide treated water. If the cooling pool is deep enough, you'd probably need to set up safeguards for drowning. Would you like to check the cooling pond's chemistry (pH, chlorine, etc.) weekly as is done for swimming pools? The list goes on and on.

The original premise is flawed and poorly thought out.

Read what I said earlier in the thread. Whenever you come up with a new idea your always going to have problems. You need to overcome problems because effectively it's a prototype.

I said container the size of a swimming pool as an example to point out you don't need a fan due to the sheer volume of liquid even in the heat of the summer months, not actually use a real swimming pool filled with water. Read post #19 carefully

You will not be using water. You could use current solutions already on the market or something else.

 

[cvaldes]

I read your post carefully. Having a cooling pond the size of a swimming pool (or even a bathtub) for a standard consumer PC is not practical. I'm not saying that it can't dissipate the generated heat, I'm saying that it can't do it in a currently practical sense.

Let's say you can dissipate 500W TDP between a CPU and GPU with a kiddie inflatable wading pool with some sort of special liquid. Is that possible? Sure. Is that practical?

Hell, I can probably cool my car's engine with a non-fan driven radiator by towing some sort of weird liquid cooling pond. But does that make any sense? Would it be expensive? (You better believe it.)

Sure, with an infinite budget and no consideration for any factors yes, it's possible.

Remember that radiator-assisted cooling isn't a "new idea" [sic]. Cars have been cooled like this for about a century. The ISS has radiators filled with ammonia. Radiator-assisted cooling is a worthy solution in some situations and far less so than others.

One thing I can always count on from the Internet is a certain number of individuals who are completely divorced from reality. You have not disappointed.

QED

 

[JCL]

Just sharing this to show a possible idea spark. Of course it isn't practical for most but it is my benching rig and works great for me.
I have an ASRock Z77 OC Formula as my daily and a Maximus V for some photo work and back up.
Both of those containers I put blocks of ice in while benching. I can remove the resulting thaw and add more ice as I go to keep temps low. Now though it is our monsoon and the dewpoint makes anything too cold very risky due to condensation. Maybe in a month or so. Even then it has to be watched closely. That is the reason for the fans on the 5700, there is a space between the card/water block that needs to have good air flow. When the air is dry enough for a lot of ice I pull a hose and bypass the rads, running the water straight to the reservoir . Even with no ice setup like it is now it hold temps well. There are different ways to do what you are after, it just depends on what you have to work with. Submerging the rad in water though, there is not a way to throw off the heat as others have said, and basically becomes a heater for the water.

 

[delshay]

I read your post carefully. Having a cooling pond the size of a swimming pool (or even a bathtub) for a standard consumer PC is not practical. I'm not saying that it can't dissipate the generated heat, I'm saying that it can't do it in a currently practical sense.

That's not the point I'm trying to put across. The point is at a given volume of liquid you will not need a fan. A container the size of a swimming pool was clearly OTT.

@Thread

Someone pointed out when radiator is submerged you need a fan. Post #09 has already proven this works without a fan for 15 mins. The liquid in the container is in direct contact with the radiator. So my question is, why would you need a fan. If the user was to submerge the radiator in a larger container, it's going to take longer to heat up.

@JCL

If you have just water in that container with fans off on the radiator, how hot does the water get in the container? Max temperature.

EDIT: If you have never done this before, don't do it. I don't want you to end up with a broken PC.

 

[eidairaman1]

If you're going to go that far, why not just use phase-change cooling?

Or a fridge tower

 

[JCL]

That's not the point I'm trying to put across. The point is at a given volume of liquid you will not need a fan. A container the size of a swimming pool was clearly OTT.

@Thread

Someone pointed out when radiator is submerged you need a fan. Post #09 has already proven this works without a fan for 15 mins. The liquid in the container is in direct contact with the radiator. So my question is, why would you need a fan. If the user was to submerge the radiator in a larger container, it's going to take longer to heat up.

@JCL

If you have just water in that container with fans off on the radiator, how hot does the water get in the container? Max temperature.

EDIT: If you have never done this before, don't do it. I don't want you to end up with a broken PC.

I haven't done that but if it ran long enough it would approach that temp of the chip almost with no removal of heat. Fans are important for a rad to work.

 

[MentalAcetylide]

Hello,
Is it a good idea, has anyone tried?

instead of cooling the rad with fans, can we just use longer tubes, take the rad away from chassis and sink it under water?
Also can we use another gpu wb as a rad for this purpose? since I think gpu block is less prune to corrosion and no fans no fins.
What are the cons?

A submerged radiator, as was already pointed out, doesn't make any sense since now you need to put a lot more energy into moving the medium that's removing the heat. Its not like you can just toss the radiator into a cool still liquid. The liquid has to transfer the heat from the radiator and move it away and get cooled again, which will not happen without having to deal with the pains of plumbing, pumps, and refrigerants or a watermill next to a river.

 

[Deleted member 24505]

A submerged radiator, as was already pointed out, doesn't make any sense since now you need to put a lot more energy into moving the medium that's removing the heat. Its not like you can just toss the radiator into a cool still liquid. The liquid has to transfer the heat from the radiator and move it away and get cooled again, which will not happen without having to deal with the pains of plumbing, pumps, and refrigerants or a watermill next to a river.

As i said, convection in a pool of liquid is not enough imo to move the heat away from the radiator, you would need a pump to make the fluid move around. otherwise the radiator would just end up surrounded by the heat it is producing. you need to "stir" the heat into the volume of cooler liquid.

Tbh i think a standard loop with rads and fans is pretty good for a PC. for long term extreme cooling, i don't think there is anything practical, as LN is not practical for anything but suicide runs, and iced water will constantly need the ice replacing.

 

[nguyen]

As i said, convection in a pool of liquid is not enough imo to move the heat away from the radiator, you would need a pump to make the fluid move around. otherwise the radiator would just end up surrounded by the heat it is producing. you need to "stir" the heat into the volume of cooler liquid.

Tbh i think a standard loop with rads and fans is pretty good for a PC. for long term extreme cooling, i don't think there is anything practical, as LN is not practical for anything but suicide runs, and iced water will constantly need the ice replacing.

Or just build a thermally isolated box for your PC with its own air conditioner . A cheap 8000Btu window AC can handle 2000W of heat no problem and no need to worry about condensation either.

 

[Sachingare]

They do it for mining farms (as an example)
Submerging in water isn't the greatest idea, other liquids are more useful

 

[MentalAcetylide]

Or just build a thermally isolated box for your PC with its own air conditioner . A cheap 8000Btu window AC can handle 2000W of heat no problem and no need to worry about condensation either.

heh, I'm pretty sure that would me more hassle than what its worth.
Yes, unfortunately you would still need to be concerned about condensation since the AC is going to be pulling more humid air from outside and blowing it into the case. While AC can remove some moisture from the air, you're still going to have to dry the air further before it gets inside of the case. It doesn't take much condensation(it doesn't have to be enough to be visible) to screw something up when it comes to electronics.

You would practically need to custom-make a hermetically sealed case with intake/exhaust hoses, portable AC unit, and a proper dehumidifier connected to a large intake chamber of silica gel. That's a hell of a lot of $ just to have sub 60 deg. F temps inside a computer case.

 

[Lei]

Water internal heat capacity 4200
Air internal heat capacity 993

room dimensions 5x4x10 meters. Air volume 200 cubic meters
20 Liters of water = 0.02 cubic meters

room has 10000 more air than 20 Liters of water
water temp rises 4.23 times slower than air

if your pc makes the ambient temp rise by 1 degree, it will raise water temp by 2364 degrees (10000/4.23)

How many liters of water if we fill your room with water? 200000
A room full of air takes away heat better than quarter that amount of water

What I don't understand : hold a cup of water next to the ocean, both under sunlight. measure the temp, why the water in cup is not boiling? Sun must rise cup temp (ocean volume/cup volume) times more than the ocean. hmm, sun is the cpu and cup is the bathtub/bucket

 

[Mussels]

This old idea coming back made me want to extend some tubing out the window and run an external radiator for summer...

Eliminate the rad and use copper tubing. Bury it coiled, deep enough under ground and it will naturally cool the water in the loop.
Its been done so googling might find a how to.

Thats what i was remembering! some old video on doing that

 

[Sithaer]

bruh at this point just throw a radiator in the lake and call it a day

Someone in my country whos known for unusual/crazy experiments did that ~10+ years ago during winter.
https://bacsis-tuning.hu/2010/12/a-balatonhutes

Just for the pictures even if you can't understand it.
Also the difference was only 1 celsius between the cold air and after submerging the heatsink in the icy lake.

Well ok not with a rad but its close enough considering the age of the 'experiment'.

 

[Mussels]

*Looks at spare quick disconnects, hosing, and cheap radiators i got for $5 off facebook*

If i can get the hose out without letting summer air in, i may have a project in a few months

 

[nguyen]

heh, I'm pretty sure that would me more hassle than what its worth.
Yes, unfortunately you would still need to be concerned about condensation since the AC is going to be pulling more humid air from outside and blowing it into the case. While AC can remove some moisture from the air, you're still going to have to dry the air further before it gets inside of the case. It doesn't take much condensation(it doesn't have to be enough to be visible) to screw something up when it comes to electronics.

You would practically need to custom-make a hermetically sealed case with intake/exhaust hoses, portable AC unit, and a proper dehumidifier connected to a large intake chamber of silica gel. That's a hell of a lot of $ just to have sub 60 deg. F temps inside a computer case.

Window AC or split AC don't work the same as HVAC, they don't use air from the outside, it just cycle and cool down the air that is inside the room and dehumidifying it (it is more efficient this way, same as car AC where you can choose between 2 modes). These AC unit can often act as a dehumidifier as condesation happen at the cooling radiator and go out the water drain, therefore lowering the dew point inside the room to as low as 5C.

So just build a sealed box so water from outside cannot get in, stick an AC in it and that's it, the AC is even cheaper than custom watercooling . I guess this is what they are doing with server farms.

Illustration how window AC works

 

[Grog6]

I once set up a 1000A power supply for electropolishing, and it was a lm317 with a shitload of pass transistors on a rack frame with a garden hose hookup.
The water going in from the city was ~50F, and would boil the output, but even the last set of transistors never went over 100C.

 

[P4-630]

Eliminate the rad and use copper tubing. Bury it coiled, deep enough under ground and it will naturally cool the water in the loop.
Its been done so googling might find a how to.

Making moonshine...

 

[Anoniem]

Dunking rads in water can be done, it just isn't a great idea. I've done it before multiple times Got a trash bin, filled it up with 9kg of ice and poured some water in it. Worked great for a bit, then condensation became a problem hahaha. Here's a random picture of it:

 

[ThrashZone]

Hi,
Yeah you really just need circulating cold tap water in the bucket.. that should be enough without too much condensation risk.

 

[freeagent]

Geothermal works well too! A guy on my team used it until he moved, his results were excellent

 

[FireFox]

Fans are important for a rad to work

They are.
However you can run your Rads fanless/off if you use a Chiller.

 

[JCL]

"When the air is dry enough for a lot of ice I pull a hose and bypass the rads, running the water straight to the reservoir ." At that time the res. contain ice. Fans or going through the rad just accelerates the increase in amount of ice used. It works great with just water during times of high humidity holding temps down. That CB20 was just water. A chiller, $$$.