Impact testing

Test pieces help to learn and get used to the production process, and each new one gets better and better. However, a test piece is only half worth when not break-tested or at least loaded within its boundaries. Having invested a couple of hours of work and getting in love of that neat piece, it takes some dicipline to go further and break it. But hey, don’t forget why you built it in the first place: you wanted to gain information on its strength and stiffness. So, today was the right time to break that nice aramid honeycomb test piece I produced last week!

I started with loading the piece within its boundaries:


I measured the deflection to roughly estimate Young’s modulus E and the flexural rigidity D, important factors in designing shells/plates. The load was increased until the deflection in the center was roughly 10 mm. It looks already quite intimidating, though the theoretical maximum stress at the skin was far from being reached:



The behaviour corresponds very well to what is expected from theory, namely a linear relationship between deflection and load:


The steepness is related to the stiffness of the plate, where steeper means less stiffer. Because the skin layout is not symmetic, the stiffness depends on which side the load is applied. This behaviour was already mentioned by Herbert Funke in his dissertation (see PDF at R&G’s wiki).

Based on the usual beam theory and a second order of innertia of I = 805 mm⁴, I estimated the modulus of the skin to be E = 10.8 GPa . Herbert Funke’s assumed value for glass fibers was E = 15 GPa . This means that the usage of flax reduced somewhat the modulus. However, some of the discrepancy might be attributed to not considering all effects, such as shear deformation and Poisson’s contraction.

After measuring the stiffness, I continued with an impact test. I expect flax fibers to add some energy absorption and to keep the other fibers together in case of rupture. Well at least that’s how a manufacturer of flax fiber fabrics promotes their products. You can achieve similar properties by using aramid, but its handling is difficult. The fibers are extremely energy absorving and difficult to cut/sand. The charm of flax is that it is renewable and very easy in handling, even easier than glass or carbon fabrics. My idea is to use flax to increase skin thickness and add some damping, and in very critical sections to add aramid for some extra protection of the pilot.

Following video resumes the impact test:

How-to use honeycomb aramid cores

Last months have been quite busy. A pity it hasn’t been on building Schneewittchen… Anyway, we started to finish the mold by sanding up to 3000 grit and polishing the surface. Soon we’ll be able to build the first parts and it is time to find out how to produce the shell of the center section.

A couple of months ago I puchased about 9 m² (100 sq ft) of 5×3.2 mm aerospace grade aramid honeycomb core material. Usually this material  runs as Nomex or Kevlar honeycomb. It’s essentially aramid paper shaped in a honeycomb pattern and soaked in phenolic resin. Manufacturers produce large cubes from which thin sheets are cut off perpendicularly to the honeycomb channels. The usual sheet thickness is a couple of millimeters, where 3, 5 and 8 mm are usual thicknesses in airplanes. The amount I bought should be enough  for the center section including some tests pieces.

If you aim at building ultra light, using this core material is a must. It weights almost nothing (29 kg/m³) and is very strong. However, its handling and application is not easy and needs to be learned beforehand. If done wrong, you either end up with a heavy or fragile piece. As an appetizer, I started with a small flat piece of about 40×20 cm (16×8″):


A crucial point in using honeycomb cores is to achieve a very good bondage to the skin. Else the skin might peel off causing failure at high shear stresses. This is why usually the skin laminate is let to gel and glued to the core with a freshly wetted thin coupling layer. Vacuum pressing ensures then that the core is tighly pressed into the coupling layer resulting in good bondage. On the one hand, telegraphing might occur when the skin is still too soft, which leads to a weaker sandwich. While on the other, fully cured skins, do not chemically bond resulting in  weaker joints. You want to avoid both. It’s all about the right timing and technique.

Roughly 80% of the sandwich weight is contributed by the skin (typical layup for 5 mm core: 2×105 g/m² and 1×160 g/m² glass fabrics). There is no way to further reduce the weight of the core. Thus, optimization of weight is only possible by reduction of the skin weight. One approach is to use other materials, such as carbon. However, the available choices of fabric weights are limited and oversizing can happen resulting even in an increase of weight. Also, light carbon fabrics increase disproportionally in price: 93 g/m² is almost three times the price of 160 g/m² carbon. Another strategy, is to reduce the amount of resin in the skin to a minimum, i.e. to increase the fiber volume ratio as much as possible.

I chose to test the second strategy and used vacuum infusion for the outer skin and vacuum pressing for the inner skin. For both skins I used the low viscosity HP-E300RI resin designed for vacuum infusion. One problem was to catch the right time when the laminate is geled enough but not too hard (roughly 3 times the „pot life time“, i.e. about 11 hours). Instead of using a thin coupling layer to bond the core, I decided to use an industrial epoxy-adhesive: 3M Scotch-Weld 9323. Pricy (roughly 160 € / Litre) and adds some weight (200 g/m²), but I’ve heard it produces a much better joint that the coupling layer.

So, I started with vacuum infusing the outer skin. The laminate is composed of two quasi-isotropic layers of 105 g/m² aerospace grade glass fabrics and one layer of 100 g/m² flax fabric:


So why flax? Well because it is a renewable natural product, it has ideal damping properties and is quite strong. It adds also thickness without increasing that much weight. Take, for example, this impressive wind-turbine blade breaking test:


The laminate was vacuum infused using DD|compound’s MTI hose and was let to gel for roughly 8 hours (longer would have been better):


Then a thin layer of Scotch-Weld 9323 was gently spread on the geled laminate. The honeycomb core was vacuum pressed and cured overnight (see first picture on top ). The edges of the core were sanded with 80 grit to obtain a nice wedge. The inner skin—composed of 80 g/m² glass and 100 g/m² flax fabrics—was hand laminated and geled for 11 hours in a strong vacuum to press the excess resin. The laminate was still soft but not too sticky. Again a thin layer of Scotch-Weld 9323 was used to bond the core material and the sandwich was cured in a moderate vacuum overnight.

I’m quite happy about the result (1300 g/m²):