Interest in the development of micro vortex generators (MVGs) to control high-speed flow separation has grown in the last decade. In contrast to conventional vortex generators, MVGs are fully submerged in the boundary layer and have the potential of inducing surface flow mixing with marginal drag penalty when suitably designed. Also, they do not result in undesired reduced mass flow such as with suction methods. The flow mechanisms at the location of MVGs are not yet fully understood, and optimal designs are difficult to establish given that both numerical predictions and experiments are particularly challenged for short element heights, yet optimal MVGs are generally expected to be at least shorter than half the local boundary layer thickness. The present work aims at investigating experimentally the fundamental flow physics concerning an individual MVG element (of ‘canonical’ or simplified geometry) at a range of near-wall heights. A fully laminar base flow is considered so as to isolate the effect of incoming turbulence as well as the more complex physics that may occur when specific and/or multiple elements are used. Tests were performed in a gun tunnel at a freestream Mach number of 8.9 and Reynolds number of $$47.4\times 10^{6}$$ /m, and the basic test model consisted of a blunt-nosed cylinder which produced an axisymmetric laminar boundary layer with an edge Mach number of 3.4 and Reynolds number of $$3.2\times 10^{6}$$ /m at the MVG location. A laminar shock-wave/boundary layer interaction with separation was induced by a flare located further downstream on the model. Measurements consisted of time-resolved surface heat transfer obtained in the axial direction immediately downstream of the MVG and along the interaction, together with simultaneous high-speed schlieren imaging. The height ( $$h$$ ) of the MVG element used in a ‘diamond’ configuration (square planform with one vertex facing the flow) was adjusted between tests ranging from $$h/\delta $$ = 0.03 to 0.58, where the local undisturbed boundary layer thickness was $$\delta $$ = 1.75 mm. The effect of planform geometry was further assessed by performing tests with the MVG used in a ‘square’ configuration (one edge normal to the incoming flow). Results show that MVG height drives the intensification of heat transfer fluctuations in the wake of the element. The optimal MVG operating conditions, where downstream boundary layer separation is avoided and minimal flow interference is produced, based here on heat transfer unsteadiness, are $$h/\delta \approx 0.3$$ with a ‘diamond’ arrangement.