We start programming from the TEMPLATE_SENSOR.urp program provided by Robotiq in their drive package for the FT on UR.
We first have the robot go to the measuring position with no object and we zero the sensor value (Call SubP_rq_set_zero)
Then we simply pick the object, go back to measuring position and check the value. We use moments along with moment lever arms to figure out the weight because moments are more accurate measures than the Fx,Fy,Fz values and we convert to pounds (weight≔((Mx/.15)/9.81)*2.2)
The measurement were very repeatable. Anyone can put a number on the repeatability of this measurement setup, let's say when measuring the same object over and over again?
Here is the logic where we have the weight
MoveJ
Waypoint_4
Wait: 1.0
weight≔((Mx/.15)/9.81)*2.2
If weight≥1.0 and weight≤1.15
Popup
ElseIf weight>1.15
Popup
Waypoint_5
https://youtu.be/hzqWs8RrmuY
Hi pros,
we have been asked at our last seminar if we could use the FT300 to identify parts based on their weight. Below is a video showing a demo for this and attached is the program.
Here are some highlights.
- You want to make sure the sensor is calibrated such that you get good measurements. The procedure for this calibration is super easy and is described here in the sensor's manual.
- We start programming from the TEMPLATE_SENSOR.urp program provided by Robotiq in their drive package for the FT on UR.
- We first have the robot go to the measuring position with no object and we zero the sensor value (Call SubP_rq_set_zero)
- Then we simply pick the object, go back to measuring position and check the value. We use moments along with moment lever arms to figure out the weight because moments are more accurate measures than the Fx,Fy,Fz values and we convert to pounds (weight≔((Mx/.15)/9.81)*2.2)
The measurement were very repeatable. Anyone can put a number on the repeatability of this measurement setup, let's say when measuring the same object over and over again?Here is the logic where we have the weight