This paper presents a delay-type PID controller (shortened as dPID), or a PID controller implemented using two delay elements. The gains are analytically tuned; one of the delays is optimally tuned and the other is chosen to be equal to the sampling time. The poles of the discrete version of the dPID lies inside the unit circle and far away from the point (–1,0). Hence, it is stable and ripple-free. The controller is then applied to a first-order model plus dead-time (FOPDT), which can represent the majority of typical industrial processes; it achieves much better performances than the classical PID controller.