[time-nuts] Schematic capture, anyone?
attila at kinali.ch
Mon Feb 27 09:47:45 UTC 2012
On Fri, 24 Feb 2012 10:41:37 -0800
Chris Albertson <albertson.chris at gmail.com> wrote:
> A good feature to look for is autorouting. and design rule checking.
> Of course every engineer thinks he is smarter than this kind of
> software. Mostly he is but it is good to use software that simply
> will not allow some kinds of errors. Design rules generally are
> things like following the schematic and the geometry of traces and
> limits of the PCB fab like line widths.
> I've used "rat's nest" routing too. This allows you to place the
> parts on the PCB and then does the interconnects with "as the crow
> flies" traces that cross and can't possibly work but they are drawn
> in red. you click them one at a time and rout them. As you place
> components you can see the rats net and you move them around to
> minimize the crossings.
I advice against the use of autorouting. It's not that the engineer
is so much smarter than the router, but it's that the engineer has
much more knowledge than the router to know what those wires carry
and which ones should be short, which wide, which do not really matter
if they have a small trip around the PCB...
Of course, you can annotate the schematics with all that info in order
to get the autorouter to the point it can beat you. But it'll take much
longer than doing it yourself.
The usual way i do routing, is to use rats nest place the parts and see
whether i can route the important signals nicely. Then route those first.
If the PCB is complex (either too many highly interconnected componets,
or not enough space) i often do a first run to asses how the connections
will look like when routed and after that a second run with adjusted placement.
Why does it take years to find the answers to
the questions one should have asked long ago?
More information about the time-nuts