Hello, a while back I posted for information on how to limit the output of a 3 axis post to set incremental distances, and received a number of usefull suggestions that in the end consisted of performing some straightfoward math on the XYZ distances using a sum of squares method and mom_pos() command to get the axis positions. At this point I would like to build upon that approach but now for a 5 axis move and I my question is that with the 4th and 5th axes now being rotary motion I am assuming I cannot just take all five positions from mom_pos() and start putting them into a sqrt(X^2+Y^2+Z^2+A^2+B^2) calculation? In looking thru the various post commands I beleive I may need to use the mom_mcs_goto() values to calculate all the intermediate waypoints prior to having the kinematics calculated for the 5 axis?
Can anyone verify this is the correct approach? If this is the correct approach my next concern is that I will also need to create waypoints to interpolate the ijk tool vectors mom_tool_axis().
I am basically trying to create a toolpath to progam a sensor that I need to take measurements at intervals no larger than "X' points along the a path, smaller intervals are ok if required to hold path accuracy but larger intervals are where I have issues.
All suggestions appreciated.
I am looking at the same thing (by my interpretation of your post) right now. We are having problems in kinematically challenging vector ranges. NX is (of course) not aware of the kinematics as it is deciding the tolerance of the points to drive the surface. Dynamic comps on the machine should hold "posture" even when kinemtics are challenged (slow down if it has to - split the move up as much as possible - etc.) My experience is that ORIVECT on a siemens does pretty well - I am trying to find the best settings for G43.4 on a fanuc to do this. I am also trying to figure out how to have cse be accurate through this type of move as well. Anyway - if a point with a vector crosses over (rolls a table rotary up to a close to singular and then reverses) - things are really tough for the machine - especially with a table and far from pivot. So - I wonder if it is best to break up linear moves in the post (at least what Siemens would recommend.)
You can control (in typical contouring operations at least) the max distance / rotary change per step, so NX *should* be able to add points where you need them.
I'm not sure how to deal with the rotary axis singularity issue, although (in a Siemens NC control group) training class I just attended, they recommended pushing the number of decimal places for the orientation vector output to 10 (e.g. TRAORI + A3/B3/C3)
Production: NX10.0.3.5 MP5 + patch/TC11.2
I'd rather be e-steemed than e-diseaseled