Aero Experts - Radius for roofline/Kamm back curve?
 

Kamm device below inspired by jwxr7's treatment on the back of his Geo.

My question is, what radius curve can I use that will keep the airflow attached even at full highway speeds?

I put this on yesterday.
It extends about 6" from the rear edge of roof.
Results appear good, going by gas gauge markings vs. odometer. No fillup data yet.
https://home.comcast.net/~brucepick1/...mmBack_350.JPG

Note the side and center supports have a curve in the top edge, forcing the main surface to curve downwards.

I want to build this using very thin plywood, but about one foot from front to rear. With several ribs spaced across under the Kamm surface to force the curve. To be primed and painted for weather protection.



Also be aware - though this has no effect on the radius question:
The liftgate on this car has its "cuts" on the rear face, letting the gate swing up. As the Kamm back is attached to both the liftgate and the sides, I have to temporarily remove the tape at the horizontal corner joints to raise the liftgate. I've already worked out a way to do that in wood with some relatively simple hardware.

Two other pics: temporary air dam with forward belly pan and ghetto grill treatment.
Actually off-topic but I haven't posted these here yet.
https://home.comcast.net/~brucepick1/...oplast_350.JPG

https://home.comcast.net/~brucepick1/...oplast_350.JPG

Looks like you have come close to jwxr7's design. His seems to follow his curve of the roofline more closely, both front to rear and side to side.

jwxr7 should have some thoughts on this. Is his sealed at the joint with the roof and is yours?

I can't really tell from your pics but your roof may be considerably flatter than his. I would think you would want to continue the roof angle a bit with your device rather than dropping it too fast. My air pressure/flow gauges have shown that attachment comes close to totally disappearing at the transition from roof to back glass on my car.

I am glad to see a few aero mods showing up that rival the radical appearance of my and basjoos' front fender skirts.

IIRC; ideally a Kammback should decrease in size until it is about 50% the cross-section of the body. I wouldn't make the decrease in cross-section happen over a slope of more than 20%, because above that you risk creating flow separation.

Aha! Sort of an answer to the radius question.
OK, this says that a slope >20% is trouble.

I'm going for a curve. That is, there's never an angle. However I'd think you need a gentle curve - question is, how gentle? I'm presuming a curve is better than an angle but anybody who knows more than I do, please speak up.

Maybe I'll start it with a downward curve and then continue at a steady downward angle. Awaiting more input from you guys out there.

As for the 50% cross section, I read that too. My understanding is, the size of area of turbulence is what counts and the greater the reduction the better. I'll end up lowering the separation point several inches and also making the separation area narrower, becaue the two side panels are angled in towards each other to make a taper.

A square backed wagon has a huge turbulence area; whatever I can reduce it is good. I'm not building a boattail (sorry basjoos etc.), it's just too much to graft onto the back of a wagon and too radical looking for me.

Thanks for the good word!

Yikes - on your car with it's nearly seamless visual flow from roof to rear glass, you say the airflow detaches nearly completely? That's scary; if your car won't keep that flow attached there I can't see how mine could. Clarify?

This one should be temporary - a design/build excercise for the plywood one I want to make.

This one is taped right down to the sheet metal, both on top and on the sides. There were some forced irregularities due to the hinges which are external (see photo) and about 3/4" thick.

Since it's taped onto the sheet metal, it should match and follow the body contour. That was the plan anyway. There's the "step" where the air has to get over the bump onto the coroplast. I didn't seal the roof-to-tailgate joint (yet).

Yes, my roof is just about dead flat front to rear. It's a decent sized wagon. Has a bit of a dome curve going from side to side though.

I'm definitely not an expert, mine turned out curved the way it is by luck of assembly. I tried to attatch the sides and top in such a way that the tape would hold it so there isn't a step transition. BTW Mine is taped along all exposed interfaces. It seems like a curve would be good for a transitional area. My rear end curves already just before the kammback begins but a larger radius seems like it would be better. It is pretty hard for me to get curves on purpose with coroplast but thin wood may be better. I was fortunate enough to have sides to my lift gate area to attatch to so it opens normally. If mine were like yours, it would have been that much harder. Looks good so far.

Good luck, keep us posted with pics :) .

Ditto on jwxr7's statement of being no expert. I'm not very sure the supposed experts are very expert in this area either.

I am not sure that you should want try to keep the airflow attached at the rear of your wagon. You may just need a clean break for detachment of high speed/low pressure airflow to rid your car of the turbulence it causes at the rear as quickly as possible. I believe that is the determination of testing done with large box trailers. The hollow extension you are using is similar in purpose to the hollow extensions tested at the rear of trailers.

The reason I asked about sealing is that I tried a contrivance on the rear of my car where I left about one inch open above the roofline and had the device sloped to the rear covering about half of the rear window trying to guide air from the roof into the low pressure area. What happened was that the low pressure got worse because the air from the roof I was redirecting was lower pressure than was already there.

I have the aforementioned trailer testing files stored on my computer and could either send them or come up with a link if you would like to see them.

CO ZX2 -
Thanks. Maybe a pic of the device you tried? Sorry that I can't picture it. Sometimes I'm thickheaded on that kind of thing.

Anyway, my thinking is I need to keep the flow attached as it goes onto and over the Kamm thing. So I don't want too much of an angle of change or too tight a radius, to keep it attached to that. Then when it detaches at the rear end, well - it's going to detach sooner or later, but if I give it a smaller turbulence area and smaller cross section where it detaches, that should be a help.

No pic, was not on long enough to worry about.

It was a scoop, more or less, that picked up air at 1 inch above the roof, one inch from the glass at the top and dumped out the bottom about a foot down on the glass and about 6" away from the glass at the bottom. Solid from top to bottom.

My thought with yours would be less angle off the roof to assist in smooth flow. Too much angle would create tubulence because the air is moving so fast it cannot follow the shape. The only other area on my car that has higher velocity/lower pressure is the hump above the winshield where the air tries and does separate momentarily from the car. Tapering in at the sides would be a good idea as you would be dealing with considerably higher pressure air from the sides of the car.

If you do not already know about the trailer testing, you really should read what they have to say. Let me know and I can get it to you.

Brucepick, it seems like the back of your wagon already has a pretty sharp transition, so maybe there is nothing to be gained. As opposed to the Metro or a Saturn wagon, which has a smooth transition that keeps the low pressure air attached longer... Know what I mean?

Yes, as built it's a pretty sharp transition. Roofline drops a bit at the very end but just a little. The sides taper inward, also just a little.

So is there some gain in having the separation occur a bit further back? Either like what's in the pic in the first post (or longer front to rear as I plan), or maybe just a horizontal fin going straight back 2-3 inches off the last bit of roof edge?

CO ZX2 writes: If you do not already know about the trailer testing, you really should read what they have to say. Let me know and I can get it to you.

Well, Bruce here you are. Notice how the top of their extension is the same angle as the roof, sides tapered. I have another file with more pics if you need it.

Trailer Aero Testing link: SAE 2000-01-2209
https://permanent.access.gpo.gov/webs...00-01-2209.pdf

Zounds. That article is an eyeful.

Their "boattail" seems to be parallel to the truck's body lines. However the panels are all "stepped" inward, that is the distance between the two sides is narrower than the truck's width, and the height of the top panel is less than the truck's roof. Almost like jwxr7's concept except due to the size and proportions they were able to use a stepped transition rather than a smooth one, and apparently it worked even with the step.

I'm going to have another look at the article, maybe at work where I can print it out on paper and share it with someone I know who may understand some of the bits that I don't get. I took a decent physics course but we didn't cover aerodynamics. I'll have another look but I don't know if I'll ever be able to say honestly that I understand what is going on there.

I don't know if you've checked out metro's kammback section so here's a link. https://www.metrompg.com/posts/boat-tail-prototype.htm
It shows the back of a semi trailer with angled boattailing.

Ya, I don't know those answers. My Element and Odyssey have a sharper edge at the top, and rounded edges on the sides. So, like you, I am wondering if there is an advantage to having a little plate that extends backwards 2 or 3 or 5 inches on either the top or the sides.

I sure can have fun speculating. Next I'm going to fully read the metrompg pages quoted two posts up by jwxr7, and get some education.

I think the decision behind the Element and Odyssey design is that consumers can accept a "spoiler" at the rear roof edge but would think the same thing would look quite odd if used on the sides. (I"m assuming the sharp top edge is kind of a mini spioler or horizontal fin).

Also if the car has rear lights going up the D pillar they would be blocked by a vertical fin unless the fin had additional lighting inside. Extra cost to the mfg.

Here is what I would do..... Get some tape, and some red string (for contrast). Put a whole bunch of tufts just before the kamm, on the kamm, and on the rear windscreen. Then, drive at a variety of speeds and have someone take photos from another car.

Then, do the same thing without the modification. If the strings on your rear windscreen suddenly point up, you've found a less desirable trait of your design.
-------
Now, from research for the HPV - a teammate found a whole lot of information on diffusers and optimal angles. We found 8 degrees off a flat plane to be the best for minimizing wake.

-----
Unless you do a CFD model, the best way is by guess and check (considering the low cost of a prototype). Try several iterations and see what shows the least amount of disturbance.

All of that said - I've always wanted to try taking big PVC, quartering it and seeing the results. I recall one boat tail design (for trailers) that had rounded trailing edges.

Bruce: my 2 cents...

- a curved shape at the rear, both in profile & plan views is best if possible (look at the EV1; Prius etc.) but not required (see the truck trailer prototypes, both stepped & straight angled)

- as for what angle or radius, you're safe if you go for something that doesn't exceed those cars' shapes at the rear. The figure often thrown around is 10 degrees, but taken from what tangent is the question