This paper describes concepts for automatic stall/spin detection using in-flight measurements of air data and inertial parameters. Several approaches are investigated, requiring simple computations from sensed quantities, easily implementable on modern flight systems. The performance of alternative schemes are tested and compared on subscale spin flight data obtained from two different unmanned testbeds, as well as on previouslypublished spin flight data from a full-scale general aviation aircraft. The algorithms reliably detect spins early in their development, within a fraction of the first turn. This early detection could be used to trigger a pilot alarm or engage an automatic spin recovery controller, with potential to significantly reduce altitude loss, and the risk of ground collision. More testing on subscale and full-scale flight data will be done to further evaluate the performance of the detection schemes.