Les capteurs sont des éléments qui peuvent percevoir le passage d'un train. Il peuvent être simplement un élément de voie isolé, un interrupteur magnétique encastré dans la voie, un faisceau infra-rouge ou même un lecteur de codes barre qui lit le code barre collée sous une locomotive ou wagon.
Rocrail a besoin des informations de ces capteurs pour contrôler les trains. C'est sa seule manière de savoir qu'un train quitte une partie du réseau.
Il existent différents type de capteurs et chaque capteur dépend du système de pilotage et du type de signal retour souhaité. Certains capteurs détectent simplement l'occupation d'une voie (que ce soit une locomotive, un wagon ou même un tournevis !), d'autres réagissent au passage d'un aimant au-dessus d'eux ou encore lorsqu'une locomotive tire du courant d'une portion du réseau.
Mais l'utilisation de capteurs impliquent 2 choses : certaines parties du réseau seront sensible au passages des trains, et des decodeurs connectés aux capteurs devront renvoyer l'information vers le poste digital ou Rocrail. Le schéma classique est basé sur un rail de contact connecté à un decodeur S88 de Märklin.
Un rail est isolé du reste et est raccordé à un capteur de courant. C'est la solution habituellement utilisé en système 2 rails. Les trains et les wagons à faible impédance (= permettant à un peu de courant de traverser leur essieu au passage d'une résistance) génèrent une connsomation détectée par le capteur de courant. Les locomotives qui roulent utilisent également d'office du courant et génère ainsi un signal. Si vous raccordez les deux roues d'un wagon avec un résistance, il y aura une consommation statique du wagon lorsqu'il sera sur la voie.
Ce système peut également être utilisé en 3 rails en isolant le rail central et en raccordant à cette section isolé le même capteur de courant. L'ancienne voie M de Märklin est très simple à utiliser pour la détection de courant car le rail central est très simple à isoler du reste de circuit, même avec une simple feuille de papier.
Les contacts reed sont des interrupteurs qui se ferment au passage d'un aimant…
Reed contacts (or switches) are closed with a small magnet. The switches shown below will be closed if a loc or wagon with a magnet on the bottom passes over them. The magnet on the rolling stock must be mounted low enough to close the switch embedded in the track.
Reed contacts can be very hard on the pickup shoes in 3 rail locs, and very hard to use with Marklin M track, where the magnet on the bottom of your loc is constantly pulling towards the metal track bed. Magnets can fall off your locs and cars, causing all kinds of problems. And once a reed switch that is embedded in your layout breaks, you must tear up the track and replace it, it cannot be serviced.
A signal is generated when something breaks or disturbs the light beam between the sender and the receiver, usually a train rolling between them. The sending and receiving LED's are usually concealed in track side structures like small houses or signals.
NOTE: Infra-red systems will fail if the LEDs are not covered with hoods and the room is flooded with strong light, such as sunshine coming in a window at Sunrise or Sunset. If you use Infra Red systems, make sure you put hoods or covers over your sensors, try to keep your light levels constant and watch out for direct sunlight.
There are a number of systems that try to read each loc by using short range detection, including the LISSY infrared system and the BARJUT barcode reader system. All these involve placing a marker of some kind (a radio, or barcode sticker, or other identifier) on the loc. As the loc passes over a sensor, the sensor not only knows that something is there, it knows exactly what is there by reading the marker.
Train detection also gives loc address information. The sensor does not just send an OCCUPIED signal back to the Central Station, it sends the address of the device crossing it.
(Rocrail supports the Lissy receiver in combination with the LocoNet protocol.)
A plastic toggle on the 3rd rail is activated (pressed to the left or right) by the pickup shoe on a loc. These tracks are usually bidirectional and have 2 outputs, one to show each direction. You will need to bridge the outputs for Rocrail, otherwise they look like two separate sensors, each of which is “one way only.”
For 3 rail systems, one outer rail is isolated from another, and the axles on a loc or wagon shorts both rails. This is the most common of sensor tracks, and is easily fabricated with Märklin C or K track.
|Märklin C track with 74030 and 74995. The bridge between the outer rails can be easily cut without causing to much damage. You have to make sure, that there is no joint between the rails itself.|
The number of sensors can vary between one and five per block. More sensors per block will give better results. Each sensor generates an event that Rocrail uses to determine what to do next for the train crossing the sensor.
One sensor per block is low cost and easy to install. The single sensor is in the center of the block and generates a single event. (enter+in) You can define a time delay for the in event (which is set in the block properties) so the loc will slow before stopping. However, the disadvantage of one sensor is the relative short stopping distance. Your block must be long enough to allow your fastest loc to stop after crossing the middle and before running past either end of the block.
[TK – single sensor track diagram]
With two sensors per block you get a longer stopping distance and an earlier confirmation if the train has to stop or to drive on. Two sensors are placed at each end of the block. Crossing the first sensor slows the train (by sending an enter event), and crossing the second stops it (by sending an in event).
[TK – two sensor track diagram]
Note that all this can be modified changing the settings for a Block. Crossing the first sensor sends the enter event which makes Rocrail determine if the train has to stop or can roll on.
If it must stop, the velocity is set to V_min, as defined in the train properties. This will start slowing the train down so it can stop at the next sensor. Otherwise, the train will continue to move at its previous speed.
Crossing the second sensor (in) will tell Rocrail that the train has arrived in the block. The previous block it occupied will now be free and other trains can move into it. If the train must stop the velocity is set to zero.
A third sensor (exit) can be used to monitor the correct stop of the train. A Ghost Train exception can be generated halting the layout when the train does not halt within the distance between in and exit
[TK – three sensor track diagram]
If the block is used in both directions the enter sensor at one end is used as the in sensor at the other, i.e. sensors can function as the enter sensor for a train coming from one direction, and as the in sensor for trains coming from the other direction.
In automatic mode the starting block for a locomotive is only set free after an in event is received from the destination block. This means that when a loc leaves block A heading for block B, it does not free up block A until it triggers the in event for block B.
A block is only lockable if it is set free and all sensors reports low. Unused block sections can be electrically combined to signal block occupation (occupied). This means that even if a loc leaves block A and arrives in block B (by triggering the in event for block B), block A will not be released if it still shows occupied because some of block A's sensors are still being triggered.
This is very useful for catching a lost wagon set, i.e. when the train loses one or more wagons while traveling. The lost wagons will still send occupied events and prevent Rocrail from moving other trains into the blocks where the wagons got dropped.
|enter||The train enters the block; a decision is made if the train has to wait or can roll on. The velocity is set to V_min in case the train must stop|
|in||The train is completely entered the block; velocity is set to zero in case of stop.|
|exit||The train leaves the block. If the train should stop this event is handled as an exception: train will be stopped immediately and put back in manual mode.|
|occupied||The block is occupied at none event sections|