Fluttering Very Small Z Movements

[CarbonCAD]

I have a very shallow concave dish that is .0006" deep +/-.0002.  Using a High Speed Spiral tool path with a 1/16 flat, .028125 step over, and a total tolerance of .000015.  The problem is that being such a small tolerance we go to five decimal places, but the fifth decimal place flutters up and down .00001 continuously.  All our other machines it is not a big deal, but our Makino machine reacts well enough that the .00001 z move can be seen visually.  Rounding only makes it worse moving .00005.  Changing the tolerance too much will take it out of spec.  Any ideas as to how to get rid of the fluttering?

 

ex:

G41 X17.08485 Y-2.48875 Z-1.24429
X17.07057 Y-2.47646 Z-1.2443
X17.05451 Y-2.46662
X17.0371 Y-2.45941
X17.01878 Y-2.45501
X17. Y-2.4536 Z-1.24429
X15.4 Y-2.4536
X15.38123 Y-2.45508 Z-1.2443

[Müřlıń®]

Remodel the part with a micron tolerance..

 

Look at your system tolerances in Mastercam...chances are when you imported the model the tolerances were looser than what you are trying to achieve with your machining tolerances...

[Colin Gilchrist]

Yep. What Murlin said.

 

You need to set your System tolerance. I would go one decimal further, and make is 7 decimal places for your system tolerance. .0000001.

 

Then, in the tool path, make your tolerance 1 millionth of an inch. .000001.

 

Make sure you post is setup to output 5 decimal places, and you'll be good to go...

[CarbonCAD]

System tolerance is now at 7 decimal places.  In the tool path it does not want to accept a millionth.  Tells me: "0.00001 to +1.0E38"

Total tolerance under Arc Filter/Tolerances section.  What am I missing?

[Müřlıń®]

yes but what does your surface deviation and cordial deviation say??

 

And all those others in between?

 

Set all those to micron....

[CarbonCAD]

This is what I have. 

[crazy^millman]

Scale the part up by a certain factor and then have your post use a switch to change the scale it is programmed to back down to the real part size.

[Allan]

Your NC precision is way too tight @ 0.0000001" you are looking for code to 10 millionths 0.00001", so the NC precision should be set to 10 millionths.

 

The rounding is being done in the post currently and the MP.dll has had issues with that since X when the control definition was implemented.

 

Allan

 

 

Oh and your minimum arc length of 0.00002" is a disaster waiting to happen IMHO; that should be 0.002.

The minimum arc rad of 0.00005" will also bite you; that should be 0.005"