Air Cooling & Head Design

[randywebb]

splitting this off from an OT digression on another thread...

It began with me quoting someone (Bruce Anderson maybe):

Oil "doesn't boil or freeze"
- nor does it cool very well...

Once you get a certain engine size/specific output you are required to run oil lines up to the front of the car to a heat exchanger. So what's the point -- you may as well use a water cooled motor and run water lines up to a front heat exchanger (oddly enough, called a "radiator" altho radiation has little to do with it).

Water cooling also allows the use of 4 valve heads (and eliminates most of the need for twin-plugging). That in turn allows better valve train geometry, greater flow and mor power.

Water cooling is just better. You can use air cooling for small, wimpy motors however.
_______________________________

The resonponse re aircraft motors and Sherman tanks seem offpoint since the design goals, gearing and air flow regimes are quite different from those for a sports car. But...

by your standards a 450hp 993 GT2 twinturbo qualifies as a "small, wimpy motor"? Just like an 1100HP 917 Can-Am mill? Being a wimp never sounded that good before....

[Chuck Moreland]

Originally posted by randywebb
Water "doesn't boil or freeze"
- nor does it cool very well...

Dontcha mean "Oil doesn't boil or freeze" - nor cool very well ?

[randywebb]

oooops!! I'll edit that

[Drago]

Originally posted by randywebb
Water cooling is just better. You can use air cooling for small, wimpy motors however.

I like my small, wimpy motor just fine thank you!

[randywebb]

ficke posted a good visualization of the improved flow from 4 valve heads:

draw a circle then draw two of the largest circles you can fit in the first circle, that is a two valve head, now draw the same size circle's as the first and draw four of the largest circles you can fit into it, that is a four valve head, if you now measure the volume of the four circles and compare them to the two circle in the first drawing you will find a large volume gain in the four circle/ valve arrangement.

[slight edits below]
On an "heat source to air cooling" (as opposed to " heat source to water to air cooling") like in a Porsche 911, Harly-Davidson and Pratt and Whitney radial engine all have two valve heads for proper valve cooling, and can not take advantage of the four valve arrangement. All engines are a compromise of some sort.

[Jim Sims]

It is hard to improve on water as an ideal heat transfer fluid - something to do with the hydrogen bonds and soaking up thermal energy. If you need to get a lot of heat out of a small area or need more uniform cooling its use is almost mandated.

Oil is a poor heat transfer fluid due to poor thermal conductivity, higher viscosity and lower specific heat.

Oil can be readily frozen and boiled, just not under conditions typically encountered in normal automotive use.

Arguably, the ultimate in reciprocating piston engines (such as the BMW 801 or Curtiss-Wright R3350) were air and oil cooled.

Water is very corrosive and it freezes; ethylene glycol added as antifreeze can burn. Water can leak internally and get in the oil and ruin bearings.

The move to water cooling appears to be driven by power density and emissions requirements. Enhanced oil cooling is a clumsy solution but in some ways so is having the engine hanging out the rear. However these features are part of the "charm" of 911's

[randywebb]

specific heat = shorthand for specific heat capacity; a material property related to how much heat it takes to raise the temperature of 1 gram of the material - i.e. an intensive property

thermal inertia = an extensive property of a thing - i.e. how much heat it takes to raise the temperature of the _object_ under discussion; others have other terms for this, such as "thermal mass"

viscosity = how 'fluid' a fluid is - there are more technical definitions but I will not shear them with you here; honey is more viscous than water...

thermal conductivity = the ability of an object to transfer heat by conduction (as opposed to radiation or convection etc.) - Al has high thermal conductivity and therefore when you touch a hot Al shift knob it hurts, whereas a wood shift knob at the same temperature would be less likely to burn you

charm = a quantum chromodynamic property of quarks on uppers

[ianc]

So, like, whaddya trying to say?

Do our 911's suck now?

ianc

[randywebb]

they suck in air...

[A Quiet Boom]

Randy,

Which is why most guys here wouldn't pass up a chance to swap in a GT3 motor given half a chance, even if they don't want to admit it.

Believe me if I had the machinist talent and a larger CNC I'd seriously consider a watercooled cylinder and head design. The old rule of thumb for V8's was that switching to better cooling alumimun heads allowed you to run 1 compression point higher on the same gas. We proved this on a 5.0 going to 10.5:1 with no detonation.

And yes objectively speaking 911 motors are wimpy, good power for there size but that's just it, the largest motor possible is about 4.0 liters and that's pushing it, imagine 5 or 6 liters with watercooled cylinders and heads. Imagine the immediate boost in power just from removing the parasitic loss of the engine fan, heck even a badly designed water pump only uses about 3-5hp. I read somewhere that the 911 fan can pull as much as 17!

Bigger displacement doesn't have to work as hard to make big power. I've seen two valve pushrod V8's make 1.5hp per liter without nitrous or supercharging and for far less money than a 911 engine. I've got a friend with an honest 500hp from a 302.

Having said that I still love my 911 and it's wimpy engine.

[Hansv]

Hi there, Porschefans;

Nice to see how a 1950's quote by Der Doctor can still start a debate fifty years later.

[Hansv]

Hi there, Porschefans;

About watercooling and four valve heads: I always understood the main reason for introducing them on Porsche streetengines was because without them Porsches would no longer pass environmental laws. If power was the only objective we may have 'had to live' with aircooling and two valve heads for much longer. The horror!

[beepbeep]

Sorry to burst the bubble but two valve aircooled heads aren't good for either power or emissions. Porsche knew this all the way and finally had to make the switch to 4-valve pentroof and watercooling. They did it long before on their race engines, 956's and 962' had all watercooled heads long before consumer cars got it.

I do like aircooled cars as much as I like tube amplifiers. But in the end, they are inferior as much as they are liked by people like you and me. It was focused hard-core gradual tehnical refinement and overdimensioning that kept them in life for such a long time. As consumers demanded more and more power and emission laws tightened PAG was being painted in the corner.

They don't earn money on purist hobbyists, they earn money on yuppies that want faster car than neighbour has. We might never know how internal decision process went when they switched to waterocooling but (being an economist myself) I guess they did a market forecast and decided it was worth to invest in new tooling and delevop new motor family instead of gradually loosing customers to other brands.

Yes, there are aircooled and 2-valve engines that make insane amounts of power but it's cheaper and easier to do it with water-cooling and 4-valve job.

P.S.
Now please don't come with "top-fuel 2-valve pushrod jobs make so much power" thing once again. That is prime example of 50's technology frozen in time being refined in absurdum.

[ANTONIO]

Yeah, everything stated here sounds logical, but do they have a soul??Ant.

[jdm61]

Good comment on soul and better comment on tub amps. I have a solid state power amp for my stereo, but a tube pre-amp and phono stage. ALL of my guitar amps are tube amps and no dsign newer than around 1963. "Soul" is very important. The air cooled argument may be the only strike left against 911's since the RS Spyder uses the "old fashioned" torsion bar setup.

[jluetjen]

Originally posted by randywebb
ficke posted a good visualization of the improved flow from 4 valve heads:

draw a circle then draw two of the largest circles you can fit in the first circle, that is a two valve head, now draw the same size circle's as the first and draw four of the largest circles you can fit into it, that is a four valve head, if you now measure the volume of the four circles and compare them to the two circle in the first drawing you will find a large volume gain in the four circle/ valve arrangement.

[slight edits below]
On an "heat source to air cooling" (as opposed to " heat source to water to air cooling") like in a Porsche 911, Harly-Davidson and Pratt and Whitney radial engine all have two valve heads for proper valve cooling, and can not take advantage of the four valve arrangement. All engines are a compromise of some sort.

Randy;
I largely agree with you except this part...

if you now measure the volume of the four circles and compare them to the two circle in the first drawing you will find a large volume gain in the four circle/ valve arrangement.

The flow potential of a poppet valve isn't related to the volume (do you mean area? Circles don't have volume, but they do have area) of the valve's head. It is related to the perimeter distance multiplied by the lift. This exposed area is what largely determine the flow potential of a given valve size. That being said, once you lift a valve past about 1/4 of it's diameter, you'll find that this lift area is greater then the surface area of the valve head (aka: the area enclosed by the valve seat) and so additional lift will rarely generate additional flow in and of itself.

Technical details aside, the whole premise of this tread is kind of silly since there is no "better" or "best" design for engines. There are better and worse compromises for what you are trying to do.

1) Aircooled engines can be an excellent compromise if you are looking for 1) a lightweight power unit and 2) a compact installed powertrain design, with an emphisis on light weight. Porsche's philosophy for it's cars through the 60's at least was to develop a compact car with excellent performance. Putting the engine over the driven rear wheels provided excellent traction, and allowed the driveline to be very short and light, and still provide generous space in the passenger compartment. They were also able to dispense with all of the weight and plumping required for water cooling, not to mention the packaging issues related to large radiators with ample air flow. In spite of the limitations of air-cooling, Porsche's team were still able to generate more then ample HP for the car to be at the head of the performance car field. It is these same reasons why air-cooled engines were favored for piston powered airplane engines, especially for war-planes since they didn't have as many or as large heat exchangers to get shot full of holes.

2) Watercooled engines provide greater power density compared to aircooled engines. So liter per liter, water cooled engines can make more power. This is why water cooled engines have been the standard in F1 since the mid-60's. Porsche's air-cooled flat-8 F1 engine of the early 60's was able to just keep up by virtue of it's lower CG, and competitive installed weight. But with the advent of the 3 liter era, there was no chance. But Porsche's air-cooled engines were still able to compete effectively in endurance racing with their 908's and 917's because the engines made enough HP, but had few parts to break down, wear out or get damaged during an event. Remember, excepting the front oil heat exchanger, you can drive a 911 into a wall, and still have the motor run great. Any car with a front radiator will never withstand such a test. Note that one of the weaknesses of the water cooled 911 race cars are the water radiators in front of the front tires. One tap there and the car is done. This would never have stopped an aircooled 911. And in an endurance race, the old addage "In order to finish first, first you need to finish" could never be more true.

So each strategy has benefits. You need to pick which particular benefits you are going to chose to leverage.

[jluetjen]

Originally posted by jdm61
...since the RS Spyder uses the "old fashioned" torsion bar setup.

As do most of the current F1 cars -- for the same reason that Dr Porsche used them in many designs. Because they provide packaging, overall and unsprung weight benefits.

People have got to get over these "conventional wisdom" sterotypes and start thinking out of the box! (forgive the cliche!)

[A Quiet Boom]

beepbeep,

top fuel engines are held to 2 valves, 500CI, small helix in the Roots style blower, no EFI, etc. as a result of rules. And the reason for those rules is insurance liability as well as track length. One of our local tracks hasn't had a pro race in over 20 years because they don't have a long enough run off for even 250MPH cars. Top fuel cars are wrench cooled BTW, you take them apart to cool them, the blocks are solid. Can you image the power they would make if allowed 700-800CI like the pro-mod motors, modern turbos and EFI, as well as 4 or even 5 valves?

Even at the top of the american V8 resurgence in the 90's Ford realized that pushrod V8's where outdated and designed the modular engine series. All the Ford purists complained about the loss of the 5.0 and other classic Ford engines, they said the engines would never be as fast as the old ones. Well about a year into production Sean Hyland broke into the 9's with new racecar running a 4.6L mod motor. If I were to build another Ford dragcar I'd use a 4 valve Cobra SVT motor in an early chassis.

Matter of fact my dream engine for the '73 car is a GT3 engine, largely because it uses the 993 turbo case and has true dry sump. This engine has already been installed into early cars.

[safe]

Originally posted by randywebb
ficke posted a good visualization of the improved flow from 4 valve heads:

draw a circle then draw two of the largest circles you can fit in the first circle, that is a two valve head, now draw the same size circle's as the first and draw four of the largest circles you can fit into it, that is a four valve head, if you now measure the volume of the four circles and compare them to the two circle in the first drawing you will find a large volume gain in the four circle/ valve arrangement.

Maybe a stupid question, but why are valves round?
Couldn't you make a bigger eliptic valve with a guide to keep it centered?

[beepbeep]

Originally posted by A Quiet Boom
beepbeep,

top fuel engines are held to 2 valves, 500CI, small helix in the Roots style blower, no EFI, etc. as a result of rules. And the reason for those rules is insurance liability as well as track length. One of our local tracks hasn't had a pro race in over 20 years because they don't have a long enough run off for even 250MPH cars. Top fuel cars are wrench cooled BTW, you take them apart to cool them, the blocks are solid. Can you image the power they would make if allowed 700-800CI like the pro-mod motors, modern turbos and EFI, as well as 4 or even 5 valves?

Even at the top of the american V8 resurgence in the 90's Ford realized that pushrod V8's where outdated and designed the modular engine series. All the Ford purists complained about the loss of the 5.0 and other classic Ford engines, they said the engines would never be as fast as the old ones. Well about a year into production Sean Hyland broke into the 9's with new racecar running a 4.6L mod motor. If I were to build another Ford dragcar I'd use a 4 valve Cobra SVT motor in an early chassis.

Matter of fact my dream engine for the '73 car is a GT3 engine, largely because it uses the 993 turbo case and has true dry sump. This engine has already been installed into early cars.

Great info! I don't know much about Top-fuel drag-racing but I know that engines are stone-age.

I believe regulating displacement instead of regulating technology would do sport a favor.

That being said, there is a Norwegian guy that run 8'83 with BMW E34 steel chassis (glass-windows etc.) and turbocharged 3.8L BMW inline-six. V8 geezers were in awe.

Movie:
http://www.vsmotor.no/wmv/vsmotor-m5-tullinge-883.wmv