Secondary axis rotation is 90* off...

[Cavi Mike]

I can't figure this out. If A and B are 0, everything is fine, part is orientated correctly, however everything goes awry as soon as I try to do some side holes. I thought it was an issue with a tool/work plane in MC but it seems that as long as Z is pointing in the correct direction, it doesn't matter which way X and Y point, the code will always output the same. I'm modifying the generic Fanuc 5axis post to use in a HMC. This machine has always been run as a 3+2 but I'm trying to use it as a proper 5axis for once.

[Colin Gilchrist]

Did you change the Machine Base Matrix?

For a Horizontal, it should be:

#Machine base matrix (Base matrix to map positions into)
matb1$       : 1
matb2$       : 0
matb3$       : 0
matb4$       : 0
matb5$       : 0
matb6$       : 1
matb7$       : 0
matb8$       : -1
matb9$       : 0

WCS should be set to TOP, and A/B Zero should be FRONT tool plane.

What is your actual Rotary Configuration? (Kinematically, how are the rotary axes mounted on the machine? Is "A" the primary (platter, perpendicular with Spindle?), and "B" the Secondary? (B Rotates about Machine Y, while A Rotates around Machine Z)

 

[Cavi Mike]

I switched to those and things got really weird. XY locations are way off on one side, pretty sure the post is trying to machine from the back side of the part now.

"B" is the standard HMC axis that rotates the pallet about Y axis, "A" (which really should be C) faces the spindle and rotates about the Z axis. B is doing what it should, XY locations are  where they should be, it's "A" that's off by 90* - but only once B rotates. If B is at 0, A is fine.

[Cavi Mike]

I'm starting to understand why it's off - the post thinks this is the typical trunnion in a VMC, which it sort of is, except instead of the primary axis rotating about X, it's rotating about Y. So that's what I need to figure out.

[Colin Gilchrist]

As long as the Rotary Direction and Rotary Axis variables are set correctly, and the Machine Base Matrix is setup correctly, then the likely culprit would be the way you have your Planes setup inside Mastercam.

For A(C) on B, you should have something like:

#Primary axis angle description (in machine base terms)
#With nutating (mtype 3-5) the nutating axis must be the XY plane
rotaxis1$ = vecy  #Zero
rotdir1$  = -vecx  #Direction

#Secondary axis angle description (in machine base terms)
#With nutating (mtype 3-5) the nutating axis and this plane normal
#are aligned to calculate the secondary angle
rotaxis2$ = vecz  #Zero
rotdir2$  = -vecx  #Direction

If you look at the Primary values (vecy and vecx), you'll notice there is no "Z" in the plane being described. This means rotation happens "Around Z" (as it should, regardless of what you label the axis address letter.)

If you look at the Secondary values (vecz and vecx), you'll notice there is no "Y" in the plane being described. This means rotation happens "Around Y" for the secondary. And should be labeled "B".

If the rotations happen "Backwards", in other words, you are getting "-90.", instead of "90.", all you should have to do is reverse the "sign" of the RotDir variables. (add a minus sign, or remove the minus sign. That's it.)

'mtype' should be set to '0.'

Also, and this is important(!), make sure these are both set to Zero, when you are setting up the rotation angles in the Post:

pang_output  : 0     #Angle output options, primary
sang_output  : 0     #Angle output options, secondary
                     #0 = Normal angle output
                     #1 = Signed absolute output, 0 - 360
                     #2 = Implied shortest direction absolute output, 0 - 360

 

[Cavi Mike]

Forum only let me make 3 posts the other day.

I tried every combo of the base matrix possible but the only thing that came close was the standard VMC layout. I think at one point with your layout I could have made a good part, but the entire part would have been orientated 90* which would have just been confusing. Turns out, all I had to do was swap vectors on the primary axis and voila, I got the code I was looking for. Thanks for the help!