Autospeed Beetle Article

[Mighty Mira]

What do people make of this article here?

The relevant commentary from the author, Julian Edgar, is here:

So the problem must be at the back? And it is. But it's a different problem to that which we've seen before. Here the flow remains attached right down to the line of the guard/boot opening. And this results in a very small wake for the (it's larger than you'd think) size of the car. But it also means that the airflow wraps in one long curve from the base of the windscreen right around over the top of the car to nearly the rear bumper. Aeroplane wing, did you say? And not only will this shape have major lift (without an undercar ground-effects tunnel, anyway; and I looked under and didn't see that), but much of the force will be upwards and rearwards - creating that monster drag.

I'm not sure I buy his explanation. The way I understand the rear of a car to work, is that the top of the car is analogous to the top of a wing or airfoil. Give it too much angle of attack, and it stalls. See this page., which gives the angle at around 16 degrees, very much like what the maximum angle most low drag cars get away with. He also makes the point that the stall "speed" will vary with loading, but it is the angle of attack that stays constant wrt stalling. I've also measured the maximum rear angles of the UFE-II (15 degrees) UFE-III (18 degrees off very fuzzy picture, 17 degrees from another), EV1 (21 degrees), Prius (17-18 degrees) and Insight (15 degrees) using GIMP and google image search, looking for photos of these cars exactly side on. Then I find the steepest part and get the angle wrt the horizontal of a tangent to that region.

If the rear surface of a car will stall like an airplane wing, it seems as though this fits rather well. Especially if you consider that angle of attack is measured as the angle the chord of an airfoil makes with respect to the oncoming wind. Note that the chord is the line from the leading edge to the trailing edge of an airfoil, and so in a typical airfoil or wing there will usually be a more extreme angle to be found taking a tangent to the rear upper surface of the wing compared with the the chord line. This means that if a wing is just stalling at an angle of attack of 16 degrees, the most extremely angled tangent at the back of the wing (analogous to the boat tail of a car) will be higher, maybe 18 or 20 degrees or so.

This agrees well with the boat tail angles of very low drag concept cars and hybrids, which can be assumed to have pushed this to the limit through trial and error.

So if you have a look at the pictures of the rear of the beetle, there are also some tufts that aren't agreeing with Julian Edgar's explanation and appear to be in turbulence higher than the artist's interpretration of where turbulence is occuring. In fact, the third tuft from the top of the window is pointing back towards the front. Perhaps there weren't enough tufts to get a good idea of what is happening back there.

But maybe I'm wrong. It's where our mental models are challenged that we learn something new.

[lunarhighway]

you might be on to something here

if drag at the rear is caused by "stalling" due to a to extreme angle that would bring vortex generators back in perspective...

[trebuchet03]

If the rear surface of a car will stall like an airplane wing, it seems as though this fits rather well. Especially if you consider that angle of attack is measured as the angle the chord of an airfoil makes with respect to the oncoming wind.

Measure it as an entire foil - not just the rear of the car As you said, stall speed is related to AoA - measured from leading to trailing edge (as opposed to the rear 80% of the foil). That's not to say the ideal rear end will be shaped like the rear portion of a foil


Yes, there are a few disagreeing tufts in that photo (mainly the third tuft down photo left on the rear windscreen and the tuft in the middle bottom of the hatch, left).

For the windscreen one -- I can't explain that, there could be external forces or a pressure wave interfering. For the one just left of the rear fender on the hatch, that looks like an interaction with the flow coming over the fender - note the tuft just above it, it is being sucked to the right (towards the fender), possibly indicating a low pressure (faster flow) zone.

In general though, the straight tufts on the bottom portion of the hatch are very telling. If the flow has stalled/separated - the odds of getting a straight tuft on a photo like that are very very low. It's very likely separation doesn't occur at the same point (hatch lid), but it does appear to get pretty far down. Perhaps the bottom of the glass


I think the person that runs this site is a member on GS... There's some good information in there

[Snax]

I don't believe that tufts tell the whole story about what is really going on. Given the apparent high degree of attached airflow over the back of the car, allot of energy is actually being put into accelerating the air. Similar to wing flaps which are designed to extend the attached airflow, providing additional lift, that comes at the price of additional drag. It's not the stalling that causes the drag, but rather the smooth airflow in lieu of it!

This may seem counterintuitive to say, but going back to the simple example of placing one's hand out the window into airflow, the effect of extreme unstalled angled of attack vs. a stalling condition can be felt as reduced rearward pull at the initial point of stall. In other words, the roofline of the Beetle is TOO CLEAN. It could benefit significantly by inducing stalled airflow much earlier over it's profile - something I believe that they actually addressed with that silly but useful little motorized spoiler on the Turbo models.

[trebuchet03]

Originally Posted by Snax

. It could benefit significantly by inducing stalled airflow much earlier over it's profile - something I believe that they actually addressed with that silly but useful little motorized spoiler on the Turbo models.

Interestingly, the 2004 Turbo S has the very same cD of .38 - go figure

[Mighty Mira]

Originally Posted by trebuchet03

Measure it as an entire foil - not just the rear of the car

Why? (Not sure if you are serious or in jest. If in jest, please ignore.)The point of measuring the maximum angle at the rear of the car is that in designing your own boattail at home, you know what angle to aim for, and not to exceed. By the time the air gets to the rear of the car, I don't think it's going to make much of a difference what point of the front of the car was technically forwardmost into the incident airflow. But there is a certainly a maximum angle at the rear (mostly, the top, because it is going to aid your drag coefficient irrespective of crosswinds) you can exceed without using vortex generators.

As you said, stall speed is related to AoA - measured from leading to trailing edge (as opposed to the rear 80% of the foil). That's not to say the ideal rear end will be shaped like the rear portion of a foil

No, although I've yet to see one of the top notch low drag cars with a rear that doesn't look like a truncated foil. Actually, come to think of it, it's less a matter of being truncated, and more a matter of being extended downwards. And if you go to solar cars, a cross section looks very much like an airfoil.

For the windscreen one -- I can't explain that, there could be external forces or a pressure wave interfering. For the one just left of the rear fender on the hatch, that looks like an interaction with the flow coming over the fender - note the tuft just above it, it is being sucked to the right (towards the fender), possibly indicating a low pressure (faster flow) zone.

I suppose that we need some sort of apples to apples comparison when discussing how good the beetle's rear is in terms of drag. It has a Cd of .38, but how much of that is due to ugliness elsewhere? e.g. the fairly abrupt angle between hood and windscreen, the egg beater wheels, the angle of the fairings behind the wheels looking to be too steep but to leave a wake behind, lack of an undertray, in general the impression that an interior decorator was made responsible for making it "streamlined" instead of someone who could do math. Maybe the back isn't as bad as I would have thought?

I suppose if the rear of the car HASN'T stalled at that point (which I'm not 100% convinced as per the photos), at least you do have a situation where you are changing the direction of the air to be pointed downwards, and this will give you substantial lift and cost you energy to do so. Most of the top strata low drag cars (not including solar, because they operate outside of usability constraints and can do a proper extended airfoil) are kamm backs, with the end of the kamm back pointed near horizontal. I suspect that this does two things, by getting the air leaving the car to be moving the same direction it was as it found the car.

1) It minimizes lift by not sending the net flow of air that the car meets in a downward direction, making handling fairly stable, and
2) since it isn't generating lift, it's not costing energy to do so.

If this wasn't the case, the wake could be made a bit smaller with the same vehicle length by keeping the most extreme angle achievable with flow still attached, right until the very end of the vehicle as per the beetle. The fact that most don't is pretty telling, and even the "exception", the EV1, has the bottom coming up to the same degree and hence is still leaving the net mass of air behind the car going directly behind (although slowed a bit because of friction).

I think the person that runs this site is a member on GS... There's some good information in there

I realize that... I've browsed his site many times, but it does seem a bit sparse and geared to the layman, to be honest.

[csrmel]

it's a neat wing, but it will never "fly" with mainstream. people don't want efficient cars. they want fast powerful looking cars with moderate to low efficiency.

[trebuchet03]

The point of measuring the maximum angle at the rear of the car is that in designing your own boattail at home, you know what angle to aim for, and not to exceed

Yes, that makes a great deal of sense - and there's some good examples out there of what not to do But as far as AoA is concerned, the angle the rear end makes with respect to horizontal is not the same as the AoA.

No, although I've yet to see one of the top notch low drag cars with a rear that doesn't look like a truncated foil. Actually, come to think of it, it's less a matter of being truncated, and more a matter of being extended downwards. And if you go to solar cars, a cross section looks very much like an airfoil.

You're absolutly right... I really need to watch my negatives (I actually edited a sentence because of a double negative, and screwed up what I meant).
It should have read:
"That's to say the ideal rear end will be shaped like the rear portion of a foil."

I suppose if the rear of the car HASN'T stalled at that point (which I'm not 100% convinced as per the photos), at least you do have a situation where you are changing the direction of the air to be pointed downwards, and this will give you substantial lift and cost you energy to do so.

So that's an interesting idea to explore. Lets say that flow detaches a little above the midpoint of the rear glass. That would mean, all of the lift issues are a result of the top 1/4 of the vehicle (okay, maybe 1/3). Whereas, if separation occurs lower down - we have about half of the vehicles rear providing lift. We know it has a major issue with lift - but I guess the question is, where's it happening and to what degree? It would really be nice to see some CFD on the beetle shape. Both scenarios are possible though.

What we really need is more photos Or tunnel testing smoke screen photos

Just curious, do you think pressure lift is more significant than reaction lift?

[Mighty Mira]

Originally Posted by csrmel

it's a neat wing, but it will never "fly" with mainstream. people don't want efficient cars. they want fast powerful looking cars with moderate to low efficiency.

That's 100% dependent on the price of fuel. People still like to get to A to B quicker than you can get there with a horse or bicycle.

Investing in fuel efficiency modifications and driving is something that already has a fairly short payback period. If the fuel price goes up like it's doing now, and keeps going up, it's all gravy.

[Mighty Mira]

Originally Posted by trebuchet03

You're absolutly right... I really need to watch my negatives (I actually edited a sentence because of a double negative, and screwed up what I meant).
It should have read:
"That's to say the ideal rear end will be shaped like the rear portion of a foil."

Ah! A typo! Now it all makes sense. I didn't think you could be taking engineering classes, involved with HPVs and say something like that. What a relief! Written communications are inherently flawed but necessary. Oh well.

So that's an interesting idea to explore. Lets say that flow detaches a little above the midpoint of the rear glass. That would mean, all of the lift issues are a result of the top 1/4 of the vehicle (okay, maybe 1/3). Whereas, if separation occurs lower down - we have about half of the vehicles rear providing lift. We know it has a major issue with lift - but I guess the question is, where's it happening and to what degree? It would really be nice to see some CFD on the beetle shape. Both scenarios are possible though.

What we really need is more photos Or tunnel testing smoke screen photos

Just curious, do you think pressure lift is more significant than reaction lift?

I think that they should be the same. i.e. If you have reaction lift in a particular design, it should show up somewhere in a more pressure under the car or less pressure on an upper surface.

I was thinking about this the other day, for a long time, Bernoulli versus Newton. Both should be a different method of looking at the same thing. Ultimately the net force acting on the car should be the sum of all the pressures on the car at different points. i.e. an integral over the surface of the car of the pressure. It can't be any other way.

Newton is useful to think of because if you can figure out where you are sending the air, you have a good indication of where your object should be pushed. Someone here has a picture of the big swirl behind an airplane in the cloud. It tends to be easier to visualize with laminar flows without a stall, because you can think of a big body of air moving a particular direction. Although I guess in a stall you are actually causing a mass of air to be moving in the direction that you are going (if you have ever walked by a semi traveling at speed you will know what I mean), and conservation of momentum means that your airfoil (or car) must be slowing down as a result.

After a certain point behind the vehicle, everything decays into turbulence and the kinetic energy you have imparted to the mass of air behind the vehicle remains in the form of heat (which is just atoms of air moving around on average somewhat faster, but every which way).

I still have a bit of an aversion to Bernoulli, but that's just because of the grade school science textbooks that mangle the explanation with a wing, arguing that because the top of a wing has more distance for air to travel, that the air pressure is lower and you get lift. The logical extension of that is a wing shaped like a triangle with the base at the bottom, because then you have even more distance to travel and so more lift. And that's obviously wrong.

But in the end, there are two truths. Momentum must be conserved (i.e. send a mass of air going one particular way, your airfoil/car will be sent another, i.e. you want to leave the air just as the car found it for low drag), and the integral of pressure over the surface of an airfoil/car must be the net force (due to air resistance, every net force I'm talking about above is only regarding air resistance obviously).