We would like to briefly describe the basics of PID tuning as applicable to the Rhino Servo Drive RMCS-2301.
To begin with , we are aware that control loop systems which are also known as feedback systems use different feedback mechanisms for the same. However, the PID control mechanism due to its flexibility and reliability is the most widely used control mechanism. In fact, 95% of the industrial close loop systems use PID controllers to achieve the set values.
For instance, when we set a particular temperature for the air conditioner, the room temperature drops and achieves the set value. This happens due to close loop control mechanism in the air conditioner.
PID consists of 3 coefficients Proportional–Integral–Derivative control. PID control provides the proportional, derivative and integral response of the system to produce the optimal output. We shall discuss about the settings of PID controller in reference to RMCS-2301 tuning.
Let us discuss about the response of the individual controller and the procedure for manually tuning of the RMCS-2301 drive
The response of proportional controllers depends only on the error between the set point and process variable. The response of proportional controller output is the product of gain and measured error ε. Therefore, when the error is zero the controller output is zero. Increasing the proportional gain increases the response of the system but too high proportional gain will further increase the oscillation and makes the system unstable. Otherwise, it provides stable operation but always maintains the steady state error.
For example, when you decrease the set point (temperature) of air conditioner the temperature of the room drops immediately and then achieves a stable point after some time, this happens due to proportional controller. You can see the response of proportional controller in the graph in below image. You can observe from the graph that initially the system will overshoot and then settle down gradually to a set point with a steady error.
Proportional Controller Response
Hence, while setting the P gain, the initial output will be in the form of oscillations and these oscillations will gradually decrease to achieve a steady value. While tuning the RMCS-2301 drive, we would want to set the P gain to achieve a steady error in the drive (red light becomes stable and not blinking)
To enable the system response to be faster and more accurate, the Integral and Derivative Controller settings are done.
The integral controller integrates the error over time until the error becomes zero. The p-controller will always result in some steady state error and hence the I gain is needed. Integral response effect is to eliminate the steady state error. It means the integral controller provides force to eliminate the error and if the applied force is not enough to make error zero the magnitude of the force is increased over time. Integral controller can’t be used alone as it slows down the system. Larger value of integral gain makes the process sluggish.
Let’s look at previous example, you can observe in the graph the system oscillates before reaching the set point when p-controller is used. Now, when integral gain is used the steady state error is reduced and system will try to achieve the set value rather than settle at the output with steady error
Proportional and Intergral Controller Response
While tuning the RMCS-2301 drive, the I gain has be kept around 10% initially and then adjusted in conjunction with the D gain to achieve the set output (the red light switches off). For tuning of RMCS-2301, I gain gives the effect of providing a delay in achieving the set point. So a higher I gain would result in a higher time delay to achieve the set point.
The derivative controller is directly proportional to the rate of change of error with respect to time. Increasing the derivative gain will increase the speed of response of the system. Thus removing the sluggish effect caused by integral gain. The derivative gain should not be kept very high because it is sensitive to the noise in the process variable signal. Noise is unwanted error in the system.
It is evident from the response graph of the PD controller that the system will reach the set point faster compared to any of the above cases.
Derivative and Proportional Controller Response
Combinations of all the three controls will not only make the process precise and accurate but will also improve the response time of the process. You can observe in the PID response graph that the oscillation has been eliminated and the system reaches the set point quickly.
PID Controller Response
With respect to tuning the RMCS-2301 drive, increasing the value of the D gain , provides a damping effect to the oscillations generated due to the P gain and the set point is achieved quicker and without any steady state error. However the value should not be kept high to avoid over compensation which results in overshooting the set value.
Tuning the PID controller on RMCS-2301
So to conclude, as you can see in the above PID Controller response, each of the three gains need to be tuned to achieve a perfect response. There are various methods for tuning of the PID controller, however we shall be using the trial and error method for the same.
We need to set the P value to about 40% of the maximum value and I and D gain to approx 10% of the maximum value at the start. We have provided a chart for approximate value of the three resistors of PID gain to be set for different model of motors available on Robokits India website to provide a starting point
The perfect tuning will be done once the red light stops glowing during the running of the motor and the motor is not having any vibrations or jerks.
However for higher speeds, it would be difficult to perfectly tune the motor manually and hence the red light would still be glowing when the motor is running. If not possible to perfectly tune the motor it is better to have a steady state error, hence the red light should be constantly glowing rather than switching on and off for a smooth motion of the motor.