TEC/Peltier CPU Chilled Water Cooling

[4Ryan6]

Just an idea here, in the event that the cooler can't handle the outside presure, even with the fewer holes.
Why not use 2 chests? One inside the other. Have your current cooler as is, but then have a second one at 0.5 bar all around it. You could use your micro-controller to alternate some valves or something and make sure you keep the presure stable in both cases. That way, the outside presure is much less. Would be much bigger and more points of failure, but it might be a solution, in the event that a cooler can't handle the presure.

He'd be better off just to get a cooler that could handle it, like a Yeti, the literally best cooler on this planet!

But they are not cheap!

 

[4Ryan6]

Can we get back on this thread topic please.

 

[toolmaker_03]

good point I do have a few questions.
how are you coming on getting this to be a smaller build?
if you have some ideas, sketches, or even just a passing thought I would like to hear it.
I have noticed that there is not much on the third TEC in this system, is it because even with the high clocks on the CPU, the temps on the water are too low to be useable?
the pumps are D5's like mine so I was wondering if you have taken off one of the pump backs to see if there is any internal damage, on the electronics, after 3 years of use at the low temps?
as well you have made this system automatic now, so it is a set it and forget solution, this was one of the objectives for this system, my question is what do you have planed next as a upgrade, or improvement for this system?
yea its 20 questions, but I think that this information might be of use to my self, as well as others.

 

[4Ryan6]

good point I do have a few questions.
how are you coming on getting this to be a smaller build?
if you have some ideas, sketches, or even just a passing thought I would like to hear it.

I have been researching low profile heat pipe air coolers the same style as the ones you are using on your relays, the only problem is finding them that can handle the load, the low profile height with water block will take up less room, so far I have found what I think will work but they're 80 bucks apiece, that's 240 plus shipping.

Presently I have home projects going on, that are not only taking my free time but are also taking any extra money that I could invest into those coolers, as they are an experiment, because if they cannot handle the heat load they'll go to some other computer project, which is not in the plans!

I'm also facing the old saying, "if it ain't broke, don't fix it", and I'll have to be 100% fully committed into it to tear down my present setup and replace those coolers of a project over 3 years running, of course I could always build from scratch but that would be quite expensive.

My present setup is easily transferred to a new computer upgrade, it is already on the 2nd computer upgrade now.

I have noticed that there is not much on the third TEC in this system, is it because even with the high clocks on the CPU, the temps on the water are too low to be useable?

No I use the 3rd TEC everyday now to automatically maintain a constant 10c target coolant temperature.
You are correct there's not much at all covering the 3rd TEC Assembly, maybe I'm just getting lazy, but I tend to respond at Toms when I see an interest?

If there's no interest why waste my time covering it?

the pumps are D5's like mine so I was wondering if you have taken off one of the pump backs to see if there is any internal damage, on the electronics, after 3 years of use at the low temps?

I wore out the pump cooling my graphics using the Watercool MO-RA3 rad, and inspected everything in that loop, didn't see anything but normal pump wear but that pump was 4 years old, replaced the pump, tubing, coolant, and also the water block on my graphics card, not because of any problems inside the water block, but because of the water block making only partial contact with the graphics card.

That was about 6 ~ 8 months ago, it is covered way back in the thread.

As far as the peltier side I haven't inspected the pump but have changed from an XSPC Raystorm to an EK Supremacy EVO, broke down the Raystorm and inspected inside it, no problems inside the water block, just in case you're curious the EK Supremacy EVO cools 5c better under the same coolant temperature testing.

as well you have made this system automatic now, so it is a set it and forget solution, this was one of the objectives for this system, my question is what do you have planed next as a upgrade, or improvement for this system?

I would love to get it all inside of one huge case, the only cases I've researched that could do it is MountainMods, I have ideas on improving my insulated reservoir, but it will occupy more space than it does now.
I do have some experimentation I'm pondering, but I won't release any information regarding that unless it turns out to be 100% successful, then I'll open another thread and share it with everyone.

yea its 20 questions, but I think that this information might be of use to my self, as well as others.

 

[toolmaker_03]

ok so its 2 X 250W TEC's on all the time, and a third on intermittently to keep the water at 10C?

these TEC's have a temp swing of about 60C so I am guessing that your hot side is about the same at 60C?
I say this because with the third TEC on all the time that your water temps would be around 0C with is too cold.

I sure you have tried this but I will ask anyway have you tried changing the fans out for faster ones, on the air coolers for the two TEC's that are active all the time, to see if keeping them a little cooler would enable them to keep the system cool without the need for the third?

I ask because this is kind of what I learned from this setup, was that one could control how cold the cold side gets, by controlling how hot the hot side gets.

not that it is necessary, just wondering if you have tried it.

 

[4Ryan6]

Actually it is 2 x 200W while being supplied 12v vs 15.6v spec.

I tried numerous fans, and fan speeds until I discovered the balance ~ of output coolant cold ~ relation to ambient room temperature.

Thinking wise?

You may be thinking the 3rd peltier is constantly used but it isn't, it is constantly controlled but only activates under certain room temperature conditions.

The high overclock I enjoy daily running is the reason because the 5ghz 3770K OC can not be stabilized with ambient temperature, been there in testing.

5ghz idling is no problem at ambient, and when I am referring to ambient with this cooling that means my coolant temperature is the same as my ambient room temperature, which would be a 0 DeltaT.

To support the 5ghz OC under load I have to be at least 5c below ambient, my 10c coolant temperature target is at least 15c below ambient room temperature.

These are actual operating numbers.

Edit: All of the above is done with zero condensation.

 

[toolmaker_03]

yea, I kind of thought that you had done that, I will have to play with mine as well to find a good setting that works for both idol and load temps.

I am sure it took hours of tinker time, playing with the valves, knobs, and switches, to find that balance.

 

[4Ryan6]

I am sure it took hours of tinker time

:lol: More like 2 weeks.

The peltier is finicky, it has a tendency to stall out regarding cold production, you have to have just the right balance of hot side/cold side, if either gets too hot or cold it affects the other side, once you find the balance the cold side will constantly produce cold for you, and that's what you need to build up cold in the insulated reservoir.

The insulated reservoir is your cold storage and if your hot/cold balance of the peltier is off, your cold build up is happening intermittently like a roller coaster of high and low output, instead of constant output.

It takes time to find the balance.

 

[toolmaker_03]

well it does seem to work well, and this is a good option for many reasons, one no condensation, so no need to insulate the motherboard. for this reason alone it's a good build, than there is the cost factor, it is a low cost solution for TEC cooling, that too makes it appealing. and now it's all automatic, so one more reason that this setup gets a lot of attention. I have noticed like four setups now with the DYI reservoirs, those are very popular, I even made some, thanks for doing the R&D on that :lol: made my life a lot easier. personally over the next couple of years I think we will see a couple of these setups come out of the closet, I truly think that there are a few who are building them, or even have them built already, but like too many people they are too shy to show everyone what they built.
these types of builds are large and ugly, so I think that this plays a factor in others willingness to share there builds, and experiences. There is also the possibility that some have tried and failed for whatever reason, so they don't like to admit that part either. me on the other hand, I don't care, if it doesn't work, I say, I can fix that. I am the type to spend the money and time to do so.
so yea, you did a great job with this build, it has come a long way from the ice chest with ice in it, to now.

 

[4Ryan6]

well it does seem to work well, and this is a good option for many reasons, one no condensation, so no need to insulate the motherboard. for this reason alone it's a good build, than there is the cost factor, it is a low cost solution for TEC cooling, that too makes it appealing. and now it's all automatic, so one more reason that this setup gets a lot of attention. I have noticed like four setups now with the DYI reservoirs, those are very popular, I even made some, thanks for doing the R&D on that :lol: made my life a lot easier. personally over the next couple of years I think we will see a couple of these setups come out of the closet, I truly think that there are a few who are building them, or even have them built already, but like too many people they are too shy to show everyone what they built.
these types of builds are large and ugly, so I think that this plays a factor in others willingness to share there builds, and experiences. There is also the possibility that some have tried and failed for whatever reason, so they don't like to admit that part either. me on the other hand, I don't care, if it doesn't work, I say, I can fix that. I am the type to spend the money and time to do so.
so yea, you did a great job with this build, it has come a long way from the ice chest with ice in it, to now.

Thank You, toolmaker_03

It has come a long way since the ice chest days!

The ice chest days allowed me to discover there was an operating range below ambient but above condensation, which gives an instant overclocking edge!

All CPUs, Intel and AMD are designed to run in ambient environments around the world, some of those environments are colder some are hotter, and the CPU tolerances are designed to run in those ranges.

Think logically about what happens when you run coolant temperatures lower than the trigger points of the CPUs built in ambient protection?

Below is a direct quote from Section 2 of CompuTronix Intel Temperature Guide, (A Very Nice Piece of Work He has Done in the Guide by the Way!), Note the Ambient Range of Intel CPUs.

Section 2 - Ambient Temperature

Also called "room" temperature, this is the temperature measured at your computer's air intake. Standard Ambient temperature is 22C, which is normal room temperature. Ambient temperature is a reference value for Intel’s Thermal Specifications. Knowing your Ambient temperature is important because Ambient directly affects all computer temperatures. Use a trusted analog, digital or IR thermometer to measure Ambient temperature.

Here's the temperature conversions and a short scale:

Cx9/5+32=F ... or ... F-32/9x5=C ... or a change of 1C = a change of 1.8F

30.0C = 86.0F Hot
29.0C = 84.2F
28.0C = 82.4F
27.0C = 80.6F
26.0C = 78.8F Warm
25.0C = 77.0F
24.0C = 75.2F
23.0C = 73.4F
22.0C = 71.6F Standard ... or ... 22.2C = 72.0F
21.0C = 69.8F
20.0C = 68.0F
19.0C = 66.2F
18.0C = 64.4F Cool

When you power up your rig from a cold start, all components are at Ambient, so temperatures can only go up. With conventional air or liquid cooling, no temperatures can be less than or equal to Ambient.

As Ambient temperature increases, thermal headroom and overclocking potential decreases.

So what happens when you run lower constant supplied temperatures than the lowest design specifications?

What happens when ambient room temperature is irrelevant, and you can constantly control the exact coolant temperature you desire to run at.

Think about it!

You are soon going to discover this yourself with your build.

 

[toolmaker_03]

so I like to leave my computers on all the time, only shutting down for maintenance, do you leave this system on all the time?

 

[4Ryan6]

No I do not, I turn it on to use it, and shut it down when I'm done.

 

[toolmaker_03]

well, I guess that I will have to start getting used to turning it off, when I am done with it. thanks.

 

[4Ryan6]

It pulls too much electrical power to just leave it on to just sit there running, and yours is going to pull way more power than mine does!

 

[toolmaker_03]

ok so when you start the system up, do you activate the TEC's and water pumps first?

that is what I was doing, I was letting the system start to cool the water, than I would start the PC, I also did this at the shut down of the PC and TEC's, I would leave the water pumps on, and let the system heat up on the cold side, and cool down on the hot side after I turned off the PC.

 

[4Ryan6]

When I push the power button everything starts, that's all 3 TECs running, it takes about 20 minutes to reach down to 9.9c coolant temperature from ambient, which is where the temperature is by morning after being off all night, and the 3rd TEC cuts off and on when it is needed to maintain the control setting, the relay cuts out at 9.9c and in at 11c.

That is where it is set right now.

It really depends on the CPU overclock ~ CPU TDP wattage, and how cold the coolant has to be to support the overclock you intend to run.

If your cooling is completely successful I'd personally be shooting for a 6ghz overclock at least, to maintain that high of an overclock you may be forced to pre-start the cooling.

 

[toolmaker_03]

wow, cool I will try for that 6Ghz. (not trying for that high, but if the system can handle it, why not)
I really was thinking that you might have needed to do that, for the same reason provided, but I guess not, I probable did not need to either, just being over cautious I guess.
TEC's do cool down rather quick.

 

[4Ryan6]

You could reach 6ghz by dropping the temperature after the boot and allowing the CPU to heat up, however you will not be able to run 6ghz 24/7 stable, unless you keep it cold and running all the time, it will take sub zero temperatures to support a 6ghz overclock with any Intel CPU out there.

Probably at least -10c or lower, simply because of the CPU voltage it will take to run that low. (That's sharing what I have learned so far stabilizing a 24/7 5ghz overclock that can not be supported by traditional cooling)

When I was running ice in the cooler I tried a few times dropping the temperature real low before booting and sometimes it took 3 or 4 boot attempts to get to the desktop so then I started booting at whatever the water temperature had risen through the night and the multiple booting went away.

It is the same effect I have now with the TECS as my coolant temperature returns to ambient when the machine is off.

When I say ambient with a chilled water cooling system that is true ambient the exact same as room temperature but in traditional water cooling it would be a 0 DeltaT.

 

[toolmaker_03]

well by using the TEC calculator its says that I would be running at -10C on my water, with the clocks that I am intending to run. 5Ghz on the CPU and 930Mhz on the GPU's. but I don't what the actual component temps will be so I am cautious about going over those clocks until then.
but as long as the CPU stays below 55C at load, I would be happy with whatever clock is possible. the GPU's are maxed at 930Mhz I cant provide any move voltage than 2.1V to the GPU, so this is as high as they can go.
this is why I give myself that range of water temps from -10C to 10C I know that it will settle somewhere between there.
-10C being best case, assuming that I have done everything correctly to get the most out of the TEC's.
as for starting the PC I was letting the water get to between 20C and 15C before I would start the PC and let the water temps drop from there.

 

[4Ryan6]

What CPU is going to require you to go down to -10c to support a 5ghz OC?

 

[toolmaker_03]

none, I was only stating what the TEC calculator says my system would be at. with would mean that if it is right, I should have a little head room left to clock a little further, or I can adjust my hot side to rise the water temps a little, so things don't get to cold at idol, or I might be able to run like your are doing now, with two TEC's active and using the third to keep water temps stable at 0C. I really don't know what the best option would be right now. and this is all assuming that the system will even run that cold at load, when I am done. I can hope that I did that good of a job on the setup.

 

[4Ryan6]

Your original intention was to run sub zero, that's why you are insulating everything and going to all the trouble you are, and you are going to have quite a bit of experimentation once it is all completed to see exactly what you've created can do.

No matter what you hope will happen, until it becomes a reality you're just guessing, calculating, hoping, and working toward your end goal, I just hope for you that it all works as you're planning and that you can be resilient if it does not, and to adopt the old saying.

"If at first you don't succeed, try, try, again", and do not get discouraged!

So far you have invested a lot of time and money, and sometimes the time investment can be more devastating than the money, and that is when you learn the raw reality, and have to go back to square one and start over.

I have experienced that many times to get to where my project is now, you have to be more determined to prove your idea will work to get beyond having to start it all over again, and that sir can get so frustrating you just want to quit.

But don't give up!

I'm concerned for you, because your project is so massive, going back to square one is a long way.

So I hope for your sake it all works as you've planned!

 

[toolmaker_03]

so do I. all I need now is a little magic to make it happen.

 

[computers_are_freaky]

Quick question about higher OC with sub zero temps:
Would this hold true for GPUs as well?

I can't see any future where my PC will have an unlocked cpu (at least in the next 3-4 years) but I have reached the max I can get out of my GPU. (Managed 10% increase in score with a GTX 950 without voltage increase). I would however like to get a lot more performance out of it in the future, if and when I try this TEC cooling, but would it actually help in the case of a GPU, or only with CPUs?

 

[4Ryan6]

I can't see any future where my PC will have an unlocked cpu (at least in the next 3-4 years) but I have reached the max I can get out of my GPU. (Managed 10% increase in score with a GTX 950 without voltage increase). I would however like to get a lot more performance out of it in the future, if and when I try this TEC cooling, but would it actually help in the case of a GPU, or only with CPUs?

If you have no unlocked CPU in your future to OC, and you do not have one now, then you do not need anything but a stock air cooler!

In the next 3 or 4 years I have plans of shrinking down this present cooling to do the same job in a smaller package but that experimentation takes money and it is definitely going to be done to take OC advantage of an unlocked CPU.

A GPU gets tremendous OC advantage just from water cooling it with traditional means, meaning a radiator, etc., like my GPU is cooled, but if you cannot even acquire an unlocked CPU to overclock, how are you going to acquire even the water cooling for your GPU?

I doubt very seriously you will ever even attempt to try TEC cooling, and I think you know that as well!

There is a big difference between the sideline Ohhs and Ahhs, and actually building something like TEC cooling.

You'll be better off reading this.

http://www.tomshardware.com/forum/id-2196038/air-cooling-water-cooling-things.html

If you had a clue as to how far I have pushed my GTX Titan overclocking it on radiator water cooling vs just how much that GPU cost me, you'd have a freaking heart attack! :lol: