Greetings,

 I should say hello since I am at least in part the instigator of this 
thread - I have just started working with BatchPCB and among the many things 
I need to fix is the system we have for panelizing boards. We are trying to 
move Sparkfun off proprietary tools -  and we've made a lot of progress but 
ECAD in general and panelizing gerbers in particular (which is now my 
department) is a stumbling block. Irksome - but not insurmountable.

 In the interest of full disclosure, I must admit that I am not a PCB expert - 
in fact I have never even used ECAD before being assigned this task, so I 
have plenty to learn - forgive me if I ask foolish questions or make foolish 
statements. However, I figured that there must be some code out there that I 
can work with - but I found that the lack of gerber support is striking, and 
I can't be the only one in the world that wants to import a gerber and edit 
it.

 The automatic panelization is really a secondary concern - I will have to 
visually inspect the results to ensure quality anyway, and add a product id to 
the silk layer so that we can tell what is what when 100 1/2 inch boards come 
back from production. So really the first issue is gerber importing and 
editting.

 I have a functional prototype parser design sketched out in php for the 
initial sizing routines on the web that I am planning to then pull into C++ 
for the "real" DRC code that runs after the designs are uploaded (which is 
currently in C++ and another issue but one that can wait for now). The notion 
that I had is to release this as a general purpose libgerber (or something 
like that) - it should be able to support additions of various import/export 
formats, initially I need gerber for Batchpcb and I am thinking of targetting 
PCB for export since I may be able to use it to panelize, possibly with the 
added nicety of at least partial automation and perhaps DRC checks to replace 
our somewhat rudimentary DRC code. 

 I have looked at a number of different options now (including the python and 
perl panelizers), kicad, PCB, Electric and gEDA tools are the most likely 
candidates - PCB and gEDA are gaining weight due to issues of either 
complexity or maturity. I am, BTW a strictly open source programmer and it is 
understood here that anything useful I work on will remain open source (in 
fact, all the code I have written for Sparkfun so far has a GPL copyright 
notice on it even if it is not really likely to be useful outside of our 
context.). Even if I wind up just doing this myself, it will end up being 
released to the community.

 Breifly again, the general specs:
 1. input gerber rs274x and Excellon/NC drill artwork and translate with 
accuracy to one of the open source tools formats.

 2. be able to edit/panelize designs - minimum size 10x15 inches.

 3. output resultant panel in rs274x format - _including_ a drill layer in 
rs274x (this is a Gold Pheonix production constraint). NC drill output is 
optional.

 4. (wishlist) some ability to automate/script this process.

 5. long term wish list: a DRC library that is independant of individual tool 
formats, drawing routines, objects or data structures - ie. generic, with a 
well defined API. The gerber library I mentioned above should be able to hand 
libdrc some data and get a report. The total effect would be a small suite of 
tools to translate between formats and automate artwork validation, which 
might then be employed in other contexts.

 Thanks for reading if you made it this far - sorry about the length and 
verbosity ... eh, just my nature i guess ;}. 
--erik -vv
--
Erik Winn
BatchPCB.com
[EMAIL PROTECTED]


_______________________________________________
geda-dev mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev

Reply via email to