85watt chiller Aircooled? 4-5c drop?
 

here is my deal, i'm going to purchace an RBX water block for my cpu
i have currently a maze2, and shite piece of volcano 7+ laying around
how much of a temp drop do you think i would benifit from on my cpu, as it is the only block in the cooling loop. if i was to take an 85watt, cold side into the water loop via the maze2. and hot side aircooled with the volcano 7+, its one of those fangled thin fin copper heatsinks.
i'm looking for around 5c or 6c to make me happy in spending the 20 bucks on the pelt. granted the 85w probably wont pull more then 60 on a 12v line, do you think that air cooler could handle air cooling a 120w?

lets hear it guys!

Wont work. You'd need a much more powerful pelt and much better cooling of its hot side.

so your saying a heatsink normally rated to keep 80watts kool wouldn't keep that 60wat pelt (effectively) in order grom? how do you figure that?
2ndly. your tellen me that it wouldn't effect ANY change to the water?
nothing at all. so where is that cold filtered into the maze2 going?

make another note, LIKE i said in the title i'm not looking for sub zero water temps. i'm only looking to cool it down by a few C's

First off, what CPU are you using? If the CPU and pump add up to more than 85W then your pelt is a heater not a cooler.

Then, take a guess at your V-7+ C/W rating. How much over ambient was your CPU when using it? Divide that by the max watts of your CPU.

jesus Groth i dont mean to be insulting but do you have shit for brains?
i currently run a maze2 with a Hydor L30, and a Black Ice Prime radiator (2 year old setup)

i'm talking about replacing the MAZE2 with an RBX
i currently get temps at the most 5c above ambient.. thats under load.
i'd be adding an 80w or 110 watt tec into the mix with the maze2 i'd have left over with a volcano 7+ cooling the tec.

you need me to draw you a picture?

No need to be rude. Do you want me to walk you through how it will work or not?

Vital data needed: your CPU's heat output, an estimate of the C/W rating of the heatsink you'll be cooling the pelt with.

Grothyou do realize the radiator will be staying intact right, as for the C\W rating of the heatsink, i can't find nothing anywhere as to what its C\W listing is, but i know its a 1/4inch coper base with milled thin coper fins, about 50 of em.

and its about 2 inches tall. usual footprint

the proc is an amd 2400xp oc'd right now at 1.75v 2260
i turn it up to 1.85v 2400 every so often, but rarely.
i'd say 80watt output heat?
i'mplannen on putting the thing between the radiator and cpu, so its gonna be cooling off room temp water.
or close 2.

Groth does not have shit for brains and does need the info that he asked for. Don't make me moderate a thread in this forum as I only know one function (banhammer).

phae, groth is only anoying me because he's basically trying to say that the tec cooled with a heatsink at that wattage, as is normally run in many a system will overheat trying to cool a roomtemperature object. now tell me that doesnt sound a little screw loose.

Think about how a peltier works. It has a maximum differential across it right that is part of the TECs specs? So how cold the cold side can get is limited by the effectiveness of your hot side cooling source. So to estimate how well your shit would work he needs performance numbers for the heatsink. I'd estimate its C/W at around 0.35 if you're running that 70mm fan full bore.

even that be the case bud, i should still see some cooling effect over 1 or 2 liters of liquid, as the unit would onlybe buggern with ambient temped water anyways.
i could understand that it'd be come a heater using something like a 170wat pelt. because i wouldn't be able to tame the hot side. but most stock hsf's should be able to take care of an 80-60watt pelt.
if they can't do that. what business do they have being on a proc?

Running a pelt chiller and a radiator on the same loop will accomplish little. If the pelt is cooling the water below ambient, then the radiator adding heat to the water. If the water isn't to be below ambient, then adding another or better radiator will give better results.

Chiller or rad?

i understand your point groth, but putting it right before the CPU after the radiator would give the full cold effect to the cpu water line before it went back around and was brought back up or down to ambient by the radiator. i'm asuming back down due to the CPU heat being enough to overpower an 80wat tec with a coldplate mounted straight to an athlon.

The temperature drop across a single device (rad, pelt chiller, CPU, etc) is on the order of tenths of a degree. Order won't change whether it will work or not.

You'll get at best about 0.5C doing that. The water will pass through the block so fast it will barely cool it.

Besides, your hotside will be roasting hot. Assuming 60W pumped, you're looking at sinking around 140W off the hotside (at 12V an 85W TEC will draw approximately 6.5A, so you have 6.5*12*12 = ~80W from the TEC plus the 60W pumped). With a 0.35C/W HSF that's a rise over ambient of nearly 50C - you're looking at a 75C hotside there, ouch.

Exactly, Butcher. That's where I was slowly heading.... Can your pelt maintain a 50 C difference between hot and cold sides? If not, then the 'cold' side will be hotter than the water and heat the water.

By the way, based on AMD's model 8 datasheet, and 2280 Mhz at 1.75V, your CPU probably puts out about 79 watts.

Wrong.
Unless the laws of physics do not apply to your computer your starting off with the wrong numbers to begin with. From your other replies you obviously know nothing of basic math and suggest you read up on it first before asking someone of Groths experience if he has shit for brains. People wonder why I lost my tact over the years......

Now there's something I actually pondered, then dropped when I saw how useless it was going to be...

Adding a TEC would normally drop the temp of the coolant, albeit not at the power levels estimated here. An 80 W pelt may, very roughly, suck 60 W of heat out of a loop, and I'm being optimistic.

The problem is keeping the rad in the loop.

With a rad, if the coolant temp drops below ambient, the rad will actually warm up the water.

If the coolant temp doesn't drop below ambient, then the rad efficiency drops to dismal levels (imagine a rad with a delta T of 1 or 2 deg C). In other words, it barely dissipates any heat at all.


The original question here isn't wether or not it'll work, there's no doubt that the configuration is valid. The question rather, is: what kind of temp drop could one expect?

This thread has digressed a little bit into how this would actually work, and Groth is right on top of it, as usual. Varsis doesn't seem to grasp the mechanics behind water cooling.

Varsis: Groth isn't kidding. The temperature change of the coolant, between the hottest and coldest point, within a loop, is minimal, and is indeed a fraction of a degree. Think about how you're pumping maybe (optimistically) 1.5 gpm, and your rig runs holds maybe 0.5 gallon: your coolant runs the entire loop three times per minute.

Heck, I'll even do the math (which I rarely do in a Forum, btw):
(from this article: http://www.nordichardware.com/articl...vattenkylning/ )
P*t/(C*m)=DT
P=79W
t=60s
m=5.64 kg
C has a tabular value of 4190 J/(kg*K).

dT = 0.2

0.2 deg C is the largest temp difference, within your coolant.

That's the maximum difference, between hot and cold in your current loop.

You had tact at some point :eek:?

Seriously though, your idea will not work. You most likely do not have enough cooling power in that TEC to do anything, since the CPU probably puts out way more heat than the TEC can move, and the TEC will be relatively toasty with that Volcano7.

Let us assume for a moment, however, that by some oversight on the part of every person that has replied in this thread, your TEC has a ridiculous amount of cooling power and could potentially lower your water temperature down to the 5 or 6ºC that you are looking for. As soon as the water gets below ambient (which it won't) your radiator will start warming it up, so you are being counterproductive.

Wow, such support, now I feel all warm and fuzzy! :D

Eh, I don't blame Varsis and bear him no malice. It's all the fault of my damned Classics professor. He inflicted the socratic method upon me, and now I have the annoying habit of trying to lead people to the correct answer by asking them questions. Some people are willing to do the thinkin', some get frustrated; life goes on.

A true teacher helps someone find the answer for themselfs. :)

You're in on the in crowd now... ;) (we got your back, bro!)

Jesus man. Such attitude would be sort of understandable if not for the fact that Groth is right. Why did you even ask for advice if you were unwilling to listen to an answer?

Let me on the "very poorly thought out idea" bandwagon thats forming around this thread.

Hey, I actually had the same idea, some time back... :p

Like I said, it's a valid configuration. Maybe not the most efficient solution... (probably wasteful actually) but it's not going to hurt anything.