Exploring Below Ambient Water Cooling

Page 24 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
No need man, had you forgotten your exemption certificate?
You alone are free to do whatever the hell you like with no need to be worried about cable hassle from me,
on the grounds that your work is awesome
I can't remember where I posted it but that was the gist of it 😛
Seeing your Pc has made mine not seem as bad to the other half hehe

Moto
 
Ahh the "Mad Scientist Clause", temporary exemption from details!

Sometimes I'm afraid to show pictures, because it may be changed a couple of days later, but for now the last pictures are what I'm presently running.

Started up this morning with 2nd peltier switched on @ 5:30A water temp in the cooler was 19.1c switched off the 2nd peltier at 5.53A, water temp in the cooler was 14.4c, my target run area for the 4500mhz OC.

A 4.7c temperature drop of the cooler mass in 23 minutes, with the 2nd peltier engaged.

That 13c ~ 15c range, keeps me at a CPU idle of apprx. 18c CPU load around apprx. 40c, this setup is doing an excellent job of cooling, it keeps me on an average of 9c below ambient room temperature daily.

Not bad eh! :)
 
Something very interesting to Note:

For the past couple of days my water temperature in the cooler has maintained 14.4c ~ 14.5c obviously even though the 2nd peltier cooler is not energized, the cooling fan on the heat sink is still running, and is producing some stabilizing effect.

A radiator would be fighting against the active peltier cooler trying to bring the temperature up to ambient, because the chilled water is actually flowing through the radiator, but the heat pipe heat sink is still affecting the temperature even though it only has ambient room temperature air flowing through it.

Under these circumstances it appears to be taking the role of temperature regulator as the single peltier does have the capability of dropping the temperature of the water down to 12c.

It will be interesting to see if this continues long term and I don't see why it wouldn't, as normally once these peltier peculiarities make themselves known they stay constant.

This is a very interesting discovery that I will continue to monitor as being in a steady dependable solid operational temperature range is beneficial for so many reasons.
 


????? So now I'm exempt? 😱

Just so you know I have to RMA my M/BD due to not recognizing a new SSD among other issues.

Once I have it back I will cool my CPU and GPUs with my house water. In the winter I can get really good temps from that ground water. :sol:
 
Yes mate, you are also exempt from cable fury,
I'll let you both into a shameful secret, once I rebuild my main rig theres a high likelihood that both cables and tubes will be on show, /cries
I'm pushing 4.6GHz on my 975Be and I can't let a few wires get in the way of my cooling needs so sacrifices have to be made 😛
Moto
 
Hey guys!

Still waiting for the peltiers and the controller to arrive from china and I have a XSPC water cooling kit just sitting there.

Anyway I have a question relating to the water in the loop freezing since I won't have a large amount of heat being drawn out like a CPU or gpu. Should I add anti freeze in the loop and if so how much?
 
The best thing to do is get a thermal temperature probe for the cooler and monitor the water temperature in the cooler, because before you reach freezing point you'll have serious condensation problems to deal with.

Once you start to drop below 10c actual water temperature condensation begins to form, all that is covered in this thread.

Could it actually freeze the water of course it can!

That's why you monitor what's going on, this is not a set it and forget it cooling solution!

If you're even considering automotive anti-freeze you need to also consider fully insulating your motherboard.
 
wow, there's been a lot since I last visited this thread 8 months ago... Ryan, hats off mate. You've been actively pursuing this endeavor nicely, quite nicely indeed!

Obviously that 2nd last line is not for peeps with a CLC 😀 who think blowing air through them rad is a tonne of work 😛

*are you a Thermalright brand ambassador? cos of all those fans and I think 2 coolers..or is that a Hyper 212 inside the case cooling the TEC.
** I still say your setup is a nice way of commendeering a room :lol:
 


Yes there has been quite a few changes since you last looked in!

Presently running 2 independent loops, the GPUs are radiator cooled using an XSPC Res/Pump, through a Black Ice 240 and 360, with 2 Heatkiller full face 580GTX cooling blocks. The radiator fans are Thermalright 2000rpm, controlled by a Sunbeam fan controller, both rads are shroud equipped and pulling the air through the rads. This setup allows a 40c GPU load vs 80c stock air cooling at the GPUs stock clock, and allows a 200mhz clock increase of the GPUs with the temp increasing to around 47c.

The CPU loop is powered by a Swiftech MCP655 pump to a XSPC Raystorm CPU water block, the return line goes through TEC assembly 1, then through TEC assembly 2, then to the Res/Insulated Cooler maintaining approximately 1 1/2 gallon of water, all tubing is insulated.

TEC assembly 1, is a switched controlled peltier assembly, (meaning it can be turned on and off as needed), the assembly heat pipe cooler is an old Xigmatec HDT cooler, cooling the hot side with a SanAce 110cfm fan. The cold side pickup water block is a modified flow XSPC Rasa, the peltier is sandwiched between.

TEC assembly 2, is the primary cooling which runs when the computer runs, the heat pipe cooler is a Thermalright 120 TRUE running the SanAce 100cfm cooling fan. The cold pickup block is a modified flow Swiftech Apogee XT rev2.

The cooler is a Rubbermaid 10qt cooler modified with an output drain, I chose that cooler because it's walls are 1 and 1/4th inch of foam insulation and since I needed low water storage volume with this new setup, ( :lol: Big change from the 54qt Coleman I began this cooling adventure with!), it was a perfect solution.

The CPUs cooling performance is pretty much whatever I want to do with it, with both peltiers energized the lowest I've allowed the water temperature to reach is 4c which is in condensation beginning territory, and it was still dropping when I turned off the switched TEC.

The primary TEC that runs all the time can maintain a 15c stable average of temperature which is usually 8c below my ambient room temperature, since I started this peltier cooling my CPU has not even reached ambient much less going above it.

If I chose to insulate the motherboard I could take it down to just above freezing or with anit-freeze below that, but that's just expressing what it can do, I have no intentions of doing that, I'm content staying above the condensation point, and freezing point.

Yes this does occupy some serious office real estate, but remember this was all an experiment to begin with to see what possibilities were actually useable.

My intentions are to time test the dependability of this TEC cooling, as radiator cooling capabilities is already established information, so my future testing is how to compact this cooling solution so as not to take up so much room, and at this point I have so many ideas.

Ry :)


 
maybe I could pitch in and reduce your load 😛...speaking of which I have alot of writeups to perform :S maybe after I work out some stuff with my build logs... 🙁

amazing work so far mate! Yeah I can recall that big tank of smurfy coolade :lol: though I think keeping the mass of water less than before might have improved your temps w/ the TEC's? 😱
 


Affirmative!

I'm considering an insulation modification of a standard dual bay reservoir, as I've said I'm trying to compact this cooling and get it down to at least being able to fit the TecBox all into one case, running 2 loops a CPU and GPU totally peltier cooled no rads at all.

I've already tried 2 loops with a common Res/Cooler feeding the pumps, but that did not work out as I had planned, so this step forward will require 2 Res/Coolers, so I'll have to make or modify whatever I need to work for these needs.

There's so much testing down the road, but the success I'm experiencing is boosting my curiosity levels, to continue!

I'm sure to a certain extent you are beginning to feel it yourself, with your ChillBox.
 
+1, Feeling that hehe,
Instance one occurs, so you come up with the (or 'A') solution, but that branches about four more ideas off, each of which has its owns effects on things which requires more development and modding and...
AArgh! its endless, but so much fun you can't sit back and say 'its done'
because you can just tweak that to , oh and reroute this, then....
but as you say, the successes feed the drive
Waiting on ram as you know but the fittings are here for the third pump and I have to turn the brain to the rebuild, and once thats done I may decide to uprate the Peltier and, oh sod it I'm doing it again! 😛
Moto
 


I have 3 highly respected individuals here at THGF, and I'm not saying this to disrespect anyone at THGF, I'm saying this because of the interest and curiosity level of the 3.

They are ArthurH, as he is not afraid to attempt something others seem crazy to gain the knowledge that venture may yield!

Motopsychojdi, as he is the modding king here and IMO, he is also not afraid to take a road far less traveled, and I'm pretty sure this TEC type cooling curiosity has also bit him in the butt!

And you sir Lutfij, because you approach objectively and can see the possibilities, and once you see those possibilities are not afraid to take serious consideration of them.

I appreciate any help, and right now possible shrinking this cooling solution down to a manageable size is my most needed input, so if any of you have ideas, please post them.

I don't mind tearing it all up and putting it all back together as I have already done it numerous times, and I may or may not go with every idea, but sooner or later we'll working together, get this thing right!

So I appreciate all and any input from you guys!
 
It's be cool if you could come of with a voltage regulator that would very the voltage given the temp of the water so you would need to watch the water temps. That'd be a project!
 


That may very well be a down the road experiment, presently,

The regulating can be controlled varying the cooling of the hot side of the peltier, the hot and cold side varying is a juggling act that can be done without adding the expense of voltage regulation.

So testing various heat sink cooling solutions may prove that a heat pipe cooler for the hot side is not even necessary, as Moto's ChillBox TEC assembly uses standard aluminum finned heat sinks?

So before voltage regulation is brought into the picture further heat sink hot side cooling possibilities need to be explored, I'm sure the company Moto bought his ChillBox from had done extended experimentation before deciding on the setup they incorporated into the ChillBox.

As far as I could see no voltage regulation is being used in his Chillbox, they've just balanced the hot/cold using the appropriate sized heat sinks and fans for the pettier wattage, to be able to market the Chillbox as a dependable sellable item.

Since in my present setup, 1 TEC by itself maintains a 15c water temperature idle, for daily use that's more than I had hoped for and since 2 TECs not only maintain a full gaming load but still have the combined ability to continue lowering while gaming as well.

Still voltage regulation may be a future consideration.

However voltage regulation can cause the peltier to stall, because if you throw off the balance of hot/cold you'll stall the peltiers performance, and that's a hard learned fact I discovered when I used a 14v capable variable power supply.
 
Yup, 48w tec with some ally finned sinks on purely wired to the powerline,
they also wired the fan motor to the same line but I'm halfway through replacing the stock motor with two 80mm fans and will wire those and the temp sensor up separately from the Tec,
I'm also keeping my eyes out for some copper sinks to replace the aluminium, no rush though just if/as I find them at a reasonable cost
Moto
 
I've replaced the installed fan (Dual spindle motor with two centrifugal blades) with two 80mm casefans (Hacking up to fit ofc 😛)
so as far as the Psu is concerned, they are normal
**Pics will be added to the log as I can, still need some glue and shops were closed today 😛
I may hook the Tec up to one of the Mobo or Pcie lines though, keep the molex lines for other things
Moto
 
So this is an airflow increase modification?

You'll be confirming whether or not increasing airflow cooling to the hot side affects the peltiers output, be careful you don't cool the hot side to the point you stall the peltier.

That won't be possible until the ChillBox reaches it's maximum cold point, in it's initial upstart it will seem OK until the cold side reaches max cold, then if the hot side cooling is too much the peltier will stall.

You're entering testing territory now that honestly truly baffled me in the beginning as I could not wrap my brain around it until multiple testing completely confirmed it.

That's why I've mentioned numerous times of the Hot/Cold juggling act required to get the maximum cooling you're after with a peltier.

However the only way to discover these limitations are to test.

The peltier is limited as it is only a 48w model, so it's only capable of so much before it stalls out, of which you will discover if it happens.

Hopefully just changing the fans won't do it, I'm sure just like CPU overclocking they have some leeway there?

 
@ Ryan - thanks for the honorable mention up top :) means alot coming from thinkerers like yourself, in fact you're a die-ing breed in this day and age, its just gimme this and gimme that with people now-> no leverage for creativity.
and right now possible shrinking this cooling solution down to a manageable size is my most needed input,
Click to expand...
I'll get about seeing what I can concoct.

Hey, as a small favor, can you take a pic of your office? something along the lines of a panoramic shot try getting as much of your office room+walls as possible? How small are you thinking of going?

I think it was you, Ryan , who mentioned giving up on copper blocks as the hotside/cold side stalled due to the copper retaining much of the cooled and then making the TEC think itself into reversing its cooling load...? If so, then modo's choice of going copper sinks will net negative results - me thinks.

@ Modo -Why not look for tall ally heatsink with as much surface area as possible. Possibly the old asus pin fin thermal heatsinks - I'm sure its available on egay 😀 have a windtunnel/shroud constructed around the sink and cool the hot side. Traditional heatsink designs have stagnation of air between their channels - when you have fins that are staggered, you get a vortex/stack(thermal) effect which should aid in more heat dissipation due to convection currents however in Modo's case - this will mean more air flow via fans as opposed to passive cooling which stack effects generally are situated in.

@ Ryan - would you be willing to get a hold of a copper drum and make that into a TEC plate? :lol: sounds silly but I figured the cooling surface needs to be increased and the volume of water needs to be decreased

edit: re-read some prev post, yeah you need reduced water mass and increased surface area.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom