Motion simulator software control

[curt5101]

Hello Everyone. Thank you for reading this.

I have a made a motion simulator pneumatically controlled. It is a bowl held by a stationary cradle with air cylinders tilting the bowl left,right, forwards and backwards. The software to control the cylinders is giving me fits. I am treating it with a simple linear function, ie;

if distance from desired position to actual position >5 give cylinders some air
if distance from desired position to actual position >50 give cylinders more air, etc.

I get a lot of oscillations and overshoots. The air is compressible and I am moving weight either up or down a curve. (A model might be a pendulum with very high friction and the weight is a heavy sponge.) I think I must take into account starting position, inertia, lag, speed, etc. in a very high function equation. Any suggestions for me would be so appreciated, I do not know where to turn to learn more for this fun practical application.

 

[Nidum]

Feedback . PID control .

Post a photograph of the set up .

 

[curt5101]

Will do when I get home. Pretty basic setup;string pot on air cylinders controlled by microcontroller and proportional valve.

 

[JBA]

You may need to add more increments than just the two feed rates with a very small amount of feed as the bowl approaches the desired position and, a certain amount of deadband space (allowable out of position error band) is going to be required at the "desired position"; otherwise, the system will not be able to establish a stable point. If speed is important use a form of geometric progression on the feed rates as the distance from the desired position increases.