Re: Circuit Board Layout Program

Hi Jon,

I think you've identified the import parts of the header I need to
parse. Certainly would be easier to parse the apertures, rather then
the hand generated table I'm temporarily using.

I've already had limited success by building an array of pads first, and
"connecting" the pads with traces (links in the table). DO3 generates a
node (pad) in the table, D02 generates a node (no pad), and D01
generates the "linkage" (trace) between the appropriate nodes.

After the array is suitably processed (ordered and linked), I can
generate outline traces by following the linked lists. This works, with
certain limitations, it is only a simple approach to the problem.
Pileups of pads to make "big pads" will fool the algorithms. My
algorithms currently only understand simple connections of pads and
traces. I was able to transverse some of Eagle's sample Gerber files,
except for the "heatsinks".

A trace width parm gives me enough information to "walk around" the nets
(linked lists), which can then be converted to Gcode (although I'm
currently only generating PLC statements for Vector).

Thanks for the comments. Any other thoughts on converting Gerber files?

Alan KM6VV


Jon Elson wrote:

>
> mariss92705 wrote:
>
> > Alan,
> >
> > This is a gerber file of a 0.9" circle inscribed in a 0.9" square
> > whose lower left-hand corner was placed at the x,y origin (x=0, y=0).
> > Most of the code must be for the circle line segments. The numerical
> > data is xxx.xxxx leading zeroes and decimal point not shown. Seems it
> > should be possible to turn it into G-Code.
> >
> > Mariss
> >
> > G04 AutoGERB for AutoCAD 14/2000*
> > G04 RS274-X Output *
> > %FSLAX34Y34*% This is the format specifier (FS) line - it says
> > coordinates
>
> are absolute, x and Y are both 3.4 digits, as you say above.
>
> >
> > %MOIN*% This one is a bit odd, bit I think sets coords to inches
> > %ADD12C,0.005000*%
> > %ADD13C,0.007000*% These are the aperture definitions, looks loke 5 and 7
> > mil circles
>
> The rest is draws and fills.
> Ignore the G54. Dnn, where nn is 10 or greater selects an aperture
> described above.
> So, D12 draws with a .005" line. D01 is a draw from current location to
> specified coordinates.
> D02 is a move without drawing. D03 is a flash aperture at that location.
> Unless you are
> using positive and negative drawing, which is a feature of RS-274X, that is
> about all
> there is to it. I have made programs that do somewhat similar things, ie.
> read in
> some variant of G-code and convert it to some other variant. The aperture
> selection
> is like a tool change. Although with an engraving point, depth can control
> diameter.
> Now, one limitation is that in the Gerber file, you can flash rectangular
> apertures, etc.
> But, the D01 is like a Z-down and then cut, the D02 is like a Z-up and move
> above
> the part. And a D03 is like a canned cycle, move and then Z-down and back
> up.
>
> >
> > G54D12*X10000Y5000D02*X9988Y5348D01*X9952Y5694*X9891Y6036*X9807Y6374*X
> > 9700Y6705*X9570Y7028*X9418Y7341*X9245Y7642*X9051Y7931*X8837Y8206*X8605
> > Y8465*X8355Y8707*X8089Y8932*X7808Y9137*X7514Y9322*X7207Y9486*X6890Y962
>
> Jon