The challenge of inducing stable and robust nonreciprocal transmission of sound is seen as an important milestone on the path towards developing active acoustic metamaterials. Structures that can transmit sound or vibration in a nonreciprocal manner may be considered analogous to an electrical diode and could be useful in many applications, such as noise control and the development of invisible acoustic sensors and acoustic cloaking. This paper focuses on conceptual development and experimental validation of an active metamaterial cell that does not obey the reciprocity principle. The structural cell considered, when activated, significantly attenuates vibration transmission through it in one direction and increases it in the opposite direction. The effect is present in a broad frequency band. The loss of reciprocity is induced by using two concurrent velocity feedback loops with non-collocated sensor-actuator pairs. Inertial accelerometers with time-integrated outputs are used in conjunction with miniature electrodynamic force actuators. The study is first carried out theoretically, using an electromechanically fully coupled lumped parameter model of an otherwise flexible active structure. The derived model is used to conduct analysis of the control system stability and performance. Given that a non-collocated transducer arrangement is considered, special attention is paid to the selection of parameters of the passive system which ensure satisfactory gain margins when the system is made active. In fact, criteria for unconditional stability are derived analytically, in terms of two simple inequalities for the system with idealised sensor-actuator pairs. For realistic transducers, however, unconditional stability is not possible. Nevertheless, if the two inequalities are respected, useful gain margins of the active system can be expected. This is validated experimentally using a dedicated 3D-printed measurement test rig. A substantial reduction of vibration transmission in the desired direction, accompanied by an increase in the opposite one, is recorded experimentally. These results suggest that the control scheme proposed could be used to design an active acoustic metamaterial, which would enable the significantly different transmission of sound depending on the direction it enters such a system.