Eagle + generate_3D_data.ulp -> IDF -> FreeCad -> Many 3D formats

Barry Walker wrote:

The generate_3d_data script has fallen into disrepair and hasn't seen

any improvements in over two and a half years. 

 

I needed to generate 3D board data for a customer, so I took the

opportunity to make some fixes.  In my research I found that FreeCAD

(exactly what it sounds like) has an IDF import function.  Unfortunately

that module (idf.py) is not installed in the default Windows

distribution, but it is easy to copy and paste.

 

The old version of generate_3d_data had many issues that were

incompatible with the FreeCAD import filter.  With some guessing and

testing I made it work.  I think the rest of the operation of

generate_3d_data is unchanged.  It may still have some pre-existing

bugs, and I may have introduced new bugs.  For my purposes it seems to

be working very well.

 

What I don't know know is if this new version is compatible with other

3D CAD packages.  I have a friend testing SolidWorks, but I don't have

access to any of the other packages.  I would appreciate help testing if

you have other systems available.

 

From FreeCAD you can generate STEP, IGES, and many other formats after

the IDF file is imported.

 

The new version 0.90 ulp is attached.

there is a thread regarding generate_3d_data vs. eagle V6

http://www.eaglecentral.ca/forums/index.php/m/150121/6377607912ccb8b76588a5b350a10e2a/?srch=generate_3d_data#msg_150121

maybe Mr. Zaffran will chime in anyway, but if not you could contact

him to obtain a copy of the updated ULP and combine it with your

changes.

--

Lorenz

Hi Barry,

Lorenz wrote:

Barry Walker wrote:

 

The generate_3d_data script has fallen into disrepair and hasn't seen

any improvements in over two and a half years. 

 

The new version 0.90 ulp is attached.

I was not aware about Freecads import capability for idf.

It took me some time (because I'm not familiar with Phyton) but I

managed to obtain the mod/idf code from githup.

The idf file generated with your version of the generate_3d_data.ulp

loads in Freecad (0.13) without issuing an error (unlike Mr. Zaffrans

1.3 Version which throws an exception 'both points are equal'), but

...

1) parts are only wireframe

2) bottom parts are placed on top

3) I have two 4mm vias (GND for test clips), but only one of them

shows up in the import

4) the mounting holes (arcs in the dimension layer of a package, that

is: coutouts no drills) don't show up at all

... waiting for the Solid2013 import to finish ...

ok, let's see:

1) in solid the parts are solid 8-)

2) bottom parts on top too

3) none of the 4mm vias shows up

4) mounting holes missing too

HTH

--

Lorenz

Lorenz wrote:

Hi Barry,

 

Lorenz wrote:

Barry Walker wrote:

 

The generate_3d_data script has fallen into disrepair and hasn't seen

any improvements in over two and a half years. 

 

The new version 0.90 ulp is attached.

while trying-out the board-outlines-and-cutouts workaround* with your

script I found that your script version no longer runs on a

stand-alone board file with packages.

Peeking into the ulp it seems you access the schematic it there a

packages in the board.

(don't get me wrong, I'm not complaining, only reporting my findings)

• board coutouts don't show up in the idf if drawn in layer 50

directly in the board.

--

Lorenz

I was not aware about Freecads import capability for idf.

 

It took me some time (because I'm not familiar with Phyton) but I

managed to obtain the mod/idf code from githup.

 

The idf file generated with your version of the generate_3d_data.ulp

loads in Freecad (0.13) without issuing an error (unlike Mr. Zaffrans

1.3 Version which throws an exception 'both points are equal'), but

...

 

1) parts are only wireframe

2) bottom parts are placed on top

3) I have two 4mm vias (GND for test clips), but only one of them

shows up in the import

4) the mounting holes (arcs in the dimension layer of a package, that

is: coutouts no drills) don't show up at all

 

 

 

1) in solid the parts are solid 8-)

2) bottom parts on top too

3) none of the 4mm vias shows up

4) mounting holes missing too

 

 

Lorenz,

Thanks for testing the script and reporting the details of your results.

The first problem with the old versions was the "both points are equal" error in FreeCAD.  The old versions basically repeated all the points twice in the .emn and .emp files.  This may be legal in IDF, but the FreeCAD importer did not care for it.  I fixed this and another error that appears after you fix that one.

The parts in wire frame issue I don't think is fixable with the script.  It may be a setting in FreeCAD.  Was FreeCAD in wireframe mode?  I am not a FreeCAD expert at all.  I do not get wireframe parts, but the IDF format doesn't really specify solids or wire frames, just points.

I confess I found the misplaced bottom parts issue shortly after my post yesterday.  It was only a problem with multi-outline parts placed on the bottom, but for me it is fixed with the v091 file attached.

Odd only one of two vias would show up.  I don't have an explanation for that one.  I have 4 large vias for mounting holes on my board, all 4 show up.  It is pretty easy to open the .emn file with a text editor and look at the hole call outs.  That would tell you if the script isn't finding the hole, or if the importer isn't importing the hole.  I did set the via limit too low once and it caught all my signal vias.  The script handled it well enough, but FreeCAD always crashed before loading the entire design.

Pretty sure the .ulp does not look at the dimension layer in packages.  If it isn't on layer 50,  57,  58, or a through hole pad it might as well not exist for this version. 

I don't use any cutouts on the board I am testing, so I have no results.  As I recall from using the old version a few years ago, there were issues with the cutouts.  I didn't touch any of the code in that section.  If I remember correctly, when I made components with cutouts on layer 50 and it handled those better than using cutouts directly on layer 50 of the board.

The new script does rely on components in the schematic so it can find the original library part dimensions and angles.  The IDF format uses library parts then placement of those library parts, much like Eagle.  The old version of the generate_3d .ulp made a new part for every rotation and mounting side of each part found on the board.  I changed that.  Possibly this is not helpful for board only designs.

I also appreciate  your comment that you are reporting not complaining.  Very kind.

Barry Walker wrote:

Odd only one of two vias would show up.  I don't have an explanation for

that one.  I have 4 large vias for mounting holes on my board, all 4

show up.

I tracked down the missing drill.

Looking into to emn file revealed all drills to be present.

So I had a look into idf.py and I think there is a bug in the line

that handles the drills.

if current_section==".DRILLED_HOLES"  and section_counter>2 and float(emnrecords[0])*emn_unit>ignore_hole_size:

         drills.append([float(emnrecords[0])emn_unit,float(emnrecords[1])emn_unit,float(emnrecords[2])*emn_unit])

changing 'section_counter>2' to 'section_counter>1' produced the

missing drill.

--

Lorenz

Hi Barry,

there's still a problem with packages, that have 3D data on the bottom ECAD layer 50 "bCAD". Here's an example of a THT female connector. In the tCAD layer 57 there are the beginnings of the pins and in the bCAD layer 58 there are the endings of the pins. generate_3d_data_v091.ulp generates wrong angle data, if your package is places in other angles than 0° or 180°. You see in the picture on the left side packes placed in 0° and 180°. No problem here. The pins on the bCAD layer are at the same x/y-Coordinate as on the tCAD. On the right side you see packages rotated to 90° or 270°. The bCAD-pins are rotated in the wrong direction.

Data in *.emn:

[..]
AMTEK-5PS2RDB63-207_BR38.1 AMTEK-5PS2RDB63-207_BR38_SMD-FM.1 X2.1
22.00 11.00 0 90.00 TOP PLACED
AMTEK-5PS2RDB63-207_BR38.2 AMTEK-5PS2RDB63-207_BR38_SMD-FM.2 X2.2
22.00 11.00 0 90.00 TOP PLACED
[...]
AMTEK-5PS2RDB63-207_BR38.9 AMTEK-5PS2RDB63-207_BR38_SMD-FM.9 X2.9
22.00 11.00 0 90.00 BOTTOM PLACED
AMTEK-5PS2RDB63-207_BR38.10 AMTEK-5PS2RDB63-207_BR38_SMD-FM.10 X2.10
22.00 11.00 0 90.00 BOTTOM PLACED
[...]

Should be:

[..]
AMTEK-5PS2RDB63-207_BR38.1 AMTEK-5PS2RDB63-207_BR38_SMD-FM.1 X2.1
22.00 11.00 0 90.00 TOP PLACED
AMTEK-5PS2RDB63-207_BR38.2 AMTEK-5PS2RDB63-207_BR38_SMD-FM.2 X2.2
22.00 11.00 0 90.00 TOP PLACED
[...]
AMTEK-5PS2RDB63-207_BR38.9 AMTEK-5PS2RDB63-207_BR38_SMD-FM.9 X2.9
22.00 11.00 0 -90.00 BOTTOM PLACED
AMTEK-5PS2RDB63-207_BR38.10 AMTEK-5PS2RDB63-207_BR38_SMD-FM.10 X2.10
22.00 11.00 0 -90.00 BOTTOM PLACED
[...]