This paper presents an experimental method for tensile testing of unidirectional carbon fibre composites. It uses a novel combination of a new specimen geometry, protective layer, and a robust data analysis method. The experiments were designed to test and analyze unprotected (with conventional end-tabs) and protected (with continuous end-tabs) carbon fibre composite specimens with three different specimen geometries (straight-sided, butterfly, and X-butterfly). Initial stiffness and strain to failure were determined from second-order polynomial fitted stress-strain curves. A good agreement between back-calculated and measured stress-strain curves is found, on both composite and fibre level. For unprotected carbon composites, the effect of changing specimen geometry from straight-sided to X-butterfly was an increase in strain to failure from 1.31 to 1.44%. The effect of protection on X-butterfly specimens was an increase in strain to failure from 1.44 to 1.53%. For protected X-butterfly specimens, the combined effect of geometry and protection led to a significant improvement in strain to failure of 17% compared to unprotected straight-sided specimens. The observed increasing trend in the measured strain to failure, by changing specimen geometry and protection, suggests that the actual strain to failure of unidirectional carbon composites is getting closer to be realized.
Keywords: carbon fibre composite; static tensile testing; strain to failure; stress–strain curve; test specimen.