Mythbusters test golf ball dimpling effect on car

[overst33r]

http://www.autoblog.com/2009/1...dimpling-mpg

Pretty cool

 

[Bignate603]

Somehow I don't think that huge golf ball dimples is going to be catching on any time soon.

 

[PlasmaBomb]

That shouldn't have worked...

watching video now...

 

[The Boston Dangler]

i'm skeptical. testing should have been done in a sterile environment - but hey they have a show to make. runs should have been performed in both directions and for a greater distance.

i thought the dimples needed a spinning ball to be effective.

 

[Rike]

Isn't shark skin rough for the same reason? Reduces drag making the shark faster & a more efficient swimmer?

Originally posted by: The Boston Dangler
i'm skeptical. testing should have been done in a sterile environment - but hey they have a show to make. runs should have been performed in both directions and for a greater distance.

i thought the dimples needed a spinning ball to be effective.

I don't think so, or at least there is still an effect if it isn't spinning. Check out the "large golf ball" wind tunnel test @ about 18:00.

 

[CurseTheSky]

I would have liked to see them do more than 5 runs each on the smooth clay and dimpled clay (10-20 would be good), and over a longer distance each time (even 5 miles would be better in my mind). Also, having at least two means of measuring the fuel consumption (both through the weight method, and through an electronic meter) would be ideal.

Finally, I wish they did a wind tunnel test on the smooth car and the dimpled car to see if there is any noticeable change. 11% isn't huge, but it's significant enough to start raising some eyebrows if others can produce similar results. The statement that we may see something like this in NASCAR in the near future may not be that far off.

Edit: on an unrelated note, I thought the hangover tests were even worse. A better test would be to find two people with very similar builds (overall mass, fat %, etc., twins maybe?). In one night, give one strictly beer and one beer + hard alcohol, and on a different night switch who gets straight beer and who gets a mix. Then, as part of the "hangover test," also measure their blood-alcohol content. That way you should at least know if the type of alcohol makes a difference, or if each person simply fared better on the second night of drinking. Finally, they should have somewhat regulated their actions. For example, if they spent the first night of drinking running around and smacking into things, it could have a different result than if they spent the second night mostly sitting.

 

[Bignate603]

Originally posted by: PlasmaBomb
That shouldn't have worked...

watching video now...

What is your reasoning for why it shouldn't have worked?

 

[chin311]

i saw this the other day on Discovery, pretty crazy, wasn't expecting these results.

 

[PlasmaBomb]

Originally posted by: Bignate603

Originally posted by: PlasmaBomb
That shouldn't have worked...

watching video now...

What is your reasoning for why it shouldn't have worked?

Essentially the reason that golf balls are dimpled is that with spheres are not aerodynamic and the wake contribution to drag is dominant.

Cars are much more complex shapes and have additional factors which contribute to drag, such as the wheel arches, wheels, wing mirrors, undertray (or lack there of), and sharp transitions for stylistic purposes etc.

Cooling package (including radiator, intercooler, oil cooler, etc) ----- 33.4%

Exterior ----- 31.7%

Front wheels ----- 13.1%

Rear wheels ----- 6.9%

Floor ----- 6.9%

Rear Axle ----- 3.1%

Engine ----- 3.1%

Front Suspension ----- 1.4%

Exhaust ----- 0.7%

Source: Autospeed.com

Adding dimples to cars will increase the turbulent airflow, so there is only a benefit if the increase in turbulent airflow is outweighed by the decrease in wake.

It is also surprising as there was a "small difference" in the hydrodynamic test of the model. While they say that a small difference in the model should scale up to a large difference in the full sized car, they neglect to say the increase in size (say it was a 1/12 scale model) is vastly outweighed by the decrease in fluid density (air is approximately 800 times less dense than water (depending on the temperature of both obviously)).

 

[PlasmaBomb]

Originally posted by: CurseTheSky
on an unrelated note, I thought the hangover tests were even worse. A better test would be to find two people with very similar builds (overall mass, fat %, etc., twins maybe?). In one night, give one strictly beer and one beer + hard alcohol, and on a different night switch who gets straight beer and who gets a mix. Then, as part of the "hangover test," also measure their blood-alcohol content. That way you should at least know if the type of alcohol makes a difference, or if each person simply fared better on the second night of drinking. Finally, they should have somewhat regulated their actions. For example, if they spent the first night of drinking running around and smacking into things, it could have a different result than if they spent the second night mostly sitting.

The alcohol tests were fine, that is generally how you test the effect of a drug, as each person acts as their own control (a paired test). Using twins wouldn't work as while they start out with the same DNA, it can vary slightly, and they can have different expression of genes - this is an interesting area of genetics (epigenetics). Also environmental factors will have a large impact on your ability to process alcohol, so unless you could find two twins who had both been Teetotal for all their adult lives their alcohol processing won't be identical. If you can find them, good luck getting them to drink

Granted measuring their BAC would have been interesting...

They both seemed fairly active during both of the evenings - didn't the dude jump into bed off the set whilst drinking spirits? He also attacked the sofa in clips from both of the nights...

 

[microAmp]

Lexus dimpled car.
http://blogs.cars.com/.a/6a00d...01157120ed53970b-800wi

Wonder if Lexus ever tested the MPG, although the dimple size are different from Mythbusters.

 

[SonicIce]

Here is some reading

 

[Bignate603]

Originally posted by: PlasmaBomb
It is also surprising as there was a "small difference" in the hydrodynamic test of the model. While they say that a small difference in the model should scale up to a large difference in the full sized car, they neglect to say the increase in size (say it was a 1/12 scale model) is vastly outweighed by the decrease in fluid density (air is approximately 800 times less dense than water (depending on the temperature of both obviously)).

You can easily scale based off density and size based on the Reynolds number. Even with a different density and size they can adjust the speed to match the Reynolds number of the full size car in air. I get the feeling that the guys at the lab running the tests are fully aware of what you need to do to run a scaled test. This is something that you'll find in any basic book on fluid mechanics. Running tests in a different fluid and scaling them is incredibly common practice, it's not like they're running full scale tests of a 747.

Anyways, the scaled testing is irrelevant when you have the result of the full sized test. They tested the car with and without dimples. It got better mileage. Whether or not you think it should have worked they got their result and it was well outside the range of error you would expect due to small changes in temperature, wind, and the driver.

 

[Kadarin]

Didn't Smokey Yunick do something like this with good results, way back in the day?

 

[SSSnail]

But, did they fill their tires up to 100 PSI? That's the more important question.

 

[Mojoed]

Originally posted by: SSSnail
But, did they fill their tires up to 100 PSI? That's the more important question.

No, but they did grease the muffler bearings.

 

[KillerCharlie]

Originally posted by: Bignate603

You can easily scale based off density and size based on the Reynolds number. Even with a different density and size they can adjust the speed to match the Reynolds number of the full size car in air. I get the feeling that the guys at the lab running the tests are fully aware of what you need to do to run a scaled test. This is something that you'll find in any basic book on fluid mechanics. Running tests in a different fluid and scaling them is incredibly common practice, it's not like they're running full scale tests of a 747.

Anyways, the scaled testing is irrelevant when you have the result of the full sized test. They tested the car with and without dimples. It got better mileage. Whether or not you think it should have worked they got their result and it was well outside the range of error you would expect due to small changes in temperature, wind, and the driver.

As a practicing aerodynamicist, I can say that attempts to scale something from a low Reynolds number to a higher Reynolds number are usually just complete BS in the first place. You cannot easily scale from one Re to another, especially with highly viscous phenomena like this.

It's also obvious that the size of the dimples on the car are completely not right relative to the boundary layer size. They definitely should have seen an increase in drag - I'd bet my job on their results being total crap.

Dimples actually make the skin friction drag worse. What they do, however, is help prevent flow separation by keeping the boundary layer attached. While dimples themselves make the drag on a golf ball worse, you also have less separation on it, so overall a dimpled ball will have less drag.

I stopped watching mythbusters when it became apparent they had absolutely no concept of science or physics, and didn't even try to.

 

[KillerCharlie]

Originally posted by: Bignate603
I get the feeling that the guys at the lab running the tests are fully aware of what you need to do to run a scaled test.

That's a bad assumption. They're probably some random mechanics that run the wind tunnels, not engineers. I've been to NASA Ames before (wind tunnel testing) and that's how it works.

Even so, they probably knew the dimpled versions were better in the tunnel, but neglected to tell the mythbusters that they have to properly scale the dimple size when they go to full scale, which they obviously didn't.

 

[Bignate603]

Originally posted by: KillerCharlie

Originally posted by: Bignate603

You can easily scale based off density and size based on the Reynolds number. Even with a different density and size they can adjust the speed to match the Reynolds number of the full size car in air. I get the feeling that the guys at the lab running the tests are fully aware of what you need to do to run a scaled test. This is something that you'll find in any basic book on fluid mechanics. Running tests in a different fluid and scaling them is incredibly common practice, it's not like they're running full scale tests of a 747.

Anyways, the scaled testing is irrelevant when you have the result of the full sized test. They tested the car with and without dimples. It got better mileage. Whether or not you think it should have worked they got their result and it was well outside the range of error you would expect due to small changes in temperature, wind, and the driver.

As a practicing aerodynamicist, I can say that attempts to scale something from a low Reynolds number to a higher Reynolds number are usually just complete BS in the first place. You cannot easily scale from one Re to another, especially with highly viscous phenomena like this.

It's also obvious that the size of the dimples on the car are completely not right relative to the boundary layer size. They definitely should have seen an increase in drag - I'd bet my job on their results being total crap.

Dimples actually make drag worse. What they do, however, is help prevent flow separation by keeping the boundary layer attached. While dimples themselves make the drag on a golf ball worse, you also have less separation on it, so overall a dimpled ball will have less drag.

I stopped watching mythbusters when it became apparent they had absolutely no concept of science or physics, and didn't even try to.

My first sentence may not have been clear. Density, viscosity, and characteristic length are known for the full sized car and the scale model. You can adjust the speed of the fluid to create a similar Reynolds number for the model. If that wasn't possible every single aerodynamic test would have to be run full scale.

When I was watching the video I don't think they ever measured drag in the scaled testing, they didn't give any quantitative answer from that test. When they were talking about a 'small effect' they were referring to delayed boundary layer separation. They were running the scaled test in a water tank with dye streams and were watching how the streams acted as they came off the back of the car.

As for the full scale testing, what exactly did they do that would undo the fuel economy savings that they saw? Beyond some math error or issue with the fuel system that caused some unaccounted gas in their calculations I don't see how their number could be 11% off. No matter what you think SHOULD happen, you need to explain what DID happen.

Yes, the mythbusters aren't the most scientific, the set up stunts and do special effects. However, I don't see a way to explain away the difference in the full scale test besides a gross error. While they aren't scientists I think they can probably handle putting a fuel cell on a scale.

 

[Bignate603]

Originally posted by: KillerCharlie

Originally posted by: Bignate603
I get the feeling that the guys at the lab running the tests are fully aware of what you need to do to run a scaled test.

That's a bad assumption. They're probably some random mechanics that run the wind tunnels, not engineers. I've been to NASA Ames before (wind tunnel testing) and that's how it works.

Even so, they probably knew the dimpled versions were better in the tunnel, but neglected to tell the mythbusters that they have to properly scale the dimple size when they go to full scale, which they obviously didn't.

Ok, they didn't optimally scale the dimple size. They did what they did whether it was the best way to attack it or not. That still doesn't change the result on the full scale test, they got what they got with the dimple size they used. Either it used less fuel or they screwed up how they were accounting for the fuel.

 

[LordMorpheus]

Dimples help a golf ball fly by tripping the boundary layer into a higher energy, turbulent mode. This boundary layer doesn't separate from the trailing end of the golf ball as easily as a smooth laminar boundary layer would. To achieve the same effect on a car, you could just place small features along the back of the roof, back of the trunk lid, etc. Anywhere immediately forward of a back-facing surface, and you wouldn't have uglied the car up.

Actually, I'd imagine most automakers already have payed some attention to this. Even if it's just not making the back glass seal completely smooth, that'll trip the boundary layer without adding anything.

Of course, cars probably operate in Reynolds number ranges that are turbulent anyway and wouldn't need any help making the boundary layer turbulent.

 

[fleabag]

Originally posted by: LordMorpheus
Dimples help a golf ball fly by tripping the boundary layer into a higher energy, turbulent mode. This boundary layer doesn't separate from the trailing end of the golf ball as easily as a smooth laminar boundary layer would. To achieve the same effect on a car, you could just place small features along the back of the roof, back of the trunk lid, etc. Anywhere immediately forward of a back-facing surface, and you wouldn't have uglied the car up.

Actually, I'd imagine most automakers already have payed some attention to this. Even if it's just not making the back glass seal completely smooth, that'll trip the boundary layer without adding anything.

Of course, cars probably operate in Reynolds number ranges that are turbulent anyway and wouldn't need any help making the boundary layer turbulent.

I was thinking this too, though not to this extent. Though wouldn't there be a benefit to covering the entire roof of the car so that if there winds hitting the side of your car, the car would be less likely to tip over?

 

[KillerCharlie]

Originally posted by: LordMorpheus

Of course, cars probably operate in Reynolds number ranges that are turbulent anyway and wouldn't need any help making the boundary layer turbulent.

Exactly - you can achieve laminar flow, but it's a lot harder at higher Reynolds numbers. As far as dimples on golf balls go, they do trip the flow, but I think the main effect is not that they just trip it, but they go a bit beyond that and keep the boundary attached longer than if it was just simply tripped. They are paying a drag penalty just for having the dimples there though (which is small compared to the drag savings from less separation). Any kind of surface roughness almost always makes your drag worse.

 

[GoatMonkey]

Interesting, but the segment about beer before liquor has a more immediate affect. Time for some shots!

 

[PlasmaBomb]

Originally posted by: Bignate603
As for the full scale testing, what exactly did they do that would undo the fuel economy savings that they saw? Beyond some math error or issue with the fuel system that caused some unaccounted gas in their calculations I don't see how their number could be 11% off. No matter what you think SHOULD happen, you need to explain what DID happen.

Yes, the mythbusters aren't the most scientific, the set up stunts and do special effects. However, I don't see a way to explain away the difference in the full scale test besides a gross error. While they aren't scientists I think they can probably handle putting a fuel cell on a scale.

Assuming that the 11% increase in fuel economy was brought about by an 11% decrease in drag... that means the body drag would have been reduced by 33%.