Migrating off Eagle to open source PCB tools

Thanks.

I hope it goes well for them.

I'm not a fan of Eagle, never could quite get along with it.. I use gEDA myself, so would obviously have preferred if they'd went that direction    I can see that the Cern contributions to KiCad will probably mean it'll evolve to be the more prominent and capable package though, so perhaps it's time to give it a try.

Either way, having the Olimex OSHW designs available in a less restrictive format can only be a good thing, for us, for Olimex and for KiCad !

selsinork wrote:

 

I use gEDA myself, so would obviously have preferred if they'd went that direction    I can see that the Cern contributions to KiCad will probably mean it'll evolve to be the more prominent and capable package though, so perhaps it's time to give it a try.

Do you know whether the two can import each other's library formats?  Interoperability between open source tools is pretty important, given that the community has to rely on its own means and resources.

Morgaine.

This is very good news.  FLOSS often suffers from a chicken/egg problem where users won't use FLOSS because it's not as good as FiaB and developers aren't motivated to perfect FLOSS because there aren't enough users to provide the helpful feedback and especially reproducible bug reports so that FLOSS can achieve high quality.  But once you get over the threshold it's gangbusters.  I'm delighted to see eCAD reaching the threshold.

Today PCBs, tomorrow FPGAs! :-)

Morgaine Dinova wrote:

Do you know whether the two can import each other's library formats?  Interoperability between open source tools is pretty important, given that the community has to rely on its own means and resources.

Not sure if either can do it directly, but I've heard various discussions on external tools to do it over the years.  Never really had a reason to investigate though. The gEDA tools essentially use text files, so it should be quite possible even if you had to start from scratch.

I found this one with a very quick search https://github.com/upverter/schematic-file-converter/ which appears to take care of sematics at least, not sure about pcb layouts.

John Beetem wrote:

 

Today PCBs, tomorrow FPGAs! :-)

Haha, indeed!

I wonder if it's worth pointing out the conceptual similarities between component libraries for PCB tools and Verilog/VHDL cores in the FPGA world.  Although PCB components are a much cleaner "black box" from the perspective of someone wishing to combine them, nevertheless there is an element of "select component from list" to HDL cores as well, and in principle open source fans could be adding to FPGA core repositories as easily as to PCB components libraries.

There's a rather large gap between principle and practice of course, but I think perhaps it might be closing, slowly.

Morgaine.

Addendum:  Top level of tree of FPGA cores and core projects at OpenCores.

So here's my experience:  A couple of years ago I really needed an EDA package; I looked at gEDA, KiCad, etc and what I found was that while the commercial packages had fairly complete component libraries, that the FOSS packages seemed not to.

In my experience, if there's something going to go wrong in a design, it's the mechanical details; I've seldom been able to create a component without a mistake, so having the parts libraries was the selling point for the commercial packages.

I ended up using Eagle based on their pricing and availability of libraries.  Even then, I still had to create a component or two.

It seems to me that what the FOSS community needs is the libraries; whichever software has that will be adopted.

I'd have to agree. I end up creating most of the schematic symbols and footprints I use in gEDA myself. The standard stuff (resistors, caps, generic footprints like SOT23, TQFP44 etc) are there, but other things need creating.

gEDA has http://www.gedasymbols.org/ that quite a few people contribute to, but even so the coverage is relatively small.

One advantage to learning how to create your own is that I have specific footprints for things like the Raspberry Pi and Beaglebone Black that have all of the connectors in the right places, silk screen markings for cutouts around any tall connectors etc.

Bruce Prewit wrote:

 

It seems to me that what the FOSS community needs is the libraries; whichever software has that will be adopted.

Yep.  There's absolutely no arguing with that.

It's very good news though, because in open source, libraries just grow and grow, and only very rarely regress.  At some point, an actively developed system like KiCad is quite likely to catch up with the bulk of "normal" components, and is then perceived as having "everything" most of the time.  From then on, it's all coasting downhill, just needing to keep up with new products.

It's a bit like the situation with device drivers in Linux.  For a long time, peripherals support in Linux was regarded as really sparse or threadbare, and if you bought a random PC there was a very high chance that it would contain one or more components that lacked Linux drivers.  But Linux caught up.  Nowadays nobody laughs at Linux driver support, and the drivers are mostly maintained forever instead of bit-rotting on the next O/S release as happens in Windows.

I expect that good things lie ahead for open source EDA tools, at least at the PCB level.

For FPGAs, well, John has had a few harsh things to say on that subject in the past.

Morgaine.