Water Doubts

[DeadHeart]

I can't decide what radiator to use! Damn! I like the Big Momma, but it is so small... I was looking for some 35ºc (full load) action on my TBird 1GHz @ 1.44Ghz (1.94v), but a friend of mine told me to forget it, because I would not achieve this kind of temps. My system will run with 1/2" tubing, using a DD Maze2 and an Eheim E1250. However, I don't know which radiator I should use. I'm gonna do the voltage mod on my mobo, therefore this is gonna get hot! I will probably buy a radiator here in Portugal because I don't want to pay taxes... What should I look for? I really wanted 35ºc. Do you think I will be able to run at this temp will a radiator like Big Momma? My ambient temp is always around 26-30ºc. My friend told me that I would have temps around 45ºc. What do you think?

[GuyBFF]

I'm not sure if you'll hit 35c with ambiant at 30, but you surely won't be at 45c. Of the units available, the Big Momma style heater cores continue to out perform almost everything else. I'd also watch for the new Black Ice, but currently the Big Momma style will beat these smaller units.

Usually you will find water temps 5 over ambiant and CPU at 5 over water.

When you order the Maze 2 ask for the 1/2" "Upgrade" version. It has more surface area above the die and also flows much better.

VModing doesn't add that much temperature when watercooling, it tends to scale much better than other forms of cooling. Example a 1500 might only be a couple degrees over a 750, Vmodding only a couple more. If your system was air cooled, it would surely increase greatly.

I run soft jumper on a ASUS A7V133, one consideration is don't use to low a resistor, as when you need to boot default your voltage will be very high. Try to find one that will max at 2.2v with maximum voltage settings. If your using jumpers, then don't worry.

[TNT2bluz]

double post

[TNT2bluz]

Deadheart, given that EU has more or less a unified tarriff system, the HWLabs Black Ice will be available through NGA Sistemas http://www.ngasis.com in Spain.

I hope this helps.
http://www.hwlabs.com/products

[Darkhorse]

GuyBFF,

Usually you will find water temps 5 over ambiant and CPU at 5 over water.

Does this rule work for any speed Tbird?

[GuyBFF]

No, obiously not, I'm giving an example, but from owning the following:

duron 650 @ 800
duron 750 @ 1000
tbird 1.333 @ 1550
tbird 1.4 @ 1700

And having tried an AC Heater Core in the DD Cube Configuration and BE Aquacoil Configuration, and now a Big Momma style heater core, I've seen that differences in the CPU don't affect the water temperature in the ways that they do in an air cooled system. The example should be close for his configuration.

It seems as though the pump and CPU both have a fraction in heating the system, and differences in the CPU's are more subtle than in an air cooled situation. V'modded my CPU's used to cook once the voltage was raised with a WBK38, now it's again only a couple of degrees. Seems like the hotter you go the better the system cools (sorta makes sense).

Right now I'm at 1702mhz 148FSB, 2.0v, 37c with a thermister on the side of the slug, and 25c ambiant. I have one 107cfm fan with a paper shroud on a Big Momma style heater core, 1/2" hose, 2 ounces of water wetter, a danner 350gph mag-drive (I think a 250 would be better) and a Danger Den Maze 2 1/2" "upgrade"

[BladeRunner]

GuyBFF What's your input coolant temp to max temp difference?

My temp difference is far more than 5°C.

I have coolant temp of 14°C atm very stable as it is cooled by the underground temps. I'm using an overclockers hideout copper waterblock which I modified so the base area is less than 1mm thick, (this actually improved max load temps by 6°C), as it was 4.6mm thick. It is a 1 gig AxiaR @ 1596mhz 1.95v 145fsb. It idles always about 2°C above water temp, (16°C), and max load it gets to 35°C. so that's 21°C difference between coolant temp and max CPU load. I'm using Artic Silver II between the CPU and block and the sticker on the underside of the CPU is removed, the thermal diode bent so it contacts the centre of the ceramic also with a dob of Artic silver II on the tip.

I thought there was something wrong originally, but a friend also running similar low coolant temps with a 1.4 @ 1.6 gets a similar 22°C difference. Flow rate is not a problem because slowing the flow doesn't make it any hotter.

[GuyBFF]

I have no way of measuring the temperature of the water entering the waterblock. According to a couple sites that rated the Big Momma's they never found more than a 4.5 - 5 degree temperature over ambiant on the water leaving the rad.

My slug temperature is measured by a thin style thermister that is right up against the corner of the slug, and a little arctic silver applied. I continuasly get temperatures of 37 to a max of 40 with ambiant of 24 up to 27c.

It only could be possible that my waterblock is holding the temperature within 5-8 degrees of the water.

One other thing to note is I still have the sticker under my 1.4g, and am running an ASUS A7V133 with that cheap little surface mount thermister (looks like a surface mount resistor, very small). With this thermister temperatures are around 42 to 45c. It appears ASUS built in at least a couple degrees of compensation into the BIOS as I checked the slug thermister against a digital thermometer, and it appears dead accurate.

As for flow rate, I don't believe slowing the flow would have any advantage, I just note that in my system the flow is VERY fast, and I think that a smaller pump would help lower temperatures (if only a degree).

[redleader]

I like the Big Momma, but it is so small

What do you consider to be big?


Re: Coolant temp

I got a coolant probe from Dtek when I got my BigMama clone. Its 10 Kohm so I will plug it into my 5 1/4 drive bay display eventually, but for now I have it on the 3rd probe socket of my A7V. According to it my coolant is nearly always at 33C while my motherboard is at 32C with ~70watts heat.

That doesn't mean my coolant is only 1C above abient. My MB temps are not room temp, but rather from some hot part of my board. According to my external probes my room is about 29.7C. I have yet to test the accuracy of the probes, but I doubt its off by more then a few degrees.

The probe is located directly before my radiator BTW, so it measures my coolant at its hottest point. However with a 318 GPH Eheim and at least 1/2 inch tubing the whole way, the coolant moves fast enough that I doubt a delta T of more then a degree exists anywhere in my coolant.

[GuyBFF]

I would hope that the radiator would shed more than a degree, even in a fast flowing system. Maybe consider moving the probe to after the radiator, doesn't matter what the temp is off the CPU, hotter better (more heat absorbed).

How much did your probe cost you?

The ambiant temps I listed are from a digital thermometer and are room temps, not case temps. My Motherboard chipset runs at around 30-35 degrees.

[DeadHeart]

I just bought a radiator. It has more or less the same size of the Big Momma, but it's all aluminium. I heard that aluminium dissipates better the heat.. It was very inexpensive, so, if I don't like it, I'll just buy another one. I bought it to a guy who told me that (not sure if it's true) it would keep the water 2ºc above romm temperature when cooling his Duron 600 @ 1000Mhz...

[DeadHeart]

Here's a pic of the radiator:

[GuyBFF]

Try scratching a bit of the paint off, I bet it's gold under it (copper). The aluminum version of heater cores usuall look a little different in the actual core area, but it's hard to be sure from a picture. If it's copper you won't need to use as much antifreeze as you have (and therefore cool better).

[redleader]

Is that antifreeze or just green tubes (the watter in the res is clear)?

Anyway hopefully thats just a painted copper radiator because Al is a poor choice in a watercooling setup due to corrosion and its lower performance compared to Cu. If it is Al make sure you use antifreeze.

[DeadHeart]

This is not my rig... I bought, I mean, I'm gonna buy this radiator on monday, and the guy who I'm gonna buy this from sent me this pic to show me the radiator.

[GuyBFF]

Redleader, in conductance of heat to air aluminum is actually superior. Copper's advantage is it spreads the heat better, aluminum releases heat better. This is why copper heatsinks sometimes have aluminum fins. The copper is much better at transfering the heat off the tiny core, and the aluminum can disperse the heat into the air better.

Why do copper heatsinks outperform aluminum then? It's because in relation to the very small core the copper transfers the heat out further than the aluminum, beating any losses it has. In a radiator the heat is transfered by the water relatively close to every fin, and a aluminum rad will release this heat into the air much better (this is proven in auto and motorcycle racing).

This aside I would still only use copper, as again, the waterblock is copper and I agree, I don't like galvanic corrosion. From the looks of his rad it doesn't have the look of an aluminum core, probably a silver painted copper core. But I'm proven wrong all the time

[imagex]

as far as i know, they don't use copper in car rads cuase of $$$

[redleader]

GuyBFF:

Thats an old myth and is incorrect. It stems from the fact that AL has a higher specific heat then Cu and was mistakenly interpreted to mean that Cu transfers heat to air slower.

Air is a fluid. Water is a fluid. No one questions that Cu conducts heat to water faster. How is air any different?

Ever been to Ars Technica?

[H]ardOCP article posted "Aluminum releases its heat faster" myth!!!

EDUCATED... aluminum vs copper, etc...

Al is used in sinks because it is cheaper, easier to work with and lighter.

[redleader]

Just noticed this one by TNT2bluz which I assume is in reference to the BlackIce

Aluminum vs Copper (Brass): The Radiator Debate

[GuyBFF]

Imagex, only in the last couple years have aluminum radiators become standard in the automobile market place. Advancements in plastic tanks and corrosion additives have made the fears of life expectancy dissapear. Almost all domestic vehicles of the eighties have copper cores. The unit shown looks like a very common 6x6 copper core used in Chevy and Ford products.

Redleader, aluminum rads have been proven to transfer 10-15 percent more heat than copper rads. Talk to anyone in motorcycle racing. In the last couple years Honda has quoted there aluminum design has increases cooling 15 percent in motocross and road racing bikes. I don't dispute your comments on the fact that air is fluid, water is fluid. In fact that's exactly the case, and aluminum does do a better job of transfering heat from water as well.

Where it fails in water blocks and heatsink fans is it won't move this heat very far. So a copper surface spreads the heat from a CPU threw more area, but in a rad the heat from the water is relatively close to every surface, and aluminum does a better job of heat transfer in shorter distances.

Again this is why the best heatsink fans (like the swiftech) are copper base, aluminum fins. Check there tests, they found aluminum / aluminum ok, copper / copper ok, but copper / aluminum much better. What more can I say.

Again I will say I will only use a copper rad, as really where talking about a degree max here. In the waterblock copper is a great advantage as it spreads the heat out over more area to be in contact with the water, and I really, really don't like corrosion, so copper and plastic are the only materials in my system.

[redleader]

Thats not true. How fast heat moves between any two materials is solely determined by the thermal conductivity of the materials and the area in contact. CU has nearly teice the thermal conductivity of Al so it is better.

Read those links I posted.

Aluminum does not 'give up its heat' better than copper! Let me repeat this once more; aluminum does NOT 'give up its heat' better than copper. It is true that, in general, aluminum will radiate heat better than copper but radiation is such a miniscule part of heat transfer in a computer system as to be deemed completely inapplicable.

Thus, a pure copper heatsink will always outperform a heatsink of the exact same geometry of a pure aluminum heatsink assuming that both have the same contact with the heat source and the same rate of airflow over the surface.

http://www.overclockers.com/articles223/index.asp

[GuyBFF]

I think you really need to re-read those articles, they basically define exactly what the point I am making.

Copper has a much higher level of conductivity, yet due to it's mass it's ability to radiate heat is lessened.

Aluminum has a lower conductivity, yet it will not absorb as much heat in it's lower mass, and will radiate this heat more readily from it's body.

In the end of the second article "educated" this is well defined and the person defending the use of copper realises this as well... maybe read all of the posts.

In the first article the posts still appear to be ongoing with each party throwing newer ideas, but neither being able to make a solid proof.

The important thing hear is that we agree the copper waterblock is superior, and unlike air cooling we are no longer concerned with removing heat from a small area like a die and being able to transfer it throught a heat sink body, the waterblock has already performed this transfer for us. We are just concerned with which material radiates heat better from the water, as this arguement is over radiators.

I know from real world comparisons in motorcycle radiators of same replacement size. One thing to not about a radiator is although they may be required to fit in the same space the construction may be different. Which they generally are, aluminum rads are usually single or dual core due to there strength, while copper rads are three or four cores to minimize the tubing size and chance of ballooning. Aluminum in this case transfers much more heat, partly due to the fact that it won't hold any heat within itself, and partly due to the fact that there is larger tubes with more surface area in an aluminum radiator.

Lastly explain why Swiftech MC462's cool better than any almost any other heatsink on the market, and in there own testing Swiftech found pure aluminum and pure copper inferior? Again copper transfers heat better, but tends to absorb to much in its mass, aluminum isn't as efficient at conducting the heat away from a small core, but excels at radiating the heat due to relatively low absorbing.

If I had to use one material I would agree copper is superior, but with a dual material system, and a medium like water, I would definately say copper for the waterblock, aluminum rad... or in a heatsink a copper base with aluminum pins as long as the conductivity between the copper and aluminum has been properly addressed. I would stress that I only use all copper to avoid galvanic corrosion (or the losses in the waters conductivity due to adding the required glycols).

[hielko]

Copper has a much higher level of conductivity, yet due to it's mass it's ability to radiate heat is lessened.

Aluminum has a lower conductivity, yet it will not absorb as much heat in it's lower mass, and will radiate this heat more readily from it's body.

I can't believe that. It don't make sense. Can someone here post the thermal resistance of copper - air and aluminium - air. I think it will prove that aluminium doesn't release heat faster, or something like that...

[DeadHeart]

I've also read that aluminium will dissipate heat better than copper and therefore, in teory, an aluminium radiator will do a better job than a copper one...

[FoolOnTheHill]

But, in theory, wouldn't a radiator with copper piping and aluminum fins be the best? Kind of like an alpha radiator? You know, use copper to take the heat out of the water, and use the aluminum to get it into the air faster?