Workflow and dependency graph issues

[perky]

Hi,
I think I'm beginning to grasp this dependency graph stuff, however I'm a bit confused with the attached graph.

Basically I created a box lid by extruding, filleting and then applying thickness, and then cloned it twice so I could split them individually into two halves with a lap joint between them, but I'm getting a "not yet supported geometry for external geometry" error.

My understanding is that a sketch in the graph should only have one arrow pointing away from it otherwise it is dependent on two things (is that correct?). The graph appears to show sketch007 being mapped onto faces of two cuts (cut001 and cut005) which are actually from different clones. Is that the source of the error?

Thanks,
Mark.

lid_depedency_graph.png (206.01 KiB) Viewed 1618 times

OS: Windows 7
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.16.6706 (Git)
Build type: Release
Branch: releases/FreeCAD-0-16
Hash: f86a4e411ff7848dea98d7242f43b7774bee8fa0
Python version: 2.7.8
Qt version: 4.8.7
Coin version: 4.0.0a
OCC version: 6.8.0.oce-0.17

[microelly2]

You make a sketch on a clone, extrude it and cut it back.
you can subtract material using a Part Design pocket instead of a pad.

[perky]

Some of those I could have done with a pocket, but I needed a taper which you can't do with pockets or pads.

Anyway I fixed the sketch007 issue (unsure how it got mapped to two things though?) but the unsupported geometry error persists. Also my sketches sometimes don't stay with their faces, or re-orient themselves by 90 degrees. It's usually simple to fix, just remap the sketch to the correct face. Strangely the dependency graph always appears to show that they stay mapped to the correct object but the position is messed up.

Mark.

[microelly2]

can you upload your model or send a pm?
so I can understand what happens.

[perky]

Thanks. I'm under NDA for this particular design, but I'll try to replicate it on something simple. It appears though that it's possibly to do with editing sketches, maybe ones containing external geometry references, something gets broken. I can usually fix the problem by deattaching, remapping and recreating the original external references.

Mark.

[perky]

OK, I think I may have fixed the issues. I originally created an open box lid by extruding with a taper, applying fillets, then used thickness. Doing a geometry check on that yielded multiple errors. This was cloned twice so any shape dependent on them also had errors.

So I replaced the open box with one that just used two extrudes with appropriate fillet sizes and a boolean cut. This was identical physically to the original open box, but this one had no geometry errors. I replaced the clones at the base of the tree with new clones of the fixed version by remapping the relevant sketches and cuts to them, and now all the shapes pass the geometry check.

This has appeared to have fixed all my sketch re-alignment issues during recomputes. So it appears thickness produces inconsistent solid shapes that break things later on. I may try to boil this down to a very simple example to exhibit this behaviour.

Mark.