Today we present the long-awaited (I hope!) Part 7 of my “The CH&FR Goes Digital” series.
This episode dives into block detection using the Digitrax BDL168 16-input block detector. We cover track setup, wiring, connecting the BDl168 to your computer with JMRI, and provide a live demo.
Block detection is what we call “knowing where your trains are”, and it works very similarly to how the real railroads do it. The layout is divided up into electrically separate segments, or “Blocks”. Each Block is a section of track where you want to be able to tell whether there is a train on that track or not. It might be a siding, or a length of mainline track, or (less likely) a track in a yard. The Block is electrically isolated from all the other blocks, and the power feed to one rail is fed through a block detector like the BDL168.
When a locomotive sits on that section of track, even if it is not moving (under DCC) a small current flows through the locomotive’s decoder from one rail to the other, and the BDL168 can detect this current flow. When it sees the locomotive’s current draw, it reports that section of track as “occupied”. If there is no current flow, the BDL168 will report “unoccupied” for that track section.
Of course, on the prototype railroad, the trains provide their own power, but the block detectors are able to work very similarly. By inducing a voltage between the rails, the detector can watch for the metal wheels of the train to short the rails together, indicating that the track is occupied.
On the model, without a little work, we can only detect the presence of locomotives. Most model railroad cars (in N scale at least) come with plastic wheels which do not conduct. On the few which come with metal wheels, the axles are insulated to prevent the car from shorting out the track.
By adding a small resistor to one wheelset on each car, the entire train can be detected, not just the locomotive. One of my favorite videos on how to add resistive wheelsets is by Daryl Kruse who runs the UPRR Geneva Subdivision in N Scale.
For further reference, here are some links…