05-13-2006 01:23 AM
05-15-2006 02:42 AM
05-15-2006 09:55 AM
Hi, Jochen. Thanks for your answer. Now I try to express my question in detail.
System: 2_D position table (50mm*50mm)
hardware: PCI7344 + 2 voice coil motor + driver(no feedback,just a amplifier) + 0.5micron encoder(Renishaw) as the position feedback
software: Labview7.1 + MAX3.1
Aim: To get the setting time of single axis lower than 40ms (even if there is some overshot)
The document you mentioned I have read before. When I tune the servo,'Auto tune' can't get meaningful parameters. So all the time I tune it manually.
1, About Step reponse: I can get a setting time lower than 40ms(step 1000 counts + 200 sample) by accident, but the same PID I can't get the good result all the time.
2,When I use 'start motion.flx' moving 6000 counts (400000 counts/s+32000000 counts/s^2+S curve time 3) or use 'Jog axis',the setting time are more than 100ms (see axis.jpg). I try to increase the Td or Kd. It can change the overshot, but the setting time is not ideal. IF Kd too large, the system will unstable. I think maybe the aim is unpractical, right? Jochen, has you tune some table whose setting time lower than 40ms using 7344?
3,what's the relationship between 'step reponse' and 'Jog axis'.Should I tune the PID according to 'Step reponse' or 'Jog axis'? (In 'Jog axis', the velocity and Acceleration and Deceleration also play an important role.)
4,The driver have no current feedback.I guess it is an important problem.what'a your opinion?