User Tools

Site Tools


ms2000_optimal_alignment_procedures

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
Next revision
Previous revision
ms2000_optimal_alignment_procedures [2016/09/07 15:26]
asiadmin ↷ Page moved from documentation:ms2000_optimal_alignment_procedures to ms2000_optimal_alignment_procedures
ms2000_optimal_alignment_procedures [2016/10/04 15:13] (current)
Line 18: Line 18:
 The purpose of aligning the analog feed back for the drivers is to reduce the dead zone near zero speed when low voltage is applied to the motor. ​ The chart below shows typical a typical transfer function for various feed back adjustment values. The purpose of aligning the analog feed back for the drivers is to reduce the dead zone near zero speed when low voltage is applied to the motor. ​ The chart below shows typical a typical transfer function for various feed back adjustment values.
  
-[{{ :​documentation:​ms2k_optimal_alignment_1.jpg?​direct&​300 |Click to Enlarge}}]+[{{ ms2k_optimal_alignment_1.jpg?​direct&​300 |Click to Enlarge}}]
    
 As you can see, without any feedback, (blue diamonds, ''​AA=1''​ case) there is a substantial region where supplying a commanded speed to the motor results in no motion whatsoever. ​ The digital loop can somewhat overcome this nonlinearity,​ and still provide accurate positioning,​ but the dead zone will give rise to “wind-up” and overshoot in a digital loop that is well-tuned for the constant slope section of the transfer curve. ​ Increasing the feedback (larger AA number) reduces the size of the dead region and significantly “linearizes” the transfer function near zero.  Not surprisingly,​ problems associated with poor drive card alignment show up when the motors need to move slowly, i.e. when completing a move. As you can see, without any feedback, (blue diamonds, ''​AA=1''​ case) there is a substantial region where supplying a commanded speed to the motor results in no motion whatsoever. ​ The digital loop can somewhat overcome this nonlinearity,​ and still provide accurate positioning,​ but the dead zone will give rise to “wind-up” and overshoot in a digital loop that is well-tuned for the constant slope section of the transfer curve. ​ Increasing the feedback (larger AA number) reduces the size of the dead region and significantly “linearizes” the transfer function near zero.  Not surprisingly,​ problems associated with poor drive card alignment show up when the motors need to move slowly, i.e. when completing a move.
Line 24: Line 24:
 ===== Adjusting the Feedback Alignment (AA) ===== ===== Adjusting the Feedback Alignment (AA) =====
  
-All of the steps below require a connection to a computer terminal program. Refer to  [[documentation:​tech_note_rs232_comm|RS-232 Communication]]+All of the steps below require a connection to a computer terminal program. Refer to  [[tech_note_rs232_comm|RS-232 Communication]]
  
  
Line 342: Line 342:
 All those numbers may not look particularly useful, but they have a lot of information about the system dynamics in them.  Using //​HyperTerminal//​’s file capture capability, we can save the dumped information to a file and plot the results in a charting program like Microsoft’s Excel. All those numbers may not look particularly useful, but they have a lot of information about the system dynamics in them.  Using //​HyperTerminal//​’s file capture capability, we can save the dumped information to a file and plot the results in a charting program like Microsoft’s Excel.
  
-[{{ :​documentation:​ms2k_optimal_alignment_2.jpg?​direct&​300 |Click to Enlarge}}] ​+[{{ ms2k_optimal_alignment_2.jpg?​direct&​300 |Click to Enlarge}}] ​
  
 The data shows the move trajectory, the error from desired path, as well as the drive signal given to the motor mover. ​ For this move, there were six ramp-up and down points on the way to full velocity. ​ Backlash of ~450 counts was enabled, the finish error was 1 encoder count, and the drift error was 2 encoder counts. ​ During the time of maximum acceleration,​ there is also the most error. ​ The short ramp (six cycles) contributes to the error (about 17 microns peak error from the theoretical path). ​ Once the move is under way, the average error settles down to about 50 counts (4 microns). ​ You can see the retrograde motion caused by the anti-backlash routine starting at about the 59th time step.  There is a fair amount of overshoot and then the move settles in to find its final position. ​ The move would clear the busy status at about step 77, but then it drifts outside the drift error range and cannot settle for another 45 time steps.  ​ The data shows the move trajectory, the error from desired path, as well as the drive signal given to the motor mover. ​ For this move, there were six ramp-up and down points on the way to full velocity. ​ Backlash of ~450 counts was enabled, the finish error was 1 encoder count, and the drift error was 2 encoder counts. ​ During the time of maximum acceleration,​ there is also the most error. ​ The short ramp (six cycles) contributes to the error (about 17 microns peak error from the theoretical path). ​ Once the move is under way, the average error settles down to about 50 counts (4 microns). ​ You can see the retrograde motion caused by the anti-backlash routine starting at about the 59th time step.  There is a fair amount of overshoot and then the move settles in to find its final position. ​ The move would clear the busy status at about step 77, but then it drifts outside the drift error range and cannot settle for another 45 time steps.  ​
Line 361: Line 361:
  
  
-{{page>documentation:​error_codes}}+{{page>​error_codes}}
  
   
ms2000_optimal_alignment_procedures.txt · Last modified: 2016/10/04 15:13 (external edit)