Schematic presentation of the numerical model. n_sav = 10 “collisional” runs are performed, each taking as a starting point one of the n_sav parent body (PB) disc configurations, corresponding to n_sav different positions of the companion star on its orbit separated by a constant time interval Δt_sav = t_orb/n_sav, obtained at the end (i.e., steady state) of the parent body runs. N_num = 2 × 10⁵ particles are released from the PB’s positions, following a size distribution dN ∝ s^qds down to the radiation-pressure blow-out size s_RP. Particles positions are recorded at time intervals separated by Δt_sav. Each simulation is run until all particles have been removed, either by dynamical encounters with the companion star or by collisions. The collected data is then used to reconstruct the dust distribution at dynamical and collisional steady state by means of the following procedure (indicated by the red arrows on the graph): the dust distribution at a given time t₀, corresponding to a given position i_φ of the companion star, is the combination of the dust released at t₀ for the i_φ run, plus the dust particles released at t₀ − Δt_sav for the i_φ − 1 run that have not been removed at t₀, plus the dust released at t₀ − 2Δt_sav for the i_φ − 2 run that has not been removed at t₀, etc. The procedure is iterated until we reach the j_final record (particles released at t₀ − j_finalΔt_sav), for which no particle has survived until t₀. Given that the binary orbit is divided into n_sav = 10 positions, all i_φ − 10 × j runs correspond to the i_φ one. (The different disc profiles displayed in the figure are not simulation results but only illustrative sketches.)