drawing has multiple contours

[Krischu]

The calculation program of a laser cutter claims that the dxf I'm uploading has multiple contours. (parts are supposed to only have one contour).

At the moment I have no idea how to find out what is wrong with my dxf. I applied polygon from segments, searched for duplicates and searched for zero length entities.

Drawing attached. Does anyone see what's wrong with my drawing?

M_Seite_2_ob.dxf

(100.72 KiB) Downloaded 383 times

[andrew]

Make sure the polyline is properly closed (it isn't: start and end point differ).

It could be that the software does not detect that this is one part as the outer contour is "leaking".

After the polyline is closed, it can be converted to a logically closed polyline (if needed):
- Select polyline
- Draw > Polyline > Logically close Polylines

If that doesn't work, I'd try without the holes to check if that works. That's a classic debugging technique: reduce the data to find out which part causes the problem.

[Krischu]

Make sure the polyline is properly closed (it isn't: start and end point differ).

It could be that the software does not detect that this is one part as the outer contour is "leaking".

After the polyline is closed, it can be converted to a logically closed polyline (if needed):
- Select polyline
- Draw > Polyline > Logically close Polylines

If that doesn't work, I'd try without the holes to check if that works. That's a classic debugging technique: reduce the data to find out which part causes the problem.

Thanks. Your diagnosis was right on point: leaking contour made the vendor's calculation software see more than one contour.
I tried the latter (without the holes) in the first place and the error was gone, so I thought the circles would be the cause. I learnt that it was quite easy to find the culprit once I knew (from you) that start and end point were different.

How can I ripup the polyline back to segments?

[CVH]

Select.
Modify explode.

[CVH]

You probably have read about my issues with Polys an offsets.

Offsetting a poly pocket I do on a seperate layer 'Busy' these days.
Explode to Lines and Arc's.
Transfer most obvious faults to sub folder 'ToKill'.
Find duplicates, transfer to sub folder 'ToKill'.
Find near zeros, transfer to sub folder 'ToKill'.
Then I recombine with OG.
Where the new string ends too early there will be a problem.
Or I transferred something too much, simply re-transfer.
Or It was not found.
Entities that are large enough to not be detected as zeros and those
that ly exactly on or very, very near an intended but are shorter are not detected by 'Find duplicates'.
Explode the string and displace wrong end.
If I can't pick the faulty end to drag I'll step through the vertices in PropEdit and alter the X or Y coordinate.
Cut off and tranfer faulty segment to sub folder killed.
Recombine with OG untill I get a Logi Open Geo closed Pocket.
Transfer this Geo closed Poly to sub folder 'Good'.
Select all what remains on 'Busy' and transfer it to 'ToKill'.
The good one I still have to change to a real closed Poly.
Select, Explode, only they remain selected, recombine with OC.
This will not only reconbine but also concat co-linear lines and co-radial arcs.
Transfer only this back to the drawing.
After succeeding here: Kill everything on: 'Busy', 'ToKill', 'Good'.

One offset.
Next.

[CVH]

addendum:

Now I come to think about it.
I'll trow in multiple Find Zero/Dulpli.

Because some selection methods don't select Null-entities.
Use TA, Select, Select All.

But there is a one danger with Null-entities.
The might crash on you. Selecting, moving, undo...

So Save, Save, Save and yes save again.

Sometimes I have to resort to 'Alt' select to find multiples on one spot.

Finally on the OC recombined 'Good':
I'll do another Explode, Find Zeros.
If found, investigate them in high zoom, panning without changing the selection.
They might be micro self-crossers.
These where auto generated to close, these must be deleted manually.
Select, Delete the tiny and Trim Double, Find Zeros
If not your are good and Cancel, Undo.

....
If somethings else comes up I'll let you know.

[CVH]

Drawing revised

Regards,
CVH

[Krischu]

Thanks again for the detailed explanation and revision. I wondered how it happened that sometimes measures, like radii of 1.2, "deteriorate" to 1.19xxx, although they once have been constructed precisely with radius 1.2. Maybe when copy/pasting?

[CVH]

Dont know, but one day I will catch the culprit.
It doesn't serve to report it again and again.
Such posts are not read anymore.
You have to catch it in the act and reproducible.

Regards,
CVH

[CVH]

Bitte

[Krischu]

Thanks again. The "not 45°" is intentional (!).