5-axis functions in postprocessor 3

[Denim]

Hi,

We are having a custom postprocessor built for our 5-axis machine, and I need some input on how the postprocessor should activate the 5-axis functions on the machine control.

The machine has a Fanuc 30i control, and we want to use Tool Center Point control (G43.4) for simultaneous work and Tilted Working Plane command (G68.2, G53.1) for 3+2 positioning work.

What is the best, or most commonly used, method of triggering what and when these functions are output from the postprocessor? Should for example a MILL_CONTROL event be manually inserted into a variable contour operation, or can the system figure this out automatically?

I appreciate any input on this subject.

/E

 

[Bugie]

I believe that outputing TCP coordinates is possible BUT you will need to create TCL code as there are no buttons for this function in post builder

 

[jelmerra]

It is possible to automate figuring out if it is a simultaneous operation or a planar operation. Main difference is programming manner either vector or msys based. I think vector method is in favour by siemens. It depends a bit upon the work you do what saves most time when programming.

jelmerra

http://www.rouwe-analysis.com

 

[Denim]

What is the difference between vector or msys based programming? Is it whether the toolpath commands the tool axis direction by vectors VS. tool axis by actual coordinates of the rotational axes?

Or is it related to the way of setting up the machining environment and geometry in NX?

 

[wmalan]

The short answer I think is they should be able to get that setup purely at the post end. The determining factor will be: Is the tool feeding (not rapid) while moving the tool axis? The post developer can test for that and set the output accordingly. IMHO Millcontrol and insert events are a last resort for many reasons.
--
Bill

 

[jelmerra]

Msys programming is when the tool always is ZM+ i.e. optimal if many operations use the same tool orientation. Vector programming is tool vector definition in each operation separately. Vector programming is optimal for lots of different orientations.

As wmalan states setting up ude's and mill controls are repetitive tasks that should be minimised. It is a trade off between a quickly developed postprocessor requiring lots of user interaction or a more automated postprocessor.

 

[Denim]

I see.
Sounds like vector programming is what we are used to. If many operations use the same tool axis, we create a MCS for each orientation with the tool axis set to a different vector than +Z of MCS. If there are many different orientations, then we define the tool axis locally in each operation. This seems to be working good.

In order to have the PP calculate and output plane rotation commands, do we need to set up a MAIN_MCS with a "master" tool axis that all other orientations are compared against? In the past, we have just used local_mcs's without a main_mcs and probed for individual fixture offset's for each orientation in the machine. With many orientations, this get's old real fast!

I agree that a good PP should be able to automatically figure out what is happening and output the correct command for the operation. Sounds like a much less error prone method than setting mill_controls manually.

 

[jelmerra]

The mcs purpose switch main/local is one way of doing things. It is siemens preference. There are other ways to do it like using a mill control with a reference mcs at the start of a program. There are a couple of other methods too most of them have been invalidated by newer releases of nx.