Rounded corners vs. Straight edges on kammback
 

I am really considering a much more dramatic, full kammback for my Neon. Allowing the sides to taper in as much as possible without interfering with opening the trunk. Also, I am considering making a hinge system on the top, to allow me to fold it back and use the trunk. My issue is with building rounded corners on the top-to-side transition.

I know that rounded corners are preferred aerodynamically speaking, ease of construction is another story:p. To the best of my understanding, this is because with just a straight, creased edge, the airflow "trips" over the edge, causing vortices which will increase drag and severely limit the drag reduction potential, am I correct?

So how drastic of a rounded corner does it need to be? I assume the more rounded the better, but is there a known minimum amount of curve that is required?

Finally, I know people can compare the amount of roundness by talking about the size of it, for example a 3/4" radius. So does that mean that if the curve was a whole circle, it would be 1.5" in diameter? Is that sufficient for avoiding air tripping over the edge? I am just trying to get my head around what kind of curve I need, to figure out the best way to build it.

Thanks for any input, you guys have already helped me out so much in building my Neon to where it is, and I appreciate it. This site is great!

No, we're talkin 2" radius minimum. That's just a shoot from the hip thing, and it assumes a fairly close aero design. 3" would do you for sure.

Keep in mind though that rarely do panels come together at a 90° angle, it's usually more obtuse making the actual part the needs to be radiused smaller.

Thanks, so as far as a 2" radius, the curve of a 4" diameter circle? Hmm i actually have some 4" pipe laying around as a nice template. Lol.

Make the structure out of aluminum bar and bend a couple transverse ribs to shape- 6" should work quite well. Then when you put the coroplast on it will conform to the curves you want. No, it won't have compound curvature, but it'll be "good enough" without going into fiberglass.

Of course you could get close to compound curvature by making really slim V slits on the sides and curving it in just slightly like a lobster tail. That would take a lot of work, though. Go for good enough now and if it works well, make version 2 even better.

Try it at 20 degrees from level and tuft test it. If flow is not attached lift it 2 or 3 degrees and try again. It should work best somewhere between 15-20*

That is actually exactly what I had in mind! Great minds think alike ;) Im trying to figure out the best way to lay out the coroplast, and how many sections to do it in. More brainstorming time.

hmmmm........

if you have squared sides, and things are not perfect, you will induce vortices which will "fix" the imperfections.

build the thing.

keep side and top tapers to 20 degrees or less, and bottom taper to 10 degrees or less.

i would not worry about the corners.

Nice! I'd align the ribs longitudinally so it bends easily, but you probably knew that as well. :thumbup:

If you have a large 4x8 sheet it should pretty much cover the whole thing, right? Otherwise let the split be at the middle because the corners will want to peel themselves back off.

Edit- 1,111th post! Hah

Edge radii solve a problem. Ideally (perfect design) flow wouldn't need to mix between top and sides. In practice cross winds mean up/down wash can't be avoided. From Hucho

http://ecomodder.com/forum/attachmen...1&d=1345635813

http://ecomodder.com/forum/attachmen...1&d=1345635813

Using a nominal 6' body %5 would be 3.6" radius.

Take a good look at the Prius, almost all folded corners. Is it ideal? Probably not, but it has been tested and it has low drag.

Mr Miller,

I think you need to reconsider this. Creating a vortex off a c-pillar is not what you want. You may be thinking cute little burbles of air that help the air "Re-attach". This would be an almost impossible thing to engineer with a team of scientists, a wind tunnel, mission statement, and $2 million NASA grant. So I doubt a guy in his driveway with duct tape and coroplast will nail it.

The Vortex creation that needs to be avoided are the long trailing tornados like you see off the wing tips of airplanes. These are very common. The energy required to create and maintain these is tremendous and it can only come from one place, the cars engine. We all know what that means.

The C-pillar is the most likely spot for these to form and they're caused when large pressure differences are within close proximity to each other, the air tries to move sideways from high to low pressure and in doing so, sets itself spinning.

So, in rough numbers terms, if Neon were to "Just Build it" yeah, anything is likely better than nothing, and he may realize 30% of the potential gain out of building a kamm extension just due to the fact that he's "Stuffing the Wake". But if he builds it properly, and radiuses the edges, he'll have a more optimized device that will see 70% of the potential gain. So 25MPG base, if he stuffs the wake, 26.5MPG, builds it better, 28MPG. Would he be happy with 26.5? Probably, but why build something half-baked when if you take the time to do it right, the result is so much better?

Also the 20° thing is not a very good rule of thumb. One has to consider what you're starting out with. If the slope at the top of rear glass is already at 10° or so, then yeah, 20° works, but if it's flat, 20° would be too drastic a change and would not be optimal. I really don't think you want more than a 10° change at any one point. This is why a blended curve works so well when adhereing to the template, it allows the air to transition as undisturbed as possible optimizing the result.

he is talking the tail of the car, right?
so the pic of the van is misleading.

we can argue about the transition from the body to the kammback, but suffice it to say if we are going from straight to 20 degrees, I'm willing to bet a 4 inch radius woudl be great, and if you had no radius, just a straight angle, you would lose a very small amount of efficiency.

On the sides and top of the Kammback, there will be no vortex created if the air was perfectly proportioned between sides and back. However, we all agree the odds of that are about zero.

However, if the top and sides are both at 20 degrees to the centerline of the car, then there will not be any huge pressure differentials in the air. What there will be is minor pressure differentials.

If there are minor "holes" in the air being turbulently attached to the car, a small vortex will help the air "fill the holes", just like a golf ball's dimples.

You and your strawman are of course welcome to disagree, but I again applaud the OP for DOING something and TRYING something where most of the rest of us are nitpicking his great efforts.

But you'd need a windtunnel to get it right.

The simple aerodynamics of yesteryear can be more easily applied and have good effects.

If you want to openly contradict me on a post in the future, be prepared for some major push back on my part.


So let's break down what you posted last.

We weren't even discussing the transition from roof to top of Kamm, you seem to be off track here. We were talking about where the sides of the Kamm itself meet with the top of the Kamm.

But while you're on the subject, going from straight to 20° would have dire consequences. I base this on the chart available to you here in this paper.

Drag Reduction of a Pickup using Add-On Devices

Look on page 85, what I see is that when you approach 20° you really lose the effectiveness of the addition and you start to generate a ton of lift. Is it still a net gain? Yes it is. Stuffing the wake is a net gain. But why go to 20° when for the same effort you could use 10°-12° and have better results? This is why I said what I did.

Feel free to back up your "Just put it in at 20° and it'll work great" statement, I'd love to see data on it.

How do you know this??? Without knowing what is upstream, you can't possibly make this statement. If the roof line was already substantially tapered to where it was at 10° prior to the added Kamm, but the sides are still flat, when you make the 20° all sides addition you speak of, there will be a serious low pressure area created on the sides with a higher pressure on the top. This will set the air spinning in a large undesirable vortex.

Anyway, your poo-pooing the vortex problem fails to recognize how significant it can be. You seem to think it a 5% impact on a bad day where you have created a monster vortex. Fact is it can be a 140% negative impact on a cars aerodynamics, and even a minor error would render an otherwise good design bad due to failing to consider this Vortex Generation. Radiusing the corner goes a long way towards reducing this effect.

Page with lots of Vortex discussion

This statement here seems to illustrate that you shoot from the hip and talk in pseudoscience terms. What "Holes" do you even refer to? I mean seriously? Would you please show me something, somewhere, that backs up what you're saying here? It sounds to me like you've got vortex generator babble mixed up with parasitic vortex creation and now you think any vortex is a potentially good thing. Again, support with something other than "Cause I say so." or "Cause I've built a million vehicles, so I know what I'm talking about" or "Here's what the Vortex Generator Support Group claims".

OK, here's the deal, go back and carefully read the first post written by the "OP" Let's call him 2000neon. You'll notice he is asking for advise and doesn't really have a preconceived notion of what he wants. So we gave him our best advise straight up, and he seems to appreciate it.

So far, it looks as though 2000neon likes what he hears in here.

Everyone here is proud of 2000neon for making an effort to do something. I for one am super heartened by the fact that he wants to get the best advice possible before he settles on a design, he wants his first effort to be a good one and not waste his time designing something which is not optimal. 2000neon is a frickin genius in my book.

The only "nitpicking", was against you. Until you educate yourself better and post things which are based on science, I think you can pretty much expect this is going to happen to you from time to time. I notice a number of people have spent a serious amount of time trying to discuss things with you rationally. I know I have been a bit terse here, but I just want to put you on notice, if you openly contradict what I say, you better have some facts and research to back it up.

"full"
 

I think that I'm going to need some help with respect to "full".Will it extend to the original length of the car,converting from a notch back to K-form roof,or are you planning to extend it beyond the bumper,combined with boat-tailing to go for really low drag?
If you maintain factory length you could mimic the architecture Chrysler did (as GM has done with today's VOLT),maintaining the same roof section curvature and greenhouse tumblehome,intersecting with an identical crease as on the standard car.
If you go longer,you could gently increase the roof/greenhouse edge radii as you progress back.
The 'Template' curvature is meant to prevent radical pressure variations which lead to the bleeding into the low pressure area,which creates the viscous shearing forces which sets up the nasty attached longitudinal vortices.
Over at the full-boat-tail trailer thread there are some pictorial drag tables.One of them demonstrates the difference edge radii can play with respect to drag coefficient.(It will show a bus of different configurations,with corresponding Cd).
The really low drag cars are circular,or semicircular when viewed from behind (this is excepting solar race cars with their PV arrays).

While it's entertaining to watch people come at each other on Internet, I have been wondering about why nobody invokes the Coanda Effect, named by Theodore von Kármán after Henri 'my tails on fire' Coanda: Coand

Maybe it's because it can't be invoked passively; the main aero example being blown flaps. If only there was a car with handy supply of hot, pumped engine cooling air in the tail. :)

Sorry for using 'invoked' in two different senses in a single post.

As penance I will refute myself: There is such a car, the Renault R8.

http://i.imgur.com/UdzN8.jpg

I thought the cooling air exhaust created an invisible spoiler that flipped air downward with a vortex. But as the vented air probably bleeds straight back, the reported gains [citation needed] might be due to the Coanda effect.

great study, and totally supports my argument on the angles.
if you look at page 75, they ran a CFD with 18.77 degrees, and a hard cut angle at the front, and hard angles at the sides, and the
FINAL RESULT IS EXACTLY AS I PREDICTED.
The minimum possible drag with no vortices.

Regards,
Douglas Robert Miller

Hijack?? I believe this has been discussed? This thread is about rounded corners on a Kamm tail.

Thank you all very much for your input, it is all greatly appreciated. If it helps, my name is Steve, it might be easier than "2000neon". You are definitely right that when (not if) I build this, I want to aim for as close to perfect as my skills and materials will allow for, so that's exactly why I have been doing a ton of reading on 3 wheelers tail extension, sven's probe, the aerocivic, weather spotter's tail, the new aerofocus, the cargaero trailer, and of course as much reading as I can on everything from Phil, your truck, and you yourself are a huge asset and inspiration. I'm sure that I am still leaving some people out too.

If you look at my other thread (which I just realized that I have to update), I have already gotten most of the simple mods done to this car already. I am now also planning on bringing the rear wheel skirts farther down to cover more of the rear tire, but that is for a different time.

I will try to address as many questions as I can.

Aerohead: What I mean by "full", is that I will be aiming to build it to the existing length of the car. I am debating whether or not to end it flush with the trunk lid, or take advantage of the extra few inches I can gain rearward of the trunklid, out to where the bumper is. I called it a full kammback because I have already done a small kammback, just over the top 2/3rds of the window. Although the side transitions were too sharp and definitely limiting my potential gains. Although I did have some small gains with it. This will be it's big brother version, and hopefully fight the itch to build a boattail (for now). :p

Based on a lot of the reading from other's build threads, I am going to try and aim for 12-15 degrees on the top. I am going to try and maintain access to my trunk, so the side taper will be limited to how far I cam bring it in without blocking the trunk lid. I'm going to try to keep everything as smooth as I can to prevent any significant pressure variations like you mentioned.

ChazInMt: I really appreciate the input, and the nice words. Same to you Dr Miller, and Sven and everyone else. It is actually incredibly helpful and motivating to have the support and input when going about something like this, especially detail stuff like construction and angles.

As I was saying before, the build plan I am aiming for will be an aluminum bar structure, mounted longitudinally, with 2, 3 or 4 vertical ribs, with some perpendicular bars attaching them to eachother. The exact number and placement will be determined when I can get a cardboard mockup built on the car. I am hoping to be at least that far by the weekend. I usually average between 55-62 hours of work a week, plus time for my girlfriend, so free time is hard to come by.

When I first started building my Neon I was very hesitant, even a grill block had me worried. Now that I have slowly become desensitized to even some of the more drastic modifications, I am actually really looking forward to taking this next step. It has gone from simple money savings, to a whole lot of fun and a great way to let my creativity come to life. I can't give enough thanks to everyone here for your build threads, discussions and input you're giving me.

It has been since 73 that I had an R10. I always thought that the air entered the grill work in the top and went forward through the radiator and out underneath.

COcyclist --
As in 'threadjack'?

[citation needed]

I thought it was about round corners vs square edges.

I try to not be a problem. Please don't take my going grammar nazi on myself as license to dog-pile on.

Varn -- I hadn't consider that possibility; but wouldn't that be wake-subtracting?

Hahahaha, the fights between the real aero guys and the black magic aero guys never cease. I think I shall start a Sensible Aerodynamics thread so perhaps we can enlighten the black magic guys, and remind the real aero guys that you don't have to make aero sound scary just so you can prove your credibility of knowledge on the subject. Aerodynamics principles are easy to understand, and to a certain extent can be utilized easy as well.

On the other hand, if you are not an engineer or physicist, and have never had wind tunnel or CFD experience, then you better shut your mouth instead of 'eye-balling' vortex physics.

2000neon, I can come up with a construction method that will allow you to easily create the radii you want! If you post pictures and dimensions of your vehicle, I'd be happy to work on designing the Kammback with you! :)

drmiller, you are silly.


Here's the page you use as "Support".

http://i46.tinypic.com/2mcx4bl.jpg

Bigger version of pg 72 to contemplate

Nowhere does it state that the 18.77° angle is optimal, and on page 84 Feysel concludes that at 10°-12° the drag coefficient is optimal and as you go beyond 12° it increases. Just for kicks, on page 9 Feysel (the author) talks about the evils of the vortex generation I speak of, it's the last paragraph of the page.



http://i49.tinypic.com/5a7g9.jpg

Bigger version of pg 85...for your consideration.

Here is the page 85 which "Summarizes" the results. It seems pretty clear to me that the 18.77° is decidedly not the minimum drag. Now, I'm no math major, but the number .3091 looks to be greater than .2892, unless I'm not reading the graph right, it is decidedly less than the "minimum possible drag" you speak of.



So, Mr. Miller, it seems to me that you have used something completely out of context to try and bolster your stance. Science and engineering principles require that we look into all aspects of a study and not just focus on the ones that help us out.

Tell you what, since we've made a complete mess out of Steves thread, I have a proposal: I'd be more than willing to remove all my posts here regarding your posts 6,11, and 17 if you take them down. Like it never happened. See if you can salvage some dignity here or do you want your clever mental judo on display for all to see?

That's all fine and dandy except for the black magic guys that LIKE black magic and don't want to entertain reality.

"You can lead a horse to water but you can't make him drink."

Learning > Being Right
I just want to see a Kammback Neon, maybe I can help, and maybe this guy drives away happy from the gas station until his Neon bites the dust. :P

Here's my analysis based on this template overlay.
Angle at front of Kamm 5°, overall 8°, rear of Kamm 11°. All angles are based on the ground being 0°. The top of rear window to trunk rear lip looks to be 17°. The car is "Leaning forward" in its design, the body side molding is 2° and the top of door line is 4°.

The Kamm looks to be about 7-8 inches tall at the back of the car based on a guess of 16" wheel size for this car, not exact but the best I can do without measuring a door er sumthin. The neon green line pictured is 8°, the red line is the template, the yellow orange thing is my quick & dirty side view of what the Kamm ends up looking like if it follows the template.

Hope this helps. Something to start with anyhow.

http://i50.tinypic.com/2i27jbo.jpg


Since ecomodder isn't able to provide bigger pictures, I made a link to this so it's easier to see wassup.
Bigger Pic to LΘΘk At

Hey Chaz, I have no problem with the added discussion at all in here. That picture looks great and pretty much what I have in mind as far as a side profile. I am hoping to be able yo be able to keep it as close to the template as possible, and maintain the curvature all the way through it. The roof of the car already has a pretty dramatic curve to it, and so I think by keeping it a smooth curve it will work better than a flat edged transition down the top of it.

As far as kamm height at the rear. I think you are pretty close. I'm pretty sure that the rims on that neon are 15" diameter, but with those hub caps, the total diameter of the caps is probably close to 16", so your guess of 7-8" seems pretty close. I would prefer the rear to be higher than lower, at least in the beginning, because it will probably be easier to take away height, than add it back on later. Plus rear visibility will be improved as height increases.

MTrenk, thanks a lot, I am glad to have any and all help I can get.

To make things a bit easier to visualize and understand where I am coming from. Here are some pictures of my car.
The earlier "partial" kammback
http://i282.photobucket.com/albums/k.../Image0068.jpg

Side profile, Due to space I couldn't get any farther back, and it is a little bit too high of a view for using the template, but still gives a rough idea of what I'm working with.

http://i282.photobucket.com/albums/k.../Image0067.jpg

You can see how the rear bumper extends another few inches past the rear of the trunk lid. I might try to angle it rearwards at the back like Chaz has in his picture of the white Neon, I can go a few inches past the trunklid, without going past the bumper.

Thank you for the pictures Steve. I will try to have some drawings by tonight. I've never scanned anything to this computer before, so I might have to do the first rendition in Solidworks.
The plan in my head is to have the Kammback bolt to the trunk, yet still extend to the sides of the vehicle, hopefully mounting flush with the roof and C-pillars. In order for the trunk to open, the Kammback will have to be able to rotate upward at some location, and perhaps it can be as easy as using an aluminum door hinge from the hardware store.
Get back to you soon.

Here's a nice visualization of what will happen with the full Kammback:
http://www.youtube.com/watch?v=6b36ZP3PGis

And here's what you get with a full boat tail: (and I have a lot to say about how this design works not just by modifying the top, but the bottom of the car as well)
http://www.youtube.com/watch?v=cPQS2...feature=relmfu

The flow off that boattail is impressive, of I didnt have to parallel park the car at school that would be a serious consideration.

I was thinking about this more today at work and revised my plan a bit. Instead of hinging a small section of the kammback to open to the trunk. I am now leaning towards making the kamm one solid piece, and allow the whole thing to hinge upward. Doing this allows for much easier building of the structure, and should be easier to make much more solid. I can also then taper the sides in farther, and have the sides of the kamm go overtop of the trunk. That should allow to gain more plan taper, and round off the back of it as much as possible.

Overall, I think that this will be easier to build, be sturdier and function better. I can then use a simple pin system at the back to keep it centered, and keep it down when I don't need in the trunk.

MTrenk, as for this idea, I am thinking if I use some thin (1/8th or 1/4") by about 1.5" aluminum to build the base of it, that will sit on top of the trunklid, and get me the basic shape at the bottom. Then I will be able to build the vertical transverse ribs to get the rest of my shape, and build sufficient support and a hinge at the top of the rear window to allow the whole thing to pivot upwards to get into the trunk.

I should be able to use some adhesive foam or rubber strips to prevent squeaking and help seal the kamm to the car.

What I am thinking for the hinge is two hinges spaced reasonably far apart at the top of the window. The two problems I see with that is that they will be in the airflow, and since the roof is bubble shaped, and is curved across the top of the rear window. I am afraid that with two hinges spaced apart, they will both be slightly curved different directions and may bind against eachother? Or I could just be overthinking things, this may even be a case where have "sloppy" hinges may be better and allow some wiggle room between them.

You could use two Heimz joints as your hinges, then binding wouldn't be an issue. :)
Sounds like you are brainstorming quit a bit on this! Keep going, and don't be afraid to draw anything thing out to sort your brain out. Even if you think you suck at drawing, still draw what you're thinking.

I've got pressing Formula SAE matters to take care of tonight, so maybe drawings tomorrow. :P

That would definitely work, but would complicate finding parts to use, and possibly budget.

Thats cool, I appreciate any help, whenever you could do it. I am trying to brainstorm as much as possible, I always like to at least work through things in my mind one step at a time and see what problems I run into. It has already helped out a few times in the planning of this kamm already. Its easier to fix something as an idea in my mind than when its metal on the car.;)

Are the junkyards still full of cars from the 40's? The tulip panel on those probably had more curve than your Neon, and external, chrome teardrop trunk hinges.

on the pics of the boattail, the second pic shows the bottom angle is much too steep. The air is not staying attached to the bottom. Study after study has reported 10 degrees is about maximum on the bottom of the car - not as much energy below the car.

As to the previous study, I'm having troubles keeping up. First we were talking about angles of sides to top, and the study provided shows sharp angles don't hurt anything. Then we were talking about kammback to body, and study provided shows angles there don't matter much up to at least 18 degrees. Then we were talking about angles of kammback, and study showed 18 degrees stays attached.

Then we got switched around to a specific case which showed there is a possibility too much angle doesn't have the correct trailing vortices, and if you have STRONG vortices, in just the right place, you can have less overall drag because the vortices fill teh wake better. Predicting the lateral vortices is more than I can do off the top of my head, but the fact is they exist, and the fact is if you do things just right, you can use them.

Lateral Vortices are bad in almost every case.

drmiller100 said part of the study constituted 'proof'.
If you made it past Geometry, you know that the only real proof that exists belongs in mathematics. Science provides evidence, and by gathering this evidence, we obtain knowledge and understanding. You merely have one reference point and have not expressed any desire to further your understanding of the subject matter. There are certain aerodynamic principles that can be used for eyeballing and brainstorming. Vortices do not belong in this category. This forum is supposed to be about learning, and instead of learning, you read something that barely gives you want you want and treat it like gold. The only gold here comes from the people who have dedicated their time to learning, and sharing what they have learned. You have learned how to read a graph and make what you want out of it; statistics can be manipulated very easily.
I suggest you take a positive learning approach to this, instead of arguing with people who have the knowledge you should be seeking.

Coanda
 

I looked into Coanda's work.It appears that none of his aircraft were capable of flight.His work is associated within the UFO community with respect to secret Luftwaffe projects.
Coanda did go on to make contributions within the aeronautical community,however none were associated with the 'effect' which bears his name.
Suction slots can do wonders with aft-body flow attachment,but so far,they require more energy to operate than they save.I believe that Georgia Tech is currently investigating this field of study.