Wrong positions after updating revolution counters

Terzi

Hi

I had to update the revolution counters after unplugging the control cable. I jogged the robot to zero positions and updated the counters. After this, positions were a bit displaced. Did I do something wrong? The robot is 6400R with S4C+ controller.

Terzi

When I move the robot to the zeros with moveABSj command after the calibration, all axes are at zero lines except the second axis. I tried this many times and the second axis is always a bit at the positive or at the negative side of the scale. Is this a bad thing?

Anders S

Hi,

If you not have changed the fine calibration it must be the counter for axis two that is zeroed one resolver turn wrong. Try to update that counter again and see if it helps.

 

Terzi

Thank you for your answer.

No I didn't change the fine calibration. I was just updating the resolvers. Is there any way to verify that the resolver is zeroed wrong? I want to make sure that the rest of the axes are fine.

I tried to update the resolver of the second axis many times but the result is always the same: displaced positions. I did a test. I tried to align the calibration line to between the zero and the next line at the plus side. After this I used moveABSj and the second axis moved few lines to the plus side. Then I tried to align the calibration line to between the zero and the next line at the minus side. Then the second axis moved few lines to the minus side. No matter what I did, I was never able to align the second axis with the zero line.

Anders S

If the robot is calibrated correct the "calibration marks" on each axis should be placed at the "zero line" for that axis. But the only way to be sure that the calibration is correct is to use a calibration tool such as "Dynalog" or similar. If you contact your local ABB Service office they should be able to help you to with that.

Terzi

Thank you Anders for your advice.

There is some deviation at the second axis but I was told that it is acceptable so I guess that it's just going to be that way.

I did some tests. I tried to create a new tool when the second axis was at the plus side. The error was quite huge. Then I tried with another tool when the second axis was a bit at the minus side and I got good results. If I understood correctly, it is not possible to have good error values for the tool if robot's calibration is bad. Please revise me if this is not correct.

About my actual problem, I have a good quess what caused it. It might be possible that the person who worked with this robot before me didn't update the revolution counters correctly (He was new at robotics). So basically when I started to work with this project, I taught the positions with wrong calibration. Luckily, my program is quite short at the moment so I was able to correct it quite easily.

In case that I have to update the revolution counters again in the future, what sould I do to prevent this problem happening again?

Simotion

When you perform a moveAbsJ for all axes to 0°, what is then the maximum permissible deviation of each axis to conclude that the fine callibration might be ok.
With deviation I mean the number of marks deviating from the zero mark on the robot axes?
Or what is the best way to find out that fine callibration is ok, when you don't have the possiblility of a calibration tool?

PerSvensson

Also check that the values for motor calibration offset are correct (found in system parameters\manipulator). Should be a sticker on axis 2 with the values for all axis.
You shouldn't loose them just for unplugging but you never know :-)