Carbon fibre materials become more and more important in many applications. Unlike to metal the technological parameters and certificated quality control mechanisms are not developed at present. There is no efficient and reliable testing system for an inline inspection and a consecutively manual inspection of the Raw Carbon Fibre materials (RCF) and the post laminated Carbon Fibre Reinforced Plastics (CFRP). Hidden defects during the production can cause far-reaching consequences with high-risk about material fatigue and even higher costs due to scrap coming out after machining. Based upon the multi-frequency eddy-current device developed at the Fraunhofer IZFP it is possible to detect structural an hidden defects e.g. missing carbon fibre bundles, lanes, suspensions, fringes, missing sewing threads and angle errors. With the help of an optimized sensor array and an intelligent image pre-processing algorithm it is possible to allocate the complex impedance signal to different carbon fibre layers. This technique enables the possibility to detect defects in the depth of 5 layers with the option of free scale resolution and measure frequency. Appropriate parameter lists for an optimal error classification are available. The sizes of the smallest detectable defects are in the range of few millimetres. A prototype of a special single sensor and even an eddy-current sensor array are developed and establish the chance to move the prototype into an industrial application.