What's new
What's new

New to CNC, G-code practice or straight to Fusion?

MachineAmateur

Plastic
Joined
Jan 11, 2021
Hi,

New hobby machinist calling in. I have some backround with manual lathes and feel very comfortable with them, but i eventually grew tired in its natural limitations (contour shapes, large material removals, threading etc.) so much, that it removed the fun out of it all eventually. So i went out and bought Brother TC-S2B (mint condition) and generic Kanmen CNC-320 lathe to revitalize the joy.

I am now wondering should i start with raw G-code or is it preferable to skip the old school methods and stick to Fusion 360? I would want to be able to produce some basic items without spending time in front of a computer first, like a plate with holes at specific locations. CNC Lathe is much more straightforward without 3d modeling, but for milling the learning curve will be steeper from what i can gather.
 
Learn code! Just because the magic box spits out a program doesn't mean it works. Fusion is a cheap buggy software and if you cant read the code before running the program those machines are in for a hard short life. G43 (Tool Length compensation) called after every tool change needs to match the tool number. Ive heard horror stories of Fusion not having H-T match and slam a tool into the table.

Being able to make edits at the control makes your value much more also. Dont get me wrong, knowing CAM is great but being able to spot errors is where the moneys at.
 
Not sure what your actual question is?

Cam will be more straightforward as is will handle the processing of the G-code, but learning what/when/where the G-codes do is important too to *know* what is happening. I would (if you) pick up Fusion, learn a bit about programmin' and go from there.

Most G-code will be "standard" across controls, with a few outliers....

Here is a starter
G-code - Wikipedia
 
What they said...it's no use to have the code when you have no clue what it's actually doing. Especially when the machine spits out a code related alarm.
 
Thanks for the replies. There are lots of tutorials on G-code, but as it is ancient and IMO not very well thought out scripting system, i have wondered whether it is more common to use mostly modeling software nowdays rather than bother with writing the code to reach acceptable solution.
Reason why i am particularly looking at F360 is that i am very familiar with that environment and the simulation aspect of it looks very intuitive. I currently have full inventor 2015 licence (which was hideously expensive), i could perhaps use additional freeware tools as a layer in between G-code and 3d model?
 
In my opinion you should try to understand the g-code and be able to bang out simple programs easily. There are essentially 20 codes that are used most frequently. Once you have a template of the basics, you can build from there. I have a simple sample Brother program with tool changes, offset calls, common canned cycles (drilling-tapping) and basic contour milling with cutter comp. PM me your email and I will email you the sample program. Lathe program will be based off the mfr. and you should try to get a sample program for it. It will have significant differences and some similarities to a mill program.
The CAM is a must for anything other than the most basic programming, especially for milling. If you have a good post processor you will rarely program at the machine. You should still know what is going on with the code though.
Another recommendation is a USB interface for your S2B. It will make transferring programs a little nicer and you can drip feed from a stick. Something like a Calmotion box.
 
Not sure the OP is understanding how things work. Choose whatever CAD/CAM, simulation, post processor etc.-etc. you want. None of that changes the fact that the code running past your eyes on the machine tool monitor will be G Code. There's no getting around that. Like others have said, knowing what that code means will be useful in more ways then you can imagine here now in the OP's virgin machinist state.
 
Last edited:
My take is to learn G code. Relying on a CAM system can cripple you. if it's doing something you don't like, it would be difficult for you to fix it. Plus, even for simple shapes you would be dependent on it.
 
You should re-read brotherfrank's advice - I think it is spot on.

Take care with regard to the simulation as you watch it on the screen. It is a cartoon, always remember that. It can fool you as to how big a bite the tool is taking. Also, it really says nothing about getting your machine to run the way you want it at the start and end of the program that you post. Often you will need to adjust or change the G-code at the beginning and end of your program to get it to run exactly as you want. Thats the reason you need to know some of the often-used codes.

You might find that you want to take a snippet of G-code that is generated by your favorite CAM program, and add your own front- and back-ends to it to make your own sequence of machining. Knowing the basics will get you a long way. You might even appreciate this old scripting language after that! :)
 
Learn code! Just because the magic box spits out a program doesn't mean it works. Fusion is a cheap buggy software and if you cant read the code before running the program those machines are in for a hard short life. G43 (Tool Length compensation) called after every tool change needs to match the tool number. Ive heard horror stories of Fusion not having H-T match and slam a tool into the table.

Being able to make edits at the control makes your value much more also. Dont get me wrong, knowing CAM is great but being able to spot errors is where the moneys at.

I agree with learning G-code, but I have lots of hours running Fusion generated code on my Fadal, and not once did it’s crash due to a tool height error from Fusion that wasn’t a direct result of me forgetting to select the right offset. Maybe I’m just overdue.


Sent from my iPhone using Tapatalk Pro
 
One option not mentioned.
Use the Cad/Cam to produce the code.
Go through this output line by line trying to understand each and every move and command.
Be the cnc control, understand each code or look it up and then get it.
The CAM shows you what is going to happen and staring at the code long enough it will make sense.
Knowing what is going to happen and looking at g-code that does it to decode is easier that a blind guess when writing.
Bob
 
All of the above info is good, but to really dig it, you should write some small routines and run them, then you understand and appreciate what you have done. Good luck.

JH
 
Hi,

New hobby machinist calling in. I have some backround with manual lathes and feel very comfortable with them, but i eventually grew tired in its natural limitations (contour shapes, large material removals, threading etc.) so much, that it removed the fun out of it all eventually. So i went out and bought Brother TC-S2B (mint condition) and generic Kanmen CNC-320 lathe to revitalize the joy.

I am now wondering should i start with raw G-code or is it preferable to skip the old school methods and stick to Fusion 360? I would want to be able to produce some basic items without spending time in front of a computer first, like a plate with holes at specific locations. CNC Lathe is much more straightforward without 3d modeling, but for milling the learning curve will be steeper from what i can gather.


I just graduated from a machining program at local vocational college. I had similar questions to you. I took 2 Solidworks classes and asked the instructors whether I would need to learn code when I can just use a CAM software to convert the CAD (solidworks ) design. All of them said you have to some understanding of code , because you will need to tweak it and will need to understand what is going to happen when the machine tries to make your part.

We mostly didn't write our own code at college, but did write some simple code , just 3 axis linear and circular interpolation.
 
Thanks to you all, i also received some sample code from Brotherfrank for my particular machine. I am going to refrain myself from milling and get familiar with lathe first in order to get better in touch with the basics of g-code.
Modifying a generated code is most likely the optimal solution, but i do not understand the fundamentals enough yet. Will be buying some tools for the mill in the mean time and reading manuals cover to cover.
 
If you don't understand the G-code how can you know how good your post-processor is? I find the post-processor is by far the weakest link in the whole chain. By the way, I think you need to learn it all, so dive right in!
 
Learn both.

Seriously. You have to understand at least the basics of your Code to be able to, say, edit out an operation, or copy and paste one in.
I ran a bunch of programs that had huge file size due to sloppy imports of vector graphics, and in a lot of cases, the files were able to be split up at a convenient point along the way, into sections that would fit in the memory.

On the other hand, even a low end CAM system can do things that would take you hours to hand code, doing say, blended curves and 3D surfacing ops.

Unless you are doing only the most basic stuff, you pretty much have to have at lest a passing familiarity with about every way you can access, to make the machine do what you need it to.

Also, it won't hurt to look over the conversational programming, if your control offers that.
 
I know this is repetition of what has been said here a;ready, but think of plunking a part in a vise, facing it, end milling it to length and drilling two holes. Kinda overkill for CAM I'd say. But an important skill to master. Not so much the code as the ability to set up your machine. Cam is great for long winded programs and fancy geometry, but there still is a lot of really basic stuff you can make money with just simple G code.
 
So, here's a question. Do those machines both have post processors for Fusion? You can get most of the way with a generic Fanuc or Siemens post, but it probably won't be perfect.

I'm actually having to pretty heavily modify a post from a different vendor's Fanuc-based lathe to get the correct output for live tooling, subspindle, etc. Hopefully, you'll be covered with the post library because I can't imagine most people are out modifying posts. Maybe if you need a small change I could help out.

Also, the lathe functionality seems pretty weak in comparison to the mill side.

Fusion has a web site with more post processors you can look into. Just search for "Fusion post processors", or just go to Post Library for Autodesk Fusion 360 | Autodesk Fusion 360.

Once you can trust the post processor, I like the suggested option of using Fusion to generate code and then look through it and learn from that. Well, maybe after looking through whatever examples might come with whatever vendor-specific documentation you have. Everyone's right that you'll need to be able to at least glance through the generated code and verify it's not doing something dumb.

... although, the post processors are generally going to do things the brute force way unless they're highly optimized. By that I mean generate a lot of movement commands instead of a one-line cycle... that kind of thing. So, there's that to think about.
 
Thanks for the replies. There are lots of tutorials on G-code, but as it is ancient and IMO not very well thought out scripting system.
I wouldn't say it's particually ancient or bad, ultimately any fancy pants CAD CAM software is spitting out G-Code and that's what the machine control wants, it works just fine if you understand what the various codes do and how to use and manipulate them to get the machine to do what you want it to.

I would consider understanding basic G-code programming to be essential, I started out trying to use CAM software to generate programs but quickly learned that for what I was doing (basic 2 axis lathe stuff) I could generate a more efficient program more quickly in notepad once I understood what was what. I think in the past 5 years I've used CAM once for a two axis lathe part and I've never used it for any 3 axis mill parts, granted I've only ever needed to write fairly simple and short programs for that.

Realistically I think you'd find most of your lathe parts would be easier to program with fingercam if/when you learn g-code programming and the milled parts will typically be easier with CAM and a good post processor then a bit of fingercam to refine the programs.
 








 
Back
Top