Jamie Baumgart commented:

>     Most problems that I have seen with artwork fabrication relate back to
>     problems with the aperture lists, e.g., mistyped, X and Y reversed,
>     thermals not defined properly, etc.

And I agree 100% it is absurd that we still have those stupid aperture
lists that harken back to when the plotters had wheels and "flashed" the
damn apertures. So I cannot understand why there is no rasterized output
with most CAD files, and why this new RS274-X is not taking over like
sliced bread. If I had our CAD software vendor nearby I would yell loudly
in her ear.

This whole stupid process creates delays and mistakes, and thousands of man
hours of effort to check film and correct the foulups.

>     In a previous life, most, if not
>     all, of these problems vanished when we received RS274-X (embedded
>     Gerber, extended Gerber, whatever you want to call it) data.
>
>     1)  Don't most CAD systems export this format?

Can't say for sure, but don't think so. We use Cadnece Allegro on a
SparcStation here, and there seems to be no clear cut answer as to whether
this is a standard output, our experienced CAD operators did not know?!
(just another example of how designers seem to almost know nothing about
how boards are made).

>     2)  Don't most CAM/photoplotters/laser plotters handle this format?

They seem to, based on my contact with vendors, I would be anxious to hear
from some. Hadco says they love it.

>     3)  Are there concerns about the reliability of this format?  (Based
>     on my experience, this would surprise me.)

Excellent question I am anxious to hear about this.

I would add this question:

Recently one of our vendors sent us a board that had three adjacent pads
shorted together when they built this design the first time.

Luckily, our assembly operators noticed the problem, and it was traced to
the mis-entry of an aperture list, the length was so long that pads simply
became one.

I questioned how the netlist information had failed to find this at
electrical test. Was told that the fab vendor created the netlist with
their own CAM tools, and therfore the "error" looked correct to the bare
board test generated from the PWB fab house CAM output.

Later I found that our CAD dept did not even provide netlist data
routinely. Our CAD people told me that they had been told by one of our
previous vendors they would charge a considerable fee to make their test
fixture from our net list, rather than from their CAM outputs.

This is a heck of a problem, IMO. Again, I was told by the fab vendor that
we needed to provide our netlist data in some IPC (D356?) format that our
CAD program cannot (I was told) output.

If anyone could tell me this is not the case I would like to know - one way
or another I want to change this situation for new designs...(almost
embarrassing to me).

thanks

Jerry Cupples
Interphase Corp
Dallas, TX