Advice & thoughts please on this 5-axis design

[Plaxton]

I'd like some advice please on the merits (or otherwise) of a 5-axis machine of the design shown below;

I'm completely new to the world of CNC (although have used 2D plotters & cutters extensively) and am considering building (or buying) something along these lines for machining medium-sized (max 1m x 1m x 250mm) shapes out of rigid foam and machinable wax. I'm sure what I really want is a rotating head machine but other than things like the 5axismaker these just don't seem to exist for the home/small business user, which is something I don't really understand.

The nature (size and bulk) of the things I want to machine don't lend themselves to a tilting-bed type machine, so I've come up with the design shown below which (it seems to me) would be fairly straightforward to build using off-the-shelf components. BUT I have no idea about whether a machine like this could be controlled & programmed using regular CNC software - specifically whether, instead of having a fixed vertical spindle, having the entire spindle body rotate either in the Y-plane (as shown) or in the X-plane (which would work just as well) is something standard software could handle. Any advice regarding this would be gratefully received - I suspect and hope the answer is "Yes" because I know that (for instance) Mach3 can drive all sorts of different types of machine, but I'm still at the bottom of a steep learning curve with this stuff and haven't yet come across anything exactly comparable to this.

Thanks in advance for any comments, and feel free to point-out any other major flaws in what I'm trying to do that I probably haven't even thought of!

<img src="https://magazineexchange.co.uk/cw/MagEx_Design/5-axis-design.jpg">

[AndyUK]

I'd be cautious jumping into five axis CAM before having experienced three axis CAM, little bit of running before you can walk type situation. Similarly it's quite an ambitious first build - I don't see why it can't be done, it's just not the easiest way to climb this mountain!

That's also probably why you're not finding commercial machines for the hobby user with large bed 5 axis capability - the high level of programming and understanding the concepts behind the CAM in combination with the mechanical limitations of a large bed hobby priced machine mean it's only ever going to be able to work on material like foam. Fine for you - but not a big market.

I'd suggest starting with a cheap premade three or four axis machine to understand the CAD CAM process first before spending a large amount of time and effort building a five axis machine to realise it isn't quite what you wanted or needed.

[Plaxton]

Thanks for that, and I was certainly aware that I was jumping-in at the deep end with this concept and had already thought of buying a fairly simple machine to start with just for learning - one of the attractions of my design I think though is that it could/would be operated purely as a 3-axis machine most of the time anyway, so hopefully could itself act as stepping-stone in the learning process. In fact, with a bit of fore-thought, the entire machine could be initially built only as a 3-axis but constructed in a way that allowed the 4th and 5th axis capability to be added at a later date

Thanks again for your comments!

[m_c]

Most things are possible.
The big issue is getting rotary motion with minimal backlash. It can be done easily if you have lots of money, but not so easily if you're on a more typical budget.
Harmonic couplings are a good solution, but new they're expensive, and although there are usually plenty available second hand, they'll of generally have been replaced for a reason, so second hand ones can be a bit hit or miss.

Machine controller software wise, there are two possible options.
One, you do all motion generation with axis/tool offsets within CAM. The big benefit of this, is you can run any machine controller that can control 5 axis, but the big disadvantage is you need to setup all your axis/tool offsets in CAM.
Two, you use a machine controller that is capable of handling the axis/tool offsets. The big benefit of this is you only ever need to generate the code once, but not many controllers are capable of handling the required kinematics.

I know Dynomotions KFlop and KMotionCNC come with various Kinematic examples, which you can customise and recompile for your specific machine. TomK (the main man behind Dynomotion) seems to love this kind of thing, and is always willing to help anybody looking to customise kinematics.

I'm sure LinuxCNC can be configured/compiled for such things.
Mach3/4, I'm not sure on their level of kinematic customisation. They certainly support enough axes.


To give you some ideas of what's possible, here's some videos of 5-axis machines running KFlops-
https://www.youtube.com/watch?v=PuIdtZKdYfA
https://www.youtube.com/watch?v=XVrsO89hoV0
https://www.youtube.com/watch?v=wBrl31IrFbE

And although not 5 axis, this video demonstrates what's possible.
The hexapod takes XYZABC coordinates, and the KMotionCNC kinematic library converts that to the required motor motion.
https://www.youtube.com/watch?v=t7KMGkfuvfM
Or there is Dynomotion's 'Geppetto' 6-axis demonstration robot-
https://www.youtube.com/watch?v=MSONYPXe3bE