Cancel
Showing results for
Did you mean:

# Re: Rotary FEEDs settings

Legend

No, I did it without any simulation model.

I had only to write the procedure of calculation of distance for each motion. My machine was turnmill with C-axis on lathe spindle and XYZB on mill spindle. One more thing: I calculated deltaX/deltaY/deltaZ not for tool tip point as NX does but for machine zero point (rotation centre of mill spindle).

You are right that the real position of workpiece affects this calculation too. To take this into account, I use UDE with params delta-centreX, delta-centreY, delta-centreZ to define offsets from MCS to rotary centre. These values are also included into my formulas.

Highlighted

# Re: Rotary FEEDs settings

Solution Partner Phenom

basic example of calculation FRN:

example:

the delta move = 1 mm,

F=1000mm/min

s=v*t,

time=1mm/ 1000mm.min-1=0,001min

FRN=1/0,001=1000

??

it is OK?

---------------------------------------------
#♫ PB, 5ax, itnc, nx, vericut ♫ #

# Re: Rotary FEEDs settings

Legend

Yes.
Then you should output the lines like:

`G93X.. Y.. Z.. F1000`

When you use the inverse mode, the address F is non-modal, so F-value should be in each line.

# Re: Rotary FEEDs settings

Solution Partner Phenom

ok thanks,

but you said that it require account rotary delta moves according to the pivot?

So my example is no good?

nx gives me this:

(F250mm/min)

`N239 Y50.734 Z149.696 A18.722 F34.843  `
`N240 **dist 0.81325852521647679 (sqrt (delta x^2 +delta y^2+ delta z^2))N241 ****rotdel 0.28899999999999998 by:  set rot_del [ASK_DELTA_4TH_OR_5TH 4]      set x [ASK_DELTA_4TH_OR_5TH 5]      if { \$x > \$rot_del } {         set rot_del \$x      }N243 ****frn 34.602076124567475 ???N244 Y51.433 Z149.28 A19.011 F34.602and by uncommenting the fanuc cmd  gives me something simillar------------------------------when I calculate it by previous post I wrote:`
`  dist 0.81325852521647679 F=250mm/mintime=s/v=0,003252minFRN=1/time=307,5..`

I realised there is calculation for maximum and it overides it.

Can you give me your opinion?

thank you

---------------------------------------------
#♫ PB, 5ax, itnc, nx, vericut ♫ #

# Re: Rotary FEEDs settings

Legend

See my story about 5-axis machining of blades above. How did I work with inverse mode? I specified some cut feed in the dialog of NX Operation, then postprocessed, then run the program on the machine and looked at the values on the NC panel to compare the original feed with real one.

At first, when I used the NX procedures (like you did), I could not get the stable feed value on the machine as I need. I thought "why" and tried other algorithms to calculate "delta distance". After all I got rather good results. See picture. This is kinematics of my machine.

When you use the NX procedures, you get delta distance for tool tip point (A) through variables mom_pos or mom_mcs_goto. However it maybe situation when tool tip point left at the same position, but spindle rotates enough much, and linear delta is great too because of values (tool_length+pivot). What distance does NX calculate? About zero. Then you get wrong time, wrong inverse feed, etc.

My idea is calculate linear deltaXYZ for another point (B) - machine zero in the centre of head rotation. To do it for this point I have to operate with several params - mom_mcs_goto, tool orientation vector, tool length, pivot. It is not simple, but the solution can be obtained.

P.S.: Of course, my picture above can be applied only to this kinematics. May be other rotary kinematics, for example, rotary table, with different formulas. Anyway my idea is to calculate what distance XYZ between motion start and motion end during rotation in the real machine space. That's why you need pivots and workpiece offsets.

# Re: Rotary FEEDs settings

Solution Partner Phenom

Thank you. Nice explanation. I thought about simillar stuff.

So when the tool tip stay the same and B rotates, the pivot is moving in X and Z - so this is the delta distance. (example)

"Anyway my idea is to calculate what distance XYZ between motion start and motion end during rotation in the real machine space."

If I got it right you mean not xyz of the tool tip but xyz of the axes motion.

Nice and simple.

thank you.

---------------------------------------------
#♫ PB, 5ax, itnc, nx, vericut ♫ #

# Re: Rotary FEEDs settings

Legend

Yes, you understand my idea.
To use the inverse mode, the time of each motion must be calculated. Alas PostBuilder's templates for FRN only operate with "delta distance" for tool tip point, not taking into account kinematics of the real machine. They are corect for 3-axis machinig, but I never met any problem with dynamics for 3-axis, so I don't see the need to use this mode here.

5-axis machinig is another thing. Sometimes we are machining relatively smooth surfaces (where no need of the inverse mode too), sometimes - surfaces with sharp edges, like blades, and users switch on FRN option to have more stable dynamics of machine.

My idea is not to calculate tool tip moves as NX does, but to calculate moves of components as the real machine does or simulation model does. Any 5-axis machine has linear and rotary components, so I calculate how much time is required to move component XYZ (as I wrоte above), how much time is required to rotate component A/B/C. After all I have clear view how long is this motion to output the inverse feed.

Of course, for calculation it is required more data, because my postprocessor doesn't refer to any simulation model and setup. So some data I get from mom-variables, some data are defined as constants inside postprocessor (for example, pivots), some data may be defined through UDE (for example, workpiece offsets).

# Re: Rotary FEEDs settings

Solution Partner Phenom

Ok maybe last thing,

Are you sure that machine doesnt need delta distance of tooltip and doesnt account this value to adjust feed of moving component through the rotation move?

Doing it itself because machine know pivots, tool lenghts, zero position, etc according to actual position of machine.

---------------------------------------------
#♫ PB, 5ax, itnc, nx, vericut ♫ #

# Re: Rotary FEEDs settings

Legend

Ok, I know that my solution is not universal. I can imagine many toolpaths without need to re-calculate rotary moves, I can imagine "clever" NC units which are able to optimize dynamics and to avoid feed "jumps". However 3-4 years ago I had to work with the machine which had dynamic problems on machining blades so that I had to search for some solutions. Exactly this time I had heard about inverse time feed. Using of NX procedures was unsuccessful, and then I found my own way.

You opened this discussion about inverse feed. On my experience I tried to explain what it means, where used, and how to use it. If you don't have this need for your machines, I'll be glad

# Re: Rotary FEEDs settings

Solution Partner Phenom

this discussion helps me. Now I know what to expect from machine and how can frn works.

It is mazak avriaxis i700t,

And I will see what customer and me figure out.

He is trying to find some machine parameter to fix it but If he wont find any such stuff

we will do something with feed.

Or some setting with multiaxis positions per minute or per step may can slow down machine, we will see.

---------------------------------------------
#♫ PB, 5ax, itnc, nx, vericut ♫ #