First, the Performance Numerical Model, developed and used by CNES for the engineering studies, had to be validated and calibrated by comparison with real measurements. This which was done using breadboard simulations and measurements, did allow to declare the PNM relevant for the instrument itself. Besides, every time a design choice had to be made, the numerical simulations were checked by experiments. By instance, the basic choice of splitting the beam splitter was made when proved acceptable on the breadboard. An important task was to acquire data to test the instrument algorithms. These are, for example : the complex radiometric calibration, the Zero Path Difference detection, the misalignement retrieval, the Instrument Spectral Response Function (also called Instrument Line Shape) model. An other work was to find and to test a process for the fine alignment of the interferometer. The breadboard helped it to validate the process that was used with the instrument during the industrial AIT phases. Finally several scientific campaign have been organised to provide the scientists with atmospheric data. These are used to run the scientific algorithms, such as the radiative Transfer model, the temperature and humidity profile retrieval model, and to see the effects of natural phenomenons as clouds, temperature inversion and so on. During these campaigns, Météo-France launched some sounding balloons to acquire real measurements of the atmosphere. Comparison with the breadboard data are a key point for the IASI scientific validation.