4 axis surfacing problem

[Carbonwerkes]

Hey guys

Im having a heck of a time with a 4-axis surfacing path that ‘should’ be simple. The part is NDA so i cant post an image, but conceptually, in the spirit of Superbowl weekend, it is very similar to just a rectangular section of a football, 6" x 3" by about .5" thickness, which includes the laces mid-body.

So, yea- I have a 3D smooth surface with an island that has fillets around it.

I am just not able to find a baseline toolpath approach that works, where I can run it 4-axis with a bull endmill for clearance (speed and smoothness), and then deal with the normal cleanup of the periphery of the island later.

My understanding of the Multiaxis Triangular Mesh is near zero, having little experience with it. I read several posts here which imply it is a 3axis path, so Im confused as to why it is listed under Multiaxis Pattern, and why it has options for 4-5 axis output. Regardless, even with it- just not having a lot of luck. I can get 3 axis output from it which looks wonderful- constant cusp is about a nice as you could hope for. But, the only time the output will call an A axis move is if the collision control is enabled and tilt is required to prevent a gouge at an edge of the part etc. It doesn’t cause the tool to run tangent to the surface, even when Tool Axis Control is set to 4-axis set to ‘Surface’ for its 4^th-axis bias…

Multiaxis Application->Roughing works reasonably well but it seems I cannot lock the 5^th axis to 0, so I end up with nasty gouges as the surface (tangent) bends further and further away from the XY place.

Any ideas/direction? I have not really had to deal with a 5axis surface before- but it just seems to me that if I can emulate it via 3-axis, I should be able to emulate it better with 4 axis- reducing cut time at least for that section of the part which is more or less concentric with the axis of rotation.

Thanks guys! Always appreciate the help.

[motor-vater]

There is alot to that toolpath for sure and it can do almost anything. But more info is needed. Are you on a 4 axis or 5? as in 3 axis with a rotary? or a horizontal? are you looking to machine all 360 degrees of the football? what shape is the stock? etc... As much info as you have given still really doesnt draw a clear enough picture for me to try to get you going. I have spent alot of time with this toolpath lately, its very versatile.

[Carbonwerkes]

Hi Motor!

Yea, sorry, it is tough to balance too much with not enough info for a new thread post…

So the actual part is a knife handle- in 3D. It has a very shallow radius- about 5”. It has a protrusion flat/island to support a pocket clip.

The op is simply to convert a 2D waterjeted blank so a shallow dome, for later texturing. There is a tiny amount of radius in the lengthwise direction- call it a 10” radius or so. This seems to be the aspect that is causing trouble.

The machine is a 3-axis VMC with a 4^th axis trunion oriented along the X axis. The handle being cut is oriented in the same way- to permit 4th axis work on the primary radius of the dome-

The goal is to try to get the handle smooth-enough to sand, where I can live with some cusps, but not with leaving 0.01” material on the extreme corners so that the mid-body can be cut to within a few thousandths. The material is 6AL4V, so hand-sanding it sucks.

Multi-axis roughing seems to work about the best of anything I can get to work. It seems to care a whole lot about the geometry of the containment boundary, which seems to like to be elevated (as with a more typical 5 axis Flow op). I don’t know what that is- perhaps it is limiting the A axis rotation and also the X/Y travel to a good average…

It just seems really weird to me that there may not be a basic multi-axis surfacing routine. It seems that I should be able to just define a 3D surface, define keepouts in whatever way is required (collision geometry, chains, whatever), and then define a pattern where the cutter remains as normal/tangent to the surface as possible, without a whole lot of effort otherwise.

Thanks so much for any insight you can share- Im 2 days into trying every trick I know, and it just seems all those tricks are pocket-related!

Best

[motor-vater]

how it it being located on said trunion, what planes do you have to use if were just to attack it as a 3 axis? multi axis paths are tricky, containment plays a role, how you drive it via curves, splines, surfaces, or lines also plays a role. Most of the toolpaths like triangle mesh have multiple places asking you what axis it rotates on, what axis you would like to retract on and so on, lots of things to experiment with so none of that helps either. Some times you have to create extended geometry and different boundaries to trick it into doing what you want. Still struggling to picture it here but if you need the tool to hit multiple angles along a pass to define a feature, you might need to drive it with a chain or spline that can make the tool tilt away from said feature to keep the angle you want. Or sometimes you create lines perpendicular to the features that would represent the angle of the tilt and drive with path with lines instead of a chain/spline. Sometimes you get lucky and just driving it off the surface works but not often enough. I'm just throwing things out there to get your wheels turning but if you find time whip up a simple version of a similar part and post the file, you will find all the help you ever wanted, seems like everyone on here enjoys a good challenge.. lol

[Carbonwerkes]

The part is aligned lengthwise along the X axis of the machine. It is running as Top plane where the handle is parallel with the table as X/Y. The rotary allows me to do far more efficient cutting in this orientation, since most of the radius/material removal can be addressed in that orientation.

Attached is a generic example of the model. The handle in this example is basically domed in one axis. In the model Im working on, there is a slight amount of radius over its length also- so it is slightly more thin at the midline of the tail as the midline of the center of the handle.

As you can see in the example here, there is an island which is proud of the rest of the body. This is very similar to the island in the model here- and poses a bit of a problem. For the opposing side, there is no island, so a Flow toolpath seems to work quite well. Flow has given me a ton of problems with this island, so I was looking for an alternate pattern/strategy. And all I have tried seem to be really unhappy with this slightly domed surface in the longitudinal axis. They are all fine if this is just a pure 4-axis (i.e. a cylinder).

So maybe I just need to really focus on Multiaxis Flow to try to get it to work better with the island. But it just seems like a very open surface like this with no hollows, no sharp notches etc should be very simple to toolpath. But- nope.

[Colin Gilchrist]

The thing that is hard to wrap your head around with the 5X paths, is that you often get better results by having a separate set of Drive Surfaces, which are not the same as the Compensation Surfaces.

For a part like yours, I would create a very small diameter cylinder (ruled surface from 2 arcs), that fits completely inside the handle surfaces, and is centered on the rotary axis.

When you drive the path, the "drive surface(s)" provide the "cut pattern". The cut pattern is the motion of the tool, as it makes cuts.

The "compensation surfaces", are whatever you want the cutter to "leave behind", or be tangent to.

Finally, you have Tool Axis Control. This is the method for orientation of the tool, in relation to the shape being cut. In the case of using a 4 Axis machine, I would use the "Surface" method, and pick the same Cylindrical surface for Tool Axis Control. 

 

For Roughing, "5X MultiSurf" is the oldest "5X roughing" path there is.

In general, the Triangular Mesh path is a 3X path. That is why you are having so much trouble getting it to wrap around a cylinder. Now, what I mean by "3X path", is that the Cut Pattern, how the path looks to "see and cut" the material, is generally from the Toolplane direction. The reason it has multi axis control, is to tilt the tool to avoid a collision. Even though you can tilt to avoid a collision, it still doesn't change the fundamental pattern that is generated on the surfaces or solid faces being cut.

Honestly, many paths are capable of generating a cylindrical cut pattern. Try putting a straight line down the center of the rotary Axis, and use the 'Cuts along curve' path to generate the Cut Pattern. Use a Cylindrical surface as the Tool Axis Control, and use the 'Collision Avoidance', with the handle surfaces, and the method set to "retract along tool axis". The "Cuts along curve" path takes any wireframe chain of geometry, and steps along the chain (based on your stepover distance), then it creates a perpendicular plane (at the intersection point of the chain). Wherever the plane intersects with the surfaces, tool motion is generated.

For the fillets around the boss, make a Toolplane which is aligned with the rotary Axis, and has the boss facing "up" to the spindle. Create some wireframe chains that are offset from the edges of the boss face, and use Surface High-Speed Project. Drive the curves, and compensate to the Surfaces. This will clean up the fillets really nicely, and Project gives you total control over the tool motion (you create the Cut pattern with wireframe above the part).

[Colin Gilchrist]

Could you make some of your own generic (fake) surfaces as an example? 

For 6AL-4V, cutting the part by spinning the rotary, is not the best option. The part will cut better with more rigidity. I'd recommend 3 + 1 positioning, where you index and clamp. Use Dynamic Opti-Rough, from 2 sides, with the narrow body facing up to the spindle, and step-ups enabled. Having the rotary brake engaged while cutting will make the 4th axis significantly more rigid. There are probably 5-10 different 3X paths that would finish each side of the knife really well and they would leave an awesome finish. Are you trying to leave the ridges in the surfaces, or do you want it smooth?

[Carbonwerkes]

Hi Colin

As always- I appreciate your insights and support here.

Thanks for the clarification on the Triangular Mesh stuff. I don’t know why this is placed in the Multiaxis- though I suppose the ‘tilt’ avoidance benefits of those paths can only be realized on a 4+axis machine.

For this particular project, the setup has 3 billets along the Y axis. That offsets 2 of the billets from a conventional rotary definition, since the surface cannot be defined by a radius from the axis of rotation. But I suppose I could replicate the cylinder you mention by the same offset seen with the billets, and re-use the path approach? I don’t know if the cylinder must be on the axis of rotation or if it just serves to generate vectors for the tool which are later resolved in 3D space.

I have had some difficulty getting the multi-axis patterns to leave the island alone. They don’t seem to respect the idea of a check surface- at least not very precisely. Is there a particular pattern/strategy you think I might be better off applying? Flow works fairly well for the opposing handle which is absent that island- but I just have nothing but trouble trying to apply it to the example here. The Roughing Multisurf creates a very nice-looking path- but yea, it just is stuck in 5-axis mode and there is no way out of that it seems.

If I can get the island to be left unscathed- really neat idea about doing an offset chain from the edges of the island top (mine is flat/parallel with the table/raw stock). I have had great luck with Project in applying textures to shapes like the handle in question- and it had not occurred to me to leverage that as a 3D fillet cleanup method for the island…

Kind regards

[Carbonwerkes]

44 minutes ago, Colin Gilchrist said:

Could you make some of your own generic (fake) surfaces as an example? 

For 6AL-4V, cutting the part by spinning the rotary, is not the best option. The part will cut better with more rigidity. I'd recommend 3 + 1 positioning, where you index and clamp. Use Dynamic Opti-Rough, from 2 sides, with the narrow body facing up to the spindle, and step-ups enabled. Having the rotary brake engaged while cutting will make the 4th axis significantly more rigid. There are probably 5-10 different 3X paths that would finish each side of the knife really well and they would leave an awesome finish. Are you trying to leave the ridges in the surfaces, or do you want it smooth?

Hi Colin

I can do that in CAD. Im not very comfortable with MCAM’s Cad functionality, so modifying the part enough to not violate NDA, while maintaining enough of the problem to be useful for an example, is outside my comfort zone within the MCAM environment. Some details, including the shape of the island in question, are uniquely identifying, and that would all have to be obfuscated. If I cant get something to work a little better than Ive got the opposing face cutting, it may well be wort the effort to build a model up which demonstrates the same core problems, and post something up so perhaps someone could have a look.

For the surface, normally Im just roughing, and it doesn’t matter if there are cusps etc because the patterning (which it typically ball mill) cleans all that up. And that is why I have not encountered this problem until now- I knew there were some cusps being left from the 5x Rough, but I didn’t realize how bad they would get at the corners when I started removing more material (read: creating a tighter radius). Now it isnt just a few thousands error- it is a hundredth or two...

For this particular variant, the handle is to be smooth, with a much less dense pattern applied (more like an engraving). So, Id like to get the surface close to final both to save time in hand sanding, but also to keep tolerances tighter. If Im +.001 mid body and +.01 near the 4 corners- it is difficult to sand that such that the patterning doesn’t give you away.

Best