Roofline/Kammback transition - sharp or rounded?

[Piwoslaw]

Say the roofline in the back of a car is more or less level (black). Should the Kammback abruptly start dropping at a 11 degree angle (green), or should there be a rounded transition first (red)? (The angles are exaggerated)



Intuition whispers that the sharp edge between roof and Kammy would cause detactment of airflow (blue):



while a rounded transition would help guide the airflow smoothly down the Kammback:



On the other hand, maybe intuition isn't always the best tool in aerodynamics, and maybe the 11 degree angle isn't enought to detach airflow?

[Frank Lee]

*Opinion Alert* Seems like the guys that built aerocaps for their pickups report similar gas savings whether they did a nice job "feathering it in" or not.

[Deezler]

Rounded whenever possible! But your area of flow detachment would probably be pretty small even with a sharp edge. Remember that your boundary layer has some thickness to it, and can cushion these edges.

[Bicycle Bob]

I think we should call that a roof extension, and yes, the transition should be gentle. Prof. Kamm discovered that it is best to have a sharp transition to a flat back, after roof extensions, etc, have been carried as far back at a reasonable taper as other considerations permit. That sharp transition is one of the few places we don't want a curve in this game. On another thread, someone wondered about a saw-tooth edge, copied from nozzles that need silencing. That does not increase efficiency, it just puts the sound into mutual interference. The proper refinement of a kamm back edge is just a mild lip, which can trip the wake into a single large, stable vortex instead of many competing ones.

[Frank Lee]

Originally Posted by Deezler

Rounded whenever possible! But your area of flow detachment would probably be pretty small even with a sharp edge. Remember that your boundary layer has some thickness to it, and can cushion these edges.

By that point on the car the flow isn't really laminar anyway, right? That boundary layer gets bigger and fatter as it moves on back.

Remember this car?



Of the three PNGVs, it had the 2nd best Cd. Looks to me like the backlight was designed so that flow would reattach at the "trunk lid". So if the geometry downstream supports it, seems to me that transition isn't as critical as we'd like to think.

[Bicycle Bob]

That seems to be using an attached vortex to pad out the gap. That's a handy technique for open-top racing cars. A single large circulating pattern is set up to avoid chaotic flow and turbulence.

[winkosmosis]

Rounded of course, just like the transitions at the front

[IsaacCarlson]

.

[Frank Lee]

You don't know that... do you?

[Hasbro]

Function or fashion?

Thanks, BB