Hello everyone, I am currently working on a project involving an MPAI linear electric piston, Kinetix 5300 Drive, and Studio5000 V35. The motion system operates in a Point-to-Point mode with a high loop response and rigid load coupling. A tensioned belt is utilized to drive the mechanism in the "extend" direction. The servo's accuracy, with an Effective Resolution of 2,079,152 feedback counts/rev, is suitable for the application without the need for dual loop feedback. The critical aspect of the system is positional accuracy, with components directly attached to the piston end for dimensional stability. The servo is responsible for moving mechanical components during a recipe change, with quick movements not being essential for the process that can span over days. However, positional accuracy is crucial. Despite having a set constant "Servo_Retract_Pos" at 35.000 mm, the servo sometimes ends up at a different position (35.1xx mm) when retracting, which is puzzling. This issue doesn't occur when extending the servo. I have conducted tests where the average error while extending is minimal (<0.01 mm) and the target setpoint is maintained. I am trying to understand why the setpoint changes when retracting and how to ensure the servo consistently reaches the correct position without manual adjustments. Any insights or suggestions would be greatly appreciated. Thank you for your help in advance. Best regards, OhNo!
Please provide information on your position and velocity loop parameters such as gains, integrals, and bandwidths. What is your position error tolerance? One possible reason for spikes at the end of your movements could be interference from the brake system. What method are you using to engage the brakes? It is recommended to avoid using MAM.PC for this purpose.
It's great to see your attention to detail and high resolution feedback. Many users struggle with this. Have you considered removing the axis from closed loop control once it reaches the 30mm set point? This is important as the command position matches the actual position in open loop, even if it's just for one controller scan. It's worth checking if the target position changes to 35.1mm. It's strange for the actual velocity to peak and then drop to 0, but this could be due to the limited update rate of the trends. Increasing the trend update rate could provide more insights.
Robertmee inquired about the position and velocity loop gains, integrals, bandwidths, and position error tolerance. Observing spikes at the end of a move, which may be caused by brake interaction, he questioned the brake initiation method, hoping it was not MAM.PC. In response, the poster expressed gratitude for the engagement and mentioned using autotune settings. Concerns were raised about potential overwriting of Position Error Tolerance Limits and efforts to improve performance. The Position Loop settings included a Bandwidth of 47.66 Hz, Integrator Bandwidth of 18.6 Hz, and an Error Tolerance of 1.73 Position Units. Changes were considered, particularly due to the relatively high Error Tolerance compared to the previous setting of 0.01. For the Velocity Loop, settings comprised a Bandwidth of 122.03 Hz, Integrator Bandwidth of 0.0 Hz, and an Error Tolerance of 27.6 PU/s. The use of a sequencer for command execution after servo positioning was described, with a request for advice on a potentially more efficient approach. Overall, the discussion highlighted adjustments to improve servo performance while addressing concerns about brake interaction and optimizing control loop settings.
Peter Nachtwey praised the excellent trend and feedback resolution, noting its superiority compared to many other posters. He inquired about the closed-loop control axis when it reaches the 30mm set point, questioning if it disengages at that point. He pointed out that in open-loop control, the command position should match the actual position, even if it's just for one controller scan. Peter suggested plotting the target position to check for any deviations and also recommended a faster trend update rate for more accurate results. In response to Peter's queries, the user expressed interest in learning more about the servo's monitoring capabilities and planned to investigate the target position discrepancy. The user also mentioned adjusting the sampling rate to address the issue and promised to update Peter on the findings.
Hello Peter, I believe this is what you were looking for. The addition of .CommandPosition is indicated by a white line. The sampling rate is set at 1 ms. It appears to me that there are sudden changes in CommandPosition, leading to a rather aggressive reaction from the system (across 0.1 mm) in order to comply. Moving forward, I need to investigate if there is a shift from a closed to an open control loop and how this transition might occur. Could the engagement of the brake, as hinted by Robertmee, be causing this issue? I am unsure how the brake activation could impact the CommandPosition, but the timing seems to align at this level. Perhaps, I should replot the data while including the brake signal and zoom in to get a clearer picture. Robertmee, I am eagerly awaiting your insights on the drawbacks of using the MAM.pc. Your feedback is much appreciated. Thank you! -OhNo-
Hi OhNo! It sounds like you're dealing with a tricky issue with the retract positioning. One possibility to consider is that there may be some backlash in your mechanical system, especially if you're using a belt drive, which can lead to discrepancies when reversing direction. Additionally, it might help to check your servo tuning parameters, as they can impact how it responds during retraction. You could also try adding a small amount of ramping in the motion profile to see if that helps stabilize the positioning. Having consistent feedback from your system is key, so ensuring everything is tightly coupled and minimizing any play in the setup could make a difference too. Good luck, and keep us posted on what you discover!
✅ Work Order Management
✅ Asset Tracking
✅ Preventive Maintenance
✅ Inspection Report
We have received your information. We will share Schedule Demo details on your Mail Id.
Answer: Answer: This issue could be due to various factors such as mechanical play, tuning parameters, belt tension, or encoder resolution. Further investigation and troubleshooting may be required to pinpoint the exact cause.
Answer: Answer: To improve positional accuracy during retraction, consider checking the mechanical components, tuning the servo parameters, verifying encoder counts, ensuring proper belt tension, and reviewing the motion profile settings in Studio5000 V35.
Answer: Answer: The issue of positional accuracy during retraction could be influenced by factors such as backlash, mechanical compliance, or tuning settings. It's essential to conduct a thorough analysis of the system components and parameters to address the discrepancy.
Join hundreds of satisfied customers who have transformed their maintenance processes.
Sign up today and start optimizing your workflow.