The thermal efficiency of an MGT (micro gas turbine) system at Stellenbosch University was intended to be improved with the incorporation of a locally developed recuperator. A clamped plate, counter-flow design was chosen due to its simplicity. An effectiveness of 0.92 was desired and formed the design constraint on which to optimise the recuperator geometry. From this geometry a detailed design was performed. The recuperator was assembled, attached to the system and tested. To compare and validate the experimental results of the standard and recuperated MGT system, a Flownex model was developed. This large effectiveness value was not achievable, both practically and financially. An effectiveness of 0.6 was designed for instead. During experimentation, the MGT system could not self-sustain due to the additional pressure drops incurred by the recuperator, which were attributed to flow-maldistribution within the headers. A decoupled mode of testing was performed to evaluate effectiveness and cold-flow pressure drop of the recuperator. Close correlation of the experimental results was achieved analytically and with Flownex. Partial recuperation, with a bypass on the hot and cold-side, was modelled in Flownex and revealed that the system could self-sustain in this manner. An improved recuperator design is therefore proposed.