[rweatherly]

I am trying to machine a terrain model, and have the x, y & z coordinates in G-code to do it. The linear interpolation does good for as the mill moves across in a straight line, the same for the next line, etc., but of course there is no interpolation between the lines. Is there a simple way to smooth out the machining?

[ger21]

Turn on constant velocity mode. Either put a G64 at the start of your code, or set it under Config>State.

[Verfur]

Are you talking about litle tiny line movements? If so the only way is to make the increments smaller but this realy adds to the legnth of the program some of the more $$ machines will follow nurbs or curves or so I have bin told. I get stuck with the same issue a million and a half little bitty lines. That is why 3D work cost so much more it realy ties up the machine. There is the posibilty if it is a standard arc to machine from a diffrent plane Like XZ isted of XY just a thought.

good luck

John

[rweatherly]

I am using a .25" mill. I machine one line, and increment mover over .25", and machine the next line. The result is a smooth z variation along each line, but between the lines, there is a step between the z values of one line and the next.

[ger21]

Sorry, misunderstood you. Where did you get the data from? Do you have a model?

[Verfur]

if you are using a .25 mill then a stanard step over of 40% of the diam is a good rule of thumb. when loaded you may be pulling the end mill side ways wich tilts the bottom of the mill. try less cut and slower feed.

unless you are using a ball nose mill for 3D. still tring to picture it.

See if that helps
John

[rweatherly]

Gerry-
I down loaded DEM (Digital Elevation Model) data from the USGS, which is in longitude, latitude and elevation, scaled it down, and created the G-code for the model. All of that works good. I guess the best way is to just sand it down to approximate the terrain.

[rweatherly]

John-
Right now I am cutting foam to check things out and am using a standard end mill. If you picture trying to machine the slanted roof of a house, you get a stair-step. Like you said, there are ways to cut down on the steps.

Richard

[Verfur]

Ah got you now, sounds like a place for a ball nose End Mill.

John

[ger21]

John, I think what he's saying is he has the gcode, but can't recreate it with a smaller stepover.

Rweatherly, not sure of the overall size, but using a larger diameter tool will smooth it out, but you may lose too much detail.

[rweatherly]

John-

Ball nose would be better, as the regular end mill is (on the roof example) only accurate in the Z axis at the centerline. At the edge of the mill you have either excess material or have cut too deep. The ball nose reduces this, but does not eliminate it. I guess there is no easy answer, but it looks like a generic problem with any 3-D object like this.

Richard

[rweatherly]

Gerry-

I wrote the conversion in Excel, so I can manipulate the stepover. Smaller steps will help, so will going to a smaller mill, which will make the steps smaller.

The fundamental problem seems to be machining a rounded object with a 3 axis machine. It looks like to me that it can't be done with any type of end mill, but it would be better with a ball nose.

Richard

[ger21]

I meant to say ball mill. Always use the largest size ball mill you can without losing detail to get the smoothest finish.

[ViperTX]

Either is going to give you a ridge of some type...whether it's a step or a peak. You'll still need some post-finishing operation. It's almost like you need a ball mill with a built in flap sander.

[whiteriver]

On DEM's I cut them with ball mills from 1/4" to 1/2". The one in the attached pic was done with a 1/2" at a 20% stepover its 22"x36"x3" approx.

Donny

www.whiterivermfg.com

[rweatherly]

Donny-

That looks great. I have gotten far enough to scale the DEM data & write the code, but I am going to a coordinate and machining the "elevation" at the centerline of the cutter. Do you do any processing to account for the cutter away from the centerline? (For example if you are machining at the base of a sheer cliff, the centerline of the cutter would be at the base of the cliff, but OD of the cutter would be taking some of the sheer wall off)

[whiteriver]

I export the dem into a stl file and open in Visual Mill. Then just use 3 axis roughing and finishing to generate the code. It takes into count were the bit is in relation to the solid model so it cuts correctly. At the base of a cliff the radius of the ball mill would be left uncut. So it wouldn't cut into the wall. Just leave a little material at the bottom. You can go back and clean it up with a smaller bit.
I take it you are writing your own code.

Donny

www.whiterivermfg.com

[ger21]

You might want to look at MeshCAM. http://www.grzsoftware.com For $100 it will offset the tool paths by the tool radius, and do roughing and finishing passes, with different tools if required.

[rweatherly]

Yes I am writing my own code. I have Mach2 now, and am doing the code in Excel. I may try writing some code to calculate the the offset to get the best approximation of the terrain surface.

[Phule]

Can you set your tool paths perpendicular to the ridges in the terrain? By not cutting each elevation contour individually all at once, but cutting across the slope rather than along the slope may alleviate your problem.

I don't know if your software will allow this though.

my $0.02

jim