Some traces missing when exporting .pcb to .step

Hello all!

This is my first post here so if I used the wrong tags or posted in the wrong place, please do let me know and I will change it asap. I also have not been able to find this specific problem but if anyone knows of older threads that discuss it please do let me know!

I am currently designing some PCB coils, and when I try to export the .pcb file to .step (marking the options to export everything), some random traces are missing, as shown in the images below.



The missing traces seem to always be consistent, but I am completely at a loss as to why.

I would really appreciate any pointers!

If you attach the project in this thread we may look into the board and see if the behaviour is reproducible. If your description is correct then you have found a bug.

And not totally random - every fourth is missing.

1 Like

I have also tried with fishbone patterns and there traces are also missing. Strangely enough they always seem to follow some pattern.

I attach the project, its really only the schematic where I place the single component (the coil). I typically keep the footprints in a separate library but in case it is needed I have attached the footprint needed in the .zip too. Please do let me know if I should share the project some specific way, this is the first time I share KiCAD projects online.

Thanks for everyone’s help!
PCBRogowskiCoilA.zip (755.4 KB)

1 Like
  • the copper traces are not “tracks”. These are only graphical copper shapes embedded inside a footprint
  • the current footprint definition will give you 68 drc clearance warnings, because all these copper shapes don’t belong to a proper net
  • the step export dialog states that there are some faces omitted (wire self interference check failed - face skipped)
  • because of the last point I don’t know if it’s a limitation of the step export or if it’s a real bug

But I think I think it’s worth a bug report on gitlab, if you want to take the time.

1 Like

I would like to fill a bug report, but because I have not done it before I will try to take some time to look into the process and do it properly.

In the meantime, is there something you would recommend me to try to get it to work? Could I perhaps create some useless components to create a proper netlist?

regarding bugreport:

  • login to gitlab (you need a gitlab account)
  • then use the kicad builtin command Help–>report bug
  • this opens a gitlab issue, where the section for “kicad version” is already populated with the correct version information
  • there are two important sections to fill out: issue description and reproduction steps
  • don’t forget to attach the example

If I get time I will look (and confirm) the issue.

regarding solutions to your issue: I thougt about combing your THT-pads and the graphic shapes to custom pad shapes, but this seems also not to work. There is probably also a bug in the “edit pad as graphic shape” command. So for now I have no more ideas.

after playing a bit around you will find the solution/workaround in the attached project:

  • all THT-pads should get the pad number 2
  • place a small SMD-pad on top of each of the radial spokes (on top + bottom layer), all these smd-pads should also get pad number “2”. disable mask+paste layer for these pads.
  • select one pad at a time → CTRL+E (invokes “custom pad shape” mode) , followed instantly by a second CTRL+E (which finisches the “custom pad shape” mode, effectively combining the circular smd-pad with the connected spokes)
  • do this with every smd pad on every spoke → you will get 17 custom top copper pads (spoke-shape) and 17 custom bottom copper pads (spoke+ small arc).

pcb2.zip (135.6 KB)

mf_ibfeew:

Many thanks for this solution!!! I will use this workaround for the meantime to get the .step files necessary for FEM simulation.

On Monday as soon as I arrive to my workstation I will fill the bug report following your instructions, once again many thanks for your assistance!

I have just created the bug report!
Once again, many thanks for your help!

1 Like

Could you place a link to the GH report in here, so that people interested in the topic can easily ‘thumb up’ and follow the developments? Thanks!

Yes, sorry I had not thought of that: https://gitlab.com/kicad/code/kicad/-/issues/20515

1 Like