Water cooling system with Peltier thermoelectric support

[Silver Surfer]

Here is the Odd thing. You seem to have accomplished the very thing I am talking about here: http://www.overclock.net/t/1633988/chilled-water-cooling-vs-3-0-build-log Isn't that you??? You claim to have accomplished this with 300 watts of power on 3 TEC. Am I so wrong? Why are you coming at me with a pitchfork?

Please disregard the Sock Puppet comment.

Yes that is my setup, and I am not trying to pitchfork you, it seemed odd that you were defending someone that's pretty much discounted all attempts to aid or advise or show an operational system.

But there's good reason for that as this section is actually covering traditional water cooling.

There is an Extreme Cooling Section in this forum that covers what you're interested in as far as TEC cooling, but it's not very active at all.

I suggest you hang your hat with others that are actually using this type of cooling as they have the experience to share.

You may get advice here from those that have dabbled with it in the past but not actively using it today, that said.

You'll get more advice than you probably would ever need from the Specialized Cooling : Peltiers/TEC, cooling section at Overclock.net as there are many there using just about any way you can dream up to run peltier cooling.

As far as efficiency is concerned that word doesn't belong in the same sentence with Peltier or TEC, but my present setup is the most efficient setup to date that I have built.

Most use TEC builds for stable high CPU overclocks and efficiency they could care less about.

 

[denali]

Thanks for that wise reply. Everything you've said makes great sense and I'll follow it. I'm really going to take this slow, even though I have such a great urge to order everything right now before I truly understand what I'm doing and before I have an actual plan. But I am not even going to order the MB, CPU, or RAM as prices seem high and I don't even know how to put together a custom loop, or even order parts for it. So that's where I'm at; Very High Enthusiasm, very little experience, much desire to get everything exactly right, and a need to be gentle with my family's budget. Thanks for the sage advice!

 

[Silver Surfer]

Thanks for that wise reply. Everything you've said makes great sense and I'll follow it. I'm really going to take this slow, even though I have such a great urge to order everything right now before I truly understand what I'm doing and before I have an actual plan. But I am not even going to order the MB, CPU, or RAM as prices seem high and I don't even know how to put together a custom loop, or even order parts for it. So that's where I'm at; Very High Enthusiasm, very little experience, much desire to get everything exactly right, and a need to be gentle with my family's budget. Thanks for the sage advice!

I am very familiar with the urge to order and with the constraints of budget, but in your experience situation, I urge you to spend some time learning and get that wisdom under your belt. You really don't want to buy something on an impulse and it gets shelved to possibly use later but does you zero good at the time, been there done that. Experimenting with non traditional cooling of PCs can get quite expensive, especially if the idea is a failure and turns out to be a lesson learned the hard way with no cooling to show for the money spent, also been there done that! If you have any questions feel free to post in my 3.0 build at OCN or PM me there, Good Luck to you! Ryan

One caution for you at OCN, you'll get a massive amount of ideas thrown your way, some are good and some are ridiculous, my advise is take your information from those running fully operational and tested systems, otherwise they're just sharing untested ideas.

 

[denali]

Nah. I'm just going to do exactly what you've done. Why reinvent the wheel?

Seriously. Sweet Rig. Nicely Done.

Thanks

 

[Lochekey]

I agree with what Silver has said above he will be one of your best resources in this adventure.

My suggestion though would be to build your system based around a traditional water cooling loop first. You will need this experience before attempting something like Silver has done and you really will not be wasting money as most of the parts can be reused in the TEC system. You may also find that a properly set up water Cooling system is all you need for good temps and quite operation.

 

[Silver Surfer]

I agree with what Silver has said above he will be one of your best resources in this adventure.

My suggestion though would be to build your system based around a traditional water cooling loop first. You will need this experience before attempting something like Silver has done and you really will not be wasting money as most of the parts can be reused in the TEC system. You may also find that a properly set up water Cooling system is all you need for good temps and quite operation.

I have to agree, that's a very wise suggestion, he needs to fully understand water cooling first, as there's a lot to learn, and everyone of us has learned something the hard way!

Nah. I'm just going to do exactly what you've done. Why reinvent the wheel?

Seriously. Sweet Rig. Nicely Done.

Thanks

Thank You for the compliment, however you proceed, I wish you the best!

 

[Deous]

Important update about TEC project status

Well....lot to talk about - my TEC project took a lot of time to find out that water cooling is not very efficient way of cooling lol.
Anyways I started with 3x180W TECS, huge radiator and tank - turned out water was cooled but for few minutes and then actually got warmer and warmer.
So what I did was separating cold loop from hot loop - 2 circulation pumps. Well it still just prolonged cool water time a bit but everything went back to same shyt.
Added 3 more TECs - just increased time of cooling but temporarily. Water was pretty hot in loop II.

What I inferred is:

Water cools good but also retains heat too long to dissipate it.
Water needs to be cooled longer to certain temperature to effectively retain low temperatures when mixed with warm/hot water
There is a ratio of cooled water that can be mixed with hot water (from CPU/GPU/mb) to achieve cooling effect.
Therefore only efficient way is to cool water with refrigerant/compressor based system and loop it normally to PC.

So what I am doing now is taking apart old fridge with condenser,compressor and evaporator and placing condenser in the
small water basin that will supply water to CPU.
In the future I probably will just hook up the copper lines directly to CPU block - needs welding copper and refilling refrigerant.

Also I must add that TECs were devouring lot of Amps - 12715 Chinese models were drawing 7A-9A/12V (should be 10A+) - poor quality
That took 25A+ and with 3 more you can easily exceed 50Amps on your PSU which gives you 520+ watts of constant power just for TECs.


Cheers

 

[Silver Surfer]

Well you've had quite an adventure haven't you?

 

[Deous]

Well you've had quite an adventure haven't you?

More like finding out how effective water cooling is and why it is so trendy these times.
I don't see any bigger advantages with that except it looks neat and maybe it is more quiet when you move pump and radiators away from case.
All those w.cooled systems are a simple lie because they get hotter over time no matter what.
Water is a bad medium for instant cooling. Better use liquid gas with low pressure to liquefy it.
I'd try to use a simplest of simplest - compressed liquefied air in simple ac system.
I don't know what engineers do today but apparently they just rip masses off with useless bs.

 

[Lochekey]

More like finding out how effective water cooling is and why it is so trendy these times.
I don't see any bigger advantages with that except it looks neat and maybe it is more quiet when you move pump and radiators away from case.
All those w.cooled systems are a simple lie because they get hotter over time no matter what.
Water is a bad medium for instant cooling. Better use liquid gas with low pressure to liquefy it.
I'd try to use a simplest of simplest - compressed liquefied air in simple ac system.
I don't know what engineers do today but apparently they just rip masses off with useless bs.

I am confused as to how you are drawing this conclusion. A properly set up water cooling system can be quite effective. My personal system can dissipate around 400w while mantaing a deltaT of 2-3 degrees C. If I am willing to accept a deltaT of around 10C my system will dissipate 900watts of heat. It sounds like you are trying to build an advanced system without having a good understanding of the basics first.


You say your heat soak of your water was presenting an unacceptable deltaT. What was the gain in water temperature over ambient? I would also like to know what type of heat load you were putting on the system and how much raddage you had to dissipate that load. A general rule is a 120mm radiator will dissipate 100 watts of heat at a deltaT of 10C. If you want a lower deltaT you need to have more radiator surface area per 100 watts of heat. If you are amiable towards a deltaT of 10C then you just need to make sure you are having enough radiator to handle your load at a rate of 120mm/100 watts. If you exceed this you will see a drastic increase in your deltaT that will eventually stabilize at some unknown point based on your inputed wattage vs your raddage.

One last point that I stated earlier in the thread. If you are not keeping your hot and cold side of the loop seperate when you are trying to use a TEC to reduce cold side water temp you are shooting yourself on the foot.

 

[Silver Surfer]

Deous, You are invited to click the link in my sig., of my Chilled Water Cooling 3.0, as it is successfully cooling my Intel i7-8700K 6 core CPU with hyper threading enabled and overclocked to 5ghz all cores. It is P95 stable with the highest core temperature of 70c on a CPU that has not been delidded, cooled by a triple peltier setup and it is doing everything you're looking for. SS

 

[Soulcatcher668]

Important update about TEC project status

Well....lot to talk about - my TEC project took a lot of time to find out that water cooling is not very efficient way of cooling lol.
Anyways I started with 3x180W TECS, huge radiator and tank - turned out water was cooled but for few minutes and then actually got warmer and warmer.
So what I did was separating cold loop from hot loop - 2 circulation pumps. Well it still just prolonged cool water time a bit but everything went back to same shyt.
Added 3 more TECs - just increased time of cooling but temporarily. Water was pretty hot in loop II.

What I inferred is:

Water cools good but also retains heat too long to dissipate it.
Water needs to be cooled longer to certain temperature to effectively retain low temperatures when mixed with warm/hot water
There is a ratio of cooled water that can be mixed with hot water (from CPU/GPU/mb) to achieve cooling effect.
Therefore only efficient way is to cool water with refrigerant/compressor based system and loop it normally to PC.

So what I am doing now is taking apart old fridge with condenser,compressor and evaporator and placing condenser in the
small water basin that will supply water to CPU.
In the future I probably will just hook up the copper lines directly to CPU block - needs welding copper and refilling refrigerant.

Also I must add that TECs were devouring lot of Amps - 12715 Chinese models were drawing 7A-9A/12V (should be 10A+) - poor quality
That took 25A+ and with 3 more you can easily exceed 50Amps on your PSU which gives you 520+ watts of constant power just for TECs.


Cheers

I need a diagram of this set up. It sounds like you are dumping the heat into a tank of water and expecting it to just magically disappear.

 

[Deous]

I am confused as to how you are drawing this conclusion. A properly set up water cooling system can be quite effective. My personal system can dissipate around 400w while mantaing a deltaT of 2-3 degrees C. If I am willing to accept a deltaT of around 10C my system will dissipate 900watts of heat. It sounds like you are trying to build an advanced system without having a good understanding of the basics first.


You say your heat soak of your water was presenting an unacceptable deltaT. What was the gain in water temperature over ambient? I would also like to know what type of heat load you were putting on the system and how much raddage you had to dissipate that load. A general rule is a 120mm radiator will dissipate 100 watts of heat at a deltaT of 10C. If you want a lower deltaT you need to have more radiator surface area per 100 watts of heat. If you are amiable towards a deltaT of 10C then you just need to make sure you are having enough radiator to handle your load at a rate of 120mm/100 watts. If you exceed this you will see a drastic increase in your deltaT that will eventually stabilize at some unknown point based on your inputed wattage vs your raddage.

One last point that I stated earlier in the thread. If you are not keeping your hot and cold side of the loop seperate when you are trying to use a TEC to reduce cold side water temp you are shooting yourself on the foot.

Yeah right, I used huge high quality car radiator and 14" curved fan - not helped much unless the fan runs 90W of power and sounds louder than a busy airport.
TEC efficiency is not good at all for such power. Fast refrigeration of water with compressor is superior to water cooling.
I saw an article they constructed WC system but with compressed hot steamy water that dissipates heat very good.
Besides radiator has limited efficiency and heat at the end must be given out into air anyway.
I am only interested in quiet systems below 40db - otherwise is just bad idea.

 

[Lochekey]

Yeah right, I used huge high quality car radiator and 14" curved fan - not helped much unless the fan runs 90W of power and sounds louder than a busy airport.
You can not compare a cobbled together car radiator and fan to a typical watercooling radiator and fan setup nowadays. The 2 radiators are designed for different purposes and perform better under different conditions. A car radiator is designed to handle a high heat load at a large delta T in a system with the liquid under pressure. The water cooling radiators are designed to be used under minimal pressure at a low delta T. The performance oft he 2 are going to vary widely under different circumstances. IF you start talking opld style heater cores the lines get a little blurry but that is for a different time.


TEC efficiency is not good at all for such power.
I understand your thread is about a TEC system but I am tryting to specifically adress the issue of a properly designed watercolling systerm. Lets not blur the point of my posts by bringing the other components of the system into the mix at this time.

Fast refrigeration of water with compressor is superior to water cooling.
When you say "Superior" what is your measure that you are basing this on? Total power consumption, temperatures achieved, degree reduction per watt? You need to put this into context for me.


I saw an article they constructed WC system but with compressed hot steamy water that dissipates heat very good.
Can you link this article I would be intereseted in seeing this. Unless they are doing something abnormal the water would quickly cool beyond the boiling point and you would lose any advantage that is happening due to phase change. Now if the system was under a slight vauum thus allowing for a lower boiling point they may have been able to achieve something but I don't know if such a system woiuld be possible without more research.


Besides radiator has limited efficiency and heat at the end must be given out into air anyway.
The same could be said for any cooling system. The heat eventually has to be transferred to the surrounding enviroment. Every piece of equipment is going to have a specified efficiency. The triuck is understanding what the is and how to properly design a ssytem around it.


I am only interested in quiet systems below 40db - otherwise is just bad idea.
This goes back to my earlier statement about a properly designed system with the correct components. A WC setup can achieve a properly cooled system at under 40db if you have sufficient radiator space and money.



If you could post a diagram of your system layout, I think it would be very helpful in seeing where the weak link in your system chain is. Some pictures of your overall setup would do wonders for this discussion as well.
Thanks - Lochekey

 

[Deous]

Water cooling just by itself without tec (fan rad.,pump,block) is going to cool to ambient/room temps. or a +-2C below
WCS on the market are a tricky lie - here example of honest guy

When it comes to TECs - I wish the industry was making the whole metal elements - not only the ones with ceramic ends.
I wait for compressor cooling era when WC is finally finished

 

[Lochekey]

Water cooling just by itself without tec (fan rad.,pump,block) is going to cool to ambient/room temps. or a +-2C below
WCS on the market are a tricky lie - here example of honest guy

When it comes to TECs - I wish the industry was making the whole metal elements - not only the ones with ceramic ends.
I wait for compressor cooling era when WC is finally finished

Well I never said you will get below ambient. The best you can do is hold water temp at about 1 degree above ambient.

I do not disagree with the statements made by the "youtuber" but once again a larger custom loss will not have the same level of heat soak that a small AIO has. If you also listen to him he is talking about the misunderstanding of most consumers and people who are not testing properly before posting online. Good reviewers and manufacturers are not putting out any information that is contradictory to this video.

I doubt we will see ohase change systems overtake WC any time soon unless there is a major breakthrough in phase change design/efficiency.


Like I said earlier some diagrams/pictures would be nice.

 

[Silver Surfer]

I am only interested in quiet systems below 40db - otherwise is just bad idea.

I wait for compressor cooling era when WC is finally finished

Compressors aren't quiet, and you're not going to see the end of the water cooling era, if anything you'll see more.

 

[Deous]

Compressors aren't quiet, and you're not going to see the end of the water cooling era, if anything you'll see more.

High quality compressors certainly are quiet. When WC is finished I'm sure the industry will make the super quiet and efficient ones.
Only problem is the pressure but some plastic pipes easily withstand 8-10 MPa where AC needs around 300psi (2MPa)
Maybe a simple PEX tubing would do here fine except it would require better sealing and strong good quality PVC. (not quite sure)
Once they make nice cooling compressors for plastic pipes with refrigerant cooling PCs - WC will disappear in one day.

 

[GTXJackBauer]

Can a mod move this thread to the 'Extreme Cooling', thanks. Don't need this negativity on custom liquid cooling in the water cooling section from someone who doesn't know what they are talking about.

 

[Alaric]

The equipment and skills needed to pull a vacuum in a just built refrigeration system alone will keep the majority of people from taking on a project like that, I would think.