Skip to content

The roster ​

The Roster page lists your locomotives, your cars and the trains made up of them. Each locomotive has a name, the DCC address the throttle drives, function labels, what the locomotive is (road name and number, manufacturer, model, owner, notes), its length and its decoders.

Cars and trains ​

Add a car for each car that runs in a train, with its Length in millimeters, measured over the couplers. Give each locomotive its length the same way, on its page. A car is named by its road name and number when it has no name of its own.

Make up a train names a train; on its page, add its locomotives and cars in order from the front, the end that leads when the train goes forward. Up and Down move one along the train, and Faces turns one around: a locomotive facing backward runs the train backward when its throttle goes forward, as a locomotive running cab last does. The train's length is its locomotives' and cars' lengths added up, so swapping a car changes it at once. A locomotive or car is in one train at most; take it out of one train to put it in another.

Train tracking follows a train through the decoders of its locomotives. RailCom detectors see only the decoder, so the train's length and where its locomotive sits in it say where the rest of the train is. A locomotive in no train is followed on its own, at its own length. Schedules will run these trains.

Decoders ​

A locomotive can have more than one decoder: the one fitted now, and any you swap in and out of it. Each decoder has a definition (the model, from the decoder definitions), the identity it reported when it was identified, and the CV values last read from or written to it.

One decoder is marked as fitted. Its address and function labels are the locomotive's: the throttle drives that address, and writing a new address to the decoder on its settings page moves the locomotive with it. When you mark another decoder as fitted, the one coming out keeps the address and labels it had, and the one going in brings back its own. So a locomotive that runs with a sound decoder one week and a test decoder the next keeps the right address and function names for each.

Settings on a decoder opens its settings page with the saved values already in place, so nothing has to be read before you look. Every value read or written there is saved for that decoder, and changes not written yet are kept until they are written or discarded. A page shows when the values were last saved; read it again if the decoder may have been changed elsewhere since. Values read or written in the current session carry green lamps; older ones have no lamp. Before the first read or write on the programming track, Interlock checks that the decoder there is this one, by its model, its address and the values saved for it. Decoder profiles keep named copies of a decoder's values to load again later.

To add a decoder, choose Add a decoder and identify it on the programming track or search for the model. A decoder opened from the Programming page on its own can be kept too: Keep in the roster on its settings page adds it to a new or existing locomotive with the values read so far.

Which locomotive is on the track ​

Identify the decoder on the Programming page also reads the address the decoder answers to, then looks in the roster. A decoder at the same address whose saved identity matches is named first, with a link to its settings; a decoder that is saved but not marked as fitted can be marked from there. When the decoder named first checks out, its session begins. A locomotive at that address with no decoder yet can take the definition you pick.

Several locomotives at one address ​

Locomotives may share an address, such as two that never run at the same time. To the command station they are one train: driving either one drives whichever locomotive is on the track at that address, and the throttle shows the same speed for both. An automation that drives the address lists both names in its trace.

Consists ​

Locomotives that run together are kept as a consist, listed below the locomotives on the Roster page. Trains covers making one up and driving it.

Importing a JMRI roster ​

Import from JMRI reads the roster JMRI keeps in its profile folder: roster.xml and the locomotive files in the roster folder beside it (on Linux, ~/.jmri/<profile>/roster). Choose those files, or a zip of the whole profile folder. Interlock shows what it found before anything changes:

  • Each locomotive comes in with its name, address, function labels (functions JMRI marks as not latching become momentary), road name and number, manufacturer, model, owner and comment.
  • Its decoder comes in with the definition JMRI used and every CV value the locomotive file holds. Those values are marked as imported, not read: the settings page shows them but counts them as not read yet until a read confirms them. A change to part of a CV (one bit of CV 29, say) reads the decoder before writing, so its other bits are what the decoder holds, and Write page asks before writing imported values over what the decoder holds. A decoder model that the definitions no longer have is marked; choose its definition later on the locomotive's page.
  • Entries with the same address, road name and road number are offered as one locomotive with several decoders: the most recently changed one is fitted, and the others join it, named by what their names add ("Amtrak P42 ESU" and "Amtrak P42 Digitrax" become Amtrak P42 with an ESU and a Digitrax decoder).
  • A locomotive already in the roster under the same name is left out.

Each row can be changed before importing: a locomotive of its own, a decoder of another locomotive (one in this import or in the roster), or left out. Names can be changed too. A locomotive listed in roster.xml without its own file comes in without CV values.

The files ​

The roster is trains.yaml in the data directory; see Trains for its fields. Each decoder's CV values are in cvs/<decoder id>.yaml, with imported values not read yet and the changes not written yet, which Interlock rewrites a second after values stop changing. Both are in backups.

GPL-3.0 licensed.