Hi c/FreeCAD, totally newbie here! I’m having a ton of fun learning FreeCAD, but I have a small question. I know the toponaming problem is going away soon, and maybe that makes this kind of irrelevant, but I’d still like to know.

Sometimes when I’m watching or reading guides on avoiding the toponaming problem, the person will say something along the lines of: “actually this technique is also more professional/proper/correct anyway, real engineers do it this way.” Basically that the methods that avoid the problem are also just best practices in general. But they always say that as kind of an aside, and I wish they’d say more! What makes those methods better? Does anyone have any suggestions for articles or videos about this?

For one example, there was one guide that suggested you should use a datum plane instead of referencing one of the object’s surfaces. I understand the toponaming problem well enough to get why referencing a surface can cause it. However, the person in the guide used the same surface that would have been referenced, as the attachment point for the datum plane. Why does that not produce the same issue?

  • TeaEarlGrayHot@lemmy.ca
    link
    fedilink
    English
    arrow-up
    9
    ·
    5 months ago

    When I’m using FreeCAD, I’ll use a spreadsheet for all of my variables, so everything is linked together by formulas. Next, I’ll use datum planes rather than drawing directly on object faces (which completely solves topological naming), although I wouldn’t attach the plane to the surface–just geometrically define the distance and angle in the spreadsheet and reference it directly (e.g. if I want a datum plane halfway up my object, I’ll make its height h/2)

    • kersploosh@sh.itjust.works
      link
      fedilink
      arrow-up
      8
      ·
      5 months ago

      I’ll use datum planes rather than drawing directly on object faces

      I haven’t had a chance to try FreeCAD, but this is generally good advice in other parametric CAD tools, too. Create “skeleton” features early on (planes, datums, simple sketches) and define later features by referencing the skeleton as much as possible. It avoids creating a long chain of feature dependencies where Feature A changes and breaks Feature B, which breaks Feature C, etc.

    • JustEnoughDucks@feddit.nl
      link
      fedilink
      arrow-up
      6
      ·
      edit-2
      5 months ago

      It is good to note that the datum plane technique completely breaks sometimes when changes on earlier shapes when dealing with angles and rotation, and non-planar geometry unless you reference it to a face (in which case topo naming again) because rotation of a datum plane is rotated and shifted with respect to the origin and not the desired face. Not to mention getting an exact face takes a long time of tweaking with complex parts.

      Also be careful with spreadsheets. Is one variable is referenced more than 15 times or so, every mouse click in the spread sheet starts taking minutes to calculate. I had an outer shell dimension in a PC case in a design that I referenced a lot and clicking a spreadsheet cell would trigger an 8 minute recompute before I was even able to enter the cell to edit the variable (so then another 8 minutes after editing)

    • Remy Rose@lemmy.oneOP
      link
      fedilink
      English
      arrow-up
      3
      ·
      5 months ago

      Ohhh, that’s brilliant! I’ve been using named sketch constraints and then formulas referencing those so far, but AFAICT you can’t seem to name the parameters in pads/pockets/etc? Using a spreadsheet sounds much better, thanks!