Re: Finishing Pass Problem:
Hi Bruce, I think it may be that the algo in your CAD/CAM that generates the coordinates, which are I think incremental (?), for some reason is generating a Z coordinate that has a small and varying error. I'm surprised that it is not just generating G02/3 commands that produce circular interpolation, but seems to be approximating circles with short chords generated by G01 moves.
I would be inclined myself either to write an Excel sheet that will generate the g-code for using circular interpolation; or using Excel to clean up the coordinates such as you list above, so that for each segment that ought to be at a fixed Z level it replaces each Z value with the mean of all the Zs in the segment. I would have more confidence in the first as one has more control, but you need to understand a bit about how to construct g-codes in Excel.
One thing that might work is to actually loosen the tool tolerance so that it is larger than the Z variations - that might persuade the CAM to generate a constant Z level.
Re: Finishing Pass Problem:
1 Attachment(s)
Re: Finishing Pass Problem:
Hi John,
I actually loosened the tool tolerance and it smoothed out 99% of the z variation. Turning on control points in Rhino helped visualize this greatly. Sky Greenwald had a helpful article of tweaking madCAM toolpaths:
https://betterlivingthroughcnc.com/2...y-with-madcam/
I believe the G02/G03 exists in the 2.5D axis part of the program and is planned for 3 axis sometime down the road.
I was experimenting with using a 1/2" ball to do a fine finish on a 9" round using a .005" stepover to try for a super smooth finish. The finishing pass post alone went over 100,000 blocks of line segments (but only about 2.6 Mbytes). That was enough to exceed the Okuma P200 controller buffer. (School MV2040 VMC). It is probably a good 8-10 yr old. I looked up specs on the latest Okuma controller ... P300 ... and surprisingly the buffer there is still the same! As memory is so cheap relative to big buck, heavy hardware, why they don't have at least 100Mb for a buffer is a good question I would like to hear an answer to. I would like to know how big of a post other mill controllers can handle. Maybe Okuma is just exceptionally small. Dunno. If this is typical for the industry, CAM packages without G02/G03 capability will be precluded from running larger area, low stepover jobs because of the mass of line segments created ... unless I am missing some other workaround here being relatively new.
I ended up raising my stepover to 0.025" for my 9" round to get the post size to fit the VMC buffer. I skipped going for a smooth surface and chose a scalloped finish using a 1/8" ball to get it moving. It produced a finish without the problems I had before and was still comfortable to the touch. Maybe even a non-slip texture which is really pretty good to have in my idea tool box.
https://www.cnczone.com/forums/attac...d=393962&stc=1
I would like to see madCAM implement a spiral machining strategy for this type of facing work. Joakim helped me get the tool lifts correct for each concentric milling circle by adjusting the 'Cut-Link' parameter, but if you look real close at the finished round you can see the start/end point for each circle. Very faint, but it's there on close inspection. Probably more due to the variability in the machine than the math in the software ... but a spiral toolpath would cover a multitude of hardware sins. :^) Maybe there are other ways to handle that as I am just on the learning curve.
Adding Circular Interpolation would probably be the most bang for the buck to avoid having to break a larger workpiece/small stepover operations down into regions and separate posts to fit a small machine buffer ... at least from my limited viewpoint right now. I will come back to a smooth finish surface project later as I want to find what the real limit there is with the tooling ... not computer memory. It's good to learn what the limitations are before I plunk hard earned $$$ down for a machine.
Much to learn yet.
Cheers!
B
