TEC/Peltier CPU Chilled Water Cooling

[Ruwed]

my peltier does not cover the entire block, and my heat sink does not cover the entire peltier

The peltier not covering the entire water block for cold side pickup is not a problem at all, but the heat sink to cool the hot side not fully covering the peltier is bad, really BAD!

You need a cold plate to make sure the hot side of the peltier is fully covered, or the uncovered part of the peltier will burn up!

A cold plate is the terminology of the peltier coverage plate you need but you will be using it as a hot plate!

These are cold plates!

No load, but with NO fan the heat sink never really climbed above 32c°

***********************************************************************

Pretty sure that the Cooler Master Evo's aren't going to cut it with that bad boy, but we'll see before I go spending even more money on a bigger cooler.

You're cooling the hot side with a CMH 212 EVO, "Good luck with that!"

You do have a lot to learn, I just hope you don't destroy something in your learning, there are facts you cannot get around, first off you have to have a lot more cooling capability than the CPU is outputting overclocked.

You have got to keep the hot side of the peltier cool enough and fully covered, or else you'll destroy it.

Running any air cooled heat sink on a peltiers hot side without a cooling fan running is literally hardware suicide!

But after all it is your hardware!

In all my experimentation I have not destroyed any hardware, but I fully researched everything, before I did it!

I did go against some advice of those telling me I could not cool the hot side with an air cooling heat sink, but even I did not attempt it passively.

This one is easy to answer.
It's absolutely not suicide to run it without a fan, I have temp gun and can monitor temps closely while I test.
The EVO is rated at roughly 200 watts of cooling, the 12715 that I have under it is 230 watts, at 15.4 volts. At 12 volts? Yeah, closer to 120 watt's I'd bet, quick math says that it should be 140, but the resistance / wattage on TEC's is not linear so you can't trust quick math.
Which the EVO can handle all by it self without a fan. A fan brings the cooler clear back to room temp with the TEC on full (12v is full in this case).

If I was running three of these (which was the original plan) I would have had a total of "extra" cooling capacity of around 240 watts off of the EVO's. Since my CPU is 140w TDP, that left 100w of overclock without ever touching the cold storage as a buffer (And it's current overclock is running around 216w total CPU useage). Not ideal, but for the money I paid for these coolers, not bad at all. I will be looking into a larger cooler for the TEC I have on the way as it is 320 watts at 13.4 volts, which is 250 watts at 12.... I just haven't been able to find a cooler actually rated high enough yet. I've been researching air coolers, but there are so many out there, and so many of them have sketchy data.. I was trying to not purchase multiple coolers because the first one won't cut it.

I haven't actually touched the plate to see if the ceramic is having heat leached from it, but a hot plate is a very good idea and one that I will probably check into! The edges being exposed bothered me as soon as I measured the cooler and realized it, but I hadn't sat down to come up with a solution yet.

I have done TONS of reading and research. I won't say that I fully understand everything, as I don't have a math or engineering degree and once you get into some of the calculations, I can muddle through but would never be able to come up with them on my own.... But I would say that I have a much better understanding than 90% of everyone else. (Edit: This was not meant to sound as condescending as it did... it was simply meant that 90% of people have not done any research into peltier cooling or even the thermal principles that keep their computers running!)
The only hardware I've destroyed in this experiment was completely my faulty and it wasn't due to lack of research, it was from just plain old not paying attention to the polarity on electrolytic capacitors. The FET I blew up, I knew it was going to pop, but sometimes you just have to sacrifice a poor FET in the name of science.....

Edit: So someone over at Overclock.net was nice enough to point me to a data table that he assembled from another review. It's pretty amazing stuff!
Cooler performance up to 345w

I found a Noctua NH-U14S for a reasonable price (any cooler above it's performance level is over $100 so far) so I picked one up, I can use it for the big boy TEC, and I can also use it to test the difference between it and a EVO on the smaller TEC's and see how much of a difference it makes.

 

[4Ryan6]

This one is easy to answer.

*********************************************

I haven't actually touched the plate to see if the ceramic is having heat leached from it, but a hot plate is a very good idea and one that I will probably check into! The edges being exposed bothered me as soon as I measured the cooler and realized it, but I hadn't sat down to come up with a solution yet.

What I posted to you, was not a question?

************************************************

I am about 2 1/2 years away from experimenting as my setup described in this thread is fully operational I am just sharing information to help you and hopefully keep you from making costly mistakes as this cooling is quite expensive just to get up and running.

Whether you listen to what I tell you, that I learned the hard way or not, is up to you, I did my part!

************************************************

Solid and even clamping pressure, (at about a minimum of 100psi), across the peltiers full surface affects the cold/hot transfer of the peltier, how can you be 100% solid on that point?, without fully covering the peltiers surface, that you're clamping down, IMO that's a No-Brainer!

Of course you haven't addressed the clamping pressure yet have you, here or in your other thread.

Seems I recall in your other thread you're still working out how to clamp it together?

Isn't that Correct?

You will eventually discover, clamping pressure does matter!

Regarding hot side air coolers:

When I discovered the Thermalright TRUE would do the job of cooling the hot side, I never researched the supposed cooling capabilities of each cooler because that changes if you use higher CFM cooling fans, I learned that a long time ago, so tested chart data is actually useless IMO, if you understand all the coolers were not tested with the same cooling fan rpm and cfm output.

Heat sink cooler comparisons seem great when you read the summary over, but not to the hardened overclocker that pushes every aspect of the cooling to get all they can from their cooling solution.

Put the coolers all being tested with the same rpm and cfm fans and you'll get different test results, the linked results used fans from 2700rpm ~ 1200rpm and 122.5cfm ~ 51cfm, such a broad range of fans were used, the only way to properly class each coolers solid raw cooling performance capability, is if they were all tested with the same fan outputs.

Back in my testing days I tried various coolers because I had them on hand, those that failed were the Tuniq Tower, (original) CMH212 Plus (original), and the Xigmatech S1283, (original, and it actually worked, but was hotter operational than I wanted it to run), but then?

I clamped the Thermalright TRUE on the assembly and it did the job using 110cfm cooling fans so Wallah, I was in business 2 1/2 years ago, no more need for research on that particular hot side cooling because it worked!

The key is to run a cooler that was originally designed as a passive cooler, like the Thermalright TRUE, because a passive coolers performance will improve with strapping on higher cfm cooling fans, where as a cooler that needed a cooling fan from the beginning like the EVO that is not a passive cooler doesn't improve cooling performance much at all adding higher cfm fans.

That's one reason Cooler Master has never released a dual fan CMH212, the 2nd fan doesn't make enough difference in cooling performance to warrant the 2nd fan cost.

Many overclockers that have experimented with the CMH212 have already learned that adding a 2nd fan or higher cfm fans, doesn't improve the cooling performance worth warranting the extra expense of the fans.

By the way you can still acquire a Thermalright TRUE because that's the actual heat sink of the Venomous SB-E, not a bad sale price either.

http://www.performance-pcs.com/thermalright-venomous-sb-e-universal-cpu-cooler.html

If you're interested that is!

 

[4Ryan6]

(Edit: This was not meant to sound as condescending as it did... it was simply meant that 90% of people have not done any research into peltier cooling or even the thermal principles that keep their computers running!)

I am attempting to help you, and share what I have learned from this cooling and that's all I am doing, and many will tell you I sometimes come across hard core and hard nosed, when trying to make an important point otherwise I wouldn't even respond.

I see you making some same mistakes I did, so I'm not attempting to insult or belittle you, I actually applaud you, as I hope your endeavors are 100% successful.

The only thing I see in your beginning efforts in the other thread that may take you back to the drawing board is making the voltage controller without having the peltier load to test with, don't you at least think that was putting the cart before the horse?

You may even change to more powerful peltiers before it is said and done, and if you do will the controller handle the added load?

You won't know that until you put it into full operation.

 

[Ruwed]

Then take my 'it's not suicide' as a return statement rather than an answer to a question

Believe it or not, I have read all of your thread. I've read threads from tons of other people using experimental cooling methods and I've found lots and lots and lots of math.... I did think about the sizing of the heat sinks prior to purchasing. These heat sinks will run the 12715 TEC's passively at 12v easily. They cannot go up in TEC size though as the 12715's plus CPU heat will come close to maxing them with a fan on there at mid range RPM, which is why I was trying to find larger..
And I appreciate your help believe me! I wish I would have asked you before I bought my others and I would have learned about the thermalright venomous cooler ....... I looked and looked and looked and could not find a T.R.U.E. cooler. *sigh* go figure. If I could find someone to buy these EVO's I would gladly replace them with those coolers, but I can't justify another $100 yet. My R&D costs are a bit up there at the moment.

I did pick up a much more powerful TEC, 40x35mm that is 196w Qmax at 13.7 volts. I did grab a larger cooler for that guy as I know that the EVO will not be able to handle it at 12v. That TEC will be the first one in the loop. I will use it's much larger cooling capacity to take the brunt of the heat from the CPU and then use the two smaller TEC's to cool the water farther as it continues it's travels to the reservoir!

I have researched clamping, it's the post before the pictures of the installed peltier. I listed out the formula for calculating torque, the torque coefficient of the screws I am using, the expected PSI of TEC's and what that equates to for minimum torque, max torque and average torque 🙂 I even posted a picture of the correct way to clamp peltiers vs using stock clamps that come with heat sinks as using four screws and clamping the corners (like a CPU) is a no-no!

As for the controller. I guess you're not actually reading the posts, just skimming, which is fine since I know I type quite a bit. But the controller as it sits is designed to handle around 25 amps. My only concern is the inductor carrying capacity as I did not purchase those but rather scavenged them from an old PSU. But that's easy enough to find out, hook it up and run it!
The design is solid, it's a circuit that's been in operation for much longer than you or I have been alive. Dropping voltage levels by using a chopper is pretty basic stuff. I repair electronics for a living so I'm not scared to let the smoke out 🙂 Worst case, I'm out $3 in components and I know I have to buy larger inductors.
The design phase of trying to size inductors and capacitors was messing with me a bit as I don't normally design, I troubleshoot. So once this one blows up, I'll know just what to do to find out why!

As for designing the circuit without the TEC's to test with.... You don't put your expensive equipment on a circuit to 'test' it. You design it first, test it using a simulated load, once the circuit (or anything for that matter) has been proven to be functional, then you hookup your equipment. In this case I don't have $40 to spend on a 250 watt 1 ohm resistor, so I tested at a lower overall load, proved the concept and will start low and slowly turn it up, monitoring voltages and temps of the board as I go. (It sounds all complicated, in reality it will take about 5 minutes to reach full voltage and then let it run for about 30, if the temps are stable we're all good!)

 

[4Ryan6]

Sounds like you have a good handle on things, but don't wait too long on those TRUEs you may need, I don't know how much longer Performance-PCs will be stocking those, and I'm pretty sure Thermalright won't be manufacturing more of those?

It's kinda like once the stock is gone, it is gone for good, it is amazing PerfPCs still has some of them, you cannot find them anywhere else, at least I haven't found them anywhere else.

 

[Ruwed]

Sounds like you have a good handle on things, but don't wait too long on those TRUEs you may need, I don't know how much longer Performance-PCs will be stocking those, and I'm pretty sure Thermalright won't be manufacturing more of those?

It's kinda like once the stock is gone, it is gone for good, it is amazing PerfPCs still has some of them, you cannot find them anywhere else, at least I haven't found them anywhere else.

I'm hoping I have a good handle on things....

Please, don't think I'm shooting down any of your suggestions! Writing out an answer to someone helps me to clarify my thoughts and sometimes catches mistakes that I've made. I would much rather have someone go over my design and question things then just read and giggle waiting for the wreck they know is about to happen! Sometimes I have a good reason that maybe you haven't seen, and sometimes I just plain screwed up

Yeah, I really want to snag at least two, but like I said, funding is short ATM. I'm scrounging and seeing it I can find anything else to sell to fund my addictio..... err, hobby. 😀

 

[4Ryan6]

I'm scrounging and seeing it I can find anything else to sell to fund my addictio..... err, hobby. 😀

ROFL :lol: You had it right! It is an Addiction!

 

[4Ryan6]

Update: For anyone interested, no pics, just information.

Since my last addition to the cooling setup of adding the 3rd peltier switch controlled and on it's own power supply, many times I almost screwed up and shut down the computer and forgot to cut off the 3rd pelteir assemblies power, and that is bad, Really BAD!

Without the main machine running which powers the pumps there would be no water flow, and the peltier still energized would freeze the water in the water block and literally split it open!

Seeing as how I came close a couple of times to forgetting to cut the power, and one time I had already left the room and remembered it and went running back, and shut it down.

I decided to just add the 3rd peltier assembly to the 1000w power supply that already controlled the other 2 TEC assemblies, which automatically starts and stops when the main computer does.

I didn't do it in the beginning of adding the 3rd TEC assembly, which I should have!

But?

I was not sure the 3rd TEC assembly was going to do what I had hoped it would? (Which was to hold a certain water temperature when running Intel burn Test, which it does! )

I was reluctant to splice it into the main cooling even after discovering the 3rd TEC worked as hoped, because of how much work was involved in tying it into my existing cooling setup, which took the better part of a day as there was much wiring and rewiring that had to be done.

So now the 3rd TEC assembly is wired into the main cooling setup properly, and on it's own switch, and when the main computer is powered down, so is the 3rd TEC assembly, so if I had forgot, and left it switched on, no worries!

I hope everyone is having a good weekend! Ry

 

[Ruwed]

Much better to be safe than sorry! Good call on taking the time to do it right!

On a side note, now that you're already done with all the work 😀, you could have just tied a 5v coil relay to a molex connector off your PSU, put your original power in to the switch side of the relay, ran the power from that to the manual switch you have now. Turn off computer, no more power to switch!

Just an idea I've been floating. I'm going to use the exact above setup, except with the ground and enable wire hooked to the relay to turn on my second PSU for my cooling box when I hit the power switch for my computer.

 

[TweakGames]

Hello Ryan.

I just wanted to throw my current test setup in here! I hope you don't mind. After 5 years of wanting to do this, after reading your thread, I finally got the courage to try it! This current setup is in NO WAY optimal, and still in the testing phase, but I have been using it solid for about 3 weeks now with zero issues. The only down side I have right now is the space it is taking up, and it is just barely a little too loud for me. I was using a modded dell server power supply commonly used by the RC community to get 12/24 volts for their chargers, it was very loud. I tried some fan mods and such, but it still wasn't to my liking. I bought a compact adjustable 50 amp power supply:

http://www.amazon.com/gp/product/B007WMOM4O/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1

and it has been wonderful. The fan is off most the time, and only comes on if I go over 13 volts after about an hour. I think I am going to swap out the 80mm fan for a 120 with the 80mm to 120mm fan adapter I have. Should make it cooler and perfectly quiet.

Again, please don't judge, the random mess of wires and junk! 😀 It's about to move up onto a shelf above the computer here shortly after I am completely confident in it, and happy with the noise. I will be shortening all the hoses and finishing the insulation. I am still borderline wanting to try water cooling on the hot side. I know you said you had bad results with yours, but I have been eyeing a dual sided 120x9 or 180x4 radiator, another water pump, and two more water blocks. You're bad results have been holding me back from clicking the buy button. 😀


Anyways, here are some pictures of my test setup:

Power supply mentioned above:

Fan next to it to show how small it is:

First peltier setup:

2nd peltier setup kinda crammed in there:

Pump:

Inside computer:

My racing suit coolshirt ice chest used as a res:

Crappy looking overview of the setup lol:

After an hour of Battlefield Hardline, 41 C max at 4.5ghz :

List of stuff:

Using the same peltiers as Ryan I think: 12711-9M31-24CW
http://www.shop.customthermoelectric.com/12711-9M31-24CW-Thermoelectric-Peltier-Module-12711-9M31-24CW.htm

Processor: 4690k @ 4.5 Ghz
50mmX50mm water block : http://www.ebay.com/itm/380915421216?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Aluminum block cut in half for hot plate unmodified. Wasn't lapped or smooth really...: http://www.ebay.com/itm/350829617891?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Swiftech MCP35X 12 VDC Pump : http://www.amazon.com/gp/product/B004XM5TK2/ref=oh_aui_detailpage_o07_s00?ie=UTF8&psc=1

XSPC G1/4" to 3/8" Barb Fittings : http://www.amazon.com/gp/product/B00NODEC8Q/ref=oh_aui_detailpage_o07_s01?ie=UTF8&psc=1

ATP Vinyl-Flex PVC Food Grade Plastic Tubing, Clear, 3/8" : http://www.amazon.com/gp/product/B00E6BD948/ref=oh_aui_detailpage_o07_s02?ie=UTF8&psc=1

Cooler Master Hyper 212 EVO : http://www.amazon.com/gp/product/B005O65JXI/ref=oh_aui_detailpage_o07_s02?ie=UTF8&psc=1

EK-Supremacy EVO CPU Water Block : http://www.performance-pcs.com/catalogsearch/result/?q=EK-Supremacy+EVO+CPU+Water+Block

Using a generic fan controllers, thermal paste, and some wrong sized insulation from ace hardware. lol

Some things I noticed already. I didn't measure the voltage I was getting from the dell server power supplies, but 12V on this power supply makes the hot side much hotter. At 12V I am pulling 31.5 amps according to the gauge on the power supply. So about 380W.

The 212 Evo heatsinks are only hot the the touch at the pipes, the aluminum fins are not even warm. The aluminum hot plate is pretty hot, but I can pinch it between my fingers for 3 or 4 seconds before I need to let go. I assume that isn't anywhere near 200C* so I should be fine.

I am content with this setup so far, but the reason I want to consider the water cooling for the hot side with a massive radiator, is I can feel how hot the aluminum hot plate is getting, and I really would like to know how much cooler the hot side could be with a water block directly on the peltier.

Anyways, that's my results so far! I don't know how people could say this doesn't work Ryan, your setup is 100 times better than mine, and mine does the job just fine. I have been careful to adjust the voltage, as our dew point was getting pretty high (47*F today) and the temperature of the water in my cooler was getting to the low 50s F.

Yay!

 

[Ruwed]

Hello Ryan.

I just wanted to throw my current test setup in here!

Spoiler

I hope you don't mind. After 5 years of wanting to do this, after reading your thread, I finally got the courage to try it! This current setup is in NO WAY optimal, and still in the testing phase, but I have been using it solid for about 3 weeks now with zero issues. The only down side I have right now is the space it is taking up, and it is just barely a little too loud for me. I was using a modded dell server power supply commonly used by the RC community to get 12/24 volts for their chargers, it was very loud. I tried some fan mods and such, but it still wasn't to my liking. I bought a compact adjustable 50 amp power supply:

http://www.amazon.com/gp/product/B007WMOM4O/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1

and it has been wonderful. The fan is off most the time, and only comes on if I go over 13 volts after about an hour. I think I am going to swap out the 80mm fan for a 120 with the 80mm to 120mm fan adapter I have. Should make it cooler and perfectly quiet.

Again, please don't judge, the random mess of wires and junk! 😀 It's about to move up onto a shelf above the computer here shortly after I am completely confident in it, and happy with the noise. I will be shortening all the hoses and finishing the insulation. I am still borderline wanting to try water cooling on the hot side. I know you said you had bad results with yours, but I have been eyeing a dual sided 120x9 or 180x4 radiator, another water pump, and two more water blocks. You're bad results have been holding me back from clicking the buy button. 😀


Anyways, here are some pictures of my test setup:

Power supply mentioned above:

Fan next to it to show how small it is:

First peltier setup:

2nd peltier setup kinda crammed in there:

Pump:

Inside computer:

My racing suit coolshirt ice chest used as a res:

Crappy looking overview of the setup lol:

After an hour of Battlefield Hardline, 41 C max at 4.5ghz :

List of stuff:

Using the same peltiers as Ryan I think: 12711-9M31-24CW
http://www.shop.customthermoelectric.com/12711-9M31-24CW-Thermoelectric-Peltier-Module-12711-9M31-24CW.htm

Processor: 4690k @ 4.5 Ghz
50mmX50mm water block : http://www.ebay.com/itm/380915421216?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Aluminum block cut in half for hot plate unmodified. Wasn't lapped or smooth really...: http://www.ebay.com/itm/350829617891?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

Swiftech MCP35X 12 VDC Pump : http://www.amazon.com/gp/product/B004XM5TK2/ref=oh_aui_detailpage_o07_s00?ie=UTF8&psc=1

XSPC G1/4" to 3/8" Barb Fittings : http://www.amazon.com/gp/product/B00NODEC8Q/ref=oh_aui_detailpage_o07_s01?ie=UTF8&psc=1

ATP Vinyl-Flex PVC Food Grade Plastic Tubing, Clear, 3/8" : http://www.amazon.com/gp/product/B00E6BD948/ref=oh_aui_detailpage_o07_s02?ie=UTF8&psc=1

Cooler Master Hyper 212 EVO : http://www.amazon.com/gp/product/B005O65JXI/ref=oh_aui_detailpage_o07_s02?ie=UTF8&psc=1

EK-Supremacy EVO CPU Water Block : http://www.performance-pcs.com/catalogsearch/result/?q=EK-Supremacy+EVO+CPU+Water+Block

Using a generic fan controllers, thermal paste, and some wrong sized insulation from ace hardware. lol

Some things I noticed already. I didn't measure the voltage I was getting from the dell server power supplies, but 12V on this power supply makes the hot side much hotter. At 12V I am pulling 31.5 amps according to the gauge on the power supply. So about 380W.

The 212 Evo heatsinks are only hot the the touch at the pipes, the aluminum fins are not even warm. The aluminum hot plate is pretty hot, but I can pinch it between my fingers for 3 or 4 seconds before I need to let go. I assume that isn't anywhere near 200C* so I should be fine.

I am content with this setup so far, but the reason I want to consider the water cooling for the hot side with a massive radiator, is I can feel how hot the aluminum hot plate is getting, and I really would like to know how much cooler the hot side could be with a water block directly on the peltier.

Anyways, that's my results so far! I don't know how people could say this doesn't work Ryan, your setup is 100 times better than mine, and mine does the job just fine. I have been careful to adjust the voltage, as our dew point was getting pretty high (47*F today) and the temperature of the water in my cooler was getting to the low 50s F.

Yay!

How many TEC's are you running?

I'm amazed that the EVO is managing to keep that even remotely cool

I think that you would see MUCH better results if you lapped that aluminum plate. I know that with a 12715 (about half the power of your TEC) at 12v my EVO gets reasonably toasty with no fan. With a fan I can keep it roughly room temp, but with no load. I think that you are probably losing quite a bit of heat transfer. It might even be worth going and finding some 1/8 or 1/16" thick copper to make a hot plate out of. Al has a thermal conductivity of 205 and copper is 401, almost twice as efficient!

I'm really happy to see another setup besides Ryan's! I'm building an almost identical setup, just a bit...... more complicated I think I put a link to my thread somewhere around here... if not and you're interested shoot me a PM, I don't want to take away from Ryan's thread! His thread was the inspiration that got me started on my journey (which has turned into a trip around the world.....)

Grats on a successful test though man! Really nice to see other people doing it!

 

[TweakGames]

Hi Ruwed,

Just 2 peltiers. I would get a 3rd but this one isn't even close to running full speed, and I am already close to our dewpoint. They each have an Evo 212 on it. They both have 2 120mm fans on them in push pull, but they are throttled way down to where I can't hear them.

If yours are getting very hot, either that means two things. Mine is working really well and dissipating the heat, or it isn't working well at all and the peltier's surface is still really hot and holding the heat. I am guessing number 2. lol If I could just find a heatsink with heat pipes and such with a 50mm base, I could get rid of the stupid aluminum hot plate and be perfect!

Update:
Oh I just reread your comment. You said yours gets hot with NO FAN. Yeah, mine would get SUPER hot very quickly with no fans going, definitely. I can put my hand in front of the fans and feel the heat coming off it. So maybe mine isn't working as bad as I thought. I don't think there is any way to cool these passively, no. I was using a tiny tiny 50W unit from Fry's I picked up for $12, and it couldn't even be cooled passively by a 212 without becoming untouchable for example. Have to get the heat out of the heatsink.

 

[Ruwed]

With no load (just an empty waterblock clamped on the cold side) and a TEC that is only running about 11 amps you can feel the heat blowing off the heatsink, with the fan. My EVO itself at the base is cool, but the top is warm and the heat is noticeable.

I would really really look for a copper plate to use as a hot plate. You can find them locally at a fabrication shop, I picked mine up for free, just 1/16 thick scrap pieces laying on the floor that they gave to me, then lapped them by hand with 400 - 600 - 800.

As for the cooler base, I *think* that the Thermal Right Venomous cooler plate may fit them. I know that the Noctua NH-U14s won't, I have a 40x35 TEC that is the same power as yours, but the Noctua just barely covers the entire thing. We'll have to wait for Ryan to let you know if the base works on those 50mm or not 🙂

If you look up a page or so, Ryan posted a link to purchase the coolers that he is using from performance PCs. Reasonable price too! (Shipping blows though...)

If you are unable to hold onto the AL base for more than 3 or 4 seconds, it's probably upwards of 60-70c° (about 150F°). If the system is working for you, by all means do what works! I'm just a bit OCD and always strive to get it working as efficiently as possible. Your cold plate side is probably holding around a 0-5c° temp. With a better hot plate, and or larger cooler you could drop your input power and increase efficiency, giving you more head room to turn the computer up!

Edit: You should also look into the clamping. It makes a huge difference

The link to my thread is up a page or so in this thread. Feel free to drop in, I've posted quite a bit about the calculations and mistakes I'm making on my build 🙂 Maybe you'll have some insight that will help me along my way!

 

[4Ryan6]

Hello Ryan.

I just wanted to throw my current test setup in here! I hope you don't mind. After 5 years of wanting to do this, after reading your thread, I finally got the courage to try it! This current setup is in NO WAY optimal, and still in the testing phase, but I have been using it solid for about 3 weeks now with zero issues. The only down side I have right now is the space it is taking up, and it is just barely a little too loud for me. I was using a modded dell server power supply commonly used by the RC community to get 12/24 volts for their chargers, it was very loud. I tried some fan mods and such, but it still wasn't to my liking. I bought a compact adjustable 50 amp power supply:

http://www.amazon.com/gp/product/B007WMOM4O/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1

and it has been wonderful. The fan is off most the time, and only comes on if I go over 13 volts after about an hour. I think I am going to swap out the 80mm fan for a 120 with the 80mm to 120mm fan adapter I have. Should make it cooler and perfectly quiet.

Again, please don't judge, the random mess of wires and junk! 😀 It's about to move up onto a shelf above the computer here shortly after I am completely confident in it, and happy with the noise. I will be shortening all the hoses and finishing the insulation. I am still borderline wanting to try water cooling on the hot side. I know you said you had bad results with yours, but I have been eyeing a dual sided 120x9 or 180x4 radiator, another water pump, and two more water blocks. You're bad results have been holding me back from clicking the buy button. 😀

Using a generic fan controllers, thermal paste, and some wrong sized insulation from ace hardware. lol

Some things I noticed already. I didn't measure the voltage I was getting from the dell server power supplies, but 12V on this power supply makes the hot side much hotter. At 12V I am pulling 31.5 amps according to the gauge on the power supply. So about 380W.

The 212 Evo heatsinks are only hot the the touch at the pipes, the aluminum fins are not even warm. The aluminum hot plate is pretty hot, but I can pinch it between my fingers for 3 or 4 seconds before I need to let go. I assume that isn't anywhere near 200C* so I should be fine.

I am content with this setup so far, but the reason I want to consider the water cooling for the hot side with a massive radiator, is I can feel how hot the aluminum hot plate is getting, and I really would like to know how much cooler the hot side could be with a water block directly on the peltier.

Anyways, that's my results so far! I don't know how people could say this doesn't work Ryan, your setup is 100 times better than mine, and mine does the job just fine. I have been careful to adjust the voltage, as our dew point was getting pretty high (47*F today) and the temperature of the water in my cooler was getting to the low 50s F.

Yay!

Hi and welcome to the thread, I have no criticism just suggestions, I know how much clutter you end up with in the beginning, you should have seen my disaster in the making, it took some time to get it like it is now.

The drive comes from realizing it works and wanting to improve on it's layout and design, but you now know it works and it can be improved on and shrunk down to become more tolerable for you.

My first question to you is how much coolant is that reservoir holding?

From my own experimentation I started with a reservoir about that size and had to keep dropping the volume of coolant in the reservoir because the more volume you have in the reservoir the longer it takes the peltiers or TEC assemblies to drop the coolant temperature.

I experimented down to close to 1/2 gallon which is a very effective storage volume, which brought another problem to the table finding an insulated reservoir that was efficiently insulated to run that low of a coolant volume, and I couldn't find what I needed so I made one from scratch, and insulated the exterior, and that's what I am running now.

So how much volume of coolant are you running and how long does it take that volume to reach a cold equilibrium?

************************************************************************

The cooler master heat sink is just not adequate for cooling those peltiers the Thermalright TRUE will do the job with 110cfm cooling fans and they're still attainable through Performance-PCc, as the Venomous is running a TRUE heat sink.

http://www.performance-pcs.com/thermalright-venomous-sb-e-universal-cpu-cooler.html

These replaced the Scythe Slipstream 1900rpm and they still output about 108cfm.

http://www.performance-pcs.com/120m...mm-x-25mm-pwm-fan-1900rpm.html#Specifications

An even better price is the link below but they are presently out of stock and may not get anymore?

http://www.performance-pcs.com/thermalright-ultra-120-extreme-am2-rt.html

********************************************************************

Clamping pressure is very important on the peltiers and the clamping surfaces have to be level that's why I lapped all my peltier contact surfaces cold plates and water blocks so they were perfectly flat, that is important!

Regular extruded aluminum flat stock is slightly bowed in the center and has to be lapped flat.

**********************************************************

That's all I got right now, I haven't had my coffee yet!

 

[TweakGames]

Sorry for the late response. Literally the day you posted the above response, I went to lowes before, and bought a bunch of 4 inch PVC pipe and made a huge 4 ft long res. It held almost 3 gallons. Lol. It didn't ever get hot, but it also never got cold, even over night. Just too much water. I too have just downgraded back to a small container and it is working a lot better again.

I have noticed that my system gets colder at 8V, than 12V. I assume it's because the hot side is getting too hot, the 212's can't handle it. I do however like the 8V from my power supply because the fan never turns on! 😀

I am stuck right now between buying 2 more peltiers, water blocks, another power supply, and 2 more venomous heatsinks you mentioned above, and having all 8 just run at 8V. Or, switch to the giant 9 120mm fan radiator to cool it all. I have about a month to decide.

I am thinking however that I will get a 40mm peltier unit, so that I can get rid of the copper hot plates all together. I purchased some, lapped them until they were mirror, but it doesn't seem to make a great deal of difference. I had no way of knowing what temp the hot side of the peltier was though.

Next, I am going to buy a switching power strip the will turn the peltier power supply on automatically when the computer is on, and this 110V temperature relay to turn the power supply off and on automatically when the water gets close to dew point. Will help save on power cost too.

http://www.ebay.com/itm/110V-LED-Digital-Temperature-Controller-Temp-w-Sensor-Thermostat-Control-relay-/380665672071

Thanks for the response and tips above. Onto test stage 3!

 

[Ruwed]

Sorry for the late response. Literally the day you posted the above response, I went to lowes before, and bought a bunch of 4 inch PVC pipe and made a huge 4 ft long res. It held almost 3 gallons. Lol. It didn't ever get hot, but it also never got cold, even over night. Just too much water. I too have just downgraded back to a small container and it is working a lot better again.

I have noticed that my system gets colder at 8V, than 12V. I assume it's because the hot side is getting too hot, the 212's can't handle it. I do however like the 8V from my power supply because the fan never turns on! 😀

I am stuck right now between buying 2 more peltiers, water blocks, another power supply, and 2 more venomous heatsinks you mentioned above, and having all 8 just run at 8V. Or, switch to the giant 9 120mm fan radiator to cool it all. I have about a month to decide.

I am thinking however that I will get a 40mm peltier unit, so that I can get rid of the copper hot plates all together. I purchased some, lapped them until they were mirror, but it doesn't seem to make a great deal of difference. I had no way of knowing what temp the hot side of the peltier was though.

Next, I am going to buy a switching power strip the will turn the peltier power supply on automatically when the computer is on, and this 110V temperature relay to turn the power supply off and on automatically when the water gets close to dew point. Will help save on power cost too.

http://www.ebay.com/itm/110V-LED-Digital-Temperature-Controller-Temp-w-Sensor-Thermostat-Control-relay-/380665672071

Thanks for the response and tips above. Onto test stage 3!

You'll have to let me know how the temp controller works long term. See if there's a hysteresis that you can program. I was in the same boat as you, temperature controlled peltiers. But after doing a whole bunch of reading I found out that they are pretty sensitive to thermal cycling and break down quickly. I ended up making my own PWM controller that controls a voltage regulator I built to try to limit the on / off. I still don't even know if it was worth it...... But it's been a fun ride as I try to get it all figured out!

For not cooling well at 12v, I'd bet that the 212's just aren't up to snuff on cooling that large of a peltier at 12v. You have two TEC's each of them rated at 220 watts max on the heat side + the heat load from your CPU, which is about 130 watts, split between two TEC's, call it roughly 300 watts per heat sink. The 212 EVO is only rated to 200 watts of cooling, so it is getting overloaded.

Even with the 40mm TEC, you require a hot plate with the EVO's (I have 40mm TEC's and EVO's). I'm not sure about the ThermalRight cooler base size. I know that a Noctua NH-U14S does not require a hot plate for a 40x35mm, but it would require it for a 40x40.

Can you link the power strip you are going to use? Right now my current plan is to power up a relay off of the 5v from my computers PSU and have that relay control the enable switch for another PSU. A power strip might be a simpler answer....

 

[TweakGames]

Hello Ruwed,

Yes, here is the power strip I have used in the past.

http://www.amazon.com/Smart-Strip-SCG5-Autoswitching-Technology/dp/B000L9FQUO/ref=sr_1_6?ie=UTF8&qid=1431557762&sr=8-6&keywords=power+strip+energy+saver

I lost it when I moved to my house somehow, so I will need to purchase another one. Basically it has a "control" plug (blue in this case). When the control plug is energized by having the computer on, it will activate the other plugs (not colored ones). The 2 red plugs in this example are always hot. It is also adjustable via a simple screw driver on the side to adjust sensitivity. Hope this helps.

Yes, I feared the 212s are just getting too hot. They are doing really well at 8 volts though.

As for the heatsink base dimensions, I have found that thermalright has most of them listed on their website, under the specs tab!
Example:
http://www.thermalright.com/html/products/cpu_cooler/venomousx_rt_rev_sb-e.html?panel=1
The Venomous X RT Rev.SB-E is 40mm x 42mm!

http://www.thermalright.com/html/products/cpu_cooler/hr22.html?panel=1
The huge HR22 is 40mm x 53mm!

It appears the 212 Evo is only 37.6mm wide.

Yeup! Hope that helps.

Thanks.

 

[Ruwed]

Cool man, thanks for the info!

 

[4Ryan6]

Yes, here is the power strip I have used in the past.

http://www.amazon.com/Smart-Strip-SCG5-Autoswitching-Technology/dp/B000L9FQUO/ref=sr_1_6?ie=UTF8&qid=1431557762&sr=8-6&keywords=power+strip+energy+saver

Interesting device that would have saved me some trouble if had discovered it existed earlier!

So the blue is the control plug that whatever is plugged in there and powered up, the red plug sockets power up as well, is that how it works?

Sweet!

 

[TweakGames]

Ryan, basically. The blue is the sensing plug. If you have something plugged into it and it flows enough electricity (depending on how you set it) it will turn on the other ports. The red ports are ALWAYS hot.

This larger picture might help. But yes, you got it.

 

[4Ryan6]

OK I see, the picture I saw was so small you couldn't read the labeling, that is cool!

I can't say I'm too comfortable with that warning though!

Warning: Automatically Switched Outlets May Turn On Without Advanced Notice

That would put it in a not rock solid dependable category for me, especially if it was energizing a power supply directly controlling the TEC assemblies and came on all by itself without the pumps running, that would be bad, really bad!

 

[4Ryan6]

The reason I use a single 12v rail PC style power supply dedicated to powering the TEC assemblies is when I power the main PC it auto jump starts the 2nd power supply using a dual power supply linking cable.

It allows one motherboard to power 2 independent PC style power supplies simultaneously.

http://www.newegg.com/Product/Product.aspx?Item=9SIA2971AW2565

And It is rock solid dependable!

And you just cannot beat the convenience of pushing one button to start it all up, and when the main PC powers down, so does the 2nd power supply controlling your TEC assemblies.

 

[apelord]

' Cooling the Hot Side:

Cooling the hot side of the TEC is a challenge in itself, seeing as how it's best performance comes from allowing the hot side to get hot '


Er.. maximal performance of TEC comes when junction temperature is 0.

 

[4Ryan6]

' Cooling the Hot Side:

Cooling the hot side of the TEC is a challenge in itself, seeing as how it's best performance comes from allowing the hot side to get hot '


Er.. maximal performance of TEC comes when junction temperature is 0.

Even if that is the operational range goal, you still have to cool the hot side!

Thank You for your input!

 

[apelord]

' Cooling the Hot Side:

Cooling the hot side of the TEC is a challenge in itself, seeing as how it's best performance comes from allowing the hot side to get hot '


Er.. maximal performance of TEC comes when junction temperature is 0.

Even if that is the operational range goal, you still have to cool the hot side!

Thank You for your input!

I'm just saying you seem to imply in your article that letting the hot side get hot (or hotter) is a good thing but maximum COP / Qw is achieved when the difference between hot and cold approaches zero ( unless you are talking about how it is necessary to turn it on..). The more aggressively you cool the hot side the better it performs (as you probably know).

From this nice article:

http://www.tec-microsystems.com/EN/Intro_Thermoelectric_Coolers.html