Back on the original topic, but with a slant, from original heatsinks that present a weight and bulk penalty, the addition of a fan to force the cooling action of the heatsink, but with the first generation (powerpc) G5 MACs they actually went for liquid cooled chips - immediate block attached to the chip, but with a water jacket then pumped off to a radiator elsewhere so they could run faster without the impending bonfire scenario. All was fine until a year or so when the system got a bit leaky, then game over when the cooling system had boiled dry - I've got one such machine at work in this state of disrepair. The surface radiation mentioned by xpensive could be used in some part as there are still pipes connecting the radiators - if these are developed to present a bigger surface area then some slight additional cooling could occur to/from the rads, at least to help a bit, not likely to be a massive improvement.
IMO.. I haven't heard any teams saying that the present radiators are not cooling enough. So the aim of a new design would be: to cool at least as good as the present radiator and have another advantage.
Alexbarwell, You mentioned water cooled computers..those would be like the present radiator systems.
We know the regular Computers are air cooled heatsink cooled like a VW bug.
So what are we saying? I am a little confused to the topic.. Air cooled F1 engine? Or Water cooled with a different shape and position to the heat exchanger..
It would be nice for some one to make a drawing i solid works.. of their new idea.
Whether he meant to or not, wesley123's original post actually makes a very interesting point. F1 engine cooling circuits use water coolant because of rules, and aluminum, tube/fin, liquid-to-air heat exchanger cores because they're efficient and lightweight. Some parts of the engine (like pistons) are oil cooled, and parts of the cylinder head are cooled by the latent heat effects of the intake airflow and fuel vaporization.
But wesley123 brought up the subject of copper heat exchangers. Amazingly enough, even though a copper heat exchanger matrix would be about 3 times as heavy as a similar sized aluminum core, it would also be about 3 times as thermally efficient. So in theory, it could be about 1/3 the surface area to achieve the same heat rejection rate.
Since F1 chassis designers are very concerned about drag penalties due to things like heat exchangers, I wonder if any of these clever guys have ever done a trade study to ascertain whether a smaller cross-section (but slightly heavier) copper heat exchanger core is more aerodynamically advantageous than the conventional aluminum unit.
What do you guys think, is copper worth a look?
"Q: How do you make a small fortune in racing?
A: Start with a large one!"
are you proposing a copper radiator over an aluminum one?
if so... what about 2 small copper ones low down in the chassis... I remmeber when the cars used to collect cooling air from low down in the chassis whereas the sidepod openings are pretty high now, always wondered why they made the switch. Best answer I came up with was to undercut the sidepods and get more air to the rear bridge wing.
Radiators or heatsinks of copper or whatever, at the end of the day you need to find the easiest, lightest and most efficient way of SOMEWHERE conveying thermal energy from a metal surface to the passing air-stream. No way around that one.
The parameters involved in said conveying, or convection, are basically: surface-area, temperature diff and air-speed.
Surface-area being the dominant pain-in-the-arse here, requiring those drag-inducing radiators.
I believed that gold was one of the best heat exchangers, that could be a possibility.
I nearly dont know how any of the cooling works so my first idea was a quick idea.
I think that copper or gold arent allowed by the rules as they cout under the expensive exotic materials, i aint sure and i blieved it was only for the negine, so actually think it would be possible to make the radiator etc. out of gold, smaller radiators means smaller sidepods thus less drag and more downforce, so it is at least worth a shot and it looks good xD
There is a limit to size and weight and surface area though. Remember that even though one metal maybe X times more conductive, the manufacture of that radiator may not be X times smaller.
But i like the direction where this is heading.
Here are two pages with the thermal conductivities and density of some metals.
Copper 385 W/mK @ 8933 kg/m3 is better than pure gold 318 W/mK @ 18000 kg/m3.
I used the data from the web page and made a table. I divided the conductivity by the density to show which metal will give the lowest weight for the same heat transfer. But the only way to really see which will give an advatage in a race car is to actually draw the radiator and simulate. Because looking on the present F1 car the Radiator inlet holes are in an elevated position.
That chart illustrates the situation very nicely! The aluminum used for radiator cores is usually 5052, since it can be welded and has decent strength after welding without a post-weld heat treatment. A brazed copper core would have about 2.8 times the thermal conductivity of a 5052 aluminum core, but it would also be about 3 times as heavy. The real benefit of the copper core would be that it would have a much smaller cross section than the aluminum core for a given heat rejection rate. So it would have a lower drag penalty.
So my question remains: Is the weight penalty of the copper core justified by its reduced aero drag penalty?
Regards,
Terry
"Q: How do you make a small fortune in racing?
A: Start with a large one!"
As you point out, there are also other practical issues to be considered. For example, a straight material switch from aluminum to copper would be a 3:1 weight difference. But would the material thicknesses used for an aluminum core (typically a tube/fin type core construction)be the same as those required for a copper core? Copper is more ductile than aluminum, so maybe the tube walls can be drawn thinner? But at the same time, 5052-H32 aluminum has better mechanical properties than annealed copper, so maybe a thinner aluminum tube wall is possible?
There is also the question of the trade-offs of weight versus drag. I'm no aero expert, but I've been told that a well designed heat exchanger core and ducting actually has very little (if any) drag penalty. Apparently, this is mostly due to the fact that the airflow passing thru the heat exchanger core picks up a substantial amount of energy which is put to good effect as it exits. But most F1 cars are already under weight, and must use ballast. So a few extra pounds in the radiator would seem to be a good trade-off if it reduced drag. The car would still weigh the same, but would have less drag. And drag losses increase exponentially with speed.
Current F1 reg's only allow water to be used as a coolant. And the coolant circuit pressure is limited. But as far as I know, there is no regulation prohibiting copper heat exchangers.
Lots of things to consider here. But it makes for a great discussion topic!
Regards,
Terry
"Q: How do you make a small fortune in racing?
A: Start with a large one!"
To the best of my experience, you are perhaps missing the point here, thermal conductivity of the material itself is very often not the most critical parameter, when getting enough surface-area is.
This is one of the reasons why we today use thin-walled-plate heat exchangers from SS316(with a terrible 13.5 W/mK) for different cooling purposes rather than the old tubular copper ones. http://www.alfalaval.com/solution-finde ... OC-BHE.pdf
On an F1 car however, which is far more optimized, durability probably matters very little, while weight does, why aluminium seems to be the obvious choice.
However, there's only so much energy you can transfer to the volumetric flow of air, can't get it warmer than the cooling-water itself, why this is perhaps another relevant angle of discussion?
"I spent most of my money on wine and women...I wasted the rest"
Yes, aluminum is 1/3 the weight of copper and copper is 3 times more efficient in conducting heat compared to aluminum. This is all very well if we were only talking about material properties. What we really need to consider is the properties of radiators constructed from these materials installed in the car and ready to race.
When we add coolant to similar sized radiators of both materials and compare weighs, the copper version now weighs much less than 3 times that of the similarly filled aluminum unit but still retains the tripled heat transfer efficiency.
There is something else to consider in regard to aluminum verses copper as a radiator material. With two radiators of similar size, the copper version would only require 1/3 the amount of coolant to extract the same amount of heat.
Let’s say our aluminum radiator cooling system caries 4 gallons of coolant weighing 32 lbs and the radiator weighs 5 lbs for a total of 37 lbs. Our copper version is only required to carry 1.3 gallons weighing 10.6 lbs with a radiator weight of 15 lbs for a total of 25.6 lbs, a savings of 6.3 lbs. Further weight reduction could be had through reduced plumbing diameters. This scenario applies to weight as a primary concern. If aero is the primary concern, restore the coolant capacity to 4 gallons and reduce the radiator size by 2/3rds. Of course any combination of the above could be done.
When it all has to end with convection to the passing air, one way or the other, the following numbers might be of use.
With an engine output of 500 kW and 25% efficiency, total power would be 2000 kW, meaning 1500 kW is wasted heat.
How much dissipated through the cooling circuit? Let's say 300 kW (300 kJ/s), for the sake of argument.
At 50 m/s (180 km/h), with 5 kg/s of 25C air passing through at a density of 1.2 kg/m^3 and thermal capacity of 1.0 kJ/kg*K, to exit the same air at 85C, you would need a radiator inlet area of some 0.08 m^2, or say two ducts of each 20x20 cm.
Aerodynamic powerloss, P=Cv*A*Rho*v^3/2, depends on the resistance of duct and radiator, but with a Cv of 0.6, you will lose about 15 kW (20 Hp) through those ducts at 300 km/h.
Last edited by xpensive on 28 Apr 2009, 06:53, edited 3 times in total.
"I spent most of my money on wine and women...I wasted the rest"
n smikle, the observation of water cooled computers was more about the original notion of heat-sinking, that water cooling is a more effective means of removing the heat from the component than the metal body heatsink, so teams are not likely to abandon the more effective system used for years. Essentially that the instigator of this discussion was not seeing the situation quite clearly. Solid (albeit finned) metal heatsinks ultimately rely on the temperature gradient to draw the heat away, as would just a liquid body applied to the same device. If however you constantly apply cold liquid to the component, then remove it when warm (pumping to/from a remote radiator) the effect is hot body to cold liquid, namely a steeper temperature gradient with a bigger transference of heat energy. To sum up, circulated cooling system 'pulls' the heat from the component. As this heat is also actively pumped into the radiator another greater temperature gradient is developed and again more heat is transferred from the radiator/coolant to the atmosphere. Remember how bulky air-cooled engines can be? Material thermal conductance is a factor, within practical limits; turbo engines at some point running sodium-cored valves...
