I am priced out of making my own model railway and I am investigating making my own locomotive out of mini drone parts so the power comes from the locomotive rather than the track.
I want the camera from the drone too, and an app where you get the driver's view and a few extra controls such as a 'dead man's switch' for modern era or lots of steam things for steam trains, with acceleration and deceleration simulated. This will make the minimal running time due to the battery acceptable.
Putting this together with parts from places like PCBway appeals more to me than normal scenic modelling. I am not looking to make my first billion from making model trains, I just want to see if it can be done.
The MVP is a basic locomotive, however, if it has the input data, then it can 'know' where it is on track to wait for signals, stop at stations or keep to a timetable.
The view could be augmented reality to paint in those stations and signals if the model railway owner can't afford the model ones.
I think that model trains need a reboot with an affordable locomotive that does not need hundreds, if not thousands of $$$ to get started with. Only boomers with their grandchildren's inheritance can afford model train sets. Children always want toys that are too old for them, rather than dumbed down for some child younger than themselves. The current entry level trains are generally Thomas the Tank Engine and that is not for everyone.
Please shoot me down in flames, or, if you know of any sensible 'mule' electronics I can borrow from a toy car or drone, please let me know.
I like this reboot idea. Let's say all engines (and many railcars) could have an ESP32, and all engines have rechargeable batteries. Then savings could be found by using commodity track that does not need to carry control signals and does not need to conduct electricity (neither AC nor DC).
Optional conducting couplers might enable DC power from engine to cars, and from reserve batteries (in boxcars or whatever) back to engines.
An issue is how to transmit control signals to engines, accessories, signals, etc. Bluetooth? Wi-fi? Something else? It should be something built into the cheapest boards. And ideally it could support streaming video from engines.
There's already a robust ecosystem around digital control of model locomotives. It's actually impressive that they managed to get to more or less unified industry-wide standard (DCC) -- locomotives can be addressed individually and told to go at specific speeds or blow horns/lights.
The track just carries a fixed AC voltage overlaid with the digital signal.
I understand the problem is that it's not really amenable to bidirectional comms-- so returning video might need to go wireless.
Perhaps I'm simply suggesting DCC transmitted over the air - again, to take advantage of cheap boards that have integrated Bluetooth and wi-fi, trading off for track that is significantly simpler and cheaper. 3D printed, even.
Switches and signals would still need DC power sources, but this could be the same simple DC feed supplying other accessories too. Put ESP32 boards everywhere and have (for example) some very nifty industrial sites on the layout. Heck, maybe functioning scale models of simple industrial processes (fermentation & distillation, anyone ?).
Not sure what is the state of motorized/RC HO-scale truck & construction vehicle models. I know that (for instance) Faller makes trucks that zoom around the layout, but they probably do not have additional remotely-controllable functionalities.
I'm going to do the opposite of shooting you down in flames.
If there were a website where people with VR headsets could login to being a passenger on a model railroad running through someones home, or simulate being an ant-sized train engineer trying to keep to schedule on the old Kitchen Carpet and Foyer line, that would be killer. It also doubles as a nice gimmick to show off a local business. Imagine getting a model rail passenger tour of the novelty hotel room you just booked.
This is brilliant and you should do it. Stop being intimidated by electricity. A drone controller is overkill. A model train is just a DC motor and speed control is just how much voltage you apply to the tracks. Wire up the tracks to a Raspberry Pi and just script it all in Python, don't waste effort trying to jerryrig a drone into something that its not. For the camera, a standard home security IP cam will do the trick, or just an APP that turns any old disused phone into an IP cam.
The Eureka I am plucking out of your kind words is to make it work as an online app, something that can be shared.
I am going to enjoy the iterations on this. It would also be good to innovate 'the Nintendo way' by using tried and tested hardware that is affordable.
Maybe Raspberry Pi is also worthy of study because that has kick started a whole sector and not been overrun by competition. Maybe it is possible to do the same for model railways, to have a standard locomotion unit that people can build their projects around, to share as you described.
I have decided against powered tracks. I like the idea of minimal tracks that look more like an aluminium curtain rail, embedded in the floor, and I don't want track cleaning to be necessary at the electrical level. Imagine a helix hidden under the stairs or in a disused chimney that enabled a train to climb upstairs, albeit taking quite a while to get there. It could exit into a properly modelled station but the journey there could be an AR trek across the countryside as far as the shared view is concerned.
I can also start with a 'reference design' kit for some of these cool controllers that have the bits on that I want. It is going to be fun getting there, but 'shared' is a must!
I'm going to double down on the suggestion of powering the rails simply on account of the fact the main product is VR camera on a train and a web service built around it. Reinventing the way model rail systems work from the ground up will get you side tracked. You can always figure out how to construct battery based, wifi enabled smart locomotives later on. There's already a ton of standardization and interoperability in model train building. You don't have to reinvent it unless you insist on avoiding track electrification.
That said, if you're really committed to designing from scratch and are looking at the ability to climb steep grades like stairs and chimneys, and if you happen to be in San Fransisco (as many here are), I strongly recommend the cable car museum. Admission is free and its extremely educational. The cable cars are a marvel of engineering. A wholly modern rapid transit built before there was even electricity. The way they accomplished this was by having a cable underneath the tracks that is constantly moving (powered by steam in the main depot). When the cable car operator wants to start moving, they pull a lever which holds the car to a fixed point on the cable (in the exact same manner as the friction brake holds it to a fixed place in the ground). The cable car museum shows in stunning detail how this system accomplished things all the basic railroad operations like turns, tracks crossing, and track switching. You might find some inspiration in seeing how they did it.
I saw a letsplay of it once, and on some weird youtube niche documentary rabbit hole I learned that train simulators have been entwined with the model building community since the start (late 90s) and the concept of having track layout planning, control over signals and interlockings, and first person simulation, all wrapped into a single piece of software is nothing new.
I would suggest prototyping stuff in lego, it allows you to quickly try an idea, see if it has legs and if not discard it and do something else in a couple of minutes. I've prototyped various industrial machinery like that and it saved me years cumulatively.
I want the camera from the drone too, and an app where you get the driver's view and a few extra controls such as a 'dead man's switch' for modern era or lots of steam things for steam trains, with acceleration and deceleration simulated. This will make the minimal running time due to the battery acceptable.
Putting this together with parts from places like PCBway appeals more to me than normal scenic modelling. I am not looking to make my first billion from making model trains, I just want to see if it can be done.
The MVP is a basic locomotive, however, if it has the input data, then it can 'know' where it is on track to wait for signals, stop at stations or keep to a timetable.
The view could be augmented reality to paint in those stations and signals if the model railway owner can't afford the model ones.
I think that model trains need a reboot with an affordable locomotive that does not need hundreds, if not thousands of $$$ to get started with. Only boomers with their grandchildren's inheritance can afford model train sets. Children always want toys that are too old for them, rather than dumbed down for some child younger than themselves. The current entry level trains are generally Thomas the Tank Engine and that is not for everyone.
Please shoot me down in flames, or, if you know of any sensible 'mule' electronics I can borrow from a toy car or drone, please let me know.