The Neo900 team decided to move away from proprietary software Eagle and converted its schematics production to the KiCad open-source Electronics Design Automation (EDA) suite.
Summer time! People travel more, roast a bit in the sun, enjoy some time off. Well, except Werner, who lives in winter, and the rest of the Neo900 Team who can't stop making progress, as good samaritan workaholics we are. Before jumping to our main feature on KiCad, let's review Neo900 project updates since June.
Last week we completed the migration of Neo900.org services, including Neo900 Shop to a new server. The old one was running out of space, and dangerously approaching capacity, especially when slashdotted.
If you encounter any problems with the new server, please report them!
After dowsing for a while, our boots in China confirmed the source mentioned previously! We already received 20 more N900 units responding to our quality criteria for enduring the metamorphosis into brand new Neo900 units for you lucky (and patient) early birds. 20 more units are on the way, and we expect more to come. We're already at 70 units and counting.
Again, if you know where to find some affordable stocks of N900 units, please contact us!
While we're at it, and to continue building momentum for our main feature today, like if you were too early at the movies, let's have a look at what our mad scientists concocted since last June.
We have updated the SIM Switch document in section 4 to reflect the added counter-surveillance monitoring capacity to the SIM cards, and in Appendix A to discuss the physical placement of the second SIM card.
A new paper discusses IO expanders, that provide more IO than the CPU has, and come with the added benefit of simplifying the connections between LOWER and UPPER boards.
Finally the Infra-Red Subsystem paper was updated to define the IR transmitter control logic, using a mixed-signal array similar to what's used for power selection in the SIM Switch, and to simplify the LED driver, making its design more robust.
You can always follow our whitepapers updates directly from Werner in the Neo900 Announcements thread on talk.maemo.org, and access the whole up-to-date whitepaper collection from the Resources section at neo900.org.
And now, without further ado, our main feature!
When Joerg took charge of the Neo900 project, the electronics design was made with Eagle, and was updated using that tool ever since. Nikolaus Schaller was the Eagle virtuoso, with his own homegrown set of Eagle extensions to cope with the demanding task at hand, but the tools and workflow associated with this solution didn't allow for collaborative editing of the schematics and layout.
In 2016, Nikolaus faded away from Neo900, absorbed by the finishing touches to our cousin project Pyra. Just a few weeks ago he confirmed that he couldn't follow up on the layout for Neo900, which prompted Joerg and Werner to consider alternatives.
In the EDA market, besides Eagle, there's Altium. But Altium, although it's probably powerful enough to serve our needs, shares a major flaw with Eagle: it's proprietary, and moreover, quite expensive. Our rationalizing mind wants to say there's cognitive dissonance in using non-free software for a free hardware project. And in hindsight, this sounds like a good rationalization.
Among the open-source alternatives to Eagle, Fritzing didn't match our need for multilayer board support; between gEDA and KiCad, the choice was easy: the latter is much more poweful and popular, backed up by CERN as part of the Open Hardware Initiative, and the community is more active.
The only major downside comes from the reduced access to Nikolaus' OMAP know-how, although we hope he will be able to review our work. On the other hand, we're no longer slowed down by uncertainty with regard to the future role of Golden Delicious in Neo900: this used to cause change requests to pile up, and we used white papers as a means of documenting what we couldn't change in the schematics in a timely manner.
That gives us wings: with KiCad, we can now provide a more transparent development process and operate in a more schematics-centric mode, using white papers only where something actually needs explaining.
KiCad's routing capabilities are superior to Eagle's:
Moving to KiCad proved to be quite an improvement over the corresponding Eagle experience.
Progress is surprisingly fast. We already completed the bulk of the conversion, and are now fixing bugs (some discovered during the conversion, and also in KiCad.)
In the coming weeks we're going to work on incorporating material parked in whitepapers (see above), e.g., define the BB-mX interface for prototype v2 that would save on costs and enable testing our circuitry all the same.
Being able to put our schematics under version control moves us away from the Stone Age, into the present. You can watch the evolving contents in our eletronics engineering Git repository.
Thank you for your attention,
– hellekin for the Neo900 team
P.S.: Feedback is welcome! Did you enjoy reading this post? What else should it have covered? What do you want to read in the news? You can tell me: hellekin at neo900 dot org, or in the comments.