Wiring and running

Why our shunters stalled at walking pace, and what fixed it

For two winters we blamed the locomotives, and for two winters we were wrong. The shunters stalled because of how the layout was wired and how rarely it was cleaned, and putting both right took a fortnight of evenings that nobody enjoyed.

At the bench: 22 April 2025
Typed up: 14 June 2025
Reading: 8 minutes
Room: Amsterdam, Noord-Holland

Stop 01 of 04

The symptom we misread for two winters

A locomotive that stops at walking pace and runs perfectly at speed is telling you something quite specific. A mechanism coasts over a dead spot at a scale forty and does not coast over the same dead spot at a scale eight, so the model stands there with its wheels not quite turning while somebody reaches for the controller. We read this as a pick-up fault and spent several evenings taking small mechanisms to pieces, cleaning wipers, and replacing one set of wheels that turned out to be entirely innocent.

What made it hard to see was the weather. The stalling was bad in November and January and almost absent in June, which pointed back at the mechanisms again, since oil thickens in a cold attic and everyone knows it. What was actually happening is that damp raises the resistance of every unsoldered rail joint in the room at once, and we had a very great many of those. A layout with one feeder per board runs beautifully in a warm month and lies to you for the rest of the year.

Stop 02 of 04

Feeders every metre, and a bus that is really a bus

The rebuild was thoroughly unglamorous: a pair of heavy wires running the length of each board underneath the track, and a dropper soldered to every rail section longer than about a metre. Rail joiners now carry trains and not current. That is the entire principle, and every other fix we had tried was a way of avoiding it. Three boards took four evenings, most of which was spent on my back under the baseboard with a torch held in my teeth, which remains the least pleasant modelling I have ever done.

Two details matter more than they appear to. Droppers should be soldered to the outside web of the rail, below the railhead, where a ballasted shoulder will hide them completely, and each one wants to be long enough to reach the bus without tension. Wire under tension eventually breaks at the joint you cannot get at. Colour coding one rail red throughout the whole room, without a single exception, has prevented at least three reversed sections since; we found the fourth by smell.

Written on the card

One dropper per rail section, not one per baseboard
Solder to the rail web, below the railhead, where ballast hides it
Leave slack; wire under tension fails where you cannot reach it
Colour code one rail across the entire room
Joiners align track, they do not carry current

Stop 03 of 04

Points, frogs and the switch nobody wanted to fit

Our points arrived with plastic frogs, which is a sensible compromise for a train at speed and a poor one for a four-wheel shunter at a crawl. A short-wheelbase vehicle can have both wheels on one side sitting on insulated plastic at the same moment, and at that moment it stops. Feeding the frog from a switch on the point motor cures the problem entirely. We knew this for the better part of a year before doing anything about it, because it meant lifting points that were already ballasted and apparently working.

The job was every bit as tedious as expected and produced the single largest improvement this layout has ever had. Slow running through the terminus throat went from a game of chance to something you can do while holding a conversation. Two of the points needed their frog rails cutting and re-gapping before they would take a live feed safely, and one of those we got wrong at the first attempt, with results that were loud and instructive. Test every frog with a meter before the first train, not after it.

Stop 04 of 04

Cleaning as a habit rather than a rescue

The attic is damp from October to March and the rails oxidise whatever we do about it. Cleaning used to happen when something stalled, which meant it happened under pressure, badly, ten minutes before people arrived. It is now the first ten minutes of every evening: a hardwood block along the railhead, a track rubber through the pointwork, and one locomotive run slowly over the whole line before any stock comes out of its boxes. Doing it calmly and early has reduced mid-session stalls to almost none.

We avoid solvents on plain track. They clean beautifully and leave behind a film that collects dust within a fortnight, and the film works its way into pointwork where nothing can reach it. Hardwood and a light hand do the same job more slowly and leave nothing at all behind. Wheels get attention every few months with a strip of hardboard laid on the rails and the locomotive driven against it, which lifts more black off a shunter than anybody would care to show a visitor.

Written on the card

Clean at the start of the evening, not when something stalls
Hardwood and a track rubber; no solvents on plain track
Run one locomotive slowly over everything before the stock comes out
Clean driving wheels every few months, against hardboard