About 15-40% of all interactions in the cell are estimated to be peptide-protein interactions and consequently designing a scaffold able to interact with a target peptide through a conserved binding mode has potentially many applications in the field of research, diagnostic and therapeutics. The Armadillo project aims to generate a modular binding technology based on natural Armadillo repeat proteins (nArmRPs). The idea is to build a repetitive protein based on modules able to bind a specific dipeptide each, thus simplifying the problem of generating a binder for every single possible peptide combination to just generate modules binding all the possible combinations of a dipeptide. This thesis was generally devoted to develop and apply new approaches to study designed Armadillo repeat proteins (dArmRPs) in solution by NMR. First, an automated iterative approach to refine structure of dArmRPs based on pseudocontact shifts (PCSs) was developed, thoroughly tested and applied to obtain the first ever dArmRPs solution structures. Next, the obtained structures were characterized with a particular focus on the supercoil, which was the structural aspect most altered in crystal structures while also representing a key design feature, and complemented by detailed investigations of binding dynamics. Finally, the N-terminal cap was redesigned from its yeast 3rd generation sequence to a new artificial 4th generation model to overcome stability issues documented during the previous analysis. In conclusion, this work contributes to the progress of the Armadillo project by supplying structural knowledge and binding dynamics of dArmRPs in their native soluble environment to the already existing X-ray-based information.