RE: 5 axis gurus, toolpath example request.

[Max_Marz]

Hello cam friends, I have this part I'm working on that I'm trying to lay some of my first 5 axis toolpaths down on and I'm having quite a hard time wrapping my head around 5 axis in fusion.
Would someone be willing to download the f3d file for this part and drop a toolpath on it that will cut these contours correctly? Being able to interrogate the provided toolpath would help me immeasurably if anyone's got a sec.

Bonus points for a 4th axis limited path (in a VMC orientation) as that is how I will be attempting to cut this geometry.
Much appreciation in advance.

[johnswetz1982]

If you look under your data panel near the bottom there is a folder called CAM examples. These are the default ones from Autodesk. You can take alook at them and see how there selections and options are setup for multiaxis work as a beginning.

[Max_Marz]

Just went through all of the cam examples in that folder, Lots of 2d wrap examples which is good for roughing out my part but what I'm struggling with is a swarf toolpath to keep the tool co-planar with the walls of the bosses in my part, the X shaped contour.

Unfortunately the part I have I dont think can be cut exclusively with a 2d wrapped toolpath I'm looking for examples of swarf toolpaths limited to 4th axis (for non-centerline Y axis stepovers)

I know this is tricky with fusion right now but if anyone sees this part I have and thinks to themself, "I know exactly how I'd do that" I would love to see what you come up with. 

[Anonymous]

Closest I can get to 5 axis

Not sure if this is worth anything but idea is to cut blank on lathe including groove, mount it in 4th axis and machine back side by interpolating A axis.

With open slot at A0 and in line with Z axis, machine front side.

[Anonymous]

I coordinated axis in both setups to same point and direction, XYZ zero is center line and part face.

I added finish tool to back contour and limited tool path to cut only required section, also removed stock left in axial direction to blend tool path with groove done on lathe..

I should have done this in same setup but because I split body to contain tool path in front setup I decided to separate the two.

I am not yet fully comfortable with components so I work within my boundaries, if someone can build on this and show better way, it would be a bonus.

[Anonymous]

In recent weeks I was told that in 4th axis work, simulation is off by 1 degree in Y axis but actual result at the machine should be good so walls will be strait, not tapered as shown in simulation.

In my first post, I am doing something that would not add up in real world, using toothpick ball EM extended out too far to clear part, only looks good in theory.

So, in on going effort to gain some more knowledge in mill work, I came up with new scheme which I believe is good way to get it done.

XYZ zero is center line and face of finished part, front and back using A axis wrap, slot done at A0 and 3D work.

My ears and eyes are open for any input from people doing intricate mill work.

[Anonymous]

Post for mill or mill-turn.

[Max_Marz]

Alright! Monday morning... back in the saddle.
Vic that is an impressive amount of work you put into this problem. Splitting the body to control and extend contours is a really good idea and I especially like the pencil toolpath you pulled off in the last operation to do the groove, that path matched the geometry pretty perfectly. 
The toolpaths you laid to cut the rear face does violate the model pretty significantly but maybe this is to the tune of the 1 degree error you were speaking of? I'm sure you're aware that getting parallel walls in a pocket with a wrapped tool-path is geometrically impossible without significant 3d surfacing as you showed for the front. 

I'm really trying to get into the 5 axis stuff here with this part, I know in my head that it is possible to run single sweeping contours along these faces with a XYZ-A only machine there is just so little info or examples out there on five axis in fusion, ESPECIALLY with the "limit to 4th axis" option for swarf. 

[Anonymous]

Hi, don't sweat the "amount of work I put into this", these are my learning projects as well. Now if I ran this on mill-turn, walls would be perpendicular to 1.5" diameter because side of EM is perpendicular to its face, when tool is at Y0 position and sitting on 1.5" diameter,  C axis rotates along with Z axis +/- move, that fact doesn't change.

I believe same is true for A axis work, with tool at Y0 position, A axis rotating and X axis moving  in +/- direction.

Fusion simulation result looks distorted due to 1 degree error in Y axis, so I've been told.

I do lathe and mill-turn work so this is my way of expanding into mill work as I have been dragged into dealing with some projects on mill lately due to lock of better option in shop.

If what I presented is not quite on the mark, I am curious as to what seasoned mill guys have to say.

[Max_Marz]

Walls will be perpendicular when the face of said wall is co-planar with the Y axis, 
Walls will not necessarily be perpendicular when co-planar with the X axis (because of the A axis rotation)
I believe this is why the rear face is overcut slightly, tool is always pointing to centerline so when you rotate A the wall angle changes.

If the swarf toolpath can be limited to fourth axis properly and the tool can step over and down the side of the cylinder in Y I believe this can be eliminated.

[Max_Marz]

Explained much better here, but I'm sure we're just misunderstanding eachother. 

https://help.autodesk.com/view/fusion360/ENU/?guid=GUID-7FF25192-AED7-4230-92DC-E22ADF15D498

Things get weird when you transition from traveling along the axis of the tube to traveling around the circumference which is essentially what this curve is but "projected" from the top down, so the walls are vertical when looking at the part top down... hmm.

[Anonymous]

Well, the way I see it is, when tool is at Y0 and sitting on 1,5 diameter , Z axis is perpendicular to X axis.

If I now rotate A axis 90 degrees and move X axis to any point away from part origin and into direction of the part,............. did my Z axis get tilted?,........... I think not, it still points strait down and it's still perpendicular to X axis.

So by moving X axis and rotating A axis, how can wall not be perpendicular to 1.5" diameter if all along the path EM never got tilted ?

I don't think computer simulation is showing realistic result or if it does, how do you explain above logic?

If there is a science in this that I am not fully getting grip on I want to know how else to machine such feature.

I'll check out the link, never too late to teach old dog new tricks.

[Max_Marz]

This is what I'm talking about, and if you look at the part I provided top-down it looks like the part on the left, the walls are parallel to the Z axis, as soon as you rotate the part in A the walls are no longer vertical like the part on the right.
What I'm describing about changing directions can be seen in the corners of the part on the right in the top down view, the walls on the y axis are straight up and down, the walls in the X axis are tapered and the transition between Y and X you sort of end up with a kind of lofted shape.

As soon as you rotate the A axis you are losing perpendicularity unless you can step over in Y which is what swarf limited to 4th axis should be able to achieve.

[Anonymous]

Ok, this must be a "lathe guy syndrome", I have Haas mill-turn with Y axis so I can make window or pocket on side of a tube that has OD and ID of same width if measured across Y axis.

My other machine only has C axis so I cannot make same window or pocket, ID width measured along (imaginary Y axis) will be narrower because tool always points to center of the part.

So we are talking about same thing here, right?, which now poses a question, how do you machine back contour that has sharp corner but you cannot use flat EM ?

[Max_Marz]

Yeah, you got it now I think and you're right, the geometry on the rear face is essentially impossible to machine and the solution you provided for the rear is probably as good as it gets.

I think the fillet in the front makes it possible though, not only to machine as a typical 3 axis top down type part but also with the 5 axis limited to 4th.

Just working through this mentally with you I think I've gotten some ideas for the front... 

[Anonymous]

I never run from a challenge but at some point I have to ask myself, is it that critical for part function to be absolutely perfect or can there be reasonable deviation that facilitates practicality and cost ?

We are only talking about .105" high wall, considering margin of error in tool path, I'd say most critical thing is blending tool paths to hide steps.

Now, if you can 3D the whole thing in one setup, great, but getting those face contour ends is going to be rough.

I hope I gave you some ideas, but in the end I'll stick to "Lathe guy shall do no harm" policy, it's all up to you, I don't do 5 axis (yet),.... I can only hold one ax in each hand.

[Max_Marz]

It was only until you suggested "can this even be done with a square endmill?" that it all fell into place in my brain. 
This is a one-off part that was quickly designed by a colleague of mine blissfully ignorant of the nuance to manufacturing processes like this.

If it makes you feel any better, it took me about ten minutes of back and fourth with my boss (Tom lipton, from oxtools on youtube, a very smart dude) to explain to him how this geometry is actually impossible to cut. 

Spoiler: We had the part 3d printed and Tom came to me and said "I want you to try and machine this as a matter of exercise" 

Still hoping someone will come around and drop some 5axis magic to look at but this has been great in any case, thanks again. 

[Max_Marz]

When you split the body at an arbitrary angle through the centerline chopping the tip off those contours you can clearly see the problem. 
Still, as far as the front contour is concerned, you can still hit that angle with XYZ-A only.
Say you place a ball endmill tangent to that fillet and hold it there in space, now rotate your A axis and have the rest of the machine follow keeping that ball coincident with the same point, at *some* angle of A axis rotation the wall *will* match up. Going to try and demonstrate this.. 

[Anonymous]

I have another idea that may be the answer. What if you left .01 radial stock in finish pass then repeated finish passes each time lifting wrapping diameter by .005 and not leaving any radial stock.

If for instance you used bull nose end mill with .01 or .015 corner radius it would essentially produce finish of a ball end mill on a vertical wall.

You would have to construct 20 models each having 1.5" diameter increased by .005 and used as wrapping cylinder to  produce finish path until done at 1.7" diameter.

There has to be a way to beat this into submission or at least minimize the error margin to insignificant value.

[Max_Marz]

We can ignore the rear wall as its impossible geometry from the start but the filleted wall would be nice to have a better solution for. 
We can exaggerate the diameter larger to have a taller wall to work with for the purposes of demonstration too.