is there 5 axis

[aaron]

Hi Adam
I have zipped up a very crude model in freeCAD of my machine (the assembly workbench is A2plus) plus a photo.
Please bear in mind I am a woodworker at heart, in no way an engineer, and this is still a work in progress.
I have a video of all axis working but it is too big for the post.

Stuff you may need:
Max movement n X = 800mm
Max movement n y = 750mm
Max movement n X = 400mm
Rotation in C = 360 degrees?
Rotation in A = 240 degrees?
Distance from end of spindle collet to rotation point of A axis = 160mm

Let me know if you need any more info. If you think I am wasting my time let me down slowly!

But keep up the amazing work

Best wishes

Aaron

[aaron]

This should be a link to a video of the machine in action
https://www.instagram.com/p/CRBgPpmHSn5/
The 3+2 simulation looks brilliant
Thanks
Aaro

[adamLange]

aaron, that's an awesome looking machine! I just viewed the image in the zip file.

[truehybridx]

Hi All,

Okay I have read through this thread a handful of times and I think I am ready to try and be useful if it helps.
Got a Pocket NC V2-10 back in 2018, and only used it a handful of times due to work getting in the way, now that work from home is a thing I could have spent more time using it if it wasn't for Fusion 360 doing its thing. I work as a software engineer for a college (mostly PHP, NodeJS, C#, basic-like something -_-), with no exp in python but it's not like that ever stopped anyone. Primary Windows user, with Linux exp and can spin up a VM and try some experiments if needed.

I haven't used the Path Workbench in FreeCad as of yet, I will start watching sliptonic’s vidoes about it and see what I can make happen to get some exp using it.

Is there anything I can help with? Testing gcode, testing workbench changes, video, measurements, etc.

[sliptonic]

Welcome! Glad to have the help. There's a ton of things that need work but central control and management doesn't work on an open-source project. The thing to do, especially if you're a programmer, is to pick a project and dive in.

If you're an experienced programmer familiar with continuous integration and test-driven development but not knowledgeable about Python, a great place to start is building tests for our current code. Our test coverage is abysmal. It's thankless work but will make you familiar with the code base. But most want to work on features and I understand that.

In my mind, the 5 axis effort has two parts. First is the 3+2 milling which is just our existing 2.5D operations with the ability to rotate the model so that any face is 'up' and can be worked on. I've been working on this and it's slow going. Actually rotating the model (and stock) to calculcate a toolpath for a position is easy. Path, unfortunately, has some assumptions about the model position built in at a pretty low level so a lot of things need to change to make 3+2 rotations intuitive and stable.

The harder part is continuous 5 axis machining. What Adam and others are doing might get us there but we still have a ways to go because there are no good 5 axis libraries available.

If you just want to chat about how Path is built and where you might dig in, catch me on gitter: https://gitter.im/FreeCAD/Path

[truehybridx]

Okay this sounds good. This will be a decent excuse to actually setup a dual boot with Linux and coax myself into getting involved.
CICD I have average exp with as I manage our TFS setup at work.

As for TDD, I wish I had more exp with. We’re trying to push for more of that in theory and on paper, but timelines always get in the way.

I’ll take a stab at looking at building tests and see if I can make sense of what’s going on, once I get my stuff setup to build it.

Anyway, it should be fun and hopefully things will start to click.

I’ll check out the development areas in the forum and see if I can start doing something.

[nesquik011]

guys , should i build 5 axis cnc to help testing ? i can let any developer watch the machine while he is testing it over webcam

the cnc i would build will be XYZAB , i will make cheap as possible but make something really works , and tell me if you want mach3 controller or what ?

[Kunda1]

guys , should i build 5 axis cnc to help testing ? i can let any developer watch the machine while he is testing it over webcam

I love the maker community

[maxwang531]

[/img]

Hi Harshmacwan,

It looks like modelica is failing to solve the first pose. The most common cause of a failure to solve is when a target-to-tool spring approaches zero length. The problem might be that the first pose to solve in pass_1.py is coincident with the initial conditions in the modelica model. You can double click any joint in the modelica model and alter it's initial conditions. First try starting the x-axis joint to 0.001 m and the A-axis to 1 degree.

Please try that and let me know how it goes.

Adam

Thanks Adam

I am on it today will update you asap.

We are getting segmentation fault issue when the simulation trying to call the python APIs.

Recently I am working on a work on 5-axis toolpath generation. After seeing Adam’s modelica model, I also want to try on this basis, but I encountered the same problem as the picture above. I still I don't know how to solve it. I have installed python3.6 in the ubuntu system and linked modelika into the python path, but it did not solve the problem.
Maybe you have some better methods?

[Michaelv]

First some background.
I come from Sketchup, including developping CAM, and more custom tools, and I was going to go into 5axis. Also I'm a mech engineer familiar with software devt (not top notch software coder, but I can figure some out)
Due to sketchup limitation (not parametric, not super great for NURBS , etc..) I had slowly started to thinking about moving to fusion (taht and FEA and CFD). Now fusion (and onshape for that matter) are not friendly to us, so here I am, and happy about what I've found so far.
I don't have a 5axis yet, only 3, but working on getting a concept demonstrator (a simple light weight and cheap (less than $500) CNC, that does 5 axis machining. It won't do metal (more like foam and maybe wood), but once the concept works, beefing up the hardware to do anything is easy, only a matter of cost). It may be a couple years as much design has to occur (very prelim so far) and so many fascinating projects . Opportunities change schedule though.
On the other hand and in a similar vein, I'm also interested in non planar 3D printing which eventually will also evolve soon into full 5 axis (non planar is possible in 3, but the limitation is quickly attained (interference with toolhead sides, extrusion at angle from nozzle))
So far I focus on swivel 5 axis, which means cartesian part/blank motion (XYZ) and 2 rotational axis toolhead. Trunion and rotating table should or may be easier, but not my objective for now.
Actually three point Z table plane for 3D printers is also a very interesting engineering concept and problem, but I think XYZ-BC is more flexible, designable for now. There is some progress on 3 point Z table, but we'll see: https://www.youtube.com/watch?v=bgkK7Fez8VU

Now a bit of out loud thinking.
5axis machining vs 5 axis 3D printing:
Machining is more complex, issues of interference, tool side and geometry (use the point of a mill or the side, etc..)
Also issues of rigidity and efforts

Maybe 3D printing is easier first at it is essentially a point tool. It also has essentially no force applied (or minimal) so the structure of the machine can be much lighter and easier to design.
3D printing is even easier in the sense that it can use the classic 3D slicing for roughing the shape, and also to resolve for problematic sections (interference, convexity, etc..) and then just print the outer shell in 5 axis.
That may make for an easier intermediate milestone (surmising at this point) in CAM development
i am thinking left and right about a 5 axis 3D printer. Some design ideas have already been sorted. The printhead has to evolve (reduce bulk, change shape, etc..) I'm currently working on it somewhat.
I also thought of a reverse robot arm (hanging from the top) and thus full 5 axis access to the whole print volume essentially. But that's more complex and kinematics is probably a nightmare, so baby steps. Some have done it though https://www.youtube.com/watch?v=HaOvKdOUBPE