All the point motors are linked back to non-latching two way switches on a panel and are manually driven. This is shown above along with a Lenz ccontroller, the switch panel for the layout lighting and its ammeter, and the DCC emergency stop button. The point switches are also linked up to Heathcote Electronics point indicator modules which feed the status of all the points into an industrial digital I/O card fitted in a PC. I have developed special software in Visual Basic which reads the state of all the points every second and displays this on a mimic panel (see track plan). In normal operation the routes that have been set are shown in solid with the other tracks dotted. The software indicates any conflicting settings and will also eventually drive all the signals. I am currently building the interface electronics to drive the colour light signals but I have already hooked up two lower quadrant signals that protect the branch leaving the main station and these are driven from the computer using Embedded Controls servo motor drivers (more info). Once complete the signals and points will be interlocked. Apart from the DCC track power, all wiring is done with solid multicore wiring as used in telecoms. This is all terminated on to ex-BT krone blocks under the layout and connections are then made as required between the blocks - as is standard telecoms practice - see the picture below. Krone blocks are the standard telecom connector as found in green roadside cabinets and telephone exchanges etc and are basically 20-way connectors on which up to 4 cables can be terminated on each 'way'. In normal telecoms use they are arranged as 10 pairs (or phone lines) per block. I've not seen them used on model railways before but they work very well and are able to handle the high currents required for point motors. They do require a special termination tool but can be obtained quite easily secondhand. When I first installed the Heathcote indicator modules, I had problems with spurious indications, and this was due to the fact that I had used a common single return wire for multiple point motors. I have uprated this return path to thicker cable and connected all the returns to earth (the PC is also earthed) and this fixed the problem.
The most recent development is to the fiddle yard. I am changing the arrangement here so that the points are driven from Lenz LS150 accessory decoders. I intend to use KAM software modules embedded in my mimic panel software so I can select a fiddle yard road on the screen and the points will be set accordingly. I will retain manual control on the visible parts of the layout as I like using it. I have also purchased some MERG accessory decoders that I may use to drive the signals soon.