Identify Analog Input reaching steady state

[Moosetracks]

I have a analog laser distance sensor. It measures a pneumatic actuators position. I want to know when the actuator has stopped (its used as a clamp so I can't use a prox sensor because the position varies.

I have done this before using a timer. Each time the timer is done I compare the last value to current value and if the change is less than a set limit I call it stopped.

However sometimes the measurement update rates are not exact or the cylinder is moving a larger load and all the other real world variables get in the way. I am about to add a FIFO so I can look at several past measurements to get better performance, but I was curious how would you guys code this to decide when its actually stopped.

 

[keithkyll]

Setup a counter to 5 or so. Compare current read to previous read. If in range, increment counter. If false, reset counter.

When counter reaches preset, you are stopped.

 

[AustralIan]

Reached steady state is usually max(FIFO) - min(FIFO) < Tolerance
Length of FIFO buffer is the time it needs to be in tolerance for.

I have been using the CompactLogix 5370 lately and they have a moving standard deviation function built in which could be a nice easy one to slap in if you feel others that need to look at the code will find function block or structured text ok.

 

[GTUnit]

The RSLogix ALMA instruction has a Rate Of Change alarm function for analog signals.
Maybe that can be useful.

 

[Mill_Control]

If the actuator is double acting and speed controlled pressure might be a better option. When the pressure on the exhaust side of the cylinder has reached atmospheric it has stopped moving unless your cylinder is marginal for the "larger loads". And if the cylinder is marginal whatever you do will work a whole lot better if the cylinder is upsized.