kicad-developers team mailing list archive
-
kicad-developers team
-
Mailing list archive
-
Message #09827
Re: CERN plans for Kicad
On Sat, Apr 06, 2013 at 09:21:16PM +0200, Javier Serrano wrote:
> This has been already discussed over some length with Jean-Pierre, Wayne
> and Dick. I really hope the donations people get their act together this
> week and we can go public. In the meantime, if you have any comments I
> will also be very interested in reading them. Looking forward to working
> with you.
I'd already seen that list before... anyway some things are already in
motion (I'm doing the layer limit removal thing, for example).
The push&shove and the differential pair routing are big feature
requested many and many times... they are *just* too complex to think
about (for me at least :P:P). Differential pair is more easy than the
push&shove router (its not recursive, at least) but I've not enough
user-experience on the process to implement something really useful...
in fact I usually move around more amperes than megahertz:D
I've only done the occasional transmission line (USB stuff) and an
evaluation for a DSP board and I agree that at the moment pcbnew is not
exactly the best tool for the job (it can be done, anyway... there are
better tool but the price can't be beaten yet)
Another thing I feel that could be useful is the split plane support
(as a mixed signal designer); the 'best' way at the moment is to use
zone fills and keep track of the zone boundary by hand. As before its
a little inconvenient but it works. Also it needs careful inspection
before production: there are 'soft' keepout zones to keep an eye to
(like crossing a split on another layer). And more than one time
I wished some kind of '3d clearance' between netclasses: a thermocouple
amplifier on the top and a chopper power return on back are a sure
recipe for noise (true story: I had to remake the board...)
Then there would be the tool convincing mechanical designers and
management to give you a little more space for the board, so that the
above defect could be avoided in the first place: they usually go like
"it should be big as a cellphone but completely repairable and without
anything smaller than 0603 for reliability; here is the enclosure, it
has to fit there since we already made the mould". And obviously
industrial strength at consumer prices...
It happens that such a tool exists, but could be better integrated in
kicad: look on the net for clue-by-four :D best held with one hand
screaming wildy; head of the previous designer in the other hand
optional :D:D)
--
Lorenzo Marcantonio
Logos Srl
Follow ups