You wouldn't cut a complex surface with a normal end-mill, you'd use a ball end-mill as you say. You'll always have some amount of scallops in that case, you control how large they are by using a fine step-over, and if that's not good enough you'd have an after machining finishing operation to minimise the scallops. Millions of injection moulding tools and dies are made this way.
Granted multi-axis machining can be better with a ball end mill, but in that case you'd be using a 5 axis machine where the cutter is angled with lead and lag from the surface normal.
4 axis lets you do continuous machining around a cylinder
For 'prismatic solids' multi axis (4 or 5) just saves you some set up time, e.g. machining on the front face, then top face, then back face, but it's still a 3-axis (if you include helical or ramped entry) machining operation. This would be an indexed table, it can be positioned between cuts but it can't run simultaneously with the cutting path
Look for Direct Metal Laser Sintering, it still has limitations but can make solid metal functional parts in Aluminium, Titanium, tool steel, and other hard to work with materials.
What would you want from Mazatrol or other extensions?
As far as I remember Mazatrol was a 'conversational' programming format that tried to make it easier for the operator to enter simple 2d paths (area clearance, finish profile, etc.). Maybe it had some better drill cycles than standard G-Code? I think G-Code was charged as an option on their controllers and some people didn't have it, which made writting post-processors a pain.
If you have a CAM system then G-Code is all you need. Some people with old controllers like using macros, but mostly that was due to limited memory on the controller (overcome with 'drip-feed' via serial communication, the machines were slow anyway) and the hope that night-operators could change the toolpath on the control if needed, but for the most part cutter radius compensation is all they needed.
Anything with 3D (x, y and Z) toolpaths mostly just needs linear feeds (G1) and coordinates. Some machines can take a NURBS curve for the parts of the toolpath, although that never really seemed to catch on well.
Maybe controllers have moved on more since I last wrote any posts but I don't think so. If anything with 5 axis it comes back to G1 xyzuv to give the angle of the tool relative to the workpiece and even less fancy cycles for the operator.
Flat plane parts would be 2 1/2 axis, height can be set per pass but only X and Y axis work simulateousely, almost like a simple pen plotter. 3 axis allows complex contours just no undercuts without special tools.
Slashdot Mirror
You wouldn't cut a complex surface with a normal end-mill, you'd use a ball end-mill as you say. You'll always have some amount of scallops in that case, you control how large they are by using a fine step-over, and if that's not good enough you'd have an after machining finishing operation to minimise the scallops. Millions of injection moulding tools and dies are made this way.
Granted multi-axis machining can be better with a ball end mill, but in that case you'd be using a 5 axis machine where the cutter is angled with lead and lag from the surface normal.
4 axis lets you do continuous machining around a cylinder
For 'prismatic solids' multi axis (4 or 5) just saves you some set up time, e.g. machining on the front face, then top face, then back face, but it's still a 3-axis (if you include helical or ramped entry) machining operation. This would be an indexed table, it can be positioned between cuts but it can't run simultaneously with the cutting path
Look for Direct Metal Laser Sintering, it still has limitations but can make solid metal functional parts in Aluminium, Titanium, tool steel, and other hard to work with materials.
What would you want from Mazatrol or other extensions?
As far as I remember Mazatrol was a 'conversational' programming format that tried to make it easier for the operator to enter simple 2d paths (area clearance, finish profile, etc.). Maybe it had some better drill cycles than standard G-Code? I think G-Code was charged as an option on their controllers and some people didn't have it, which made writting post-processors a pain.
If you have a CAM system then G-Code is all you need. Some people with old controllers like using macros, but mostly that was due to limited memory on the controller (overcome with 'drip-feed' via serial communication, the machines were slow anyway) and the hope that night-operators could change the toolpath on the control if needed, but for the most part cutter radius compensation is all they needed.
Anything with 3D (x, y and Z) toolpaths mostly just needs linear feeds (G1) and coordinates. Some machines can take a NURBS curve for the parts of the toolpath, although that never really seemed to catch on well.
Maybe controllers have moved on more since I last wrote any posts but I don't think so. If anything with 5 axis it comes back to G1 xyzuv to give the angle of the tool relative to the workpiece and even less fancy cycles for the operator.
Flat plane parts would be 2 1/2 axis, height can be set per pass but only X and Y axis work simulateousely, almost like a simple pen plotter. 3 axis allows complex contours just no undercuts without special tools.