My whole room watercooling project - I need your help!

[HollyBoni]

For projects like this everyone recommends the 18W DDC. The D5 is quieter and more reliable, but the 18W DDC is a monster in terms of power. It's really loud at full RPM but that's not a problem here. 

[Benji_w]

You don't think it's necessary? Think again.

 

http://i.imgur.com/l9fVfbu.jpg

 

If the room temp is 20C and the humidity is 65%, you'll get condensation if the water is 13.2C. 

 

->condensation->water dipping down from the tubes, GPU and CPU block->the end. 

 

Could you give more information on this so we can all learn a bit more? Are you telling me if I put my radiator outside when it's cold (its 8c here) then my pc indoors will look like I put it in a freezer?

[AnonymousGuy]

I've got a slightly different idea of how you can reach your goal. What about a heat exchanger in the model and comprise your cooling system of two loops?

For example, you'll run the two PCs in one loop which is contained in room A. The PC loop is exchanging heat with a loop in room B (all connected to a rad in a bucket of water? not sure how it could be achieved), which is then handle the dispersal of the heat into room B or outside, using some PC rads or botching it and use a car radiator. That would be hilarious.

This would allow for relatively easy maintenance of the PC loop (given that it's carefully designed), and you won't have to worry about it freezing because it's contained inside. For the outside loop, you don't have to be so careful and could use tap water and copper tubes (not that I recommend doing this, it's just an example). This would also allow for separate pumps handling the PC loop and the outside loop. You could use D5 or DDC pumps for the inside loop, and then something like a fountain pump for the outside loop. 

These are just my thoughts. 

 

This is how I do mine and how it should be done if you want less problems.

 

http://www.ebay.com/usr/honestystillives?_trksid=p2047675.l2559 is great company for them.  I have 2 of the 40 plate ones, one for each loop in my desktop.  You'll need fittings from McMaster though to adapt 1" NPT to barb.

 

I also highly recommend a UV sterilizer on the dirty side of the exchangers (the side that doesn't go through your waterblocks).

[capawesome9870]

here is what i would do:

1. rotate the placement of the computers 90° (alt+248 for the degree logo #BarnaculesTechTip or something) so the PCs are on the wall with the window (so you do not have the water hose in the walk way (plz read farther on why). 

 

2. next have the piping flow under the PC, with the PCs having 2 large reservoirs (in/out) (either sharing each tank or 2 for each) mounted a little above the PC.  

should be organized something like this

Ceiling

tanks

PC

Piping out/in side

Radiators

Flooring

 

3.  use normal size piping (9.5-13mm) for connecting the PCs in/out of the reservoirs, but use larger 19-25mm to connect out/in the reservoirs to the radiators. 

 

4.  Connect the radiators to the large pipe using Y adapters (or 1 to 4 adapters).  each radiator will be in parallel to increase overall flow of water. 

 

 

 

reasoning:

 

water will naturally try to even out. 

you will have 2 pcs pull from a radiator (A) then push to Rad ( B) making A lower and B higher. 

the water will go from B through the 19-25mm pipe to the radiators (more or less evenly) and flow back up to A

 

No need for extra pumps outside of the PC pumps. 

 

http://imgur.com/xuyCCr9

 

 

 

i'm actually planning on doing this my self with a recycled Copper Car radiator. 

 

 

edit:

 

forgot to say any high pint you should add a fill/breather port. it will make filling and getting air bubbles out of the system.  the good thing about my idea is that there is only the reservoirs as high points. 

[vanished]

Could you give more information on this so we can all learn a bit more? Are you telling me if I put my radiator outside when it's cold (its 8c here) then my pc indoors will look like I put it in a freezer?

 

The idea is that since it is very cold outside, the water coming from the rad outside could be entering the house at less than room temperature.  If it is coming in cold enough, the air around the tubing will also get cold, and if the humidity in the room is high enough and/or the temperature of the water in the pipes is low enough, condensation will form on the outside of the pipes as moisture in the air "drops out".  I don't have a humidity chart for the ambient temperature range in front of me at the moment so I can't confirm the numbers given, but it sounds believable.  If this water is still sufficiently cold when it reaches your CPU/GPU/etc. blocks, that condensation could occur on the block and drip over the motherboard.

 

This is a legitimate concern, but I believe it should also be fairly easy to mitigate.  As I mentioned, the pipes taking "hot" water away from the system towards the rad should be insulated to avoid leaking all of the heat into the room, since that would make the project pointless.  However, the piping leading from the rad back to the computers should probably be left un-insulated, or even sent through an indoor rad in order to guarantee it has been brought back up to room temp before hitting your components.  That will ensure that any condensation that does occur happens harmlessly on the pipes and not inside your computer.

 

Update: Here, I found this chart:

Now, look on the x axis at 20 deg C and go up until you intersect the 65% relative humidity curve (just estimate half-way between 60% and 70%).  Now, from that point go straight left until you hit the 100% curve.  Now look straight down from that point - this is the temperature at which condensation will start to occur for those initial conditions (20 deg, 65%).  We can see that it is indeed about 13 deg C, so that comment was correct.  You can also use this for any other conditions you may want to test.

Edited January 5, 2016 by Ryan_Vickers

[HollyBoni]

Could you give more information on this so we can all learn a bit more? Are you telling me if I put my radiator outside when it's cold (its 8c here) then my pc indoors will look like I put it in a freezer?

 

I just found the picture randomly on a forum, but yes if you put your rad outside when it's really cold your PC will look like this. Just ask guys who use phase change coolers, or watch the video Linus did. All I know about this picture is that the guy put his rads outside in the winter.

But check the comment above me, there you can see the dew point scale. Condensation, even if it's not this extreme is a huge concern. But the solution I wrote (two rads parallel, one inside one outside with a thermostat) is a pretty simple fix. 

 

[vanished]

(two rads parallel, one inside one outside with a thermostat) is a pretty simple fix. 

 

So that would require (in addition to the two rads) a splitter piece, two electronically controlled valves, a joiner piece, and two (or at least 1) temperature sensor?  Yeah, that should do it, but I'm wondering if there is an even easier way.  I was thinking put it through the outside rad, then send it through the inside rad (in series).  That does 3 things - one, it's simpler, two, it actually would act like AC in a way, cooling the room, and three, it also ensures that the water reaches the computers at a safe temperature (not too cold).

[HollyBoni]

So that would require (in addition to the two rads) a splitter piece, two electronically controlled valves, a joiner piece, and two (or at least 1) temperature sensor?  Yeah, that should do it, but I'm wondering if there is an even easier way.  I was thinking put it through the outside rad, then send it through the inside rad (in series).  That does 3 things - one, it's simpler, two, it actually would act like AC in a way, cooling the room, and three, it also ensures that the water reaches the computers at a safe temperature (not too cold).

 

One Y or T splitter (2$), one valve (60-70$) (it's basically a T splitter with a valve so you don't need anything else) 1 water and a room temperature sensor (20-30$ max). Not very expensive stuff and I don't think money is an issue here, and it's not complicated to build at all. 

 

There might be other solutions but you can control the water temperature very precisely this way. Your solution might work but it's risky, and I wouldn't do that without a water temperature sensor so you need that there too. But you can get water temperature sensors for super cheap. 

[vanished]

One Y or T splitter (2$), one valve (60-70$) (it's basically a T splitter with a valve so you don't need anything else) 1 water and a room temperature sensor (20-30$ max). Not very expensive stuff and I don't think money is an issue here, and it's not complicated to build at all. 

 

There might be other solutions but you can control the water temperature very precisely this way. 

wait, are you saying just a manual value and a temperature sensor then?  I just assumed you meant some kind of automatic control system with a PID or something like that so the temperature would be managed automatically

[HollyBoni]

wait, are you saying just a manual value and a temperature sensor then?  I just assumed you meant some kind of automatic control system with a PID or something like that so the temperature would be managed automatically

No, an automatic valve. I think it's called "automatic thermostatic mixing valve" in english. It cost 60-70 bucks where I live. 

 

Example: http://www.langor.hu/cikk_leiras/Esbe%20VTA572%2020-43.jpg

 

You look at the room temperature, if it's 25C for example you set the valve to 25C (just to be safe) and it mixes the hot and cold water automatically until the reaches that temperature.

It's not fully automatic because you still need to set the temperature you want, but that takes about 2 seconds, you just twist a knob. But you don't need to do that too much, because in the summer there is air conditioning, and in the winter the room is heated. 

 

You don't need two. Just draw the loop or something and you'll understand.

 

I still think it's the least complicated and best solution. 

[vanished]

No, an automatic valve. I think it's called "automatic thermostatic mixing valve" in english. It cost 60-70 bucks where I live. 

 

Example: http://www.langor.hu/cikk_leiras/Esbe%20VTA572%2020-43.jpg

 

You look at the room temperature, if it's 25C for example you set the valve to 25C (just to be safe) and it mixes the hot and cold water automatically until the reaches that temperature.

It's not fully automatic because you still need to set the temperature you want, but that takes about 2 seconds, you just twist a knob. 

 

You don't need two. Just draw the loop or something and you'll understand.

 

I still think it's the least complicated and best solution. 

oh, didn't know that existed... too much work with industrial-sized solutions

yeah that sounds like a good option

 

edit: I thought of something though - that valve has to assume one path is "cold" and one is "hot".  What happens in the summer when the outside is actually the hotter rad (vs inside)?

Edited January 6, 2016 by Ryan_Vickers

[HollyBoni]

oh, didn't know that existed... too much work with industrial-sized solutions

yeah that sounds like a good option

 

edit: I thought of something though - that valve has to assume one path is "cold" and one is "hot".  What happens in the summer when the outside is actually the hotter rad (vs inside)?

I think you just switch up the connectors. You can see the hot and cold site better here:

http://netkazan.hu/data/termekek/e/s/esbe-vta-322-34-os-20-43-keveroszelep-akcio_n.jpg

 

Although my outside rad is going to have a massive 40cm car radiator fan on it, the inside is going to be passive, and I don't have air conditioning. So I still think the inside rad is going to be hotter. Or I might put both rads outside with fans in the summer, but I know for sure my room is a little bit hotter than the outside temperature in the late afternoon-evening.

 

But we'll see, I hope I can build it before summer.

[vanished]

I think you just switch up the connectors. You can see the hot and cold site better here:

http://netkazan.hu/data/termekek/e/s/esbe-vta-322-34-os-20-43-keveroszelep-akcio_n.jpg

 

Although my outside rad is going to have a massive 40cm car radiator fan on it, the inside is going to be passive, and I don't have air conditioning. So I still think the inside rad is going to be hotter. Or I might put both rads outside with fans in the summer, but I know for sure my room is a little bit hotter than the outside temperature in the late afternoon-evening.

 

But we'll see, I hope I can build it before summer.

 

Perhaps.  I just know where I am, for a good part of the year, which is colder will change from night to day each day so switching the connectors would not really be an option.  I would want to have a temperature sensor on the outlet of each rad and then a basic computer can decide which is hotter and coordinate the flow accordingly.  It's sounding more doable now than when I first read your idea, but I still think just an outdoor rad followed by an indoor rad in series would be ideal since it requires 0 extra parts (no matter how cheap or simple they may be) and would work year round, regardless of conditions, with no need for flow redirection or controls of any kind.

[HollyBoni]

Perhaps.  I just know where I am, for a good part of the year, which is colder will change from night to day each day so switching the connectors would not really be an option.  I would want to have a temperature sensor on the outlet of each rad and then a basic computer can decide which is hotter and coordinate the flow accordingly.  It's sounding more doable now than when I first read your idea, but I still think just an outdoor rad followed by an indoor rad in series would be ideal since it requires 0 extra parts (no matter how cheap or simple they may be) and would work year round, regardless of conditions, with no need for flow redirection or controls of any kind.

It might work. Someone has to try it and we'll know, until that it's just speculation. But I still think the water would be too cold in the winter. Where I live it gets down to about -10C. 

And for me it would be scary not having precise control over the water temperature. 

 

 

Now I thought about it a little bit and yes it would be possible for the inside radiator to be colder, but only when both rads have fans on them. Think about it. There is a massive fan on the outside rad, and the inside runs passive. I run my config passive at idle, and my radiator and the water temps are always much hotter than the air temperature. So I still think the outside rad (water) would be a little bit cooler, even if it's warmer outside. And the valve can bring in a tiny amount of "outside water" to cool down the water to room temperature. 

[vanished]

It might work. Someone has to try it and we'll know, until that it's just speculation. But I still think the water would be too cold in the winter. Where I live it gets down to about -10C. 

And for me it would be scary not having precise control over the water temperature. 

yeah it would be too cold  That's why we reheat it with the indoor rad.  The other benefit to my design is it keeps the whole flow rate through the outside.  With your design, the switch will send very little outside since it wouldn't take much to make the loop cold when chilling it with -10 air, which means there will be much lower flow through the cold rad, which means there is a greater chance the water will actually freeze solid.  But the main issue I'm worried about is just that on those cold days, sometimes it is 15 C at night and then 27 C in the day, and if inside is 21 C, which is hotter and which is colder will switch, and unless the valve knows that, it will be doing the wrong thing and wondering why it is not helping.

[HollyBoni]

yeah it would be too cold  That's why we reheat it with the indoor rad.  The other benefit to my design is it keeps the whole flow rate through the outside.  With your design, the switch will send very little outside since it wouldn't take much to make the loop cold when chilling it with -10 air, which means there will be much lower flow through the cold rad, which means there is a greater chance the water will actually freeze solid.  But the main issue I'm worried about is just that on those cold days, sometimes it is 15 C at night and then 27 C in the day, and if inside is 21 C, which is hotter and which is colder will switch, and unless the valve knows that, it will be doing the wrong thing and wondering why it is not helping.

Yeah I understand, my plan does the same thing. I just don't know if the inside rad can heat up the water to a usable temperature when you bring ALL the cold outside water in. I don't think the water freezing is an issue if you use antifreeze. 

 

I edited my previous post about the inside is colder than outside thingy. 

 

But really the only way is to try it out. 

[Aereau]

I have been running a watercooling system with external radiators and pump (outside the house) with my uncle at his place for years now. Before we were using anti-freeze we had water being frozen in the tubing, so at that point we started using anti-freeze for the peak-low temperatures (Netherlands). Ever since we have had the CPU running against 4 degree (around 16 deg C below room temp) and GPUs running at room temp under load for extended periods of time. However, we have never had a confirmed case of condensation. I realize that this would be the biggest objection against having outside watercooling, but we have never seen it happen. I will be posting a topic around here shortly showing our implementation!

 

Edit: You can find my topic with pictures here

[AnonymousGuy]

Mind you...this isn't with outside temperature, but this is what happens when you run on the edge and have a spike of humidity indoors (say...it's raining outside and someone opens a window too long):

 

 

This was 77F ambient and about 50% humidity.  Water temp was probably 55 degrees and dew point was 57 degrees.

[HollyBoni]

What the hell are those connectors?     

 

Also how did the water temp go below room temperature? Are you a wizard?

[vanished]

Yeah I understand, my plan does the same thing. I just don't know if the inside rad can heat up the water to a usable temperature when you bring ALL the cold outside water in. I don't think the water freezing is an issue if you use antifreeze. 

 

I edited my previous post about the inside is colder than outside thingy. 

 

But really the only way is to try it out. 

 

I was thinking about this a little more and I think I thought of a way to use your parallel method with a small change that would solve a few problems.

 

Instead of having an inside rad and an outside ran in parallel, have an outside rad and simply a pipe that stays inside in parallel (ie, replace the inside rad with just a pipe).  How this will help?

• You will always know which side is "hot" and which is "cold", since even on warm summer days, the inside pipe will be the temperature coming out of the computers and thus probably even hotter
• With the 1 rad inside idea, some of the heat is being dumped outside and some inside.  This seems like a waste since the point of this project is to dump as much as possible outside.  This idea ensures that 100% is dumped outside
• It gives the valve a more powerful heating option than using the room temperature water from the inside rad since the un-radiated water will be at the outlet temp of the computers

Honestly, I think this is the best option so far, of either of our initial ideas.  I'd love to see it built

[vanished]

I have been running a watercooling system with external radiators and pump (outside the house) with my uncle at his place for years now. Before we were using anti-freeze we had water being frozen in the tubing, so at that point we started using anti-freeze for the peak-low temperatures (Netherlands). Ever since we have had the CPU running against 4 degree (around 16 deg C below room temp) and GPUs running at room temp under load for extended periods of time. However, we have never had a confirmed case of condensation. I realize that this would be the biggest objection against having outside watercooling, but we have never seen it happen. I will be posting a topic around here shortly showing our implementation!

 

Edit: You can find my topic with pictures here

 

Interesting.  That means that

1. The tubing is well enough insulated that despite the low water temperature, the outside of the tubes never got cold enough to cause condensation, and 
2. The computer components are heating the water in the block enough that the outside of it is also never getting too cold.

Honestly it sounds risky but since it's never been a problem, I guess your setup is safe for your climate

[TheBluePotato]

Ok I have a way better Idea.

 

About 4 Years ago we installed an air conditioning system with heat exchanger. What I will do now is NOT put any radiators outside or in the room next to it BUT put it on the loft.

Well how the hell should this help you might ask. Good Question. In every living room and bedroom are air - out - tubes from the ventilation system. In every bathroom and kitchen are air - in - tubes from the system. In the loft there are the tubes for the air - in - tubes for the bathroom (in means it goes to the machine and then outside and out means if goes in the room from the machine). I can easily add some more pipes to above the Gaming room and put the radiator in a sealed in little box I will build. This way the new fresh cold (well cold means a little bit more than room temperature) will get in from one side of the box, through the radiator and then instantly sucked out again. The heat exchanger in the air ventilation system will put the heat right back to the whole house. This way it will get evenly everywhere. (which it does now already with everything in the house, so dont tell me the house will heat up it happenes right now with my computer too and its not .) I maybe wont even need fans on the radiators because of the air flow from the system. (which i dont trust on I will put fans there dont worry).

 

Also since this will be the highest point There can be the reservoirs there to get rid of the air (because air rises in water, duh) and I can easily refill the whole loop from there if I want to.

 

What do you think about this new Idea? Do you think it is practical?

 

I also attached some pictures:

 

Air -out- part for the bedroom (which will be the gaming room you can see in the sketch later on)

 

Air ventilation system with heat exchanger

 

Loft (at this end the radiators will be)

 

Loft (at this end its too dark there are the tubes for the sucking out air thing from the bathroom, even if you cant see it trust me)

[vanished]

*snip*

 

That was a little hard to follow, but if you can setup an indoor rad in a way that the heat is sent straight outside through your heat exchanger/AC system before it can warm up the rest of the house, I can't think of any reason why that would be bad  Not to mention it simplifies the whole problem even more - no more dealing with condensation, self-regulating valves, unpredictable outside conditions, drilling holes in your house to run piping in and out, etc.

[Noorquacker]

Nnnnno, don't even think about the whole room water cooling, but if you do, use a much stronger pump. Normal pumps for a single computer expect much less than going through 2 computers and then going through the tubing.

[ToastyBear]

Look at aquarium pumps. They have units that can push as much water as you want it to. I would be afraid of how much pressure these things push. They are not that much more expensive than a D5 too. For the price of two D5, you can get an aquarium pump can pushes as much water as several D5 pumps.

 

The only thing you have  to think about is aquarium pumps usually have 3/4" threading which is much bigger than what CPU blocks use but you can probably find an adapter and these things will need to be plugged into the wall.

 

If you do choose to go with something like this. I would probably go for something thats around 1000GPH. Aquarium pumps don't drop as much pressure as the D5, they are designed to push water 6 feet high and through filters. Definitely don't buy something 6000GPH, that much destroy your computer.