This paper considers an adaptive minimal observer synthesis algorithm based upon a well-known model reference adaptive system strategy. This algorithm gives state and parameter estimation from output feedback alone for single input single output processes and results are presented which illustrate its excellent convergence properties.