Fusion 360 CAM post processor for MPCNC?

Hi all,

I tried unsuccessfully to create a tool path for leveling my spoil board. The posted G-Code is completely wrong though the simulated tool path looked to be correct. I am using GRBL for the post processor. Fusion must need some other post defined.

Also, I notice that the controller didn’t like G21 for metric. What’s up with that? I am using the Mini RAMBo.

Thanks in advance.


You really should stick to my tutorials and estlcam for a while.

As far as I know you can not run GRBL on a mini rambo, if you bought it from this site it should be running Marlin. We have made a PP for fusion, found on the basics page. If that doesn’t clear it up I can help you much more with estlcam. There are simply too many ways to mess up fusion settings for me to be of any help.

Thx Ryan… I have been reading the thread about the Fusion post-processor. It looks interesting.

Yes, I bought the board from you but am still a novice on the terminology… so not GRBL… yes Marlin. This implies that both are code that runs on the board… I presume a modified Arduino.

As for ESTLCAM goes I was hoping not to have to learn multiple systems. However, if that’s where you can provide the best advice, then who am I to argue! I’ll take another whack at it tomorrow. In the mean time I have Replied in the Fusion 360 PP thread and perhaps can gain some insight there. I did CNC programming for Lockheed & Hewlett-Packard for 15 years but it’s been 20 yrs since I have written any tool paths. The CAM systems like MasterCam and their ilk have taken the fun out of coding. Given the nature of 3D I suppose the features are necessary.

Anyway, I love the simplicity of your design but I am am quickly learning the limitations of plastic. I have already broken the lower piece on the XYZ assembly and had to reprint. I started with 12" legs and have cut them twice again down to 8 1/2" and in the process have over tightened all of the feet.



Hi Gary,

Don’t worry about your firmware. Marlin is a good choice (and currently the only choice for mini-rambo). I’ve used both and there are benefits to either.

As far as estlCAM vs. fusion, you might be ok in fusion, but estlCAM is super easy, especially if you already understand CAM basics. It sounds like you’ve got the PP working in fusion and know what you’re doing.

As for Java and Perl… Yuck. :). I use C++ and Python at work, so I think we are mortal enemies. Good to see any more software developers around here though.

Fusion 360 is 5 software sort of bundled. Fusion CAM is truly difficult to use for such a simple machine as ours. If we had some tool auto tool changers it might be worth it to spend the time on. I actually find some of the things Estlcam does better…carves. Nothing you learn in estlcam isn’t transferable later. In estlcam you just pick a tool path and a depth. In fusion you have to set up your work piece and pay attention to 3 planes for every path, and I believe 4-5 tabs of settings. Estlcam will let you use you machine in minutes. Fusion…I still am not to the point to get a cut ready in under 20.


I hope you are only milling foam? 4" big for wood, 8.5" is big for 3D printing, I couldn’t even imagine trying to mill anything other than HD foam at that height. Most CNC’s are in the 3" range. If you use large pieces of wood it is best to accommodate them with the table not the machine, drop table. If you actually broke a part, you should check them all. I have never actually heard of a part breaking other than over tightening the bolts or screw, never from actual use.

I have to agree with Ryan. I do my design work in Fusion, and have done some CAM for my MPCNC with it, but it’s a real pain for anything 2D or 2.5D. EstlCAM is so easy to use for those I go back to it every time. For 3D carving, I have had good success with Fusion, but it IS tricky. The best advice I can give is to watch some of Lars Christensen’s YouTube videos.

@Ryan, what do you mean exactly by “to accommodate them with the table not the machine”? How do you drop the table?


You can make the middle of the table adjustable. Since most bits can;t handle milling much more than 1" you can keep the machine and the Z axis short, and if you were milling the top of 6" piece of wood you just move the table down 6" not the machine 6" taller. I think Dui has a 12" drop in his table,

1 Like

The brown bed moves up and down, generally you would not use a long Z axis with it but I believe he swap out the entire Z when he mills for a really short one.

hey Jeff… that’s harsh! LOL I was never a fan of Java but my big project started in 2001 and it was a web application integrated with Oracle. Since we already owned the databases (I am a DBA as well) the sensible choice at the time was Oracle Internet Server. Prior to that project I was much more fond of Perl and have written many tools with it. C & C++ are the right choices for speed but nothing I did required that. I’ve written code in both and VB.NET and C#. You just go with the tool that makes sense for the project. However as a fan of the semicolon I can’t stand Python. I had to write some JPython… don’t like it! HA!

Now back to the regular scheduled programming (pun intended).

Yes, the Fusion PP worked fine.

Ryan… since I have to redo the feet anyway how long should I cut the conduit? Right now it is at 8.5" which is as it turns out is just about the length of the Z axis below the bottom XYZ support.

Thx in advance,


3"-4" past the XYZ is typically on the normal to long side. It all comes down to what you plan on using it for. I have info in the assembly page about size as well as the FAQ’s, and Specs page.

… keeping in mind my intent is to carve out a solid body guitar and neck. Is this perhaps too much to ask of this particular design?

It is difficult to find long bits, especially 1/8". No matter what you will have to flip it and cut from both sides as there are no single bits long enough to cut all the way through from one side…plus you will likely have pockets on both sides…3-4" is still enough.

Let’s just both agree that JavaScript is the worst.

Mostly no good debug tool I think… but yes, many of my headaches came from Javascript. Twenty years is enough for anyone.

I don’t know how it is today but pointers in C/C++ were the worst to debug… especially in recursive calls… makes my head hurt just thinking about it.

Yes, so called “dumb pointers” are very dangerous. C++ (I don’t have as much experience with C) has many parts that you just shouldn’t use unless you have a very very good reason.

In the case of pointers, you should use smart pointers (std::shared_ptr or boost::shared_ptr) instead. A lot of great features were added in C++11 (a lot of very Pythonic features, IMO).

Most of the time, even smart pointers aren’t needed if you design your objects well.

I don’t actually hate Java, I really don’t like the enormous namespaces and names. It probably also bothers me how close to C++ it is, without being C++. I am just not as familiar as I am with C++. Python is the only other language where I’ve actually picked up enough to know as well as C++. Perl is very confusing to me.

JavaScript was terrible 10 years ago, and then at some point I tried jquery and thought it was neat, and then recently I tried React for the sandify program and I am pretty happy with that. I will probably love whatever comes out in 2 more years.

BTW, have you seen Sandify? And interest in contributing or reviewing it?

Our primary app at work runs Perl.

Let me just say that as an Infra guy, perl dependency hell is a thing… and carton’s inability to recognize dependencies without failing is just plain silly.