Numerical Innovations - Help Center
How can we help you today?
Only Pulsonix 8.5 Gerber files did not load correctly - there were extra circles. Could you let me know why?
1 Comment
Thanks for sending the sample files.
The problem with only the 3.5 Metric Gerber files, was that Pulsonix 8.5 is writing extremely small arcs: 10 nanometers in length (i.e. the Start and End points of the Arc are only "10 nanometers" apart). By gerber format definition (section 4.4.1)
Gerber Arcs inherently have rounding issues by their very nature (as defined in Gerber Format specifications below). Thus when FAB 3000 sees an arc defined with Start/Endpoints separated by 10 nanometers - it can be assumed using fuzziness that the intention was to form a closed circle (i.e. Start and Endpoints are Equal to create a complete circle). This is why you saw FAB 3000/DFMNOW draw those full circles which caused shorts (only for the Gerbers output with 3.5 metric format).
Note: Arc Fuzziness defined by Gerber format definition (section 4.4.2 of the Gerber Format Standard):
To illustrate the problem, here's a sample snippet from your Gerber file to create a 10 nanometer arc.
(3.5 Format Gerber, Metric)
Truthfully there's no need for Pulsonix to generate a 10 nanometer arc in Gerber files. We will have an update tomorrow of both FAB3000/DFMNOW which will ignore those "extremely small" arcs defined by Pulsonix Gerber files.
Thanks again for sending the sample files, and we really appreciate your feedback. Please let me know if you have any further questions.
The problem with only the 3.5 Metric Gerber files, was that Pulsonix 8.5 is writing extremely small arcs: 10 nanometers in length (i.e. the Start and End points of the Arc are only "10 nanometers" apart). By gerber format definition (section 4.4.1)
If the start point of the arc is equal to the end point, the arc is a full circle of 360°.
Gerber Arcs inherently have rounding issues by their very nature (as defined in Gerber Format specifications below). Thus when FAB 3000 sees an arc defined with Start/Endpoints separated by 10 nanometers - it can be assumed using fuzziness that the intention was to form a closed circle (i.e. Start and Endpoints are Equal to create a complete circle). This is why you saw FAB 3000/DFMNOW draw those full circles which caused shorts (only for the Gerbers output with 3.5 metric format).
Note: Arc Fuzziness defined by Gerber format definition (section 4.4.2 of the Gerber Format Standard):
The arc definition has a fuzziness of the order of magnitude of the arc deviation. The writer of the Gerber file accepts any interpretation within the fuzziness above as valid. If the writer requires a more precise interpretation of the arc he needs to write arcs with lower deviation.
To illustrate the problem, here's a sample snippet from your Gerber file to create a 10 nanometer arc.
(3.5 Format Gerber, Metric)
...
Y199995D01* <---- Arc Start Point
G75*
G03Y199994I81851D01* <----- Arc End Point "with same X" and only 000001 difference in Y.
G01*
Truthfully there's no need for Pulsonix to generate a 10 nanometer arc in Gerber files. We will have an update tomorrow of both FAB3000/DFMNOW which will ignore those "extremely small" arcs defined by Pulsonix Gerber files.
Thanks again for sending the sample files, and we really appreciate your feedback. Please let me know if you have any further questions.
Numerical Support Team