Several large-scale wall specimens were recently tested nnder dynamic conditions on the shakingtable at LNEC (Lisbon). The main objective of this experimental programme was to assess the effectiveness of selective interveotion techniques for repair and streogthening of reinforced concrete walls. The models were tested up to collapse using both artificial and natwal accelerograms. In this paper, a detailcd description of the set-up of the experiments is presented. Different test-rig configurations are discussed, highlighting the advantages and drawbacks of each solution, and justifying the selection of the scheme adopted. Scaling issues are also analysed in detail and their implication on choice of materials, test set-up and choice of input motion is discussed. This is followed by a summary of the procedure employed in the design of the models together with an overview of the model construction process and test sequence.