46

JUNE 2005

SCOTT SPANGLER

The Secrets of

Wood Designs Building with wood means you must leverage its strength and accommodate its weaknesses Neal Willford

P

art one of this article looked at the structural characteristics of wood for aircraft use. This month we continue the discussion by looking at the practical application of wood to aircraft design details.

There is a new spreadsheet available to download from the EAA Sport Aviation page on the EAA website at www.eaa.org that will be helpful for those who are interested in designing wood aircraft. Despite the growth of composites and popularity of aluminum, wood is an excellent choice for new airplane designs. It is helpful to study the structural details of wooden airplanes that have stood the test of time to see how their designers solved various structural problems. EAA has published the plans to a couple of wood aircraft, including the Bowers’ Fly Baby (in 1963) and the Turner T-40 (in 1965). Copies of these articles can be obtained from the er Manual, which contains the plans for the Pietenpol Aircamper. In this article, we will only focus on a few of the major structural details.

SCOTT SPANGLER

EAA library. EAA also sells reprints of the 1932 Flying and Glid-

EAA Sport Aviation

47

Rib Design

planes with dive speeds below 150 rib design where the diagonal piecRegardless of its material, the job of miles per hour. Airplanes with high- es were in compression under the the wing skin is to transmit the air er dive speeds required a reduced rib worst flight loads (as shown in Figloads to the wing ribs. The chord- spacing that can be calculated using ure 1). This is because the diagonal wise air load distribution tends to be Equation 1. pieces were still pushing into the triangular, especially at high angles Though the requirement was for joints when the glue joints started of attack, as shown in Figure 1. This fabric-covered wings, these spacing failing. Applying the chord load disdistribution can be estimated for the guidelines can also be used as a tribution as shown in Figure 1 to different loading conditions the rib would try to bend the The face grain [should] be using the spreadsheet from cantilever portions of the rib Reference 1. oriented vertically for solid ahead of the forward spar and The ribs in turn transaft of the rear spar upward. mit the resulting loads to plywood ribs and horizontally The diagonal pieces in these the wing spar(s). Certificaportions of the rib act like the tion requirements back in when lightening holes are used. struts on a low-wing airplane the 1930s required that ribs and are in compression. be designed for an additional safe- starting point for plywood-skinned The diagonal pieces between the ty factor of 1.25 for fabric-covered wings. spars are also in compression only wings and 1.2 for plywood- or metComprehensive testing in 1930 if the load is applied to the botal-skinned wings. They also defined found that the truss rib was the tom of the rib. Most of the air loads a maximum rib spacing of 15 inch- most efficient design (Reference 2). typically act on the upper skin, but es for fabric-covered wings in air- Specifically, the best was a truss when rib stitching passes around

Figure 1. Suggested rib truss arrangement based on NACA testing. Joint gussets not shown for clarity.

Figure 2. Box spar showing the neutral axis location.

Table 1. Ultimate strength (lbs./ sq. in.) for aircraft grade wood.

Sitka Spruce

Douglas Fir

9400

11500

Tensile Stress (perpendicular to grain)

165

145

Compressive Stress (parallel to grain)

5000

7000

Compressive Stress (perpendicular to grain)

840

1300

Shear Stress (parallel to grain)

850

920

1300000

1700000

Tensile Stress (parallel to grain)

Modulus of Elasticity 48

JUNE 2005

Equations:

Dive Speed(mph) 17.5 Bending Moment xc 2. Bending Stress= Moment of Inertia Torsion Moment 3. Torsion Stress= 2 x Area x Skin Thickness 1. Max Rib Spacing (inches)=23.6-

the complete rib, then it is correct to apply the load to the lower surface. If the rib stitching (or whatever attach method is used) does not go completely around the rib, then 75 percent of the ultimate air load is applied to the upper rib and 50 percent to the lower. These two percentages add up to the additional 1.25 safety factor mentioned earlier, so be careful not to apply this twice. Looking at several popular homebuilt designs showed that the rib trusses were made from ¼-inch by ¼-inch or ¼-inch by ½-inch spruce, with most being the latter. These dimensions are a good starting point, but should be checked for sufficient strength for the actual loads in a new design. Plywood gussets 1/16-inch thick are typically used on the rib joints to increase the joint strength and stability. Wings that use single-piece ribs require filler strips between adjacent ribs to account for the rib cap thickness. These strips are not counted as spar height when sizing the spars. As we will see shortly, the spar’s height has a significant impact on its strength. Using ¼-inch rib caps would lower the spar’s height by ½ inch total, so it is not surprising that some designers use a three-piece rib that allows the spar height to extend to the skin. Development of the Beechcraft AT-10 indicated that this can cause skin cracking unless an additional plywood apron strip is used between the spar and skin (Reference 3). This strip was wider than the spar cap and also had tabs that extended a short distance onto each rib, helping reduce the abrupt change in EAA Sport Aviation

49

the stress level. This apron strip was made from thicker plywood than the wing skin, so the gain in spar

height was reduced accordingly. The AT-10 had a fairly high wing loading, which likely contributed to the need for the apron strip. Several successful wooden homebuilt designs have used a three-piece rib design without such a strip, so depending on the wing loading, your design may not need one. Plywood ribs have also been a popular choice among designers. Although they are the least efficient, construction simplicity is often a

bigger consideration. Two successful wood homebuilts, the Fly Baby and VP-1, both have plywood ribs. Reference 5 recommended the face grain be oriented vertically for solid plywood ribs and horizontally when lightening holes are used. Looking at the plans for both of these designs shows that their designers followed this recommendation.

Spar Design The wing spar is one of the most critical pieces of an airplane’s structure and needs to be carefully designed to ensure that it won’t fail under normal operation. Small general aviation airplanes with a stressed wing

skin usually have a single main spar that resists the shear and bending moment due to the air loads. This spar is often located near the maximum thickness point on the airfoil or about 25 to 30 percent back from the leading edge. The rear spar on these wings primarily serves as an attach point for the ailerons and flaps. Fabric-covered wings do not have skins capable of resisting the torsional loads, so this requires that the rear spar share in the load carrying burden. Reference 1’s spreadsheet will estimate the amount of shear and bending moment each spar will experience, which largely depends

Figure 3. Ratio of ultimate compressive stress to tensile stress for wood spars.

Figure 4. Forty-five degree plywood face grain orientation for maximum strength and stiffness (when up-bending and nose-down pitching moment are critical). 50

JUNE 2005

on the spar locations. Historically, the front spar for fabric-covered wings is located 12 to 17 percent chord from of the leading edge, and the rear spar 65 to 70 percent chord aft of the leading edge. A spar’s stress at different points along the span can be calculated using Equation 2. The bending moment is for the particular wing station being evaluated. The “c” term is the distance between the neutral axis and the outer edge of the spar cap. The neutral axis is the center of gravity of the spar’s crosssectional area and can be seen in Figure 2. Under positive “g” maneuvers, the upper part of the spar is in compression and the lower in tension. The dividing line between tension and compression is at the neutral axis. At this point there is no load along the axis of the spar. The inertia term in the equation is the moment of inertia of the spar cross section. It is approximately equal to the sum of the cap area times the square of the distance from the neutral axis to the center of gravity of the cap area. Since the distance is a squared term, it indicates we want to get the spar cap material as far away from the neutral axis as possible. The resulting stress must be less then the wood’s ultimate strength or the spar will break. Table 1 shows that wood is much weaker in compression. However, the ultimate allowable compressive stress also depends on the spar geometry as shown in Figure 3 (from Reference 4). The key parameters are the ratios of web to overall thickness and compression cap to total spar height. For a solid rectangular spar, the web thickness ratio is one, and the chart indicates that the ultimate allowable compressive stress is equal to the tensile stress. The other extreme is a box beam, where the web thickness ratio is zero. In this case the bottom curve of Figure 3 is used. For example, the upper ratio of cap thickness to spar height is 0.25

for the Figure 2 box spar. A Large Compression Load on is below the wing (like on The bottom curve of Figa high-wing design), posiure 3 indicates that the ulti- the Spar Due to the Strut Can tive-g maneuvers result in mate allowable compressive an additional compression Cause additional Problems. stress would be 65 percent load. The opposite occurs of the ultimate tensile stress when the wing experiences of the spar’s material. The curves in spar inboard of the strut attach a negative-g loading. between these two are for routered point will experience an additionIn either case, the additive load spars, which some designers use al compression or tension load. is equal to the horizontal compoto help reduce the weight. This is Whether this stress is tension or nent of the load in the strut and okay as long as the minimum web compression depends on the flight is shared equally by the upper and thickness isn’t less than ½ inch. condition and the strut location rel- lower caps. The end result is that The spreadsheet for this article will ative to the wing. When the strut inboard of the strut attach point, automatically calculate the allowable compressive stress based on the ANC-18 method. Box spars are more efficient because they concentrate the cap area at the outer edge. Plywood is used to tie the two caps together and resist the shear load. The face grain orientation of the plywood shear web plays a key role in the spar’s stiffness and ultimate strength. Orienting the grain at 45 degrees allows the spar to withstand much higher shear stresses. Plywood’s shear strength at 45 degrees is greater when the face sheets are in compression (as shown in Figure 1928 Stearman C-3B, 5). This is particularly true for birch restored by Poly-Fiber plywood with poplar core. It is also true for mahogany/poplar plywood up to ¼ inch thick. Plywood with 45-degree face grain is more expensive than sheets with the grain parallel to the length. A spar web’s shear strength is also Even Poly-Fiber, the ruggedest, easiest-to-use much lower if it is oriented with the face grain horizontal (or vertifabric coating system, won’t last forever! cal) and consequently may require a However, being pretty nearly foolproof and thicker web or two webs. No matter what orientation is used, the miniconsistently yielding spectacular results have mum recommended plywood spar a lot to do with why Poly-Fiber is still the web thickness is 1/8 inch. Testing in Reference 5 indicated top-selling system out there. that the ultimate shear of a plywood spar web also depends on the spar stiffener spacing. This was particuwww.polyfiber.com larly true when the ratio of the stiffe-mail: [email protected] ener spacing to the spar web height exceeded 1.5, where the ultimate shear stress value started decreasCall for a free info pack! ing rapidly due to buckling failure of the web. If the wing is strut braced, the

800-362-3490

EAA Sport Aviation

51

the upper compression spar cap will get bigger and the lower tension cap smaller. A large compression load on the spar due to the strut can cause additional problems, though. This load, combined with the spar geometry, the bending moment outboard of the strut, and the air loads inboard can actually result in the bending moment inboard of the strut attach position increasing to a point where it becomes the critical design condition. This is called secondary bending and is more common on designs where the strut attachment is far out on the wing. This month’s spreadsheet can be used to check this special condition. When a wing spar is under a compression load due to the strut, it can act like a column and will buckle if it is not laterally stable. The leading edge and fabric covering provides some stability to the spars, but not as much as a fully skinned wing.

Wing Skins Plywood is an outstanding material for wing skins. The torsional shear stress in a wing with a “D” cell leading edge skin can be calculated using Equation 3. The area in this equa-

Plywood is an outstanding material for wing skins. tion is that bounded by the leading edge skin and main spar. This kind of wing often has an angled rib back to the rear spar that, along with the root rib, has the plywood skin continuing back to the rear spar. The equations get more complicated when the wing is a “two cell” design where the skin goes back to the rear spar. This article’s spreadsheet will estimate the skin stress for this kind of design. Like the spar web, face grain orientation affects a wing’s torsional stiffness and shear stress. Figure 5 (from Reference 3) shows a wing skin’s ultimate shear stress as a function of rib spacing and face grain orientation. Figure 4 shows the desired wing skin face grain orientation for 45-degree plywood when a nose-down pitching moment is the critical torsion case.

Fuselages Boxy fuselages are really just big spars and can be analyzed in a similar fashion. Several popular designs have longerons and cross members

Figure 5. Maximum torsional shear stress for mahogany plywood skins. 52

JUNE 2005

that measure ¾ inch by ¾ inch or 1 inch by 1 inch. Some smaller homebuilts use even smaller ones. In the early days of aviation, the fuselage longerons were diagonally braced with wires. This tended to be a maintenance headache and also proved dangerous in a crash, where splintering longerons could skewer the occupants. The de Havilland Co. is commonly credited with developing the fuselage design that replaced the bracing wires with plywood skins. This resulted in a safer, maintenance-free design that has been universally adopted by wooden aircraft designers ever since. These skins are typically 3/32-inch to 1/8-inch thick on homebuilt designs. It is important to check the members for column buckling strength and design accordingly. It is important to remember that new designs should be structurally tested to ensure their strength. Stress calculations are important in sizing and optimizing an airplane’s structure, but there is nothing like actually testing to know that it is actually strong enough. References: “Estimating Air Loads,” Willford, Neal, EAA Sport Aviation, June 2003. NACA TR-345, “The Design of Airplane Wing Ribs,” Trayer, George, 1931. Available at http:// naca.larc.nasa.gov. “Wood vs. Metal Aircraft Construction in Aircraft,” Rawdon, Herb, SAE Transactions, December 1945. Design of Wood Aircraft Structures, ANC-18, 1944. NACA TR-344, “The Design of Plywood Webs for Airplane Wing Beams,” Trayer, George, 1931. Available at http://naca.larc.nasa. gov.



more at www.eaa.org