Posts by Teddy

Hey thanks for answering.

I've done what you asked and got varying results in mostly the x and y coordinates. I started off with my part contact oriented in the longitudinal position where it was aligned with the 12 o'clock position and rotated it 45 degrees each time in the clockwise direction. My PR didn't change but I looked at the current position of the tool head/tool frame with respect to my user frame and got my numbers below.

0 degrees: Contact part was found and was able to be picked up
x: 20.291 mm
y: 23.900 mm
45 degrees: Found part but missed picking position when attempting to grab
x: 17.384 mm
y: 29.830 mm
90 degrees: Found part but missed picking position when attempting to grab
x: 19.442 mm
y: 34.124 mm
135 degrees: Found part but missed picking position when attempting to grab
x: 20.291 mm
y: 37.684 mm
180 degrees: Found part but missed picking position when attempting to grab
x: 21.183 mm
y: 36.648 mm
225 degrees: Found part but missed picking position when attempting to grab
x: 24.104 mm
y: 31.930 mm
270 degrees: Found part but missed picking position when attempting to grab
x: 24.539 mm
y: 27.277 mm
315 degrees: Found part but missed picking position when attempting to grab
x: 23.718 mm
y: 24.869 mm

Seems like my y-coordinates are really changing the most but I don't know what it correlates to.
Fanuc told me it could be with the camera calibration to that certain user frame that could be the issue due to robot bumping into it self.

So I did an automatic grid frame set using a Fanuc given 11.5 mm grid frame fixed on the feeder plate. This is my first time using it but I set up the automatic grid frame set with the settings below:
1. Robot Group Number to use: 1
2. Set UFrame or UTool: UFrame
3. User Frame Number to Set: 7 (Was set to a userframe named feeder2Cal)
4. Camera user Tool Number: 4 (Calibrated camera tool number)
5. Camera Name: RobotCam
6. Exposure Time: 8 ms
7. Grid Spacing 11.5 mm
8. Start position: (I recorded right above the feeder plate to get all the circles in)
9. R Angle: 10
10. W Angle: 5
11: P Angle: 5
12: Z height: 30 mm

My first run encountered problems cause it would lose sight of the big circles so I had to bring my camera start position high enough so that all the big circles will be in the picture while execution but it made my focus a lot worse and blurry. I got it to work and did a 2 plane calibration on my feeder plate as I have a robot mounted cam:
Application Frame: (I used my feeder plate frame)
Cal Grid Frame: (I used the one I just got from userframe 7)

I got huge errors for my points ranging from 0.3 to 12 percent errors...
Then I tried moving my camera back over my feeder plate but now the camera goes to the location but keeps going up in the Z direction, in which that's not supposed to happen.

Getting a bit frustrated

I'm fairly new to robot programming in general so if you can please make it simple for me to understand that would be great

I currently have a R30B robot set up at my company and I'm having an offset issue where it totally misses the object it's supposed to pick up. If the orientation of my object is longitudinal to my feeder plate, it picks it up just fine, but when it is rotated, the picker totally misses.

I'm thinking it's a calibration issue that I'm trying to solve by using 'ADJ_OFS' call command, but I don't know how to properly use it within my program. The manual doesn't really justify what to do as it's very vague on how I can implement it.

Does anyone know or can provide a step by step with clarification on how I can apply ADJ_OFS to my program? I'm using RUN_FIND and GET_OFFSET for my IRVISION to get the parts recognized by my camera.

I was recently tasked to work on an older model of a fanuc robot arm which I didn't have any clue or foundation to stand upon.

Our work cell has three feeders that will supply assembly material for the arm to pick up and place within the main assembly item. So far feeder 1 & 2 works fine but feeder 3 keeps giving us issues. Within feeder 3 we have small contacts (sockets) that the Tool keeps missing upon coming down to picking them up. However, if we orient the socket contact within the same orientation as the image it uses to recognize it, the arm is able to recognize which end to pick it up from, otherwise it will miss or aim right for the middle of the contact.

So far we have tried numerous calibrations to try and solve the issue:
1) TCP re-calibration, all the tools used have the same dimension and we did the test to move it around a center point and it worked fine.
2) Re-calibrating the feeder user frame
3) Re-imaging the socket contact

Can someone provide another solution or method that hasn't been listed yet?