DCC Automated Branch Line Signalling

[JST]

I have fitted Heathcote Electronics block signalling gear to my main lines and they work swimmingly well but I am now wondering what to do about my single branch line. I tend to leave a bubblecar or push pull auto coach on shuttle mode while I do other things on the mainline. Now I would like to add signalling to the branch line.

 

In simple terms the line goes from the main station bay platform to the bay platform of the BLT and the total distance is about 60 feet. I want to put a Dapol semaphore start signal on each bay platform and a 2 aspect colour light shortly after leaving the station at each end. Is there any gear out there which will operate this set up automatically without "SPADS" while the trains on "shuttle" but with the ability to have manual control? I cannot get too hung up about prototypical accuracy as it is only me my family and friends who watch the layout hence the suggested signal placement is for best visual effect rather than anything else. I run the layout via Hornby Railmaster and an Elite which controls the track and an eLink with a 4 amp supply that runs the accessory bus. I could set up a programme of signal control within Railmaster but this would be time bound rather than actual position related and I don't know if you can "loop" the programme.

 

Any ideas welcome and I am quite prepared to be told I am in cuckoo land!

[RAF96]

You can definitely loop an RM program, but as you say unless you crash the buffers now and again to ‘reset’ the time based sequence its a bit hit and miss on continuing accuracy. I used the repeat command to stress test a sound decoder, by incrementing the speed up and down every thirty seconds, whilst playing each sound in sequence on and off. The repeat command let me run it for several hours behind a firmly closed door to avoid upsetting domestic management.

 

Now if you had the Hornby loco detection system then that would solve all you problems. Unfortunately they have yet to bring it to market after more than five years of waiting.

 

Train-Tech do a signal based relay system that makes a train wait for clearance, but that would be stand alone not under RM program control, as any waiting time time at a signal would not reflect into the program and hence it would throw it all out after the first stop. See the video here http://www.train-tech.com

...links on that page to relay control and train control.

Edited July 25, 2020 by RAF96
More info

[Nigelcliffe]

It would be possible to link to the RM software, but only in one direction:  RM can tell the branch to go into "auto signalling" or "manual signalling.  Rough architecture like this:

 

DIY Accessory Decoder, Arduino based, responds to Accessory Addresses.  One address puts it into "auto" mode, other addresses used for manual control of signals.   In "auto" mode, sensors on track (magnet+reed switch, light detectors, occupancy detectors, all are possible) determine train positions/directions, and operate the signals with timer delays on resetting the signals.  When not in "auto" the signals respond to normal DCC addresses.    
The "hard" parts strike me as determining where the train position switches need to be located, how you determine train direction of travel (possible, two sensors close together).

 

 

- Nigel

[JST]

I forgot to point out that the shuttle function is controlled by ABC modules in both bay platforms and by the ESU sound decoders so I am not sure how a RM programme is affected by this. The time delay in the decoder is always bang on every time so maybe there is hope before going down the more technical route inferred by Nigel.

Ray, to save me having to quiz the RM manual is it easy to set up a loop? I like the idea of programming in all the sound effects and luckily for me, the railway room is well out of the way of the domain stalked by "Management".

I will go and do some experimenting!

[Nigelcliffe]

I'd assumed "ABC shuttle" was likely here.   

 

My approach is almost irrelevant of the way the locos move,  its just assuming the locos move in a particular way at a reasonably predictable speed in the local area:  so they could be manually driven, so long as the driver does it in a reasonable way.   The sensors would detect a train (and its direction) and change the signals.  After passing another detector, the signals switch back.    (The use of it as a parallel accessory decoder is optional - you could do without that part). 

 

You could experiment the whole thing with bits of paper and cards for rules, and try to "play" the game of changing the signals following the cards as the sensors are tripped.  Thus proving your logic and sensor positions, without actually building any electronics.   Then prototype the whole lot with an Arduino and a few LEDs and buttons (to represent the sensors), before building it "for real".      

 

Simple concept for one end, showing two signals, buffer stop.  Four sensors (squares).   The two near the buffer give you direction of travel - depending which order they trip.   Store the direction of travel in the software.   When loco heading towards buffer, start a timer to set the signals to "go".  The signals should change before the train moves, so the timing isn't critical.   The first two sensors trip, confirming direction of travel away from buffer (no action to be taken).  Then single sensors trigger to change each signal back to "stop" when passed.    When train returns, the sensors for the signals are not important, they can trip, but the signals are set to "stop" anyway, so clearing things is irrelevant.   Passing the pair gives direction into buffer stop, and re-starts the timer sequence to set signals to "go" again.    

 

 

 

- Nigel

[JST]

Thanks for this Nigel but I have just been up in the railway room and tried RM in record mode. It works well. Because ABC shuttle is in play there is no danger of the bubble car hitting the buffers at either end. I currently have the delay set for 1 minute which allows a good margin of error for signal changing but the Dapol bubble car seems very consistent in it's running times and so it all seems fine for my approximate purposes. No points changes are involved so what can possibly go wrong? 

Just need to figure out how to put RM programmes into loop mode.

[Nigelcliffe]

Its unlikely to go wrong, as you describe its a shuttle train on its own bit of track, running on the ABC auto-shuttle feature.   So that keeps running regardless of the signals.  

The signal timing will drift over time, because you can't get two clocks to run perfectly without some external reference (the two being "the time a train takes to complete a shuttle loop, variable on temperature amongst many other factors" and "the clock in the RM software").   So, eventually your signals will change "out of sync" with the trains.   Whether that drift happens sufficiently rapidly to be an issue for you, is another matter - if it takes a week or more to drift, chances are you've shut down the layout and gone downstairs as you're worried why your tea is late well before the problem occurs ! 

 

- Nigel

[JST]

Absolutely right Nigel! I will have to make sure that I don't run the shuttle for too long when one of my mates from the UK is over visiting - he is a retired BR/Amey senior signals engineer!