I have a zero at the origin of my root locus plot. I understand that I need integral control to have zero steady-state error. I have 2 questions about this. 1. When I add integral control like this, Kp (1/(Ti*s)) where Ti is from Z-N first method and Kp=1, I get a pole at the origin. I don't don't see any change in the root locus plot by changing Ti, but I do see change in the transient response. Why is this? 2. When I add the integral control and there's a zero at the origin, I have a CL pole at the origin. This means I'm not able to achieve my settling time requirements. Is there a way to achieve both when there's a zero at the origin?