TEC waterchiller

[Jorlen]

I have started the work on a waterchiller.
The idea is to use 10 of each of these:
70W Qmax peltiers, $50
Scythe Heatlane Zen NCU-1000 heatsinks, $100
Koolance CPU-100 waterblocks, $40
So far $190 spent.
I have an old 12V 300W PSU to I plan to use to power the TECs, passivly cooled.
I will update as I make progress.
Some pics of the heatsinks and waterblocks.

[Jorlen]

I plan to stick a peltier between each heatsink and waterblock. I will connect to waterblocks in parallell. The 3/16" connections on the Koolance WB makes this a must.
The whole thing will be mounted on top of my case. As the pictures above demontrates. Some kind of frame will hold the heatsinks above the top of the case so that air can flow up through the fins of the HSs. If necessary, which is likely, some quiete fans blowing air from beneath will help convection.

I am not sure how I will control the power to the peltiers.
I have been thinking of using the CPU-fan control output voltage as a control voltage for a mosfet powertransistor that would control the voltage fed by the extra PSU to the peltiers. To bad I dont know enough electronics to know if this would work.
This method should allow easy control to run condensation free, but still keep the system just a few degrees above condensation temperature.
I am open to suggestions..

[systemviper]

Wahooo

I love TEC Projects, keep us posted, lots of Pic's..


cool

[Jorlen]

Glued together silicon tubing, two 12mm/18mm tubes with ten 5mm/9mm each. The 10 Koolance blocks will go onto the small tubing.
(camera phone picture)

[Jorlen]

The option to control cpu fan speed via pulse width modulation is available via the motherboard cpu-fan connector. And I was planning on using it to control current to my peltiers.
According to one TEC manufacturer switching frequency need to be in the order of 1khz or higher, or the TECs efficiency will suffer noticably.
http://www.marlow.com/technicalinfo/...tions_faqs.htm

So I am asking if anyone know what frequency that cpu-fan speed PWM works at?

[n00b 0f l337]

I don't think a 300W old psu is going to hold you.

[ben333]

I don't think a 300W old psu is going to hold you.

Neither do I. Also, I think if you just ran hose the normal way, a lot more water would move through the blocks and therefore more cooling would get done. Also, those heatsinks don't look like they can take 70 watts, especially passively.

[Xeon th MG Pony]

I smell doom.

[n00b 0f l337]

I wish I could smell over the internet

No wait, no I don't.

[Jorlen]

I will undervolt. I could get another powersupply if i wanted to go past 300W of input power. But I think thats wasteful enough as it is. And as you can see from my post about temperature control I dont plant to get water temps down much below ambient. Atleast not until I feel like undertaking a big insulation/isolation project.

I am not sure what you are trying to calculate there Hollo. If you compare cooling power with max 70W you need to stay at 0 deg temp delta.
And when temp is constant cooling power is linear to voltage in the working range of the peltier element.
So a better calculation would be
Qc@6V (6/15)*70 = 28W (280W for all 10)
P@15V = 121W
I@6V = (121W/15V)*(6V/15V) = 3.2
Pi@6V = 3.2*6 = 19.2W
COP = 28W/19.2W = 1.458

Althoug I feel this is wrong too. I will look up correct numbers later today.
COP should be higher.
With 300W of input power a working point COP could be as low as 0.66.
And peltiers should be able to reach between 20-30 degrees delta at this temperature. That would load my HS with 500W, 50W each.

[Xeon th MG Pony]

I will undervolt. I could get another powersupply if i wanted to go past 300W of input power. But I think thats wasteful enough as it is. And as you can see from my post about temperature control I dont plant to get water temps down much below ambient. Atleast not until I feel like undertaking a big insulation/isolation project.

I am not sure what you are trying to calculate there Hollo. If you compare cooling power with max 70W you need to stay at 0 deg temp delta.
And when temp is constant cooling power is linear to voltage in the working range of the peltier element.
So a better calculation would be
Qc@6V (6/15)*70 = 28W (280W for all 10)
P@15V = 121W
I@6V = (121W/15V)*(6V/15V) = 3.2
Pi@6V = 3.2*6 = 19.2W
COP = 28W/19.2W = 1.458

Althoug I feel this is wrong too. I will look up correct numbers later today.
COP should be higher.
With 300W of input power a working point COP could be as low as 0.66.
And peltiers should be able to reach between 20-30 degrees delta at this temperature. That would load my HS with 500W, 50W each.

If voltage goes down current must go up!

What you really should be doing is wiring them in series so they can take 140V, then rectify AC to DC with a filter then feed it to the pelts through a veriac. This will keep current within sane levels and still provide full power.

If you wanted to get really fancy you could use a chopper circuite to regulate current and inject a PWM to the chopper reset circuit to control over all cooling capacity.

[hollo]

@Jorlen
sorry you can pretty much ignore my post, i thought they were 70W power consumed, not 70W Qmax
i'll delete the post since it's useless
did you get the 121.6W ones?
Dimensions: 40mm x 40mm x 3.2mm
Wire length: 34cm (13.5")
Model number: TEC1-12708
Voltage: 12V
Couples: 127
Umax (V): 15.2
Imax (A): 8
Resistance ( £[): 1.66 ~ 2.0

COP = heat-load over power-consumed
= Pload / (19.2 x 10)
since Pload is probably about 100W for a dual-core, while theoretical cooling power is about 280W (cooling power is linear with voltage while power consumption is square with voltage? that's great), it's looking good

if you don't want condensation problems and want to start straight away (before building a voltage controller) you could do 9 peltiers @ 4V (76W total power draw ((4/15.2)^2*121.6*9), 167W capacity at dT=0 (Qmax conditions, 4/15.2*70*9))
there's no way it would overload... in theory....

what CPU do you have?

[n00b 0f l337]

If voltage goes down current must go up!

No. That is only if power consumption stays the same

[Jorlen]

And here are some pics. Scythe packaged a nice copperplate that mounted according to instructions. I drilled 2 extra holes in the copperplate and threaded them. And cut up a few of the motherboard back-plates that came with the HS to make the small beam that goes across the waterblock. The beam is stainless steal 1.5mm thick sheet metal.

[Jorlen]

Thanks for the tip Xeon. My plan is to use the PWM of the motherboard cpu-fan output to a drive chopper circuit that controls the current from my PSU. Is chopper circuit the correct term?

Yes Hollo, the ones on e-bay. Though I havent got them yet, but I have paid for them. We shall see if and when the show up. Its a long way from Hongkok to Stockholm.

@NOL I can always get more PSUs if I want more cooling power. But at this point I am not aiming for maximum cooling, just really good watercooling. Think of it as watercooling with a slight performance boost. Not waterchilling to compete with direct phasechange systems.

[Sam__]

As ben said.... I would have run the hose normally. the way youve done it wont the water just rin along the bottom? as its not being forced into the blocks. just a thought.

[Jorlen]

300W if input power is just fine for my purposes. So the PSU will hold me. What is the normal way to run a hose? Are telling me to run 10 pcs of 5mm innerdiameter waterblocks in serial?
Thats just crazy.
And the water wont run at the bottom. Water will travel in parallell through all the waterblocks. Do I need to mention that i wont short-circuit the waterloop by allowing water to circulate through the system without passing through the waterblocks?

[hollo]

what is your CPU?
and will you do a 4V x 9 peltier run for me :D?
i've never seen a peltier setup with a COP greater than 1 before

[Jorlen]

Sure, I will can run them at a variaty of voltages. I dont remember the range of my PSU exactly. But i think its 10.5V to 13V. Its tunable via a potentiometer. Using it and different ways of serial/parallell circuits I should be able to test a lot different voltages. I can measure the PSU range later today when I get home to see how far it ranges.

My system is a E4600 on a Gigabyte DS3 P35, Ati HD2600 pro. I havent taken my time to figure out highest stable overclock yet. And its possible I wont overclock at all. And I dont run folding or other stuff that consume my idle cycles.

I got the tracking ID from the ebay peltier purchase, and the package is being shipped out from hong kong today.

[hollo]

i just had a great idea - gotta share it

set it up with 9 peltiers
eg. Pin from your 10.5-13V power supply:
min (((10.5 / 3) / 15.2)^2) * 121.6 * 9 = 58.0263158W
max (((13 / 3) / 15.2)^2) * 121.6 * 9 = 88.9473684W - high enough??

set up the 10th peltier between a waterblock and air-cooler, like the others, but connect it to a small amplifier circuit that is connected to a relay so that when it generates a certain voltage due to the water (water-block) temperature being lower than the ambient (heat-sink) temperature it will deactivate the entire peltier loop.
(i'm not sure what the circuit would look like, but it'd be pretty simple)
it'd be perfect because condensation is dependant on the difference between the loop temperature and the ambient temperature rather than an absolute temperature. it'd just need a circuit that was quite sensitive to (or good at amplifying the) current from the peltier so a water temperature 1-2 degrees below ambient would be enough to shut the circuit down.

it'd be a cool 'feature' for day to day usage

[Jorlen]

The idea would work. I am trying to stay away from solutions where thermal cycling of the peltier elements occur. It reduces the life span of the TECs.

PWM is the way to go. Since it can control the power much more fluently, even though in the microscale it switches on and off at kHz speed.
Either I use the motherboard, or I create my own potentiometer+thermistor controlled PWM circuit. Or rather I copy/modify someone elses.

[Xeon th MG Pony]

I have a psu I pulled from a device that used an NTC thermister controling the inverter chip, I'll see if I can't get some revers engineering don on it today, I won't make any gurentees though.

[hollo]

ahh i see

perhaps you could create a circuit so that voltage generated by a peltier when the water-block temperature is higher than the heat-sink temperature would feed the "base" of a "transistor amplifier circuit" which would have a gradual effect on the amount of current that could flow through the "collector" and "emitter" - as water temperature drops to ambient the current shuts off in a nice analogue curve.
(i've drawn the "transistor amplifier circuit" with "base, collector and emitter" as symbols of transistor function - i dunno what the real circuit would look like)

then you'd still have the 'luxury' of a water-to-ambient-dT based power control - but not a digital on/off one.

you might also want a variable resistor that you could use to set a minimum amount of power that'd flow into the peltiers regardless of water temperature - this'd allow you to set a minimum temperature near or under ambient. without it the minimum temperature would be slightly over ambient. if full load temperature is over ambient i guess you'd leave it out.

though if the ambient temperature in your computer room is fairly steady this'd be a bit overkill

[Jorlen]

Please post what you learn from that Xeon.

Hollo, its completly possible to do as you describe. The drawbacks are that a power transistor works a variable resistor in the way that it increases its own resistance to control the current/voltage. Thats why you need massive heatsinks on these suckers. So although it would work, its my second choice. Dont wanna generate more heat then I absolutely have to, you know. Its a cooling system after all.
A peltier device is one way to measure temperature difference. But its a rather bulky operation considering you need to dedicate a whole HS waterblock peltier assembly just for it. I imagine a much more space and cost efficient solution can be made from using two thermistors that are for measuring temperature in air and water. And via some resistors and transistors you can transform the difference in resistance into difference in voltage. And use that voltage differential to drive a power transistor. Maybe you need one inverted ranged thermistor if you wanna make it easier. As you wanna decrease voltage differential the cooler it gets.
I dont know.

My first priority is to measure the output of my motherboards cpu-fan socket. And see what I can make of that. There are opto-transistors that I can use to keep galvanic separation between the TECs and their high currents and the motherboard.

[Xeon th MG Pony]

ahh i see

perhaps you could create a circuit so that voltage generated by a peltier when the water-block temperature is higher than the heat-sink temperature would feed the "base" of a "transistor amplifier circuit" which would have a gradual effect on the amount of current that could flow through the "collector" and "emitter" - as water temperature drops to ambient the current shuts off in a nice analogue curve.
(i've drawn the "transistor amplifier circuit" with "base, collector and emitter" as symbols of transistor function - i dunno what the real circuit would look like)

then you'd still have the 'luxury' of a water-to-ambient-dT based power control - but not a digital on/off one.

you might also want a variable resistor that you could use to set a minimum amount of power that'd flow into the peltiers regardless of water temperature - this'd allow you to set a minimum temperature near or under ambient. without it the minimum temperature would be slightly over ambient. if full load temperature is over ambient i guess you'd leave it out.

though if the ambient temperature in your computer room is fairly steady this'd be a bit overkill

You could do this by feeding the pelts out put to a Voltage controlled Oscilatore then filter it through a schmit trigger for a more streamlined aproach.