Pocketing, why the up-down?

[carbidetooth]

I'm pocketing out what will become rabbets and dados all at same depth. .75 material,.375 depth with .375 upcut EM. Blue lines represent pocket and red the actual perimeter of part post cutout, a outside profile cut at full depth, less onion skin.

Given the nature of the task, I would have thought that toolpath would ramp down into material and simply move in X and Y till complete. But no, it plunges in and out for reasons I'm not sure of. I've tried a few scenarios, dog bones, conventional vs. climb. It appears I could raster cut it and be at constant depth, but that seems pretty inefficient too.

Help me to understand the strategy it's using and perhaps how to defeat in this case.

[highpockets]

Can you strip the file down and post the job file with just that part including the toolpath?

[carbidetooth]

Yes, but it might take me a bit to get back to this project...many on my list today.

I thought there might be a pat answer. I cut rabbets like this on a product I sell and it behaves the same. Gets the job done, but with what seems like extraneous motion.

I'll be back...

[carbidetooth]

OK John, here's a Google Drive link to stripped down file with another component that does the same with unexpected Z movement. I saved this as a CRV file which I've never done since getting Aspire. My thought is it might make it more accessible as there are no 3D elements.

https://drive.google.com/open?id=1mf7Na ... z-IB4aRDJv

I'm very interested to hear your analysis. I've chalked it up to quirks of the pocket strategy, but would like to understand why.

[highpockets]

VCP tries to use the most efficient cut path, which in the algorithm seems to be cutting the straights areas then cleaning up the small areas missed. This seems to be the case in your project. You can play around with the Stepover setting to cleaning it up a bit, I found 70° worked pretty good.
Hopefully someone with far more experience will chime in and shed more light on this.

[TReischl]

John has it right.

The "strategy" for most pocketing algorithms is to remove all the material possible without raising the head by stepping over. I think we all get that part no problem.

However, in pockets like these there are small areas leftover that would either require the tool to stay down and move at feedrate to get to them or raise the head and rapid to them. The faster choice is obvious.

If you just want to play around you can sort of simulate how the software "thinks" by using the offset tool. On pockets like these eventually you will see the small areas left by stepover. The question then is how to get to them.