naturephoenix
Member
I wanna make own PID just to practise. So far I have implemented following stuffs:
1. Sample time is always fixed and it's determed by OB35 cycle.
2. Derivative of the Error is equal to negative derivative of Input - to avoid big changes in the output once we change Reference point.
3. Smooth transfer once I change the Ti.
4. High and Low limits - for integration part and as well on output.
5.Man/Auto -> Setting Output manually.
6.Initialization -> Always when I'm switching between Man/Auto -> initialize things for a smooth transition, to keep the output from jumping.
7. Deadbeand -> By not chaning only output value if the error is in some borders(limits)
My question is what else I should do implement to make my PID better, and how of cource.
Any advices are welcomed!
1. Sample time is always fixed and it's determed by OB35 cycle.
2. Derivative of the Error is equal to negative derivative of Input - to avoid big changes in the output once we change Reference point.
3. Smooth transfer once I change the Ti.
4. High and Low limits - for integration part and as well on output.
5.Man/Auto -> Setting Output manually.
6.Initialization -> Always when I'm switching between Man/Auto -> initialize things for a smooth transition, to keep the output from jumping.
7. Deadbeand -> By not chaning only output value if the error is in some borders(limits)
My question is what else I should do implement to make my PID better, and how of cource.
Any advices are welcomed!