This is a subsection of the larger topic of false occupancy detection problems. To start from the top in terms of troubleshooting, go here:
The following discussion is about a specific trick one can do that allows a current transformer based occupancy detection become highly immune to high capacitance
The most common current transformer looks like a black tombstone mounted on the PCB board with wire running through a hole in the center of it.
The trick here is to counter the leakage current by creating a counter current at the current transformer itself using a capacitor!!. Stated another way, use capacitance to fight capacitance in a very cleaver way!!
There is a file at the bottom of this webpage that describes everything you need to do to make this trick work and restore reliable current based occupancy detection.
This clever solution was presented by jmsmith871.
Here is what he wrote:
=============================
February 27, 2007
Track Wiring Capacitance
The track bus that feeds DCC power to the rails in the Neil Avenue Switching District (NASD) on my HO scale railroad is about 21 feet long, and consists of two 16AWG wires, twisted together. An NCE BD-20 block detector placed at the beginning of the bus always showed the track occupied, whether it was or not.
The resistance of the NASD bus and track measures about 25 megohms, and the capacitance measures about 400 picofarads (pf).
In experiments with the BD-20 block detector alone, it showed "occupied" when loaded with 39 kilohms of resistance, and "unoccupied" when loaded with 62 kilohms. This level of sensitivity is acceptable. However, the BD-20 showed "occupied" when loaded with 470 pf of capacitance (only), and showed unoccupied when loaded with 300 pf or less.
The capacitance of the NASD track wiring is sufficient to make the BD-20 show "occupied."
To compensate for the track and wiring capacitance, I wired a 390 pf capacitor to the track bus so that the capacitor current traveled through the BD-20 core in the direction opposite that of the wiring capacitance. With this addition, the BD-20 shows the NASD unoccupied. Placing 39 kilohms of resistance across the rails causes the BD-20 to show "occupied". Placing 62 kilohms across the rails does not cause the BD-20 to show "occupied". These results are the same as for the BD-20 tested by itself, as described above.
Additional experiments showed that connecting 62 pf to 1000 pf as described would make the BD-20 show "unoccupied", so the value of the compensating capacitor is not critical.
7/13/15