[awerby]

Alright, you asked for it.

First, in general, I am looking to do alot of 3D work...guitar necks and carved hollow-bodies, asymmetric bowls, other strange shapes. Some of this work will require good accuracy. Alot of my questions come from this perspective. I currently have only used Fusion360 for CAM so that has shaped some of these questions.

Basic questions:

1) What are the sorts of things that people do to avoid stupid setup mistakes? I have already have a couple of instances where I have set my zero point incorrectly and that has resulted in either crashes or machining errors. Any sort of checklist for this? I would like to avoid these sorts of mistakes in the future, and I have a feeling that I haven't made all of the common mistakes, yet.

[Sure, make a checklist. Is the material securely held down? Are the zero points accurate? Is the tool tight in the collet? Is the dust collection running? I could go on, but you can do this yourself.]

2) Related to (1) above, what methods do people use for setting the zero point for a job? Clearly the centerline of the spindle needs to be calibrated to the workpiece in order to machine in the right position, but a router bit is a really weird shape and hard to get centered on anything. I have been using a 1/4" steel pin to find X and Y, but that requires that I remember to offset by 0.125" before I set the zero. I've already forgotten that once. I also have been using a piece of paper to determine when the pin is getting close to the workpiece, (i.e. it pinches the paper) then I offset by 0.003" to account for the paper thickness. This is not a new problem...there have to be tried and true ways of doing this.

[If you need to be accurate, use an edge finder to set X and Y. This goes in the spindle instead of the tool, and gives you a reference that won't break if it overshoots the mark. You do have to remember to offset by half the diameter (put that on your checklist).]

3) Do people generally mill their workpiece to size using conventional woodworking tools or use the CNC? I can already see that this will depend on stock size, clamping methods on the CNC, what conventional tools are available, etc. so I'm sure there is no single answer here.

[Right; it usually helps to have a little excess material to hold onto that won't be hit by the tool. But it depends on what you're doing, and how you're holding things.]

4) What work holding methods do people use and for what reasons? I've seen hold-down clamps, edge clamps, screwing right to the spoil board, vacuum. This may be something I just have to work with until I get more experience. Edge clamps make me a little nervous because they don't seem to have a good way to push the stock down into the spoil board. Do they work?

[They can, but it depends - on thin stock, they can bow the material upwards. Vacuum works better for things like that, if you've got it. But if there are lots of little pieces, or a lot gets removed in the cutting process, then it can start to be overwhelmed. ]

5) What sorts of tricks are out there to make setups easier and more accurate in general?

[You're right, someone should write a book...]


More complex questions:

6) What are the real-world differences between various tool path strategies? Fusion 360 (like other CAM packages, I'm sure) has a variety of strategies for creating tool paths, and many seem to have similar purposes. It is clear to me that some of these were developed simply to accommodate the geometry being machined. Others were created to reduce the time needed to machine or to somehow improve the quality of the machining. I am also confident that the best way to learn is to try different things, so maybe that is the real answer.

[You've answered your own question. But if you think about it, some of the logic behind using one or another strategy will become evident.]

7) Is it generally better to try to complete machining geometry with fewer, more complex programs or with more, simpler programs? An example of this would be carving out the inside of a shallow bowl. I can understand wanting to rough out the shape with a fairly large flat end bit, leaving some material behind for a finishing pass with a ball-end bit. This requires two programs and re-zeroing the bit in the Z direction for the second program. The alternative would be to program the roughing and the finish pass with one program using the ball end bit. Maybe the CAM software doesn't really provide good options for this, so the first option might be the real answer. Are there standard ways of figuring out this sort of stuff, or should I just learn the different schemes and start being creative with how I use them?

[There's no particular reason to rough out 3D shapes with a flat end bit. A big ball-nose will work just as well, without creating "stairsteps" that don't machine away as smoothly as cusps in the final pass. If there isn't too much detail, you can use the same bit to finish as you did to rough out; results will be smoother than if you used a smaller bit.]

8) There seem to be many ways to control certain aspects of machine operation. Specifically, to minimize the jarring motion that the machine can produce when it changes direction, the accelerations can be adjusted either in the motion controller or the G-code (via the CAM package). I have also seen settings in both that relate to how accurately the machine will follow the ideal path. Where is the best place to control these sorts of things, the motion controller or the G-code? Maybe this depends on what CAM package and motion control software is being used (among other things).

[You tune the motors in the control software; that's where the acceleration is set, which determines how fast a move is at the outset. Velocity, which is the eventual speed that's achieved, is addressed in the CAM software, but the G-code can't vary the acceleration on the fly. You can also program the machine to do "continuous contouring", which makes things smoother at the expense of absolute accuracy, since it tends to cut corners. "Exact stop" will be more accurate, but a lot jerkier.]

9) I've seen alot of discussion about feeds, speeds and depth of cut. I am familiar with concept of chip load and how to calculate it. There still seems to be a certain amount of intuition that goes into this, however. My system is probably powerful and stiff enough to run a 1/2" 2-flute bit through hard maple at 600 IPM with a 15000 RPM spindle speed with a 1/2" depth of cut. The real question is whether that would actually produce a good cut and whether the bit will break. This is just an example, but I don't have a good way to know if I should program my machine to do this. From my experience with manual routing, I would NEVER attempt this, but the CNC holds the spindle much more rigidly than I could, and I couldn't move accurately at 600 IPM anyway. Do you really have to rely on experience to figure this out? I'm not looking forward to seeing a 1/2" bit breaking off, so I have limited my feed rates and depth of cut to reasonable limits so far. Is there any better guidance for depth of cut and step-over distance? Is there a rule of thumb related to the cross section of the bit?

[These things vary with different machines. A large commercial router in the $100k range might be stiff and powerful enough to do a cut like that - a DIY machine made with skateboard wheels running on plumbing pipe not so much. It sounds like yours is somewhere in between. It's usually best to start out with a reasonable speed (not so slow the material's showing dwell marks) and a shallow depth of cut, and if it seems to be working fine, trying a slightly more aggressive cut next time - until something breaks or grinds to a halt, at which point you back off...]

10) Am I just over-thinking all of this? ;-)


Thanks,
Robert

[Not really; these things are worth thinking about, but you're still going to have to experiment to find answers that work for you.]