Spherical toolpaths and general info

[Pat]

Hello Group and especially HFD,
Hu, sorry to hear you aren't in the BobCAD forum anymore, you helped me out a lot, thanks again, and that is why I am here now.
I need to make a split mold cavity that is cut using climb milling with a 3/8" ball mill. The cavity is basically a cylinder with hemi-spherical ends. I get a lot of ugly lines at the bottom of the cavity where the cyl. and the sphere intersect. I am not well versed with BobCAD and I need some direction. Do I need to interpolate the section arcs?
Thanks in advance.
Pat Dexter

[hardmill]

If you can post the file on the ftp server we could take
a look at it. Its been awhile since i've used Bobcad
but i'd be happy to check it out.
PEACE

[HuFlungDung]

Hi Pat,

I downloaded your file that you posted over at Bobcad and took a look at it. I see what you mean, I also tried a few times and saw the mess at the bottom that you speak of. I even tried trimming the bottom part of your arc section lines so that they do not quite make it all the way down to the -90 degree zone, as this seems to be where Bobcad goes flaky. This still did not solve the problem.

Interpolating is not really necessary for a skin, because the skin function has its own settings for spacing and max lengths of toolpaths. Interpolation is good to do when your section paths do not have equal quantities of entities, but in this case, you are using simple single arcs for sections, so interpolation is not required.

You need to know that Bobcad is using some sort of a "NURBS-like" curve to create offsets. This is why things go so screwy. Even before the skin gets to the bottom, you can see some jaggies in the paths as they go around the bends. This would not look good on your part.

You could also simply go in and delete the faulty part of the paths.

For this type of job, a better option might be for you to, instead, do an interior offset of the chain of the bottle (allowing for your cutter radius), and rotate copies of this offset around the axis of symmetry. You would still have to connect the ends of the chains yourself to make a climb milling path.

It is exactly this type of crap that made me set Bobcad aside. I fiddled with your part for 3 or 4 hours (just amusing myself) and couldn't make it work the way it should. So I exported an iges version of your file and imported it into Onecnc and had it ready to machine in 5 minutes.

[HuFlungDung]

Pat, in an effort to show that I try to be fair with Bobcad, I have worked on your problem a little more. I have also emailed my version of the file to you.

Here are some new tips:

My aim was to skin the entire part into a series of continuous toolpaths, not a sectional version, such as your original toolpath showed.

I succeeded in doing this by making the following modifications: trim the entire (green) outline of the bottle into a chain, including making .1875" fillets where the neck begins to enlarge.

Sharp cornered intersections are typically a bad thing, both for cutter radius compensation at the machine and for skinning. Keep this in mind and make every effort to draw your parts with nice fillets, both internal and external, commensurate with your intended finish-tool radius.

As I mentioned before, Bobcad uses a "NURBS-like" offset to create skins. This means that your part's cross section lines provide sort of a "gravitational" attraction which affects how the skin lines will be drawn. With this in mind, you can see how it would be better to draw with filletted intersections, because this reduces the concentrated severity of the "g-force" when the skin line comes near to a sharp, unfilletted corner.

This is also why you need to draw enough section lines to represent the important aspects of your part. The section lines give the NURBS function more guidance about what you expect it to do because it will attempt to draw a flowing line near all your cross sections. Without enough section lines in the neck region, you may notice an unfinished "split" towards the bottom of your skinned toolpath.

I added fillets at the neck, so I also created two new section arcs at each end of this fillet. I also divided the curving portion of the neck into two, and added another section arc at this midpoint. This solved a skin split problem.

In Bobcad's case, the section lines must all intersect exactly with defining coordinate features of the path (or part outline).

In your specific part, I divided all the (violet) section lines into two, each consisting of a 90 degree arc. This is because the sections need to be selected sequentially all the way around the part, from the surface, pointing downwards. Only select one arc per choice, so when you are done, all the cross section arcs are selected with their vector direction pointing downwards.

Remember the "tail to tail" rule for activating the skin function: the first entity of your path must lie tail to tail with the first entity you select in the sections. After that, the order of selection must proceed in an orderly fashion, CW or CCW from your first selection.

What else: when you are initially using the function "Set path", you need to pay attention to the number of divisions and max length. I'd recommend an arc accuracy of .001 or finer. Max length should be about 1/2 of what your intended "stepover" will be. Proper choices here will reduce the amount of "crap pathing" you get near the bottom of the skin. Keep in mind that short segments are easier for Bobcad to trim properly when it gets working in close quarters.

All these short segments may make for a very long program later, but you should be able to amalgamate a lot of these straight areas of your path with the "Reorganize/cleanup and optimize" function applied to the skin. In fact, I know this will work.

As for making this skin one continuous, climb milling path, you need to watch what direction you intially "set path" in. It should be in the climb milling direction.

In my experimenting, in order to get a continuous path, I had to close in the "open end" of your bottle. This makes sense since Bobcad cannot create a continuous climb-milling path out of an open loop. I simply added a small arc at the end of your open bottle as part of the main path. The tool will cut air in this zone.

When you are making your parameter selections in the skinning dialog, select the option to "automatically remove undercutting". This requires a lot more processing time, but it helps to eliminate some of the mess that is residual at the bottom of the skin, near each filletted end. That part of the skin you will have to clean up manually with the delete key, or you could inspect it and let Bobcad cut it, it looks like it is in a safe zone and may not gouge

[keithorr]

NM

[Kookaburra]

Pat,

You could also try producing a spiral 2-d pocket from the bottle shape (top plane view) at the split point of your mould, then draw up the bottle profile in the front view down the centre line of the mould and use your project function, this has worked for me in the past and it it very quick.