If there is a double integration in the plant under control -- for instance, from a torque/force command thru velocity to position -- you need derivative action to get the damping required for stability. Proportional gain here is just a spring (no shock absorber -- the car suspension analogy is a good one here).

If there is only a single integration in the plant under control -- for instance torque/force to velocity, or voltage command to current thru an inductive load -- you do not need derivative action. The system is much more inherently stable.

Curt Wilson
Omron Delta Tau

Many thanks Curt. You provided me with very useful information. where can I read more about these about these recommendations for PT, PID and PD.
Regards,

Read modern control engineering by ogata. it is a very useful book.

I have Ogata, but I can't find these information

If there is a double integration in the plant under control -- for instance, from a torque/force command thru velocity to position -- you need derivative action to get the damping required for stability. Proportional gain here is just a spring (no shock absorber -- the car suspension analogy is a good one here).

If there is only a single integration in the plant under control -- for instance torque/force to velocity, or voltage command to current thru an inductive load -- you do not need derivative action. The system is much more inherently stable.


I am wondering about guidance for when each type is used

Many textbooks talk about the "type" of the physical system. A "Type N" system has N integrations (poles of the system at the origin of the S-plane).