A Signal Circuit To Display Occupancy
By Trains In Multiple Blocks.

Bill Payne and I have been experimenting with logic circuits for trackside signals for use on our Nottawasaga Model Railroad Club (NMRC) and our home layouts. Commercial circuits by IrDot, Logic Rail, Dallee, etc are possible solutions, but we wanted to see if we could make our own and save some money.

On the Internet we discovered a 3-light signal circuit with approach lighting by R. Paisley. Our first task was to test it on an Evaluator board that I own.

This allows one to construct the circuit without soldering or using a circuit board. It is easy to change parts. It looks messy and it is, but it works. Each block set-up is identical. The interconnecting wires trigger the signal colors.

We made the circuit cover 5 blocks so we could see what happens when moving the ground wire that represents the detector. The circuit, as shown in his diagram, has a train detected in block 2. Block 2 should then be red, block one should be yellow and block 3 is green. This represents one-way direction of travel, not bi-directional.

The choice now is to photo-etch signal circuit boards, find someone to do it, or use commercial stripboards as are often shown in wiring circuits in the NMRA's Scale Rails magazine. In the latter method one cuts the traces where needed. You can see the layout of the prototype board in the photo below.

As designed, this circuit uses LEDs. However, if the 470 ohm dropping resistors are bypassed, this can be used for 12 volt bulbs.

On the club layout we have a section of single track mainline. This will need to be bi-directional. That calls for a "coincidence" circuit. We haven't found one yet. The LogicRail boards I have been using on my Utopia Northern allow for bi-directional travel. We need to find a solution.

The other task is to develop a detector circuit that doesn't interfere with track current, like the one that Dallee sells. We've found a circuit using opto-isolators and are about to mock it up as soon as we locate all the parts. R. Paisley shows a Transformer-type, DCC only, block occupancy detector using a 555 timer.

Another set of circuits for block detection and signals is outlined in an NMRA clinic by Richard Schumacher. We haven't tried his yet. There's a lot of good information about assembly of the circuits in the articles.

At a recent Chapter meeting of our Canadian Association of Railway Modellers (CARM), one of our members, Peter Hughes, directed me to the MERG site in Britain. If you're interested in model railroad electronics (and you must be if you've read this far), check out The Model Electronic Railway Group. This may ultimately supply us with the circuits we need to build.

Bill then used a stripboard planning sheet to design a master signal circuit board for the components. The sheet was downloaded from www.kpsec.freeuk.com. This shows where to cut the traces on the horizontal rows. A multimeter is used to confirm the traces are cut. Bill used a drill bit to ream out the holes. I was using an X-Acto knife.

Bill then built a prototype circuit board using a stripboard that he cut apart with a hack saw so we can make three boards from one sheet to save money. He mocked up the set-up so we could demonstrate how the lights change as a train moves from left to right through the blocks. If this were a loop of track, block 5 would be connected back to block 1.

The green masking tape at the bottom of the photo on the ground wires shows the block number. Touching the wire to ground simulates a detector triggering the LED. This is what the prototype mockup of the circuit looks like when using LEDs. The first block on the left is unlit to simulate approach lighting. Block 2 is yellow as a warning to slow down. Block 3 is red because that's where the train is at the moment. Block 4 is green and the last block on the right is unlit because the train hasn't reached the detector for that signal yet.

Update: March 25, 2010

We began working with Bill Hudson's Model Railroad Signals boards and signal circuits a few months ago. Bill has been extremely helpful in guiding us through the process. His boards are designed to accommodate different detection methods: Twin-T, photocells or external logic. We ran into problems with the Paisley detection circuits not getting the negative logic low enough. Although the signals were changing color we kept getting both red and green when we should only have had red for block occupancy. Bill has designed an alternative circuit for us using three NPN4401 transistors. We are breadboarding enough for four test blocks. Stay tuned. We've almost got this solved with Bill's help. Here's my messy test bench.

April 15, 2010 update

Bill Hudson has continued to work closely with us to design an external circuit that will work well with DCC. We have settled on a current transducer (CT) to recognize occupancy. This is like a donut with a wire passing through it. The wire connects to the rail being used for the signals. The uncut rail acts as the common. Bill has modified his circuit several times for me as we tried various methods to get the logic to go low enough to trigger the aspects. This involved modifying his boards by cutting traces and changing components. He is posting his "retrofit" on his website (see above link). Last night Bill Payne and I ran a diesel switcher across our four test blocks on the workbench and watched all the signal heads change properly in sequence. We still have a flicker problem to reduce as the engine transitions from block to block. This is related to the sensitivity of the circuit with the CT and capacitor value we're using so a little more "tweaking" needs to be done. Then we need to add detection from the "dark" territories at either end where there are no signal heads and resolve occupancy detection when the locomotive is stationary (no current being sensed). We may also need to install lamps in the vans (cabooses) to draw current so the end-of-train is detected as it exits a block. Still work to be done but it's fun to watch the signals change. Now I need to build more signals and boards so that I can get the signals off the workbench and onto the layout.

Should you like to try your hand at this, I recommend that you check out Bill's site and look for the section about retrofitting his boards for external detection with CTs. You'll find Bill to be very helpful if you decide to try using his circuits. I'd suggest doing what we've done and try it on your workbench first. I had taken an old length of 3-foot brass track and cut one rail in three places for a test track.

The electronics aren't very useful unless you have signals to drive! Walthers had a sale on signal bridges in last month's flyer. I bought a couple. I also picked up some more Oregon Rail Supply signal kits. Constructing signals has to keep pace with making the boards. Then there's all the wiring to install. This is by no means a simple project. The end result is worth it when visitors come to the layout. Soon my operators will be getting demerit points if they run a red!

July 7, 2010 update

Work sessions have slowed down with the summer weather. However, the section of mainline from Valleyview to South Point has been wired with 5 consecutive blocks with bi-directional 3-color LED signals. All are working perfectly. Short "dark" sections have been wired to the CTs at each end. The detectors have been built on small pieces of breadboard. All that's left is to wire them to the CTs and add the "handshaking" lines to the singal heads at blocks 1 and 5. We're getting there.

---

Return from "signal circuit" to the model railroad wiring overview.

Go from "signal circuit" to photocell train detector.

Go from "signal circuit"signal construction.

Return from "signal circuit" to my Home Page.

Return to Top