Just a brief one as I'm a bit chuffed with some initial success with my DCC "box".  I've been working on extending its abilities past the normal range of a DCC generator/controller as I outlined in the last entry, and now, finally, the various elements have come together.  I'll aim to write up something more informative soon (I haven't started testing it yet), but my foray into PCB design and manufacture has worked out better than I anticipated, and resulted in the following "solution" for the controlling electronics:

The Arduino UNO has been replaced with a socketed Nano, and the single motor shield simplified into a pluggable module focused on just DCC generation.  Still two outputs per board, so the whole unit can run six DCC outputs.  Six was chosen as this is the effective limit of the Nano in terms of pins available.

I've crammed this, a power supply and one or two other bits and bobs back into the original case, and so I now have this:

As I said I haven't really started testing anything yet, but the Nano is happy and now displaying 6 districts (A through F).  All the panel components have had to condense sideways as the PSU is filling the extreme right of the box.  The main upgrade here are the three sockets, each presenting two DCC Districts with Ground and +15V

One advance on the district front (apart from each district being independently powered) is that in the event that there is a power related "Opps" the firmware will try reversing the polarity of a district briefly before actually turning it off.  The aim here is that things like reversing loops/triangles or turntables which normally need some explicit mechanism for handling the potential for shorting can simply be made their own district.

Well, that's the design goal.

There is a whole pile of careful testing before I subject an engine to trial runs, not mentioning actually rigging up a test environment where the practical testing can take place.  But this is something of a landmark that I wanted to share.

Tomorrow will see our oscilloscope (just a hobbyist one, nothing fancy) get some use as we work our way through the various use and error cases to see what happens.  I've no doubt that there are issues, how could there not be, the firmware is over 6,500 lines of C/C++