On the physics of the Portevin-Le Châtelier effect. Part 2: from microscopic to macroscopic behaviour