Some features are ignored if the stepdown is smaller than the surface of that feature about grid-apps HOT 13 CLOSED

I've reproduced the problem at this point

from grid-apps.

ok i've reproduced the bug. will fix.

from grid-apps.

fix live. i need to refactor this code to me more efficient and thus correct.

from grid-apps.

this is not the way it's supposed to work. it's supposed to find any flat Z surface and force a cut at that Z index. at least, it used to. i'll have to look into it again.

from grid-apps.

I'm also affected by this issue, and not only are shallow features ignored, but pockets deep enough to get machined often end up shallower than they should be. I'd like to help get this fixed because tweaking the gcode by hand is no fun. @stewartoallen do you have a known working commit so I can do a git bisect?

from grid-apps.

i'll look at this today. @andygarylee please (re)send me the stl that you had this problem with.

from grid-apps.

I can provide a sample which I think is affected by this issue: https://gist.github.com/lgeek/bee61124ce396b68e11f2f8683397010

Using the kiri:moto instance at https://grid.space/kiri, with 0.15mm tolerance, I get the following results:

• with 2mm stepdown, the pocket is machined in a single pass at -2.5mm depth
• with 4mm stepdown, the pocket is machined in a single pass at -3.6364mm depth (this is still out of the specified tolerance from -3.9mm)
• with 5mm stepdown, the pocket isn't machined

from grid-apps.

checked in a potential fix. version should show up as 1.2.4d. let me know

from grid-apps.

I can confirm the code for milling the pockets now appears to be generated correctly for a couple of models where I remember I previously had this issue. Thanks so much for fixing this.

On a related note, if g-code is generated with a stepdown higher than specified (for example going directly down to -2.5mm with a 2mm stepdown setting like above), should it be considered a bug? Should I open a new issue for that?

from grid-apps.

i'm not sure i understand. if you provide a stepdown greater than the distance between detected flat faces, it will ensure the flats are handled. that may cause an intermediate pass you don't expect. if the stepdown is less than the distance between flats, it will attempt to interpolate between the flats at a distance <= stepdown. in other words, stepdown is the max stepdown value, not the guaranteed stepdown value.

from grid-apps.

I'm seeing some passes (or at least the first one) deeper than the stepdown setting. E.g. if I take the stl from my gist linked above, and I generate the rough cut with a stepdown setting of 1mm, the first pass is at Z-1.2903.

from grid-apps.

that's incorrect. do you have stock set so it clears the top? i'll take a look now.

from grid-apps.

Thanks. The stock dimensions are all 0.

from grid-apps.