The controlled growth of atmospheric snow crystals presently resists the in situ tomographic approach due to several reasons: 1) the difficult temperature and humidity regulation, 2) the rapid phenomena involved, 3) the tiny structure of some crystals, and 4) the practical need of pre-selecting a unique crystal among several growing seeds. These 4 items are major obstacles to access the evolution of such structures in 3D. We found solutions to most of these difficulties and could recently obtain a short series of crystal growth using our cold cell CellDyM3 on BM18 beamline. Based on this preliminary result, we propose to grow at least 2 very distinct types of crystals using CellDyM3 and to monitor their growth using time-lapse tomography. Such an experiment would allow to access a lot of geometrical parameters that cannot be commonly accessed with 2D optical measurements, and to provide reference growth evolutions for the evaluation of 3D ice crystal growth models.