eXtreme Liquid Cooling - Build Log + Overclocking

[Ultrasonic2]

I missed this somewhere along the road. The heat exchangers you're looking at are aluminium blocks with slots milled out and copper pipes pressed in. If you're considering tec - copper sheet - pipe side of HX, I'm pretty certain you'll get better performance by omitting the copper sheet and turning the HX over. I'm not alert enough to work anything out for this though I might give it a bash tomorrow (revising thermofluids at present & tomorrow, good to have real world problems to solve), but this would be a very good thing to test.

^You mean output voltage of 3-15V, right?

Regarding condensation, I assume there's a reason why people don't just use heaters around the socket and behind the motherboard?

i'd like to know how much these water blocks cost ?

Also if you are going to use a copper sheet you will HAVE to get it machined flat as it WONT be perfectly flat and therefore useless.

You'll be 100 x better off getting proper TEC water chiller made for you.
Thats supposed to be my next project.

Im guessing those water block and copper and what not are gonna be pretty expensive ?

 

[Ultrasonic2]

Indeed, I've seen the thread. It's not unreasonable that I'd want to do my own calculations as well. I'm interested in the temperature difference between the cpu ihs and the copper in contact with the water, which I can't see anywhere in that thread. There's a fea analysis of the performance of the waterblock side. Since 12V would be looking at a load delta of 10K this number is pretty significant.

Could you clarify whether the tec facing sides of the copper block is flat, or with a ~65mm square indent to seat the tec? I'll take that as a "no" to letting me see a model of it then.

ok

The copper is 12mm + thick
and the side that the TEC goes against is flat .. the Step in the cold plate is so the block and fit over any nearby caps and what not that are higher than the cpus height

This can be seen here

EDIT i cant give you a pic of the model cos im at work and dont have access to that.

 

[Pneumonic]

i'd like to know how much these water blocks cost ?

I am assuming this is the Heat Exchangers you are asking over? They are around $60 / exchanger, works out for me at £64 each once I pay for shipping etc..

 

[Ultrasonic2]

I am assuming this is the Heat Exchangers you are asking over? They are around $60 / exchanger, works out for me at £64 each once I pay for shipping etc..

thats cheap

 

[Pneumonic]

thats cheap

certainly worth it for a project at any rate.. only slight annoyance is that they are 94mm by 300mm so width wise I am restricted on TEC size if I want to fit two rows hence why I went for 40mm x 40mm TEC as I can fit 14 on there if needs be

 

[JonJ678]

Thinking aloud...

Flat works, makes it a lot easier to lap the surfaces before mounting the tec. Thickness must be difficult to judge, thinner leads to better temperatures but deflection under load has to be negligible to avoid damage to the tec. I'm not sure the extra square on the base is always needed, not all motherboards have awkwardly placed resistors, in which case 4mm can be taken off the base thickness with no likely ill effects. Cheers for the photo, makes things clearer.

Back to heat exchangers, for the CP15G01, which is a (good) cold plate that will fit six 40mm square tecs, I was quoted £130 a pair. The OP is using a larger size that will hold twelve 40mm square tecs. Three are needed to do it "properly" in my strange definition of things, so £200 on cold plates. Twelve tecs at $26 each before shipping, so £200 on peltiers as well. Probably looking at just under £500 in total.

Based on this datasheet for the 12711-5M31-15CQ and assuming no form of temperature control and wiring in series pairs, it would move 500W at a delta of 20 while using 33A at 12V. Or 240W at delta of 30 for 360W. Alternatively in series triples, it would move 240W at a delta of 20 while using 10A at 12V. So it would be initially expensive, but the running costs fairly low. This was the result of me optimising for cop at 12V supply voltage.

The other approach would be to aim for 200W processor load moved at delta 20, six of them at 6V would move 250W at cop of 1.3 so is probably about right, 50W parasitic when insulated can't be orders of magnitude off. £100 in pelts, about 200W electrical, probably £100 in other raw materials to make the cold plates and insulation. Should be possible to completely eliminate condensation by wrapping tubing in insulation, then heater wire, then insulation again. /thoughts

 

[zipdogso]

Sweet jesus....JonJ678 the way you rattle off figures makes my head spin...LOL


hey Ultra....we've found a working forum !!!!!

Yes-oops. I'm tired. 3-15V input wouldn't be very useful...
Post edited

Well I think £95 was a bit steep £50 would be more like I pay...These supplies might not be too great in actual usage unles your undervolting 263 couple or higher TECs.....but certainly if you can get it cheap enough very useful for checking out single TECs undervolted against performance charts etc and if you can set up a TEC holder with good sensors in the cold/hot plates you can get an idea what goes on with Dt etc. I built one from a 3mm copper coldplate and a 5mm copper hotplate (about the very minimum sizes you can use.) i can bolt any size TEC to it and I have a high end asetek vapochill micro bolted to the hotplate good for 160-170w ( dirt cheap nobody wants them..) sensors are drilled in the plates. You can rack the voltage incrementally and record the temps there will be a point where the difference between the hot and cold sides jumps i.e. big Dt and that's the most efficient (sic) sorry best COP point to run the TEC.

 

[JonJ678]

Cheers

Heated tubing as a means of dealing with condensation is apparently an industry standard approach. It's called "trace heating". Eventually I'll have an idea which isn't already extensively patented

 

[zipdogso]

Jon1678....you might like this....

All TECs the world over are made from the same material it is pretty stable there isn't much variation at a temp of 27ºC it passes about 0.1227v per couple of pellets (each couple is the same stuff just one half "doped" positive the other negative.) so a standard 127 couple TEC is 0.1227v x 127 = 15.4 volts. It is the physical size of the couples that determines current.
This standardisation also loosely applies to performance charts. To get a "ballpark" figure it is not always necessary to have the exact chart just use the right current and the close value of couples and remember to allow for a variation of (couples you have/couples on chart) volts.

I like ferrotec ( some don't.) if you havent done it go to their site - thermal solutions - thermoelectric modules - single stage TEC - then pick a TEC click the model number and scroll down if you have a small screen. Ferrotec give 4 very interesting charts initially at a hotside temp of 50ºC next click the thermal modelling tab (you might have to scroll back up !).
here you can change the hotside temp, I don't know why they have suddenly relabelled it ambient temp If your using an HSF on the hotside with a known rating you can enter that careful here it can really screw results - don't know why you would want to increase the number of modules doesn't make much difference and if you want you can enter a starting Dt then click recalculate and voila new performance charts.
Very handy if you can read them and with your engineers mind I am sure you will.

 

[Pneumonic]

Back to heat exchangers, for the CP15G01, which is a (good) cold plate that will fit six 40mm square tecs, I was quoted £130 a pair. The OP is using a larger size that will hold twelve 40mm square tecs. Three are needed to do it "properly" in my strange definition of things, so £200 on cold plates. Twelve tecs at $26 each before shipping, so £200 on peltiers as well. Probably looking at just under £500 in total.

do you mean using three plates the outside plates as hot and a shared inside cold plate? or three sets?

 

[Ultrasonic2]

Sweet jesus....JonJ678 the way you rattle off figures makes my head spin...LOL


hey Ultra....we've found a working forum !!!!!



.

Yes .. for now anyway

 

[JonJ678]

I've met the ferrotec website and you're quite right, I do like their graphs
I don't suppose you know what the ceramic is do you?

I meant a shared cold plate with a hot plate either side yeah. In addition to this you could get some aluminium or copper sheet and put sides on the sandwich, screwed to the two hot plates but with a channel milled to keep it off the cold plate, or shims, or cut down the cold plate slightly etc. Drill holes in a couple of sheets to pass the hot and cold water tubes through and do the same to the ends. You end up with a largish box where the entire casing gets warm while running, but the cold plate in the middle (insulated from the hot plates and the casing) remains cold. Fill the box with closed cell foam and I think it's as effectively condensation proofed as possible.

Benefit is a smaller system as you get more tecs in less space and that insulating the cold side of a pair of plates is no longer such an issue.

 

[zipdogso]

I meant a shared cold plate with a hot plate either side yeah. In addition to this you could get some aluminium or copper sheet and put sides on the sandwich, screwed to the two hot plates but with a channel milled to keep it off the cold plate, or shims, or cut down the cold plate slightly etc. Drill holes in a couple of sheets to pass the hot and cold water tubes through and do the same to the ends. You end up with a largish box where the entire casing gets warm while running, but the cold plate in the middle (insulated from the hot plates and the casing) remains cold. Fill the box with closed cell foam and I think it's as effectively condensation proofed as possible.

Benefit is a smaller system as you get more tecs in less space and that insulating the cold side of a pair of plates is no longer such an issue.

Phew.....at first reading this seems quite complicated and a pain if your not a reasonably competant DIYer or just don't want to have to do too much work.

 

[Pneumonic]

I've met the ferrotec website and you're quite right, I do like their graphs
I don't suppose you know what the ceramic is do you?

I meant a shared cold plate with a hot plate either side yeah. In addition to this you could get some aluminium or copper sheet and put sides on the sandwich, screwed to the two hot plates but with a channel milled to keep it off the cold plate, or shims, or cut down the cold plate slightly etc. Drill holes in a couple of sheets to pass the hot and cold water tubes through and do the same to the ends. You end up with a largish box where the entire casing gets warm while running, but the cold plate in the middle (insulated from the hot plates and the casing) remains cold. Fill the box with closed cell foam and I think it's as effectively condensation proofed as possible.

Benefit is a smaller system as you get more tecs in less space and that insulating the cold side of a pair of plates is no longer such an issue.

JonJ, I had planned to mount mine inside just a normal plastic electronics kit box I can actually get one for £8 which is exactly the right size.. just filled with the silicon closed cell and a couple of holes drilled for the hot/cold loops.. if you were feeling really adventurous.. once you know you have a working HX you can buy the silicon based foam as a spray.. so just close the box and spray in the foam.

 

[surge]

I don't suppose you know what the ceramic is do you?
.

I would imagine it is alumina (Al2O3), possibly with some additives

 

[zipdogso]

JonJ, I had planned to mount mine inside just a normal plastic electronics kit box I can actually get one for £8 which is exactly the right size.. just filled with the silicon closed cell and a couple of holes drilled for the hot/cold loops.. if you were feeling really adventurous.. once you know you have a working HX you can buy the silicon based foam as a spray.. so just close the box and spray in the foam.

I think I have lost the idea a bit here...

you plan to sandwich the TECs between two copper plates and then your heat exchangers and then put all that in an insulated box...is that right ?

I can see now where JonJ678 was going with his idea...If I HAD to choose it would JonJ678's but if it was my own choice I'd choose neither. Insulating the hotside doesn't sit with me at all, not even slightly you really don't want to do it. Insulating both sides to any great degree in the same box in my opinion is a seriously bad idea. TECs are only 3mm thick on average you cannot have the heat off the hotside going the wrong way...it is only a 3mm trip to the coldside if you have sufficent heat crossover to the coldside it sets up a heat loop which will eventually kill the TEC...think about it....TECs are heat pumps, they are merrily working away removing the heat you apply to the coldside and then all the sudden heat starts arriving from the hotside as well so it does its best to pump that too and starts going into overdrive...a heat loop. A heat loop is basically what happens when you power a TEC with no cooling on the hotside but don't try it, as I am sure you know, TECs die in seconds literally.

 

[Pneumonic]

I think I have lost the idea a bit here...

you plan to sandwich the TECs between two copper plates and then your heat exchangers and then put all that in an insulated box...is that right ?

I can see now where JonJ678 was going with his idea...If I HAD to choose it would JonJ678's but if it was my own choice I'd choose neither. Insulating the hotside doesn't sit with me at all, not even slightly you really don't want to do it. Insulating both sides to any great degree in the same box in my opinion is a seriously bad idea. TECs are only 3mm thick on average you cannot have the heat off the hotside going the wrong way...it is only a 3mm trip to the coldside if you have sufficent heat crossover to the coldside it sets up a heat loop which will eventually kill the TEC...think about it....TECs are heat pumps, they are merrily working away removing the heat you apply to the coldside and then all the sudden heat starts arriving from the hotside as well so it does its best to pump that too and starts going into overdrive...a heat loop. A heat loop is basically what happens when you power a TEC with no cooling on the hotside but don't try it, as I am sure you know, TECs die in seconds literally.

Zipo,

Sorry I am talking about the outside of the HX... so the TEC hotside will still pump heat into the HX plate, where the water running through the tubes in the plate will take it out to the rads.. issue is that on the cold side the surface on the other side to the TECs will also be cold.. so I don't want condensation forming and dripping everywhere

So would have something like this....

--------------------------AIR---------------------------------
----> Chilled Liquid Loop ||||| HX Plate |||||| ----> CPU Block
---------------------------TEC Cold Side --------------------
---------------------------TEC Hot Side ---------------------
----------> Water Loop ||||| HX Plate |||||| ----> Radiators
--------------------------AIR---------------------------------

I guess its less of a problem if I use three plates, as the cold parts would all be in the middle..

--------------------------AIR---------------------------------
----------> Water Loop ||||| HX Plate |||||| ----> Radiators
--------------------------- TEC Hot Side --------------------
--------------------------- TEC Cold Side --------------------
----> Chilled Liquid Loop ||||| HX Plate |||||| ----> CPU Block
--------------------------- TEC Cold Side --------------------
--------------------------- TEC Hot Side ---------------------
----------> Water Loop ||||| HX Plate |||||| ----> Radiators
--------------------------AIR---------------------------------

 

[Frozennova]

Sorry to take this a bit of topic but Ultrasonic would you be able to send me a copies of all the solidworks files for you cpu block as I'm teaching myself solid works and cant get the flow simulation to work on my assembly I can PM you my email if you need it.

Thanks

 

[JonJ678]

If you're not keen on either layout zip, what are you suggesting as an alternative?

I'm not sure you've followed what I'm suggesting, though it's close to Pneumonics second diagram with some condensation proofing thrown in. I'm fairly busy today, but should be able to spend some time modelling tonight. so I'll knock up a sketch of my current design.

Cheers surge, that sounds familiar. Not even sure you'd need to dope it, must look into this.

 

[zipdogso]

Zipo,

Sorry I am talking about the outside of the HX... so the TEC hotside will still pump heat into the HX plate, where the water running through the tubes in the plate will take it out to the rads.. issue is that on the cold side the surface on the other side to the TECs will also be cold.. so I don't want condensation forming and dripping everywhere

I guess its less of a problem if I use three plates, as the cold parts would all be in the middle..

--------------------------AIR---------------------------------
----------> Water Loop ||||| HX Plate |||||| ----> Radiators
--------------------------- TEC Hot Side --------------------
--------------------------- TEC Cold Side --------------------
----> Chilled Liquid Loop ||||| HX Plate |||||| ----> CPU Block
--------------------------- TEC Cold Side --------------------
--------------------------- TEC Hot Side ---------------------
----------> Water Loop ||||| HX Plate |||||| ----> Radiators
--------------------------AIR---------------------------------

Gotcha....I was sure I was mis-understanding somewhere...what threw me was the put it all in a box and spray foam in idea...in that case you wouldn't have the air gap would you, it would be insulation all the way to the inside of the box. I see your way you are just going to attach neoprene to the inside of the box leaving an air gap. Yeah....I would vote for the second way with 3 HX....similar to JonJ678's idea.

Even with your HX on the outside of the TECs hotside if you insulated the outside of the HX to any great degree I think you would get some heat crossover on the TEC. To be honest I am still not sure about the whole thing being boxed in even with the air gap. A small insulated box will get warm quick I don't think your HX is efficient enough to mop up all the heat I am pretty sure some will still be radiated off the back of the HX.