To validate the experimental setting adjustments and to detect geometrical errors as sample misalignment (translation and rotation defects) and the missetting of the incident beam, some external stressed standard samples have been realized. A fine analysis of the raw robin-test results by a statistical analysis, taking into account the uncertainty of each measurement is then performed to understand the critical discrepancy of the results between the laboratories. It is shown that the most relevant uncertainty origin is induced by the inter-laboratory errors and an improvement of the stress value accuracy can be made by the standardization of the diffraction treatment peak procedure. This work is sponsored by Peugeot S.A., RENAULT S.A. and SNECMA.