--- In kicad-users@yahoogroups.com, Robert <birmingham_spi...@...> wrote: > > > Just as a matter of interest, what happens to curved tracks that are > > placed by Freerouter and then imported into KiCad from a .SES file? > > I also have need of curved tracks. When I looked into the seemingly > curved tracks produced by Freerouter, zooming in showed them to be > piecewise linear, and some way from smooth curves. > > If you do a bit of googling you'll find there is some discussion on the > developers group about problems with using arcs on photoplotters. True > arcs do seem to be a big deal, and are not widely supported in PCB > packages (at least not in the ones I was able to download and try out). > However, Eagle does support them, and it would be really nice if > kicad did too. > > Although kicad allows the creation of arcs on non-copper layers it plots > them as piecewise linear (and rather crudely too in the 2009 edition at > least, though I think that might have been improved for the 2010 edition). > > If you want to tackle this issue one other area that would have to be > addressed is creating and editing arcs. The kicad method is IMHO very > clunky. It needs a graphical method of moving end points and adjusting > the shape of the curve. Eagle seems to do that quite well. What > Eagle doesn't do is use live DRC checking to prevent the track getting > too close to another track, which of course is a very useful feature of > kicad that would have to be built into arc creation and editing. It > sounds like a lot of work to me, but if you have the time I think it > would be worthwhile. > > Regards, > > Robert. > > > No virus found in this outgoing message. > Checked by AVG - www.avg.com > Version: 9.0.851 / Virus Database: 271.1.1/3063 - Release Date: 08/10/10 > 19:34:00 > Hi Robert,
Many PCB packages produce piecewise, or chordic, approximations for arcs, especially for the Gerber output, even though the Gerber language itself does support true arcs with the G02, G03 and G75 codes and always has done. Many use 16chords/360deg, although some are programmable. Although it doesn't necessarily look that pretty at this level, the 1/16 chordic approximation only results in a maximum error of -0.64%, and would be acceptable for most applications. This kind of (code) work usually entails much more than it appears at first, and no - I don't really have time. But that doesn't always stop us, does it? :-) Regards, Tony
