I am working on actuator load-stroke control, to control force applied to actuator while test actuator extends and retracts. We have a separate load actuator with Star servo valve, load cell linking the two actuators and position feedback on the test actuator.
The test actuator position is NOT being controlled, only the force applied during movement. The test actuator is cycled using a solenoid valve.
There are several different load-stroke curves for different cases, and the curves differ when test actuator is extending vs. retracting. I think I should be using Curve Star - Advanced to treat position as a “master” with y-axis interpolation is command for force control. It is not clear how I should specify parameters for this command, or if I am using the correct command/approach. Please advise.
Equipment;
RMC75E-AA1
VC2124 V->I to drive servovalve
EXP70-D8
EXP70-A2…added to read position signal
Yes, make a curve with the x axis as the position and the y axis as the desired force values.
For the Curve Start Advanced command, here is how to set the command parameters:
Curve ID: the ID of the curve you made for this.
Master Reg: the Actual Position of the test actuator.
Cycles: 1
Options: I suggest 6, which means absolute x axis, absolute y axis, (absolute means you are using the curve values as they are) and Truncate, which means if the Actual Position goes beyond the defined range, the force will be what it was at that end of the curve.
Leave the scale and offset parameters at their default, meaning you are using the curve values as they are:
Curve Scale: 1
Curve Offset: 0
Master Scale: 1
Master Offset: 0
Status Block: optional - you probably don’t need it.
You indicate that the load cylinder will only be in force control. One thing I highly suggest is adding a velocity feed forward into the force control loop. This is because the force control varies significantly with velocity. See this topic for details: Velocity Feed Forward in Force Control for Testing Cylinders
Thank you Jacob. I will work on adding the velocity feed forward, although for this test the velocities are very slow. (approx. .5 in/sec.) As we will be moving to very high velocities in the future I will learn the technique and test out on our slow example.