Encoder Indexing

1. PID was done to ensure smooth start/stop values as per the attachment
2. The servo mode was Torque mode.
3. I think I have indexing and FERROR problems, because :
When I restart the servo drive I have two possibilities:
- Either the encoder read zero of less than 1mm and then start changing the output until reach the zero .
- OR start with a different output far away from Zero , in that case the servo motor lose the location completely and never move correctly , and in that case I have to restart the driver ( may be 2-4) times until I get the encoder within a range less than 1mm , then I being able to move it .
Question [01] : Isn’t indexing problem ? if yes how can solve it ?
4. As per the attached images, when I use CMD say[G0 X500] the motor rotate until reach 498(or near) then stop rotation while the encoder continue increase the output unit reach 499.998 then stop I used FERROR : 20 Min_FERROR 10 which is too much but once I reduce is I got joint following error
Question [02] : Isn’t FEEROR problem ? if yes how can solve it ?

Is the encoder signal coming from the drive or from an actual encoder on the motor?

Encoder feedback --> to drive --> to 7i77 Mesa

Hmm sounds like the drive is outputting many counts at startup

Does the drive have absolute encoders?

Maybe there is an option to disable this behavior

Normally LinuxCNC will clear the encoder counters at start-up
so the encoder position will start at 0

the servo motor motor has incremental encoder .
there is an option to reset the fault ...as the attached sheet ... and the driver has P2 cable connection to mesa card , and P4 which is the feedback from servo motor .
attached P2 & P1 pinout .

You might try plotting the encoder position with halscope from
when LinuxCNC starts to when the drives are enabled to try and track
down what is happening