I'd argue exactly the opposite. Why running three cases with three different grids, and different values for the BCs, and where it's more difficult to interpret results when you can non-dimensionalize ("homogenize") your problem into one single case run three times just with different Reynolds Numbers.... Anyway, going back to your question, the length of your recirculation region is determined not only by the height of the obstacle, but also by its width (Martinuzzi & Tropea, 1993), Is this a 2D simulation (as opposed to an infinitely wide obstacle)? The second thing you need to have in mind is that shedding vortices from the 'roof' of your obstacle are not going to be resolved in SIMPLE, because they are "shed" (function of time). The shedding of vortices in the roof, when the obstacle is "small enough", opaque the lateral "karman vortices" which shorten the recirculation region...