Sounds like PITZ business :) First thing to check would be the phase. Is it MMMG for both? Mare sure MMMG, switch off space charge, track a single particle through and then compare the orbits (or with two particles for comparing the sizes). Maybe some clues already from the orbit. Besides, one could refine mesh lines for the field map imported to Astra. Or simply run a particle tracking job in CST tracking solver with the CST fields you got already. Generally, the paraxials should represent 3D fields if e.g. no asymmetry-induced kicks. Cheers.