We are having a hard time with multi axis drilling using hole making and bounding box for clearance option. as usual it looks great on the screen, but will not work when run on the machine. i have tried to read more on how nx uses the bounding box option, but i do not understand how it creates the Z rapid plane without knowing the true c/l of rotaion of the machine.
is anyone switching to TCP in order to perform rapid moves?
this is a sample of the code im getting on a 3+2 operation using DFO ( G254 )
DFO is active at this point.
X-3.7219 Y-.9006 I-.0034 J-.0049
G0 Z5.905 ( this is canceling the DOF )
Z10.3551 ( no DFO makes this move un reliable )
X-3.0615 Y.2574 B50.644 C90.
X.9626 Y-.9006 B-15. C113.
G40 G3 X.9626 Y-.8887 I0.0 J.0059
X.9626 Y-.8887 Z4.7591 I0.0 J-.0119 K-.00
"is anyone switching to TCP in order to perform rapid moves?"
Yes, this is what I do in postprocessor when user wants machine motions to be identical as NX tool path verification.
I activate RTCP for 5-axis motions between holes (when tool axis changes)
And for next hole I deactivate RTCP if it was activated during reposition move and activate 3+2 transformation for each hole - to make proper drilling cycle or helical interpolation available.
This was not easy and implementing it for first time costed me a lot of hairs from my head
But even more difficult it was for machine kinematics like DMU Evolution series, where rotary axes are not perpendicular. In this case you need to switch IKS settings in postprocessor - different for 5axis simultaneous motion and different for 3+2 hole machining.
Maybe there are different methods, but this one allows me to achieve perfect NC output.
Other option which I also use quite often for reposition move between holes on different angle is:
- after machining previous hole cancel drilling cycle (or helical interpolation)
- retract to Z-maximum (rapid motion)
- cancel previous 3+2 transformation
- skip all points from tool path between holes (5 axis motion along clearance geometry)
- reposition table (rapid motion) + activate 3+2 transformation for next hole
- safe approach to next hole XY first + Z next (rapid motion)
- machine next hole
If machine isn't too big - it might be even faster than repositioning "along NX clearance tool path"
I agree with @Marek_Pawlus " a lot of hairs from my head "
Also it is possible and even for-to-point drilling.
(so, much more hairs , cant imagine all of problems you can face...)
"But even more difficoult it was for machine kinematics like DMU Evolution series, where rotary axes are not perpendicular. In this case you need to switch IKS settings in postprocessor - different for 5axis simultaneous motion and different for 3+2 hole machining."
--- So I suppose youre using axis position of perpendicular kin. like spatial angles.
and for 5ax simult. youre using real kinematics. I know that idea.
-- but switching kinematics doesnt work during operation. - that is problem.
But there is no need to switch kinematics:
--- rotary axis positions are outputed correct by PP with real kinematics in every position,
but values for tilting function are different thing (have to be calculated).
"-- but switching kinematics doesnt work during operation. - that is problem."
I was able to do this, even multiple times during one operation. The clue here is that IKS gets changed in event AFTER you actually change it.
So to make it work, I was changing IKS in the event BEFORE I actually wanted it.
Back then, priority for me was to make it universal for all machine configurations, where IKS settings makes calculations for me.
As always with NX we have several ways to achieve goal...
I have heard that Siemens development is working on new feature of handling multi axis transitioning (kinematically) which will be a big help. It is out beyond NX12. It won't help now but you may want to do things different based on knowing that improvements are coming.
Yes, this kind of news will influence on how much effort I will put in this topic. Thanks.
I can't wait just to see just which way development plan to handle it and what "kinematically" means in this case.
Only thing I know - it will require kinematic chain in MTB, not the machine model
In the past I was trying to switch kinematics by this:
global mom_kin_4th_axis_vector mom_kin_5th_axis_vector
set mom_kin_4th_axis_vector(0) 0
set mom_kin_4th_axis_vector(1) 1
set mom_kin_4th_axis_vector(2) 1
set mom_kin_5th_axis_vector(0) 0
set mom_kin_5th_axis_vector(1) 0
set mom_kin_5th_axis_vector(2) 1
(also Ive changed axis range in some case)
But it didnt work...
Is there missing somthing int such procedure?
It looks similar to what I do. I'm not sure why it is not working.
If you tried it long time ago, you could make sure if you have mom_kin_iks_usage set to 1 in that old postprocessor.