Base Coordinate Misalignment

General Application Help: How do I?
1

Hello,

I’ve been working on a UR5e application for a little over a year now, it’s a fairly complicated project, but manageable. However, something keeps happening on the robot side that completely destroys my progress and confidence. Something will happen, and suddenly my entire base frame is off by about a half inch in x and y. I measured the most recent event, and my x has shifted by about 5mm, and my y by 10mm. It’s especially obvious because I have a Sensopart camera mounted on it, and it’s clear that every capture point has shifted by the same amount, in the same way after manually moving to it.

What’s especially weird is that the camera is programmed to detect a pocket in frame, then move the tcp to the center of mass of the detected pattern to pick or place an object. However, after experiencing this offset event, the camera detects the center of mass perfectly, but the robot moves and lands off center in the same way!

My installation settings haven’t changed before or after the event, and the poses I go to have not changed either. I compared my job from 10-6 to 10-30, and it’s still the same.

I know about flexx tools/flange to help realign, and I have them. The problem is that the camera operations reference the base frame. I could apply a universal offset, as my robot program is generated, but that seems risky to me, overall.

Any insight or advice would be very appreciated.

2

are all coordinates in the base plane? and are you overwriting any planes in the program? does your camera have control of the planes.

you could try in the start og your program to overwrite the base with all 0’s if everything is in the base plane?

3

Something like that used to happen to me and it was because on one movement I had it to “turn on radius” instead of “move to this point” (I don’t remember the exact names). The entire robot would jerk and get misaligned.

The jerk was big enough that all the points would move around 2mm per jerk.

I would check that, it is a very silly mistake that can create a lot of problems.

4

Is it normal for base coordinate systems to be offset by a bump, or “jerk” in your case? I guess I’m worried about having to reteach everything if something like that happens.

5

i would assume only if the robot is able to move in its mounting, a little rotation in the base could quickly become more in the tool flange

6

we had a similiar thing, but the change was a few mm (which was a big issue for our case). I was worried as this happened after a safety stop when violating the restrictions while in freedrive. But then we found out, our mechanics just screwed the robot to the table but didnt use the bolt. So this explains that an abrubt stop can cause a change in the nounting orientation. But it cannot explain what happened in this forum topic…