TEC/Peltier CPU Chilled Water Cooling

Page 24 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
Just about every metal reacts with aluminum, but that just means you'll have to use a coolant that can handle that type of mixed metals situation.

Even copper water blocks and yes, copper water blocks are available.
 
so, it's either full ally, full copper, or special coolant...
i'm wondering if volkswagen-spec G12 coolant will be non-reactive, will have to look into that. it's a nice red colour and i have loads of it...
 


I got this one for testing, thought it might help for the next time
http://www.bing.com/videos/search?q=how+to+test+current+with+multimeter&FORM=VIRE16#view=detail&mid=231CD51EF3F91D84CDF7231CD51EF3F91D84CDF7
 
thanks for the assistance so far... just one more thing: roughly how many watts of cooling would be needed to exceed 6GHz with a 4790K?

please don't say it can't be done, someone already said i couldn't convert this ex-XPS cooler into a custom loop, and i'm already doing that; nothing is impossible, short of drinking gravel through a straw.
 


6Ghz with a 4790K is LN2 cooling range you are way past any normal type of cooling to reach those clocks stable.

In the study below their extreme water cooling is displayed even though they hyped it up, their actual coolant temperature was 4.4c, I can actually run that low of a temperature with my peltier cooling that this thread covers.

Without insulating the M/B but by adding and running high speed 60mm server fans to evaporate the condensation before it becomes a problem, I do not run that low of temperatures because I am not going to listen to all that noise those fans produce.

However I have already experimented down to 5c so I know the cooling can do it.

http://lab501.net/intel-core-i7-4790k-intel-pentium-g3258-overclocking-study/14/

Going to page 16 for the 4790Ks overclock results note the results of the first listing as that is the 4790K, all 4 cores with Hyper Threading enabled, you can see their results and the voltage they used to get there. That's with a 4.4c coolant temperature, which is below ambient, not below zero, the reason for my hype comment above.

http://lab501.net/intel-core-i7-4790k-intel-pentium-g3258-overclocking-study/16/

Below is the 4790K overclocking results cooled with Phase Change Cooling note the first entry 4 cores 8 threads and they still have not reached even 5ghz yet.

http://lab501.net/intel-core-i7-4790k-intel-pentium-g3258-overclocking-study/19/

Below is the 4790K overclocking results cooled with Dry Ice (DICE) they finally broke 5Ghz note the first entry 4 cores 8 threads.

http://lab501.net/intel-core-i7-4790k-intel-pentium-g3258-overclocking-study/22/

Below they did finally break 6ghz using Liquid Nitrogen (LN2) but they could not do it with 4 cores!
Note the voltage it took to do it.

http://lab501.net/intel-core-i7-4790k-intel-pentium-g3258-overclocking-study/25/



Many told me the cooling this thread covers couldn't be done and if I had listened to all the negativity I would not have this cooling in operation today.

On another note, my advice to you is to set a more realistic goal for overclocking the 4790K to maybe 5ghz, because 6ghz stable is way out there in expectations, you might want to read that entire overclocking study and glean what you can from their hands on experience overclocking the 4790K.



 
i will go through all of that when i have time. but it is already over a year old, and i am seeing frequencies up to 5.6GHz on ambient watercooling on other overclocking/benchmarking sites so i don't think 6 is too unrealistic. currently LN2 cooled systems are on the other side of 7GHz, and if i can come close to the clocks that phase change or dry ice can achieve, then i'll be a happy bunny.

 


I would love to see you a happy bunny but keep in mind you cannot believe everything you read on the net, there is a big difference between reaching an overclock number validation and a 100% rock solid stable system.

Examine net provided information to see whether they had to disable any cores to reach their posted results.

Those extremely high goals take extremely high voltages to run, which is fine for a simple test, but not 24/7 operation.

Without system stability you have nothing to depend on, bragging rights don't mean a thing if the machine is Blue Screening under a application load, and your HDD data is getting corrupted.

Don't take me wrong I am not attempting to talk you out of anything, go for it, I've been running 5ghz now for 3 years plus, just throwing in some cautionary logic, it's your hardware to do as you please. :)

Edit: By the way you can drink gravel through a straw if you pulverize it fine enough, but why would you want to?

 
if it will hit 6GHz for long enough to count, great. but i will turn it down a couple of notches for everyday use after i find out what it is capable of.

if you pulverise gravel, it becomes powder, and i don't fancy doing the cinnamon challenge 😉
i prefer to experiment and push back the line between what is and isn't possible.
 


There has been many lofty goals and extreme ideas proposed in this thread since it was first posted, almost none have actually reported back with any successful results achieving their proposed goals, so good luck in your experimentation adventure.

 


First of all a CPU-Z validation only means it was stable long enough to get the validation, it could have crashed immediately after saving the validation, you do not know that!

Now if he was showing the CPU-Z validation, a completed Intel Burn Test, and some benchmark scores he had completed all in a screenshot, then you could take that as possibly legitimate, but more than likely still not Crysis, FarCry, or any high end type of PC game stable.

However running 1.655v on a 22nm CPU, Seriously? :lol:

Does the poster cover how many CPUs he went through to get those numbers?

Don't you know in most overclocking competitions they usually have multiple CPUs to sacrifice and not every CPU can produce the numbers they're after?

They don't normally share that information freely.

Now if that's your next door neighbor you can go see what's going on yourself with your own eyes, that's different, but you are taking internet information as legitimate, without knowing any of the details of how it was done.

You can literally use a stopwatch to time the demise of that linked CPU!

A safe voltage at 5c is whatever will stay cool enough to keep the CPU from thermal throttling, and automatically applying built in measures to protect the CPU.

Which you have zero control over, the CPUs protective measures will activate in relation to the CPU temperature, and are not user adjustable!

Meaning there is no guaranteed safe high voltage in the 1.500v ~ 1.655v area, at all!

I would not approach or go past 1.500v with my own 22nm CPU, unless I no longer cared about using my CPU anymore and was ready to sacrifice it for some stupid OC number.

A safe 5c voltage is not reaching any 6ghz overclock level I'll guarantee you that.

Just curious?

What have you actually accomplished so far?





 


Uggh, (Rzisitance is Axeptrible!)

Me C Progress! :)

What's going on Moto?

 
working again, fighting to keep house and get money in order, applied for a nice job yesterday, cleaning data rooms with a good wedge of overseas work involved 🙂 :-D
if all goes well then Imma get my ass PAID to travel and manage cables haha, will show em my build logs if I get an interview 🙂
 


I hope it works out prosperously for you Moto, missed your input at Toms lately! Ry

 
so far, i have accomplished, not much computer related. i have recently decorated my staircase which has sucked up most of my available cash... i could have built a skylake monster for the cost of all that paint and flooring, but my wife is amazing and i'd like to hang on to her if i can...
also, the kids needed new school uniforms and my dog is going to be having puppies soon, and bloody christmas is looming... so financially i am in a hole until next year. oh well.
 


Yeah we all have higher monetary priorities over any computer spending.

 
it may take a lot longer than expected, and not reach the target speed i have set myself, but i am not throwing in the towel. it is going to happen one way or another...
i have some set aside for the small stuff, power supplies, memory, cooling components etc, and i have an i3 4150 which will get it running so the cooler can be set up and tested;
and when i can afford a proper CPU that i3 will be donated to my son for his build. so, yeah, the game is afoot.

failure is always an option, except failing to try - that just does not compute.
 


Please don't take my caution the wrong way.

I am not saying to give up, I am only saying set realistic reachable goals.

That's all I am saying.

 
hey ryan,
i completely didn't take any of your cautions or advice the wrong way, and i'm not giving up any time soon :)

i understand that 6GHz may be unattainable; it's a nice round number, though...
but even if it's beyond feasible, i'm not going to be disappointed with whatever speed i can achieve, unless i get a dud chip that won't even go past the turbo frequency...
i don't mind being proved wrong, but i believe that your cooling system that inspired me to start this project is not the weak link. the CPU is the thing i have no control over.
anyway, i have read Imarkler's thread over on OCN, and that has given me much to think about, like is 15v going to be more of a cooker than a cooler? i may substitute boosters for adjustable voltage regulators to lower the input, rather than increase it...

things are going to get experi-mental :)


 


That's great! :)

IMO a 5ghz goal for a 4790K is possibly reachable with this peltier cooling, even a modified design of it, but you are 100% correct the chip quality is the most important factor and will be the determining factor as to how far of an OC is reached, no matter what cooling you have.



 
Status
Not open for further replies.