Aerofoil layout

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
Use the Riblett method and happily scale away evermore!
The drawing I posted yesterday shows a shape drawn on tangents "the NACA way" (in blue) and scaled there way and what you call "the Riblett method" (in orange) [actually I scaled both offsets from the mean line rather than the chord line so the distortion is magnified] . The reason the NACA did it that way is because it reproduces what happens when you bend an elastic material (and they were a bunch of math nerds). As you can see from the drawing in post #16 when you scale a shape the NACA way the new shape has the same camber and thickness distribution as the starting shape but the shape derived by multiplying the y offsets does not. This may or may not have undesirable consequences (usually not) but if you need an NACA 17.5 the only way to get it is to either find software that uses their formulas and accepts decimals or scale an even numbered set of coordinates from a catalog as per the blue lines in post #16 because they didn't publish odd numbers or fractions.
 

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
I wrote a Python tool to generate NACA airfoils, and was frustrated when checking it against Xfoil — until I realized that Xfoil does not generate NACA airfoils correctly. It applies the thickness perpendicular to the chord line instead of the mean line. I didn't know the internals of Xfoil well enough or know enough Fortran to fix it, and never got a response to the bug report I sent.
Well crap. The NACA did publish the formulae for producing their curves, and a program once computers became available, but I've never tried to find it because I only use them as seeds so don't really care about authenticity. Now that the sun's up I'll type in coords and compare them to the shapes from xflr5 and JavaFoil.
 

davidjgall

Well-Known Member
Supporting Member
Joined
Oct 7, 2010
Messages
163
Location
Florida, USA
Well crap. The NACA did publish the formulae for producing their curves, and a program once computers became available, but I've never tried to find it because I only use them as seeds so don't really care about authenticity. Now that the sun's up I'll type in coords and compare them to the shapes from xflr5 and JavaFoil.
It's true, Xfoil (and, thus, xflr5) draws NACA foils incorrectly. If one searches within the 'naca.f' source file for Xfoil one cannot find a single instance of 'sin' or 'cos' anywhere.
 

Attachments

  • naca.f.incorrect.png
    naca.f.incorrect.png
    50.6 KB · Views: 0

Aerowerx

Well-Known Member
Joined
Dec 1, 2011
Messages
5,885
Location
Marion, Ohio
I'm confused.

If you look in a catalog of NACA airfoils and pick one you like, and want a 4 foot chord, why cant you just multiply by 4 feet?

Now, I can understand that if you want some thickness that is NOT in the NACA catalog, you have to do some mathematical gymnastics. You can't just multiply the Y values by some factor.

Which of these two are being discussed here? I took the OP to mean my first statement.
 

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
I'm confused.

If you look in a catalog of NACA airfoils and pick one you like, and want a 4 foot chord, why cant you just multiply by 4 feet?

Now, I can understand that if you want some thickness that is NOT in the NACA catalog, you have to do some mathematical gymnastics. You can't just multiply the Y values by some factor.

Which of these two are being discussed here? I took the OP to mean my first statement.
Yes, it started out as a simple question about plotting airfoils from coordinates and was answered early on. It has now drifted to the accuracy of various methods of plotting and refining those airfoils. All methods introduce there own idiosyncratic errors, including the NACA's own recommended hand drafting slope and radius method.
 
Last edited:

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
I wrote a Python tool to generate NACA airfoils, and was frustrated when checking it against Xfoil — until I realized that Xfoil does not generate NACA airfoils correctly. It applies the thickness perpendicular to the chord line instead of the mean line. I didn't know the internals of Xfoil well enough or know enough Fortran to fix it, and never got a response to the bug report I sent.
There's a Python API for xflr5 under development. Fortran is out of the picture. XFoil was written in Fortran77 but it was ported to C for xflr5. Since development of xflr5 ended a few years ago this looks like the only way new features can be added. Sadly it looks like it only runs on Linux for now but that's how xflr5 started too.
 

davidjgall

Well-Known Member
Supporting Member
Joined
Oct 7, 2010
Messages
163
Location
Florida, USA
...if a spec says the wing is NACA 2414 at the root, and NACA 2410 at the tip, interpolating 14 -> 10 along the wing length does not do what you want.
Lofting such a wing for flat-wrap material fabrication, e.g., sheet metal construction, by the usual method of connecting points of like slope will give adequate results that are manufacturable while still being close enough to the intended airfoil shape.
 

BJC

Well-Known Member
Supporting Member
Joined
Oct 7, 2013
Messages
15,918
Location
97FL, Florida, USA

BJC

Well-Known Member
Supporting Member
Joined
Oct 7, 2013
Messages
15,918
Location
97FL, Florida, USA
One isolated, theoretical, data point:

Some years ago, I asked a homebuilder and aero engineer, who had access to Boeing’s aero computer applications, the following question about the Glasair III wing (LS(1)-0413 airfoil): What would be the top speed impact of a flat spot about 8” long with the center of the flat spot 1/8” below the ideal profile, on the upper surface of the wing, and tapering to about 4” (the width of the spar cap) at the tip.

The computer analysis predicted a loss of 3 to 5 MPH from about 275 MPH.


BJC
 

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
Lofting such a wing for flat-wrap material fabrication, e.g., sheet metal construction, by the usual method of connecting points of like slope will give adequate results that are manufacturable while still being close enough to the intended airfoil shape.
Are you sure about that? Looks like it would require some stretching to force lines drawn on the sheet metal to hit points of equal slope as Raymer defines "flat-wrap". To avoid stretching just cut a trapezoid with the ends equal to the wrap length of the airfoils and attache it wherever it fits.

FlatWralStretch.jpg
 

Norman

Well-Known Member
Joined
Nov 28, 2003
Messages
3,313
Location
Grand Junction, Colorado
One isolated, theoretical, data point:

Some years ago, I asked a homebuilder and aero engineer, who had access to Boeing’s aero computer applications, the following question about the Glasair III wing (LS(1)-0413 airfoil): What would be the top speed impact of a flat spot about 8” long with the center of the flat spot 1/8” below the ideal profile, on the upper surface of the wing, and tapering to about 4” (the width of the spar cap) at the tip.

The computer analysis predicted a loss of 3 to 5 MPH from about 275 MPH.


BJC
Yep. Fluids don't like flat spots or corners.
 
Last edited:
Top