Kc=4 |
This is a bit oscillatory, but the thing to notice is that the
oscillations are initially centred underneath the setpoint -
integral action isn't pulling up the controlled variable quickly
enough
|
The response is oscillatory, but the oscillations are approximately
centred on the setpoint - integral action is probably OK, but the
gain needs to be reduced
|
This is clearly unstable. In cases like this I'd get rid of the
integral action and reduce the proportional gain. Once I was happy
I'd cut the gain a bit and gradually add integral action
|
Kc=2 |
There's little oscillation here, but the effects of too little
integral action are clear - it's taking ages to drag the controlled
variable to setpoint
|
Quite a nice response. A fast rise time and rapid settling. I'm
happy with this!
|
This kind of response can be difficult to distinguish from that
which we get when we have too much gain. The best thing to do to try
to improve it is to reduce the integral action - if we end up with
the sort of responses on the left hand side, then increase the
integral action (reduce the time constant) and decrease the gain
|
Kc=1 |
There's no power in the control action at all here - both
proportional and integral action are too weak
|
The proportional action isn't strong enough to carry the controlled
variable over the setpoint. If you tune the controller by sorting
out proportional control first, this shouldn't happen!
|
This looks OK, but it has a much longer rise time, and the overshoot
is slightly greater, than the properly tuned response
|