gEDA-user: Hurray for gEDA!

[John Doty <jpd@xxxxxxxxx>]

Last night I delivered the schematic design and netlist for a data  
acquisition board for the TESS space mission. 1200 components, but  
the design is only seven schematics and three connector pin list  
files. This is possible through hierarchy and my pins2gsch script  
(http://archives.seul.org/geda/dev/Nov-2008/msg00069.html).

On Wednesday, the layout guy at MIT told me that it was coming out  
smaller than expected, based on fragments I'd sent him, and that we  
had the space to use 0603 components rather than 0402. By changing  
just two symbol files, I changed 750 footprints to 0603. The power of  
the project symbol approach should be evident.

Three pin list files actually generate all of the information for six  
connectors: four of the connectors represent identical interfaces,  
and a simple AWK script expands a table representing that interface  
into pin lists for four connectors. I represent the connectors in the  
viewable schematics as pinless boxes at the ends of busses. Those  
symbols supply attributes like footprint. It is important that  
gnetlist see these before the pseudo-schematic generated by pins2gsch.

gnetlist generates pages and pages of messages like:

WARNING: Trying to rename something twice:
         X4/GND and X4/GND
are both a src and dest name
This warning is okay if you have multiple levels of hierarchy!

So it's okay, I guess.

drc2 was difficult: first I was getting stack overflow, and  
implemented the usual gnetlistrc fix. After that, lots of spurious  
warnings, with a few gems buried in there. It doesn't like the  
separate power symbol trick, reporting "Duplicated reference", and  
there are a lot of those here. It also does this for the invisible  
symbols generated by pins2gsch. It assumes hidden pins are power  
pins, so it complains of pintype mismatches for the pins2gsch pins,  
too. And finally, in the check for unconnected pins, it bombs with:

ERROR: In procedure length:
ERROR: Wrong type argument in position 1: #f

Still, drc2 was very useful, better than I expected from previous  
projects.

All in all, this is really amazing leverage. I'm doing aerospace  
preliminary design level work on the shoestring budget and tight  
schedule of a feasibility study. I've done this kind of design with  
Viewlogic for previous missions, but it took a lot more time and  
effort. The rough edges reported above largely come from "pushing the  
envelope" of design flow, but the wonder is that while gEDA  
complains, it gets the job done here. This kind of robustness is rare  
in software.

gEDA is just a phenomenal tool.

John Doty              Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd@xxxxxxxxx




_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user