This is one of those times when I don't know enough about a subject to even ask coherent questions. This is all pie-in-the-sky stuff, it'll be a fair way down the track before anything happens – if ever. But here goes with just a couple of simple (ha-ha) questions, hoping any answers will either point me to somewhere/somebook/something where I can start to learn in a cohesive way, or will nudge me in the right direction for asking further questions.
1. I've never used 3D CAD software. I've never had any association with CNC
machining. I'm an EE (retired), I've designed, built, and written the code for lots of scientific equipment that required servo/stepper controls and so forth, so the general mechanics of CNC machines hold no fundamental mysteries for me. But the control was by native code in embedded micros, so while I'm aware of G-code, I've never given it more than a very cursory glance.
2. The disconnect in my knowledge is the path from a 3D CAD image to (presumably) an input file for the CNC controller (MACH3 ?). This doesn't seem conceptually difficult for a 3D object containing only orthogonal flat surfaces i.e. pretty simple X/Y/Z stuff. But when the object contains surfaces consisting of compound curves then the machine will (obviously?) require additional axes (rotation about X, Y, and Z). Are there standard assignments that associate drawing rotational axes with machine rotational axes? Hmmm, thinking about that further, it probably shouldn't matter what axes the image is drawn around. Provided the CNC machine itself has sufficient axes, it should be possible to transform the image coordinate system to suit the machine (true or not?). Although it would probably be best if the design coordinates and the envisaged workholding method were designed to map directly to the target machine (true or not?).
3. The end result of this thought exercise is to determine what hardware (CNC mill) and software would be required to design and fabricate a single blade of an axial compressor. The blade would have a cambered airfoil section, with twist from root to tip, tapered thickness, and various planforms – hence all the compound curve questions. I think 4 axes would be sufficient (although possibly not with the 4th axis as the 'usual' (?) rotation about the machine spindle axis), but in any case I need to understand how the CAD design becomes suitable G-code for a given machine. Is the operation intended to be automatic, or does it need further directing by the designer?
As I said at the beginning, I barely know what it is that I don't know. ![sad sad](data:image/gif;base64,R0lGODlhAQABAAAAACH5BAEKAAEALAAAAAABAAEAAAICTAEAOw==)
Edited By Roger Head on 13/12/2016 14:49:48
Edited By Roger Head on 13/12/2016 14:51:58