Mythbusters tests "golf ball" dimples

[niky]

I was impressed by it at first, but it really was a very questionable result.

Mythbusters is fine entertainment, but that's what it is first and foremost, entertainment.

The Golf Ball test was done in uncontrolled conditions, with merely A and B testing. Not A-B-A. Meaning to say, while they did all they could to ensure the runs were conducted in as similar a manner as possible, they did not control for weather... wind speed, temperature, humidity... or car condition (even if the car starts "warm" or "cold" both times, the fact that it already ran earlier in the day is already a big factor...)... which they could have done simply by performing an A-B-A test to ensure the results for the second set of tests were truly representative.

This is because they overcomplicated the making of the dimples. The extra layer of clay exaggerates the effects of shut-lines and gaps. By the end of the layering, it wasn't a very aerodynamic car. And the clay made it impossible to do both tests within reasonable time of each other.

I would have gone with making dimpled body panels. Swappable dimpled body panels, with two cars. Say Fieros. Run both side by side, one dimpled, the other not... swap panels... repeat. Swap panels again, repeat again. Or produce fiberglass panels that fit snugly over the stock, dimpled panels on the Taurus.

A-B tests prove nothing. I've seen A-B tests "prove" everything from HHO injection to air bleed devices to fuel magnets. Easiest way to tell if the "modification" is real or bogus? take it off and test again.

[redyaris]

Originally Posted by ACEV

@aerohead #46,
Could you please put that into simple language so we all can understand? Your good points are very difficult to visualize.

At this point we do know that dimples in a car body surface do increase fuel economy since we have all observed it.:

Not true. The expirament did not have adequit controls to isolate what caused the change in fuel consumption.

QUOTE:
If we feel that there is something missing in the given test, then it should be pointed out. If we feel that there are other things in the test making the fuel economy better, then they should be pointed out.:[/QUOTE]

They have been... see previouse posts...

QUOTE:
Short of that, the average person should be able to rely upon the fact of the fuel mileage improvement as demonstrated. If we overthink everything, we will never move forward.[/QUOTE]

What amount of thinking is overthinking?

[MetroMPG]

Originally Posted by ACEV

If we feel that there is something missing in the given test, then it should be pointed out. If we feel that there are other things in the test making the fuel economy better, then they should be pointed out.

Several shortcomings of the Mythbuster test have already been pointed out!

A poorly designed/executed test proves nothing.

EG: How do we know the changes in fuel consumption aren't from different speeds during the runs (e.g. no cruise control used)?

EG: How do we know the results seen aren't due to a drivetrain that was gradually warming up and getting more efficient? (This is normal and well understood.)

EG: How do we know the results seen aren't due to the transmission torque converter not being locked during the "before" run because perhaps the drivetrain wasn't up to full operating temperature?

EG: How do we know the effects seen aren't from changing ambient conditions during the day (e.g. hotter out = more efficient, again normal and well understood).

How do we know the effects seen aren't from some combination of any of these things?

The only correct answer is: we do not know.

So it is a fundamental mistake and a failure to understand the scientific method (a controlled test was not performed) to say that what you saw on TV proves that dimples in clay increase fuel economy.

At best, we can say: Hmm! Isn't that interesting! Wouldn't it be nice if they could perform a well-controlled test (see ideas presented above in previous posts) so that we could have confidence in the results?

[Patrick]

ACEV, here is what you do:

1. Keep records for your car for 5000 miles.
2. Take a ballpeen hammer to the surface of your car.
3. Keep records for 5000 miles.
4. Report your results here.

[some_other_dave]

Originally Posted by ACEV

One must assume that the ideas you put forward are having to do within atmospheric situations.

Typically the field of aerodynamics is concerned with things that happen within an atmosphere, not in a vacuum. So that would be a valid assumption...

-soD

[niky]

Dear ACEV: if all it takes is another test to disprove Mythbusters... no matter how poorly either test is constructed, here you go:

Fastskinz Test Drive: Can a Golf Ball Covering Improve MPGs? - Popular Mechanics

There... despite flaws which are just as bad as the Mythbusters test, this "proves" dimples don't work!

Confounding, ain't it?

Originally Posted by MetroMPG

EG: How do we know the results seen aren't due to a drivetrain that was gradually warming up and getting more efficient? (This is normal and well understood.)

EG: How do we know the results seen aren't due to the transmission torque converter not being locked during the "before" run because perhaps the drivetrain wasn't up to full operating temperature?

This is a pet peeve of mine. When we do acceleration tests for cars for our annual COTY tests, we pay attention to the trending of the times. If it gets slower and slower as the engine and tranny warm up, abort... she's done.

But sometimes... just sometimes, with the new cars we test, a car is actually still breaking in as we do the test... each run is faster and faster... Just a few blasts at full throttle loosen everything up and let the car breathe easier.

Just the fact that a car cools down and heats up between tests makes simple A-B tests very, very variable.

[gone7]

@niky,
Thanks for the good link. However, since the results are within any reasonable fudge factor, I would say that the differences were negligable.

However, that does raise questions about the dimples used. As a thin film, the dimples would have been quite shallow, and quite likely very equally spaced. Not so with the MythBusters application; leaving all sorts of variables. However, since there results were highly significant, their test is the one that needs further study, rather than dismissal.

Further, A-B-A testing is questionable because of such little difference between that and A-B. Why not A-B-A-B..... to infinity? Would that help? At what point do we say that the improvement is real? Or even good enough?

@MetroMPG,
Please note that the test results were very significant. They were not minimal. Therefore, the small details you mentioned would not likely make any significant difference. Actually, what we can really say is that they got real results that were significant.

9% is nothing to disregard.

What still takes me by surprise is that no one seems to notice that the air on the surface of a moving object is what creates drag. Therefore it is reasonable that breaking that up would decrease the drag of attached air. Let's discuss the basics first.

[Patrick]

Originally Posted by ACEV

What still takes me by surprise is that no one seems to notice that the air on the surface of a moving object is what creates drag. Therefore it is reasonable that breaking that up would decrease the drag of attached air. Let's discuss the basics first.

From: Turbulence - Wikipedia, the free encyclopedia

Flow over a golf ball. (This can be best understood by considering the golf ball to be stationary, with air flowing over it.) If the golf ball were smooth, the boundary layer flow over the front of the sphere would be laminar at typical conditions. However, the boundary layer would separate early, as the pressure gradient switched from favorable (pressure decreasing in the flow direction) to unfavorable (pressure increasing in the flow direction), creating a large region of low pressure behind the ball that creates high form drag. To prevent this from happening, the surface is dimpled to perturb the boundary layer and promote transition to turbulence. This results in higher skin friction, but moves the point of boundary layer separation further along, resulting in lower form drag and lower overall drag.

" . . . flows at Reynolds numbers larger than 5000 are typically (but not necessarily) turbulent."

Now reread post 46, paragraph 5, above, by Aerohead:

"5. since Reynolds number is a function of a bodies length,as compared to it's velocity in a specific fluid,when calculated for automobiles it is found that all production motor vehicles will achieve critical Reynolds number near 20 mph,and from there on,their drag coefficient will be constant up to transonic flow velocities,where compression effects begin to enter the picture."

Get it now?

[Tamn]

While the basic idea is right, they added a lot of surface area by adding (2+?) inches of clay onto the entire car body. But nonetheless, cool stuff.

My father tried convincing his race team to dimple parts on their porche race car in the early 70's. They didn't buy it then,

[redyaris]

Originally Posted by ACEV

@niky,
Thanks for the good link. However, since the results are within any reasonable fudge factor, I would say that the differences were negligable.

However, that does raise questions about the dimples used. As a thin film, the dimples would have been quite shallow, and quite likely very equally spaced. Not so with the MythBusters application; leaving all sorts of variables. However, since there results were highly significant, their test is the one that needs further study, rather than dismissal.

Further, A-B-A testing is questionable because of such little difference between that and A-B. Why not A-B-A-B..... to infinity? Would that help? At what point do we say that the improvement is real? Or even good enough?

@MetroMPG,
Please note that the test results were very significant. They were not minimal. Therefore, the small details you mentioned would not likely make any significant difference. Actually, what we can really say is that they got real results that were significant.

9% is nothing to disregard.

What still takes me by surprise is that no one seems to notice that the air on the surface of a moving object is what creates drag. Therefore it is reasonable that breaking that up would decrease the drag of attached air. Let's discuss the basics first.

Read

Post #22
To follow up on that point, I found this nicely written piece from airliners.net user QantasA332:

Basically, there are three primary types of drag acting on an aircraft: induced drag, skin friction drag, and pressure drag. It is pressure drag that is the main factor involved in the dimple design's existence. Pressure drag is primarily the result of a moving body's wake. Depending on how soon the airflow separates as it passes over an object - that is, how far along the object the flow travels before no longer following the contour of the object - the size of the wake will be larger or smaller. A larger wake equates to more pressure drag (put simply, there is a larger region of stagnant air behind the body meaning the airflow pushing on the front of the body has less impeding its production of drag) and vice versa.

Now, imagine a sphere. Because its height/diameter is large in comparison with its length, it is what's known as a "bluff body." Bluff bodies such as a sphere have disproportionately large wakes, and as a result they have disproportionately high pressure drag. (This is compared to both their own skin friction drag and a not-bluff solid's pressure drag). Obviously, then, overall drag on a sphere (or other bluff body) can be dramatically reduced if pressure drag is reduced. That is, pressure drag is what you want to specifically target and minimize.

Enter dimples. Dimples turbulate the airflow over an object, thus increasing the flow's kinetic energy. This acts to delay flow separation, which then leads to a smaller wake, which in turn leads to less pressure drag. And this solves the bluff body problem! Because bluff bodies have such high pressure drag compared to their skin friction drag, what little extra of the latter drag is created by dimples is more than offset by the drastic reduction of the former drag. So a golf ball - the classic example of a bluff body - will travel farther with dimples than without, and that is of course why they have come to carry these dimples.

Now, to finally answer your question: 'normal' aircraft are very simply not bluff bodies. Dimples would create more skin friction drag than they would reduce pressure drag, defeating their purpose.

So looking at a car, only where you have flow separation, would these dimples or any kind of vortex generator help. On most sedans, the back glass is prone to flow separation, and so you may see some reduction in drag there if implemented right. The rest of the car should have attached flow and thus the dimples would only add skin friction drag.