Water Chiller Project.
 

I was going to wait until Joe got the site revamped but that is now on hold for a week or two....

So what I am planing on doing is making a water chiller to chill the water mainly for my test bench but possibly for other projects as well. What I plan to do is use 2-3 TECS sandwiched between some custom blocks that I will mill up. So the design begins. I am going to let you guys help in this project unlike many of my prior projects.

I think a 50mm TEC would be what to work with. They can be had pretty cheap on e-bay.

This leaves the design. I am thinking either channels or pins. There should be an ample bolt pattern for decent pressure.

What I need the chiller to do? Mainly keep the water temp in my test bench somewhat consistent. Secondary goal is hard core chilling for possible overclocking adventures.

I was thinking of using 3 TECs and then undervolting them to keep room temp or a little less. Then if I decide to get extreme I could ramp up the voltage and freeze some stuff.

I already have an idea in mind but would like to hear everyone else's as well.

How do you plan on maintaining the coolant temperature? Some sort of controller wired to the TECs or manual adjustments?

Either way, aim for a few TECs at low voltage. Lower voltage == higher effceincy == smaller powersupplies and cooling equipment.

Pretty sure it will control itself for the most part as room temp and the heat load from the system will be pretty consistent. Problem I have is the room temp is to hot. After equilibrium sets in water temp is 30C. Would like to get it down to 25C when room temp is 24-26C normally. Although if anyone has any idea on how to mange the temperature then all the better.

Yes I was thinking the same thing on the efficency.

The TECs will not be under idle at any point? There will be condensation if they are. Depending on heat load and size of TECs possibly under load as well.

So you want to go from a current coolant temp of 30 to 25 if I understand correctly. What is the purpose of this change? seems small to me but I am sure there is reason behind it.

It is for testing water blocks on my test bench. Consensus is 25C is the temp to keep the water. Seems BillA, RoboTech, and pH use that temp IIRC.

Not worried about condensation. Whats it going to hurt?

True....darn you and your test bench.

but still, you want a consistant temp and I am guessing 15c will mess up results as much as 30c will. 3x units will give you some flexibility as they can be in series to drop the power down.

I still think the answer to controlling TECs is to simple control the temp of the hot side. If the cold side loop is getting too cold just stop the pump on the hot side loop. Simple enough, you should be able to rig something up to keep things at 25 easily with a similar concept.

I was thinking about using a 12VDC pump such as the MCP350 or the C System Mag and using a fan controller to very the speed of the pump. Now that I think about this it might just be cheaper to buy a self controlled chiller.... Even still I think I would do the project.

have you considered a small phase change system?
using a resonable sized reservoir , might be more efficient.
increased temperature stability?
... or complexity defeats the advantages?

It would have to be very small. Which means to expensive. I don't have the time or money to work such a system out anyway. I need this done soon (as in 2 weeks).

I was pondering the design of the blocks for this last night and got to thinking why bother with a design? I am thinking about direct contact with the water on both sides?

you could try a minifridge phase change system
they're nice and tight (no pun intended ... hmm, maybe).

wouldn't a waterblock with some design increase the contact surface area...?
a simple pin grid like swiftech's older blocks , or something.

Mini fridge systems cannot handle a load and are not rated to run continuously. It would have to be able to run 12-15hrs a day while I test one block.

Yeah it would add area but I don't think it would be all that much better. Swiftech's MCW50-T is direct water cooled on the hot side and works. I think it would work on the cold side aswell. If I did that I could use something other than metal and make it a lot easier to mill (and cheaper). All out performance isn't required here either. I guess I should buy a TEC first and make the design around it.

How about using a temperature based fan speed controller (or alike system) and change that to increase/decrease the voltage on your TEC's based on the water temp?
I don't know how possible that is or how hard, but I would definitely consider that if I were aiming for a test bench setup...

That would be ideal but I have no idea how to do that. In fact that could be a commercial product if someone could figure it out. The problem would be adjusting voltage on the power supply.

Seems simple eh?

google variable peltier controller. I wouldnt mind one of their setups If it worked at 5v. May work out for your testing though. I think running three 26v TECs in series off of a 24v PSU through one of those controllers would be an excellent option.
http://mcshaneinc.com/html/5C7-350.htm

Thats what I would do.l

For a year now, Joe and I have had a working controller that adjusts tec power according to temperature from a thermistor. We could keep X point at room temperature, and when the heat source went up or was removed completely, the tec would keep at the same temperature.

For example, the controller is on, then you turn on your computer, the chip was forced to remain at the same temperature. The computer could be turned off, then the chip would stay at room temperature, or a temperature of our choice.

With creative application, it could work with your tecs in your water experiment. It has enough amps to handle a couple of them. Joe and I havent decided if we should market this or let someone else do it. We suspect there isn't enough interest to warrant us putting the money up for it right now.

why don't you just use an air conditioner to get constant ambient temps? that also controls humidity at the same time. and just use six 120.3 radiators (18 120mm fans) and blow that thing outside the room.

Its very difficult to precisely control cooling power with phase change. With TEC you can adjust the voltage and then easily hold it there. With phase, you'd have to try and adjust the CPEV or else have a second variable air flow evap in another room that you use to bleed off excess capacity.

I don't know if this is serious or not. I hope not...

That is what I am looking for. I would venture to guess they would sell pretty well and in fact open up a whole new market for TEC cooling. Which Joe are you speaking of?

Joemac or 'Joe' is from aquajoe, sorry for the confusion

This watercooling market is strange. Tec market is very small at present. It could open up a new market, but this market would be... small due to the complexity. Wouldn't you agree?

Anyway, for a word of encouragement, a controller can be done. I am interested to read of the updates and plans on your project. PM or email me if you need some tips.

edit: It looks much like this one pictured here , however it can be set to follow ambient or an offset from ambient (higher or lower) to reduce or eliminate condensation depending on setup.

No probelm on the Joe issue. I wasn't sure if you ment the Joe here, the Joe at OC'ers or Joemac. I forgot you guy were in together. I am very interested in that controller but I think it is way over my head. If you guys don't market this thing you should work with someone like http://www.wintschlabs.com/ that specializes in TEC cooling or even Swiftech if they don't already have such a product in the works. I seen a lot more intrest in TEC's lately. I think now would be the time to start getting it out there.

I have to go with the commercial product: you can find water chillers on eBay for $100-$150. They'll either be in running condition, or in need of a small repair.

I'll be more than happy to help you repair one (it'd be easier if I was there though). I'm moving to Ohio in ~two months; where are you again?

In Spokane WA. I am also moving in a few days/weeks (waiting for people to get out of the place I am going into) but only a few buildings over. I have looked over the e-bay one's and seen a few decent looking one's. If I go that route I will have to wait untill I move.

Jaydee, sounds like a fun project I have been wanting to do for a while. Bigben says you can get them for 100-150 on ebay, thats definately going to be cheaper and better for sure.
3 peltiers are going to cost around $100.


Jon

I am definatly building the TEC chiller regardless of use. I think I am going to get 2 - 245/320watt TEC's off ebay. http://cgi.ebay.com/Awesome-320-Watt...QQcmdZViewItem

That should run about $50 after shipping. I might pick up power supply as well. I already have 2 sufficent one's though.

I am going to try the direct cooling approch first and then maybe a copper block set. Will see how things go once this new ProCooling site is setup....