It seems, that in atom-style variable formula
variable EFx atom ${Field}*sin(v_ksi)
sin does not work as it should and gives 0
When I changed to
variable EFx atom {Field}*cos({Pi}/2.-v_ksi)
everything was OK.
(intel MPI fortran compiler)
variable time equal (step)*{NormalTimestep}
variable ksi atom -{k0}*cos({phi})*x-{k0}*sin({phi})*z-{omega}*v_time+{k0}*{r0}