In the original 1922 Stern Gerlach experiment the single horizontally propagating incident beam was split into two ‘up’ or ‘down’ diverging beams. An observation not consistent with predictions of the time which were that the path deflection angles in a classical model should be deflected up or down in only an even range of angles. Here it is proposed that net translational forces on a dipole in an inhomogeneous field can correctly model the observed split paths for a classical model. In that the dipoles will initially experience a range of very small path deflections via the up or down net translational forces on them as they enter the apparatus. A deflection force dependent upon the specific angle of the N-S axis of polarity of each incident dipole relative to the applied external N-S field in the apparatus. This separation of the beam into 2 paths, one up and one down is effectively a classical version of the "space quantisation" often referred to in QT. After entering the field, the dipoles will then have been sorted into two up and down paths as well as each path having a range of these very small different angled path deflections from the horizontal incident path. They will then all each experience an additional amount of net translational forces applied equally on all aligned dipoles as they propagate through the 3.2 cm length of the external field. Separating the 2 up and down sets into two distinct paths.