Desmond,

in my 23 years of experience I don't think it is very necessary. I think professor Burnham will be the right person to answer it.

hope you find the answer before Wednesday.

 best regards,

sachit

I agree with Sachit Kadavil, type 1 system already have integrator which is essential for zero steady state error if the reference signal is constant or step, however for other reference signals or disturbances zero steady error cannot be achieved, because of internal model principle.  

adding an integrator may also lead to instability due to the delay/phase that is added to the system.

Prof. Michael short has sufficiently well developed  question and I fully agree with his reply

Thanks

Mourad

Desmond  i fully agree with Prof. Michael well explained answer. why don't you try with disturbance observer to overcome the steady state error and also try with proper Pseudo filter when you add derivative function to improve/restore the phase margin and speed up the closed loop response.

Thanks 

Branesh

yes, I agree with you because pole at origin ensures steady state error. Moreover, addition of an integrator may affect stability

No - not necessary: If there are no load disturbance then integrator is not necessary  for type 1 system since the inherent integrator will track setpoint in a well tuned controller. 

Yes-required: If there are disturbances then integral control is necessary for disturbance rejection.