The apparatus used to study the buoyancy-driven convection in a horizontal fluid layer is shown schematically in Figure 5.12. The cavity is 447 mm long with a square cross-section of edge 32 mm. The test cell consists of three sections namely the top plate, the fluid layer enclosed in a cavity and the bottom plate. The top and bottom walls of the cavity were made of 3 mm thick aluminum plates. The flatness of these plates was manufacturer-specified to be within and was further improved during the fabrication of the apparatus. The central portion of the experimental apparatus is the test section containing the fluid medium. The side walls of the cavity were made of a 10 mm Plexiglas sheet. In turn the Plexiglas sheet was tightly wrapped with thick bakelite padding in order to insulate the test section with respect to the atmosphere. The height of the test section was 32 mm and was measured to be uniform to within . A window was provided in the direction of propagation of the laser beam. It was held parallel to the longest dimension of the cavity for recording the projected convective field in the form of two-dimensional images. The apparatus was enclosed in a larger chamber made of thermocole to minimize the influence of external temperature variations.  The room temperature during the experiments was a constant to better than over a 10-12 hour period. Experiments were initiated by flowing water over the hot and cold surfaces respectively. After the start of the experiments, the surfaces reached a steady state within a few minutes. The thermal fields in air evolved over a longer period of time and stabilized over 5-6 hours.  In water, dynamically steady patterns were realized in 1-2 hours. The thermally active surfaces were maintained at uniform temperatures by circulating a large volume of water over them from constant temperature baths.  Temperature control of the baths was rated as at the cavity location, direct measurements with a multi-channel temperature recorder showed a spatial variation of less than . For the upper plate, a tank-like construction enabled extended contact between the flowing water and the aluminum surface. Special arrangements were required to maintain good contact between water and the lower surface of the plate. Aluminum baffles introduced a tortuous path to flow, thus increasing the effective interfacial contact area.

For air, the temperature differences applied across the cavity walls were in the range of 5-50 K. These correspond to Rayleigh numbers of . In experiments with water, the temperature differences applied were in the range of 3-13 K. These correspond to Rayleigh numbers of respectively. The Rayleigh numbers realized in the experiments can be seen to be quite large in comparison to the critical value of 1707, indicating the presence of strong fluid motion in the cavity.