Tec cooling idea need help - Overclock.net

Death Dealer · 04-11-08

Ok I want to know if this will fail if I try. I want to get a tec, but instead of sandwitching between my cpu and waterblock would it be possible to make a plate attach the tec cold side facin the plate and attach a HSF to the hot side and run it in my loop but would I still need to insulate for condensation. The reason I bring this up is I want to cool my quad core with only a dual 120mm rad to conserve space in my case. I would use like a 100watt tec if I follow through with this would I need to worry about condensation??????????

The plate would be done on a cnc mill if that matters.

Here is a rough drawing I don't have FeatureCAM up and running yet =(

sublime0 · 04-11-08

Thats basiclly the idea Of Thermal Electric cooling. YOu know you can buy them now. From Chilltech

Death Dealer · 04-11-08

Yeah but I want the tec away from vital parts of the computer so I don't have to worry about the this dripping water all over I just want it to chill the water that passes over it so the quad gits nice cold water.

If they make them could you post the site you saw them at please!!!!!!!!!!

DuckieHo · 04-11-08

Quote:

Originally Posted by Death Dealer

Yeah but I want the tec away from vital parts of the computer so I don't have to worry about the this dripping water all over I just want it to chill the water that passes over it so the quad gits nice cold water.

If they make them could you post the site you saw them at please!!!!!!!!!!

http://www.ultraproducts.com/product...&productID=571

However, they don't perform much better (if any better) than normal air coolers.

What you describe is a TEC water-chiller. I have experimented with it (as others). They are not very efficient or effective. I was using 100w to cool my water and drop my CPU temp by.... 1-2C.

thunder_2008 · 04-13-08

Hello DD & others,

I can tell you that you have a really good point duck. I have an Arctic Web tec 437watts w/600watt p.s. unit. It gets my temps to 25-30load in 80ambient temp. I'm using true measured 437watts and I also have my northbridge and memory in the loop.

Without the north or memory I have had a steady -14idle & 0load.

I would say DD, that what makes your set-up different as per your drawings is that you have the water loop attached to the cold side of the tec as opposed to the hot side.

In the drawing you show the heatsink fan as cooling the hot side...while I can not think of any current manufacturer that makes any setup this way. I would admit it is not such a bad idea at all~!!!

As stated earlier my tec uses the hot side to pass water over, which transfers all 437watts of heat into my Aqua Computer 360XT water tower. It is not good as it's designed to only intermittantly turn on it's three(very silent)fans for a brief minute then go 3 or more minutes before doing it again.

Since I added the tec the fans are on always, also the water temps are around 50 coming in and 30 going out cold again. Before the hot water coming in was 30 and 17 and lower going out.

So, in theory to have the hot side of the tec cooled sufficiently by air and have the cold side by water would be good.

The problem here is that then you would be dispersing 437watts of cooling into not just the cpu block but the water as well, thus dissipating the benefit of direct tec cooling.

What would be best would be to either cool the water it's self with an array of tec.s'(much like the new coolit sys.); or ....

And I think the most efficient economical would be a hybrid of watercooling the hot side as well as a strategically placed heatsink and fan. Thus cooling the heatsink allowing the water to cool prior to reaching the radiator.

But then the guys that made the Arctic and the Coolit sys. . They would say people don't want to hear a fan in the computer when they already have ones on the radiator.

Still a water-jacketed air-heatsink would in the least seem to be a more efficient future design, I'm certain this alone could remove at least 5degrees(more still if a small silent fan) and with the fan much more if designed to run quiet.

I think of many people who have all water cooled computers yet still have the small accessory fan on the motherboard of Asus motherboards.

The idea is good just needs evolution...one thing I will say is that new ideas go along way. After all is that not how the modern computer is where it is now?

I have some ideas of my own on using phase change. I am selling my Arctic Web on eBay now with the power supply as well....so anyone wanting to get into tec cpu cooling at a great price check it out.

Adrienspawn · 04-13-08

Quote:

Originally Posted by thunder_2008

Hello DD & others,

I can tell you that you have a really good point duck. I have an Arctic Web tec 437watts w/600watt p.s. unit. It gets my temps to 25-30load in 80ambient temp. I'm using true measured 437watts and I also have my northbridge and memory in the loop.

Without the north or memory I have had a steady -14idle & 0load.

I would say DD, that what makes your set-up different as per your drawings is that you have the water loop attached to the cold side of the tec as opposed to the hot side.

In the drawing you show the heatsink fan as cooling the hot side...while I can not think of any current manufacturer that makes any setup this way. I would admit it is not such a bad idea at all~!!!

As stated earlier my tec uses the hot side to pass water over, which transfers all 437watts of heat into my Aqua Computer 360XT water tower. It is not good as it's designed to only intermittantly turn on it's three(very silent)fans for a brief minute then go 3 or more minutes before doing it again.

Since I added the tec the fans are on always, also the water temps are around 50 coming in and 30 going out cold again. Before the hot water coming in was 30 and 17 and lower going out.

So, in theory to have the hot side of the tec cooled sufficiently by air and have the cold side by water would be good.

The problem here is that then you would be dispersing 437watts of cooling into not just the cpu block but the water as well, thus dissipating the benefit of direct tec cooling.

What would be best would be to either cool the water it's self with an array of tec.s'(much like the new coolit sys.); or ....

And I think the most efficient economical would be a hybrid of watercooling the hot side as well as a strategically placed heatsink and fan. Thus cooling the heatsink allowing the water to cool prior to reaching the radiator.

But then the guys that made the Arctic and the Coolit sys. . They would say people don't want to hear a fan in the computer when they already have ones on the radiator.

Still a water-jacketed air-heatsink would in the least seem to be a more efficient future design, I'm certain this alone could remove at least 5degrees(more still if a small silent fan) and with the fan much more if designed to run quiet.

I think of many people who have all water cooled computers yet still have the small accessory fan on the motherboard of Asus motherboards.

The idea is good just needs evolution...one thing I will say is that new ideas go along way. After all is that not how the modern computer is where it is now?

I have some ideas of my own on using phase change. I am selling my Arctic Web on eBay now with the power supply as well....so anyone wanting to get into tec cpu cooling at a great price check it out.

Sorry, but I just licensed your post to Warner Bros. and they're gonna make a movie about it.

----------------------

Ok, on topic now:
For those who dont understand what hes saying (like the first guy who responded

) hes asking if you can cool the TUBE of water using a cold plate,followed by the TEC and a heatsink with fan. I would say this wouldn't be too effective since :

1- You might have problems freezing the tube.
2- The surface your cooling will be very small compared to the size of what your using to cool it.
3- The water will zip right by it, not having too much time to dissipate its heat.

A better solution would be to TEC-cool a reservoir, where you have much more space to cool, and you can get an aluminium one which I believe cools better than a plastic tube. Also, more water would be in contact for longer, so the water temps will be colder, and the heasink on the TEC will have to cool off a bit less.

Maybe you could even cool the radiator...hmm..

thunder_2008 · 04-14-08

Yes...I actually came up with an effective solution using just exactly what you mentioned, having the tec chill the reservoir. However, a very experienced old tec designer told me that it would require alot of water on the hot side as well.

As air cooling the hot side with radiator and fan alone would not be good enough. Really, here is a way in which it can be done....but before I share that with anyone let me get the patent rights for it first...lol.

ttyl,

Thunder

Ultrasonic2 (muffy) · 04-14-08

i must admit i still can't fully understand what everyone is saying ...
my 2 sense on the original questions is you dont have to worry about condensation unless your tec/ water is getting to dew point or less .. it would seem that the temp wouldn't get that low cause you'll have an inline radiator(?) warming the water backup if it gets below ambient. i assume you have this inline radiator on the cold side cause a 100watt TEC couldn't cool a modern cpu

"I'm using true measured 437watts and I also have my northbridge and memory in the loop."

are you say your using 437watt off electricity or your Qmax(cooling power) is a measured 437?

As stated earlier my tec uses the hot side to pass water over, which transfers all 437watts of heat into my Aqua

This comment suggests that your using 437 watts of electricity ? however it's not allowing for the electricity used and the heat load cooled (probably and extra 200watts)

The problem here is that then you would be dispersing 437watts of cooling into not just the cpu block but the water as well, thus dissipating the benefit of direct tec cooling.

This comment seems to suggest that your 437 is "cooling" (Qmax) not heat (electricity)

Your whole post has me lost ?

You do realize that electricity used/heat "Pmax" is not that same as heat cooled/moved "Qmax"

thunder_2008 · 04-15-08

I do now...lol. I meant the amount of electricity....is measured at 437watts...it is a 600watt power supply yet there's an adjustable level and I did not want to set it more than the 437watt level.

The condensation is a real destroyer if you don't use neoprene around the tec as well as the back plate, with my setup.

I would imagine it is less severe but still harse with 100watts if directly placed like my block is, however if theres water passing on the cold side it would lessen any cooling thus negating the whole reason for tec'ing a block at all.

Ultrasonic2 (muffy) · 04-15-08

i still think people try to control condensation in the wrong way. controlling it correctly would help reduce a tec's biggest drawbacks

Condensation

Electrical consumption

TEC's should be controlled to keep a variable load (CPU) at a given temp. obviously above due point. In doing so the TEC will consume less electricity at light loads while not sending the lightly loaded CPU temp into the negatives. but keeping the load at X temp when the CPU is under full load

04-14-08	#8 (permalink)
Ultrasonic2 (muffy) Extreme Cooler Join Date: Jun 2007 Location: Auckland New Zealand Posts: 765 Rep: 45 Unique Rep: 32 Trader Rating: 0	i must admit i still can't fully understand what everyone is saying ... my 2 sense on the original questions is you dont have to worry about condensation unless your tec/ water is getting to dew point or less .. it would seem that the temp wouldn't get that low cause you'll have an inline radiator(?) warming the water backup if it gets below ambient. i assume you have this inline radiator on the cold side cause a 100watt TEC couldn't cool a modern cpu "I'm using true measured 437watts* and I also have my northbridge and memory in the loop."* are you say your using 437watt off electricity or your Qmax(cooling power) is a measured 437? As stated earlier my tec uses the hot side to pass water over, which transfers all 437watts of heat into my Aqua This comment suggests that your using 437 watts of electricity ? however it's not allowing for the electricity used and the heat load cooled (probably and extra 200watts) The problem here is that then you would be dispersing 437watts of cooling into not just the cpu block but the water as well, thus dissipating the benefit of direct tec cooling. This comment seems to suggest that your 437 is "cooling" (Qmax) not heat (electricity) Your whole post has me lost ? You do realize that electricity used/heat "Pmax" is not that same as heat cooled/moved "Qmax" __________________ Watercooling System Radiators: 3x 240SGTS radiators in parallel to each other. Fans: 6x Scythe ULTRA KAZE 120 x 38 mm fans sucking though a 7cm shroud Pumps: 2x MCP655 pumps in series RES: 80x400mm Water Blocks:D-tek FuZion CPU (stock), Swiftech MCW60 GPU Controller: M-Cubed T-balancer BigNG adjusts fan speeds and pump speeds to keep my system at X temp CPU: INTEL 6600 GO GPU ATI 4850 HD Six disks in RAID 0 Last edited by Ultrasonic2 (muffy) : 04-14-08 at 08:30 PM.

04-15-08	#10 (permalink)
Ultrasonic2 (muffy) Extreme Cooler Join Date: Jun 2007 Location: Auckland New Zealand Posts: 765 Rep: 45 Unique Rep: 32 Trader Rating: 0	i still think people try to control condensation in the wrong way. controlling it correctly would help reduce a tec's biggest drawbacks Condensation Electrical consumption TEC's should be controlled to keep a variable load (CPU) at a given temp. obviously above due point. In doing so the TEC will consume less electricity at light loads while not sending the lightly loaded CPU temp into the negatives. but keeping the load at X temp when the CPU is under full load __________________ Watercooling System Radiators: 3x 240SGTS radiators in parallel to each other. Fans: 6x Scythe ULTRA KAZE 120 x 38 mm fans sucking though a 7cm shroud Pumps: 2x MCP655 pumps in series RES: 80x400mm Water Blocks:D-tek FuZion CPU (stock), Swiftech MCW60 GPU Controller: M-Cubed T-balancer BigNG adjusts fan speeds and pump speeds to keep my system at X temp CPU: INTEL 6600 GO GPU ATI 4850 HD Six disks in RAID 0

		Overclock.net - Overclocking.net > Cooling > Peltiers / TEC
Tec cooling idea need help

04-11-08	#2 (permalink)
sublime0 nVidia Enthusiast Join Date: Jun 2006 Posts: 5,244 Blog Entries: 1 Rep: 228 Unique Rep: 157 Hardware Reviews: 4 Trader Rating: 7	Thats basiclly the idea Of Thermal Electric cooling. YOu know you can buy them now. From Chilltech

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)