• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

chilled water or TEC?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

aarion

Registered
Joined
Aug 7, 2002
i want to bring my water cooling setup to another level. What would you recomment?

A 172 watts peltier. And chilled water at 0-10c. Not really sure how cold the water gonna be. Just estimate.

Pls feedback. Thank you.
 

Paxmax

Member
Joined
May 8, 2002
LOL! What do you want to cool, your toes ?? :)

On the more serious level,
'What is cooling the chilled water? What capacity has it?
Will you use chilled water AND pelt ?
TEC, specs of the TEC, the voltage you intend to run it on and such.
What cpu ? What voltage? What frequency ?

You need to feed us with much info..... /Paxmax
 
OP
A

aarion

Registered
Joined
Aug 7, 2002
Paxmax said:
LOL! What do you want to cool, your toes ?? :)

On the more serious level,
'What is cooling the chilled water? What capacity has it?
Will you use chilled water AND pelt ?
TEC, specs of the TEC, the voltage you intend to run it on and such.
What cpu ? What voltage? What frequency ?

You need to feed us with much info..... /Paxmax

I just wan to cool the CPU for now. The chilled water will be cooled by a custom made fridge water chiller. What do meant by capacity? A 172 watt peltier with maybe 12 volts.

XP 1800 @ 1885 running at 2.08volt.

Planning to run it at least 2000 Mhz. Is the info enough?
 

aenigma

Member
Joined
Jul 4, 2002
Yeah looks like someone needs to read the thread ;)
Yeah chilled water and a pelt works pretty good.But I don't know if a 172 watt pelt can handle it since I don't use pelts anymore.But if the pelt can handle the heat, you will see some pretty good temps.

He was trying to ask what kind of refrigeration you plan on using.You said a fridge, but do you mean a cheapass minifridge, or a real full size refrigerator?
 

aenigma

Member
Joined
Jul 4, 2002
Warlord2 said:
you need atleast a 226watt, a 172 watt I dont believe will do it.
Yeah probably, but since he is talking about using chilled water a 172 watt pelt might do ok...
 
OP
A

aarion

Registered
Joined
Aug 7, 2002
sorry for the miss understanding.

let me clear things up. I want to either just use chilled water OR tec. Not boht together. Can't afford both.

For chilled water, i may need to build my own reservoir with the evaporater in it to cool it down.

For peltier, just use my current setup wih it. So you guys recommend at least 226 watts?
 

The Spyder

Member
Joined
Mar 25, 2002
Location
Oregon
ok....... 1. I only got 1990 out of my Xp 1800 and it would not load windows at 2.25v on a 8k3a+ with trusted true ram~ i would get a 1600 Xp agoia that can actualy reach the OC you want....... Then
2. Get a 226w Pelt and a radiator that can handel the head produced~

Spyder
 
OP
A

aarion

Registered
Joined
Aug 7, 2002
hehehe, but i found out that the chip will go higher if i can get the temp lower. Not only the can clock higher but also can lower the voltage at the same time. In my country, there is no more
XP 1600 in the market. If there is. the core is not overclockable. It the AGOIA F batch. Thats the problem. I can have 2.6 volt on my chip also.

2. The cost of getting the peltier and make a water chiller is almost the same. But i don't know about the power comsumption.

Thank you anyway.
 

aenigma

Member
Joined
Jul 4, 2002
Ok I see, definately chiller all the way.Peltiers suck compared to water chilling.As long as you don't use a minifridge or something like that.
 

doc

Member
Joined
Jun 6, 2001
Location
Michigan
Chillers are a good way to go. Do you have a fridge already or are you looking for one. If you do not have anything yet, consider using TECs for a chiller. I have done this, and it works well. I made a small chiller at work. It used a 5 gallon res. that was well insulated and cooled by 4 40 watt TECs. The TECs were run 24/7, but at a low power most of the time. The equipment being cooled did not run 24/7, so we had a coolant temp -5C or less most of the time. The PS that ran the TECs was an old computer PS and we switched between the 5 volt or 12 volt setup to get more or less heat pumping depending on our heat load. Not to mention that 5 gallons of water-glycol can absorb alot of heat before its temp comes up 1C, a buffer of a sort.
 

Paxmax

Member
Joined
May 8, 2002
Okey aenigma, why wouldn't a TEC work for a chiller?
Doc just said "I have done it, it worked"

If properly cooled, a TEC works just like a charm, although it comsumes a bit more energy than a refrigirator phase changer.
I think it is far easier to control a TEC than a phase changer in my opinion, but thats probably due to my better knowledge with eletronics than gas flow regulating.
 

aenigma

Member
Joined
Jul 4, 2002
Paxmax said:
Okey aenigma, why wouldn't a TEC work for a chiller?
Doc just said "I have done it, it worked"

If properly cooled, a TEC works just like a charm, although it comsumes a bit more energy than a refrigirator phase changer.
I think it is far easier to control a TEC than a phase changer in my opinion, but thats probably due to my better knowledge with eletronics than gas flow regulating.

Well I am truly sorry that you don't grasp the fact that TEC's are extremely uneffecient. :)

It will work if you don't want nice temps.I know of someone that did it, I think it was 4 226 watt pelts and it didn't do a very good job considering the cost and ammount of power it consumed.
A phase change system is MUCH more effecient.Cheaper if you know what your doing, and lower temps.

If you notice, Doc didn't give any information about his system.
If loaded liquid temps aren't at least -20c, then as far as I am concerned, it doesn't work.And he already said he doesn't run his system all the time, if he did then his liquid temps would not be -5c.
Oh and it isn't a "refrigirator phase changer", it is a phase change system/refrigeration system. ;)

Numerous people have tried using pelts as water chillers, but it never works out.Most people already know this.
Go ahead and make a pelt chiller that actually works and prove me wrong. :)

Oh by the way, you can control a water chillers temperature with a thermostat.
 

Paxmax

Member
Joined
May 8, 2002
aenigma said:
Well I am truly sorry that you don't grasp the fact that TEC's are extremely uneffecient. :)
I do know that a TEC if less efficient than a phase changer, I just pointed that out specifically in my previous post.

It will work if you don't want nice temps.I know of someone that did it, I think it was 4 226 watt pelts and it didn't do a very good job considering the cost and ammount of power it consumed.

Then someone didn't take enough care to cool the hotside of the TECs, otherwize he would have nice temps. Yes,TECs are costly. Regarding the rest, we have already covered the efficiency issue.

A phase change system is MUCH more effecient.
We have already covered the efficiency issue.

Cheaper if you know what your doing, and lower temps.

You have to know alot if you plan on doing all by yourself. You need to have the skill of how so hard solder, own or have access to a torch and proper solder. You need coppertubing and stuff. It will be hard task to get it portable. Remeber folks, you are dealing with high pressure gas here, it has it's way of escaping. The only thing that I might percieve as easy would be to go really ghetto style, and thrash a fridge, bend it up and fit it into a big bulky bucket, and hopefully not got a kink on the tubes. With a darn good built phase changing system, of coz you can expect low temps. Slap in a coldplate + peltier to your waterblock and you are ready to go. Though in his case I also agree that a refrigiator waterchiller will yeild better temps. If you aren't afraid of the hazzle a re-friggin unit can be a really good choice.

If you notice, Doc didn't give any information about his system. If loaded liquid temps aren't at least -20c, then as far as I am concerned, it doesn't work.

True, Doc didn't mention what system he has. Ok, thats your definition of "works", I bet atleast 70-80% of the community would be very happy with -19.9 C loaded temps.

And he already said he doesn't run his system all the time, if he did then his liquid temps would not be -5c.

True, but he didn't exactly input alot of power into it either, probably less than a 1/4 HP 'pressor.

Oh and it isn't a "refrigirator phase changer", it is a phase change system/refrigeration system. ;)

Ohh..... pardon mee..... ;)

Numerous people have tried using pelts as water chillers, but it never works out.Most people already know this.
Go ahead and make a pelt chiller that actually works and prove me wrong. :)

Not my first priority, but my second.

Oh by the way, you can control a water chillers temperature with a thermostat.

Yeah, *CLICK* *B-R-R-R-R-R* *CLICK* *annoying silence* ;)
 
Last edited:

aenigma

Member
Joined
Jul 4, 2002
hahaha hey you were the one that wanted to control temp, why you would want to do that is beyond me.Maybe lack of insulating skills? ;)

This guy I speak of watercooled the hot side.His temps (water) were about 0c, i think they would have gone a bit lower.But for the cost and effort it is not worth it.I started with no knowledge of refrigeration, it isn't impossible.But if you don't want to learn it.Then just make a waterchiller using a window a/c unit or something.Or just don't do it if you can't handle it.

-20c loaded temps are fine for alot of people I am sure, but please note I said liquid temps.CPU temps will be much higher.But if you can use a modern day processor, 4 40 watt pelts and get a loaded liquid temp of -20c, I will be in awe :)

Now I am not exactly sure what you are talking about when you refer to the 1/4hp compressor?But I think your trying to say 4 40 watt pelts is less than a 1/4hp compressor.It will cool less, but it will probably draw about the same ammount of watts.
Pelts will also require 4 individual heatsinks a capable power supply and 4 more heatsinks or waterblocks to cool the lqiuid.All that for lousy temps.

But if you want to make a pelt chiller, be my guest.A window a/c unit/refrigerator/freezer/dehumidifer etc. would be much better, but we both agree on that.
 

Paxmax

Member
Joined
May 8, 2002
aenigma said:
hahaha hey you were the one that wanted to control temp, why you would want to do that is beyond me.Maybe lack of insulating skills? ;)

Hahaha..no. :cool: I just don't see a gain at all in having water on -40C if it's going to be -20C under load anyway. -"Hey dudes check my fancy smancy rig out it's capable of cooling water to -40C when I'm not using it!" I don't need to cool it more than to reach the lowest temp I get under load.

This guy I speak of watercooled the hot side.His temps (water) were about 0c, i think they would have gone a bit lower.But for the cost and effort it is not worth it.I started with no knowledge of refrigeration, it isn't impossible.But if you don't want to learn it.Then just make a waterchiller using a window a/c unit or something.Or just don't do it if you can't handle it.

Well, I think it is good for the both of you that you don't use TECs.

-20c loaded temps are fine for alot of people I am sure, but please note I said liquid temps.CPU temps will be much higher.But if you can use a modern day processor, 4 40 watt pelts and get a loaded liquid temp of -20c, I will be in awe :)

So will I !! :D

Now I am not exactly sure what you are talking about when you refer to the 1/4hp compressor? But I think your trying to say 4 40 watt pelts is less than a 1/4hp compressor.

Exactly.

Pelts will also require 4 individual heatsinks a capable power supply and 4 more heatsinks or waterblocks to cool the lqiuid.

No, I can do it with one heatsink/or/WB and one coldplate. I'm astounded that you think it's REQUIRED to use 4 individual peices at each end.

But if you want to make a pelt chiller, be my guest.A window a/c unit/refrigerator/freezer/dehumidifer etc. would be much better, but we both agree on that.

Thank you for your approval and yes, a phase changer has better performance/power requirement.
 

doc

Member
Joined
Jun 6, 2001
Location
Michigan
OK, it would seem I have some smoke here. You are right I didn’t provide much detail, so here is some more detail. I used a 5 gallon res, 3800 cc/gal * 5 gal = 19000 cc , water with some ethylene glycol in it is almost the same density as water and we will assume that the density does not change a function of temp. So will use 1 gram / cc , specific heat of 1 cal / gram C and 4.18 Joules / cal. So it looks like this:

19000 cc * 1 gram / cc * 1 cal / gram C * 4.18 Joules / cal = 79420 Joules / C and 1 Joule / sec = 1 watt

We get 79420 watt sec / C which is how much energy can be stored in 5 gal of water per degree C of dT.
What does this mean? If you have a 100 watt heat load it would take 794.2 seconds or 13.2 minutes to raise the temp of that five gallons 1 degree C, without a heat pump running. The 5 gallons acts as a large heat cache. My actual heat load was about 30 watts ( not a CPU ) which would give 44 minutes for a 1 degree C rise in the coolant temp. Now add some heat pumps or TECs, 4 of them.

5 volt specs
dT max = 66C
COP = 0.41
Qc = 2.15 watts X 4 = 8.6 watts
Power input = 5.25 watts X 4 = 21 watts

12 volt specs
dT max = 70.8C
COP = 0.34
Qc = 12 watts X 4 = 48 watts
Power input = 35.6 watts X 4 = 142.4 watts

As you can see dT does not change much as a function of input power, this is why the res was very well insulated ( I used Great Stuff or foam in a can ). The TECs would be removing heat 24/7 and lengthen the time that it takes to rise 1C. –5C is not the limit by any means. One time I had a 4 day weekend (yahoo!) and came in on Monday and the coolant was
almost frozen ( like a popcicle, yes a green one ), remember I only used just a little ethylene glycol to go to about –10C and ran at 5 volts, my temp was at –16C ( good insulating job ). Wanted to keep the specific heat up high and a low viscosity. In reality the specific heat is ahout 0.94 cal / gram C, which one could just multiply by 0.94 at the end of it all. However in this example here, I did not calculate for losses due to heat passing into the res thru the insulation. That is a little more than what I want to do here. I have nothing against phase change cooling, TECs should not be ruled out. When used correctly they work well, and I like the idea of no compressor and low noise.