Using the NCE BD20 occupancy detector

As I have mentioned I’ve been installing NCE BD20 block occupancy detectors for use with my signaling system on the Piedmont Southern. For the signal logic board I built a circuit that Jeff Scherb designed and published in MR back in March 2001. It uses two 2N3906 transistors and a few resistors so is a quick and easy build even with perfboard, which is what I used. I provided a circuit diagram and parts information for this circuit in my book “Wiring Your Model Railroad”.

With the circuits built and the BD20s installed I powered it all up and when the detector inputs were pulled to ground the LEDs lit up just as described. Then I made the final connections, put a loco on the tracks and the LEDs lit up. Next I got out my resistors and tried using a 10K across the tracks–no lights. Then I dropped down to a 4.7K–no lights. Two 4.7K did give me just a hint of light in the LEDs. Finally a couple 2Ks did the trick but I don’t like the idea of using resistors that low so I started experimenting with the sensitivity of the BD20.

I went up to four turns but no deal, and with a DCS200 putting out 8 amps it didn’t seem to be a good idea anyway so I went back to three turns. Finally I dug into the literature on the NCE website and found a pdf by the HUB Division, NER, NMRA describing a similar problem they had. Their solution for using BD20s with CMRI boards was to feed +5 volts into pin 4, the ground into pin 1, and used pin 3 for the logic feed and got good sensitivity using a 5K or two 10K resistors. Sounded just like my problem so I hooked mine up the same way and sure enough it works!

I ended up using two 10K resistor equipped wheeltsets on each caboose which gives me good detection. The other difference is that my signal board uses a 12 volt power supply so I fed the 12V+ through the BD20 on pin 4 and the ground or negative wire to pin 1, with the output from pin 3 going to my signal detect in connection. Here is a link to the pdf file describing the wiring used by the HUB Div NMRA for their CMRI based signaling system. A tap on the link should open the file or allow you to download it: DetectionSensitivityandNECBD20sonCMRI


  • Thanks for the explanation. That helps a lot.

    • Yes, those are the correct values. However the book was written before I replaced R1 with a 750 ohm resistor.

  • Interesting but sounds a little complicated. I guess the BD20 is wired to the logic circuit which is wired to the signals. I am still confused how the BD20 is wired as your book says it has no direct connection to the track bus. I think I will pull down the instructions on the BD20 that NCE has on their site. Also confused on whether you have to add resistors to each car in the train or just to the caboose. I just need to sit down at the computer and sort through these questions and then possibly purchase a BD20, build a logic circuit and try it all out before making a complete investment. Thanks for peaking my interest in this aspect of our hobby.

    • Yes, there are no direct connections. The little black thing that looks like a tombstone with a hole in it is a current sensing transformer. A track feeder passes through the center of the transformer 2-4 times. When a train enters the block the DCC power passing through he center of the transformer induces a current to flow out of the transformer and that activates a transistor switch. The detector circuit senses when the transistor does its thing and the signal logic turns the lights on and off. When I first decided to try Scherb’s logic circuit and the BD20 I bought a couple and put together a couple boards just to see how it works. The original article in MR shows how to set up the circuit and experiment with it to see how it works. Once I saw how it worked I went ahead and bought enough components to make a couple dozen circuits and ordered enough BD20s for the whole layout. I use Oregon Rail Supply three aspect signal kits.