The track is wired in a way that is almost certainly over-engineered. It has been done this way for reliability, and for ease of maintenance. This is for the long haul: I don't want to be fixing it before it is finished!
Each piece of track has a pair of dropper wires: I am notrelying on electrical connection via the fishplates. This is almost certainly over-engineered. The dropper wires are 50-75mm lengths of 22SWG tinned copper wire, soldered to the rail bottoms before mounting. I have used 60%:40% leaded solder: lead-free solder melts at a much higher temperature and did NOT flow well at all. It may be possible to improve the "flow" that with different fluxes, but the temperature means more heat into the track and more opportunity to distort the plastic.
My CAD design shows the rail polarity, and location of power feeds to blocks. The rail polarity is marked by the rail being "red" or blue". The "red" rails are gapped between sections; the blue rail is continuous through the entire power zone. (The railway is also power zoned, and this makes wiring slightly more complex; the "blue" rail is also gapped between zones. See the power management section for more details).
Because the railway is power zoned, there are several track buses - not just one. Each zone has its own common return wire for the "blue" rail; the "red" rail is fed from the relevant block detector output. This means a lot more wiring than for non-block detection DCC wiring, even with the block detectors fairly well distributed around the railway. This is essential for computer controlled operation; it means that a short in one area will not prevent trains running in another. Ultimately the wiring has taken a little longer, but it isn't difficult and it is still easy to follow. However, an accurate record of "what is connected to where" is essential. When planning the block detection, I developed several spreadsheets that show what track section is in which power zone and which block detector provides the power.
The common return wires are 0.75mm2 24 strand copper wire. These can see a maximum current of 2.5A, because of the power zoning. The feeds to each detected block are mostly 0.5mm2 copper wire; except when there is a short, these will only see perhaps 0.2A because there will only be one train in each section. Each is still rated (and tested) for the full 2.5A "short circuit" current.
The droppers from the track sections are joined to the bus power feeds using crimp "bullet" style connectors. The dropper wires have a male bullet crimped to them. The "red" wires are fed directly from the block detector boards, arriving ultimately at a female bullet connector. The blue (or sometimes black, where I ran out) return wire feeds a "daisy chain" of female bullet connectors. I found it very easy from under the railway to run a wire to the next position, and have two wires crimped into the bullet connectors.
I had intended to use "Scotchlock" suitcase-type crimp connectors to make the tap-off joins onto the main bus runs. There are a few used in places. However I found that daisy chaining the crimp connectors was easier, in the end. I also struggled to find these connectors that were specified to work with the cable sizes I'm using; while they may work over a wider range of wire sizes, the "specified" sizes that the manufacturer guarantees are quite restrictive.
Point frogs get power from the switch built into the SEEP point motors. The track bus is wired to the point motor, and the frog is powered by the sliding contact.
Everyone recognises that DCC does not deliver the "two wires are all you need" utopia. In this railway, the power has to be fed separately to a lot of track sections simply for train occupancy. That already implies a lot of wires.
So what do you do? Early on I wrote a comprehensive "wiring rules" document. Having wired much of the railway and looked back, much of that was rubbish so it's been discarded.
Wiring falls into several classes; I've used a different approach to each.
A consistent rule from the start has been no mains wiring on the railway. That is an absolute. The mains wiring is limited to things that plug into distribution boards attached to the wall; these are fed from a single switch. There are no circumstances where a mains wire runs around under the baseboards; it just doesn't happen.
Structural wiring refers to DCC power feeds from the boosters, power feeds to the distributed electronics. These wires go to the panels around the layout with the distributed electronics, and I established a common connector wiring scheme for a 12 pin plug-on "chocolate block" connector. This allows power to be isolated if required. The structural wiring is generally made with brown/blue figure 8 mains cable. I'm satisfied this implies no safety risk; other may disagree in which case they are free to do differently!
The power is zoned and block detected. This means there are over 100 power feeds, and 16 returns.
The returns are all lettered, and fed with a black wire. The feeds are numbered, and fed with an orange wire. (I had planned to use red, but the supplier was out of stock. I figured as long as it was consistent it didn't matter). All wires are labelled with their number or letter; I found some label sheets that could be laser printed then folded around the wire to provide an enduring record.
Points & Signals
These are all controlled by 6 core 7/0.2mm stranded alarm wire. This is used either to power two adjacent points, or one signal. The wires are labelled with an indelible marker, but that may on reflection not be very reliable. For signals, the wire colour is matched to aspect with black being the common 0v return. Where two points are powered by one wire, I established the following colour convention:
|red||THROWN for lower numbered point|
|green||CLOSED for lower numbered point|
|blue||THROWN for higher numbered point|
|yellow||CLOSED for higher numbered point|
In various places, rails need to be isolated. This can be for several reasons:
- The two tracks from the point "frog" need isolation to the next track;
- Where the different power zones join, both rails need to be isolated;
- between different block detector zones, one rail needs to be isolated.
The simplest way to isolate is to use a nylon insulating "fishplate" when laying the track. We've used those on point frogs. But for the others, the positions can sometimes only be finaised after initial track laying.
To ensure the wiring is correct, I've used several stages of testing.
Visual inspection: make sure that each track section has a power connection, and comes from the correct wire.
Isolation checks: remove the power connection from the return "blue" rails, and check they are electrically isolated from each other. (This picks up any rail sections that have been connected to the "wrong" return wire).
Short circuit check: use a coin to short each separate rail segment. Check that the power manager boards detect and isolate the short in one section only, and that the booster does not "beep" indicating a short that the power manager didn't handle.
- Block detection check: use an LED to check that each section has power, and that the block detector reports this section (and only this section) as occupied to LocoNet.
If those tests pass, then the track has been wired correctly!