Model surface quality

[Gregore]

I seem to be either not understanding how to make the surface better through mesh settings or I have something set incorrectly you can see in the picture I have random inaccuracies along the surface as well as all the paths are faceted if you look at bottom edge you can see it .... for reference the grid in the back ground is 1mm and each facet is .3mm long

I have tried all kinds of settings in rhino mesh as well as absolute tolerance of .001 mm in madcam
It seems to me that no matter what settings I use the render gets no better

I assume the render shows mesh quality

Thanks
gregore

Similar Threads:

• surface quality of 3+2 vs. simultaneous CAM programming
• Need Help!- surface model vs Solid model machining
• Newbie- how to improve surface quality?

[awerby]

It looks like you need to mesh it finer in Rhino. Madcam seems to be faithfully following the facets, but you need to boost the polygon count so there are fewer of them.

[MARV]

Is your mesh generated from a point cloud taken from a scan? My experience is that scans often come with artifacts that lie outside the form and need to be elevated. Also, it helps to decimate the mesh if it has a very high point count (in the millions), before modeling surfaces. I would take a look at how you've generated your tool paths as well. I presume your images are of a finish pass? You may want to "tighten" up your tolerance setting. A tighter tolerance will reduce faceting, but will tend to grow your file size.

[Gregore]

Thanks

I noticed that changing the cutter tolerance helped a lot , I had it at .01mm and and changed it to .0005 mm this seemed to help a lot to make the tool paths smoother .

But I still really have no clue what all the advanced mesh settings really do even after reading everything I can get my hands on.

Any one feel like taking a crack at explaining how they effect the output on a tool path , or how the cutter sees the surface

Thanks
Gregore

[routalot]

I admit to knowing very little about working with meshes in Rhino.I have had a little success by creating surface patches from a point cloud and using various surface tweaks and matching/joining to improve the situation.

[MARV]

I used to work with meshes in Rhino and, although not an expert, I learned a few things. First, its fairly easy to generate smooth surfaces or accurate surfaces from meshes in Rhino, but difficult to generate a smooth AND accurate surface. So, I would decide which was the priority and work from that perspective. Next, if you're working with a mesh generated from a point cloud from a hand held laser scanner, there will be artifacts (or digital noise) that will lie outside of what would be a smooth surface and those will need to be culled from the point cloud. Beyond that, you'll employ a variety of tools in Rhino to generate your nurbs surfaces (surfaces generated from mathematical calculations) that will facilitate higher quality tool paths.

So, my workflow typically went something like; convert point cloud to mesh (that was typically done with the scanning software). Import into Rhino. From the menu bar; Mesh - Mesh Edit Tools - Cull Degenerate Mesh Faces. This removes duplicates and poor quality triangular mesh faces. Next; Mesh - Mesh Edit Tools - Reduce Mesh. Its common that point clouds are composed of millions of points, most not needed to do what needs to be done. Reduce the count to something like 30,000 or 40,000 at most (sometimes much less is better). At this point a judgement will need to be made as to next steps. I've generated tool paths directly from meshes (not great by any measure, but sometimes just gets the job done), but usually continued to 3D model from the mesh. Using cutting planes that intersect the mesh will allow the generation of sections lines. These can be traced over to be cleaned up and used to generate surfaces from networked contours. These are the nurbs surfaces that will generate the best tool paths yielding the smoothest surfaces. Again, this may not be the way a Rhino savvy tech might do it, just what worked for me.

[Gregore]

Thanks Marv.

I have never used scanned items , though I can imagine it being useful and I have thought about it , so the above info will be very useful in the future .

Almost all the work I do is nurbs surfaces generated in rhino , and almost everything is jewelry sized . Thankfully I do not need both smooth and accurate , smooth is a much more of a priority .
I find that the biggest issue is random surface texture left from cusp height on 3D curved surfaces . Most of my cutters are .254mm and smaller ball end mills . I am still trying to sort out if it is from cad , cam , feeds and speeds or machine . Too many variables to make it easy to track down .

Thanks

[MARV]

Gregore,
I don't know if this applies here, but I remember someone on a Rhino forum saying that modeling a small item in an oversized scale and then, once done, re-scaling down to full size. Supposed to help with refinement of detail.

You also mention an issue of "cusp". I presume you are referring to the juncture between 2 adjacent tool passes? Only way I know of dialing that out is to lessen stepover. Again, refines the surface at the expense of file size and run time.

Once you find what works for you, the work flow becomes a good deal easier.

[Gregore]

Thanks for the scaling idea , I will certainly try it

Yes decreasing step over works to a point on my machine but it seems I have reached a limit that increasing more leaves no better a result . Thus I am thinking either the surface of the model in rhino or vibration in my machine , I can rule out the spindle as mine is less than a micron runout.

Maybe I can see if really slowing down surface speed helps ...... I once seen a formula that calculated cutter diameter to feeds and speeds to get the cusps to all line up like corn rows

Thanks

[Gregore]

i found it for those of you who would be interested

https://www.ultrapolishing.com/mold-...lling-finishes

Thanks

[MARV]

This may not have a bearing on metalworking, but in the area of woodworking, being able to mill a surface at something other than surface normal (90º cutter axis angle to surface) tends to improve the quality of cut, lessening underlying cellular deformation and subsequent sanding/finishing operations. More easily addressed in 5 axis operations.

[Gregore]

marv sorry I missed your post .

In response I have found very little benefit from tilted passes in metal , though if one was needed that extra little bit it should be worth it .

As for wood now that is a whole different animal .
With gem stone grinding ( which I do ) you get what’s called subsurface fracturing , which is sub surface cracking up to 10 times the grit diameter . I would think something similar happens in wood though it would be the tearing out of wood fiber from not cutting cleanly .
Because the center of a ball cutter is not actually turning thus it is plowing and it would be tearing thus tilting the cutter to the surface will allow greater surface speed for the cutter. Same issue with diamond cutters on gem stones the center of the cutter approches zero surface speed , plated diamond cutters can not handle the friction of the plowing through the material.

Because my 5 axis is not and running yet , still only 4 axis , I try to use a flat bottom diamond cutter for as much stone work as possible as the center of a flat bottom cutter almost always has the the edges cutting before center

Thanks for responding , when my 5 axis is running I will try using it to cut some of my ebony parts and see if I improve the surface .