Hi everybody,
The problem started when pulse coder batteries have gone while power outage happened, causing SRVO-062 BZAL alarm.
We replaced the batteries and reset the alarm. Of course, the robot required mastering. Since mechanical parts (motors and encoders) were not touched, I set $DMR_GRP.$MASTER_DONE variable to "1" and then $MCR_GRP.$CALIBRATE to "1"(this bit returned to "0" immediately) in order to get the robot accept the master counts and make calibration.
The robot was able to move. However, its zero position was wrong, it was extremely wrong. I checked the master counts, they were the same as previously, before the power outage.
Then, I tried to restore the original images of the controller (FROM00.IMG to FROM31.IMG and SRAM00.IMG files). After restore, I found that the master counts were identical to those that had before. That means that the master counts were not affected by power outage neither by battery outage. However, I still have the same problem - the master counts do not match and zero position is extremely wrong.
My question is how it is possible that the original master counts are wrong.
Note: none of the motors or pulse coders have been ever replaced in the robot. The same restore procedure with images made by my own is working perfectly, but this image has the robot mastered approximately, by joint marks, it's not as precise as required.
Any support would be very appreciated.