Steve Welcome to the forums.... Green Frog did a video/DVD about 10 years ago called the Apple Valley Line. It was a 6' x 10' plan based on one designed by Jim Hediger, Senior Editor at Model Railroader. It has 2 hidden through staging tracks and they used simple and cheap light detectors from Rat Shack. They placed several between the rails on each track. You could put one every car length for complete coverage in case you spot a single car. You may want to check it out. http://www.greenfrog.com/bmrr_3dvdset.shtml You can get it from Ebay REAL cheap - be sure to use this forums E Bay search engine if you do. If you put a "kit" together for sale I would suspect this type of info would come out if it became popular and expensive. I hope this helps and again welcome to the forums ratled
I think such a detector would be good. We have current sensitive detectors. The DCC current goes through the coil to the module. Wolfgang
First, welcome to TrainBoard! That's a very good idea there. I encourage you to pursue it further and don't give up - sometimes innovation takes some time to take hold. I work in electronics R&D myself, and I know how hard it is sometimes to sell a new idea to the world. It's a question of overcoming market inertia - people are content to be set in their ways and don't like to have their routines upset. For computer interfacing, it could be as simple as a USB-parallel adapter and a few extra components. The computer just has to poll the adapter once every couple of seconds or so. Keep it up!:thumbs_up:
A simple, easy to install system would be a god-send to people like myself who are aware that there is an ac and a dc current, beyond that - it gets real hazy. I'm attempting to model a short section of a railroad that was CTC controlled and have bought two of the four signals that I need but so far they are just real expensive scenery items. A system that worked reliably and wasn't horrendously expensive would be just the ticket for myself and others in a similiar situation.
I would welcome anything that didn't require me to change out "all" wheel sets to be detected. Sounds like it would work for ABS as well as CTC.
I've watched your videos and am impressed. Have to admit that on my layout, 3 - at most 4 blocks only would be involved.
I'm very impressed, Steve, and am intrigued as to the ability to know there is a portion of the train within the block. Do the IR detectors act in a similar fashion to an axle counter (counts x axles in and x axles out and any difference leaves the signals at stop)? Does it require two detector sets as a minimum?
For the right reasons too! The other Aussie Admin, Colonel, and I are signal engineers. Yes, sensor is the correct term. I see, does it have dual input/outputs? I'm thinking there would be an input from the track end and an output to the next track end and the modules are interconnected like a network. Are you referring to model or prototype, and if prototype would this be axle counters which use magnetic induction or an IR-based system? Appear to be, though the rest of the system is just as much so!
Elegant in it's simplicity! Thanks for the description, Steve. Your occupancy detector can be used with the current range of signalling modules available and doesn't need a computer. Add the turnout position indicator circuit you have the making for a stand alone signalling system that will operate like the prototype. The bonus over the present detection systems is no need for current sensing requiring the use of metal wheels and dropping resistors. A boon for N Scalers if the IR sensors are small enough.
I'm not quite with you here. If the beams are below the top of the rail, the rail will block them - I'd think ... Or do you drill holes in the rail for the fibre? There is a school of thought here that (I think) believes all future model rail control will require a PC hooked up to it. (I'm not going to take that discussion any further here as it is OT.)
Just a little warning on that, from my own experience in industrial controls - counting systems (pulse counters) have to be really fast to 'see' the pulse. It's not just how short the pulse is but also how quickly the next one follows as it needs to see the space between as well (it has to see the 0s between the 1s as well as the 1s). In your current iteration the ICs are always 'watching' their detector, but if you devolve it to a computer (or industrial PLC) the 'watching' becomes intermittent because the processor has to do other things. As a result the required cycle time of the program to reliably capture pulses is often too fast to be accommodated - maybe down to a few milliseconds (this on a PLC; you don't usually have this control and structure in a PC). So for these systems the pulse counting is done by dedicated logic in the IO module (the bit connected to the detector) and the computer/PLC just drops by once in a while (maybe every second or half) to see what the current count is. I'd strongly recommend you keep your existing system (up to a point anyway) and just report the numbers, or even just a 'block clear/not' status, back to the PC. Otherwise you'll need to research carefully to ensure the hardware/software linking back to the PC, and maybe the PC itself, are going to keep up with the pulse train - and I get the impression you'd rather not do that
Very interesting idea!! I'm not sure about the drilling holes in the rail and all that, but I wonder would simply sawing a rail gap suffice? As for the computer I'll echo Mike's advice. You do not want to get a PC in this system at that level. A PLD or a small microcontroller would be the ticket, and then provide the occupancy info or axle counts out a port of some kind (LocoNet or OpenLCB/NMRAnet come to mind, if not just USB). I applUd you, sir. Well done.
