Former Member
Guest
I must say the biggest thing in understanding CNC is translating programmed actions to movements on a machine (be it lathe or mill).
A good simulator is what is need to see the program and actions at the same time virtually (and though undesired crash virtually if needed).
Obviously Fusion, Solidworks, Mastercam as the big names. GWizardE as a small (what I have) and less costly system and Nxviewer as a cloud based free version.
By no means am I a G-code wizard, I find it an extremely poorly documented and no consistent language across platforms and machines (and this comes from one that started programming in a lot of the old languages in the early 80's including hexcode). I do let intermediate platforms do the grunt work (like those mentioned above and in my case Intercon from Centriod) and use a similuator to tweak the resulted G-code to optimize.
The second biggest secret is starting from a 0,0,0 point that and marking that on your drawings, this reference and its location to "do not hit points" determines your work holding and as all your tools reference that point. Since I figured that out I have zero crashes.
Third secret is optimized tool path, this is something that we can also apply in manual machining. The best paths provided the shortest cut time while providing the longest tool life resulting in the best finish. There is a fine balance between the three. Again good simulators are important here (and sometimes a few scraped parts for trial runs).
In my case I am running significantly faster feed rates than I ever did or could manually and achieving better results more consistently. Example fasted movement with power feed was 6 IPM flat out though cuts ran at about 2-2.5 IPM with a 0.025 depth and full 1/2 endmill cut, now I'm running 20 IPM, with a 0.050 depth same endmill, though longer and more consistant tool path which results in a more than 1/2 of the time required to complete a part consistant repeatable dimensions.
A good simulator is what is need to see the program and actions at the same time virtually (and though undesired crash virtually if needed).
Obviously Fusion, Solidworks, Mastercam as the big names. GWizardE as a small (what I have) and less costly system and Nxviewer as a cloud based free version.
By no means am I a G-code wizard, I find it an extremely poorly documented and no consistent language across platforms and machines (and this comes from one that started programming in a lot of the old languages in the early 80's including hexcode). I do let intermediate platforms do the grunt work (like those mentioned above and in my case Intercon from Centriod) and use a similuator to tweak the resulted G-code to optimize.
The second biggest secret is starting from a 0,0,0 point that and marking that on your drawings, this reference and its location to "do not hit points" determines your work holding and as all your tools reference that point. Since I figured that out I have zero crashes.
Third secret is optimized tool path, this is something that we can also apply in manual machining. The best paths provided the shortest cut time while providing the longest tool life resulting in the best finish. There is a fine balance between the three. Again good simulators are important here (and sometimes a few scraped parts for trial runs).
In my case I am running significantly faster feed rates than I ever did or could manually and achieving better results more consistently. Example fasted movement with power feed was 6 IPM flat out though cuts ran at about 2-2.5 IPM with a 0.025 depth and full 1/2 endmill cut, now I'm running 20 IPM, with a 0.050 depth same endmill, though longer and more consistant tool path which results in a more than 1/2 of the time required to complete a part consistant repeatable dimensions.