Future Electronics for Model Railroads

I hope you retire young!

@Bob posted:

I hope you retire young!

Yep, like right now, no time like the present to embark on a venture like this!

All of your wish list is feasible except the costs, which is why those things are not off the shelf.  If your computer work includes product development, you should be aware that getting a *_product_* out the door is an order of magnitude more effort & cost than the design+prototyping+coding&testing effort to get an alpha release.  At my last job, we used the engineer's estimate, tripled it, and changed the order of magnitude. That formula would mean the $3 switch machine you envisioned would sell for $90 in small quantities.  The price would come down to a reasonable point with volume but that means the manufacturer knows the market, can produce the product, and sell it at a decent profit margin.

You have to provide documentation so that even a One-Delta-Ten-Tango can install and use your product.  Then there is ongoing warranty support costs that must be built into the sales price.  Fail at these things and the Internet blogs will kill your sales. 

A few weeks ago, a forum poster asked about getting a layout he bought with servo+Arduino switch controllers working.  Since there was no support available, the solution was to tear it all out and install Tortoise machines.

Every time I wince at the price of Lionel's LCS products, I remind myself that these are "products" and not a kit for the enthusiast.  I can plug and play with a high probability of success if I actually read the fine manuals.

WRT the future of train control electronics, reliable communication is the challenge.  Bluetooth is already available and I would expect to see WiFi-based engine control to emerge.   All of these technologies have to operate in an incredibly noisy electrical environment.  MTH's DCS was announced but delivered a couple of years later - probably because of signal propagation issues that had to be solved (and partially mitigated by us re-wiring our layouts).  Lionel's IRV2 shipping was 18 months late and Dave Olsen said they had to re-engineer the product to protect it from noise.  With large layouts, range for RF comms is a limiting factor.  Add in all the devices blasting out signal in the 2.4Ghz and 5Ghz consumer spectrum and expect interference that will degrade performance of your product.  Add a couple of hundred cell phones hunting for WiFi at a train show for competition in the 2.4GHz space.  I have seen a performance gaming router obliterate the 2.4GHz spectrum and jam the Legacy to CAB2 comms.  A lot of challenges to creating a RF train control product.  Don't get me started on wireless video (without dropouts).

So, while the items on you list are desirable and worthy for product development, I don't see future products coming available at anywhere near your price points.  I would like to see the Arduino and RasberryPi experimenters come out with prototypes to merge in-cab video with WiFi control.  For G gauge using battery power, this appears feasible with off the shelf parts.  I'm not sure the electronics will fit into O gauge yet.  +1 on Bob's comment about retiring young.

Even the simple electronic products I make for model train enhancement illustrate John's points, it's amazing how small a part the actual design and prototype cost as compared to the ongoing support.

I'm going away from SuperStreets to an Automotionfx system that has greater flexibility.

Screen control of switches is great for a small number of switches. I find distributed panels around the layout easier to remember and quicker to control. That's why I have a DZ-2500C interface circuit being  made hopefully in China. Interesting note I found a photo on LinkedIn of the sales woman who is handling my circuit build. Real cute. Too bad I'm so very old and far far away!

The photos of the sales reps on the Chinese companies are always cute girls.   I've dealt with a handful of the PCB outfits, they all have the same mold.

I built a 14-channel block detector using current transformers and a PIC microcontroller.  Since this is a conventional system, it controls relays to reduce power to blocks to keep trains separated.  In the attached photo, the controller is in the center and the current transformers are mounted separately under the table.

I think this is the key to keeping the young kids in the hobby. A lot of the fun I had as a kid playing with trains was when my father and I would design,  and build a new accessory. 

Today, there is so much more available to keep the kids interested. Raspberry Pi's, MicroBit's and so many other inexpensive systems that are easy to program. 3D printers that can used to create more accessories. We need to challenge the kids to make something, do something. Let it be more than just running the train around the track.

Lionel should start having hack-a-thon challenges for new accessories, or to upgrade old accessories. There are so many manual items out there that are just waiting for someone to figure out how to automate them.

John

Like I had a chance!

The costs that I show are not a unit purchase price but the price of the parts with my (free) labor to assemble.  I have built and successfully tested on a small scale the LCC block detection circuits and the LCC block signal circuits that work with JMRI.  The main skill needed is through pin soldering.  I also built test boards so that circuit boards I built could be quickly and methodically tested.  After playing with low cost servos I decided to go with the tortoise switch machines.  There are HO layouts successfully using the low cost servo switch machines.

Just curious, what were your issues with servos? 

Brendan

I like that the tortoise switch machine has 2 built-in relays.  One to manage the power to parts of the switch and the other to power indicator LEDs.  You can add relays to the servo but that is getting to be more work and a kluge.  I could also use LCC output instead of the relays but that adds a cost of $3.  Building a servo switch machine will also take more time than installing Tortoise switch machine and I would rather spend that time working on other problems.  My impression was that the tortoise switch machines would be more reliable than the less than $1 servos from China.  It was not a easy decision.  It is one of the few where I chose a more expensive solution.  I have been successful on ebay buying tortoise switch machines for $10 or less.   

Kal's post makes me think of the complex Lego layout at our holiday show last year.  The kids help build the trains and structures and dad added Arduinos to control various automations.  When the kids get older, they can take over the automations.

@carl552 posted:

I like that the tortoise switch machine has 2 built-in relays.  One to manage the power to parts of the switch and the other to power indicator LEDs.  You can add relays to the servo but that is getting to be more work and a kluge.  I could also use LCC output instead of the relays but that adds a cost of $3.  Building a servo switch machine will also take more time than installing Tortoise switch machine and I would rather spend that time working on other problems.  My impression was that the tortoise switch machines would be more reliable than the less than $1 servos from China.  It was not a easy decision.  It is one of the few where I chose a more expensive solution.  I have been successful on ebay buying tortoise switch machines for $10 or less.   

The builtin contacts are a definite plus.  I find the servos a bit fiddly and have had to revisit several installations to re-adjust both the servos and the relays.  I have started using hall effect sensors and 2mm cyl magnets instead of relays.  Of course, it adds a bit more electronics, but being less mechanical than a relay, it has proven to be a more reliable alternative.  For me, the attraction of servos was the electronics challenge.  I recently had to add a machine to a remote turnout on our club layout - I went with tortoise.  I've started playing with LCC. I like that you can use JMRI to program LCC.  I've been using MERG CBUS for years. 

Brendan

Brendan,   If you are playing with LCC you probably got the starter set from RR-Cirkits.  You should also read Dana Zemmerli's LCC book.  I built a simple LCC input node and LCC output node that work with the RR-Cirkit nodes.  I also built a block detection board that works with my input LCC node.  The parts for the 3 boards cost about $25.  When I get home in August I can send you those nodes to play with for $25 plus $10 shipping.  

@carl552 posted:

I like that the tortoise switch machine has 2 built-in relays.  One to manage the power to parts of the switch and the other to power indicator LEDs.  

I use Tortoise switch machines on my layout, but I have not attempted to power dead rail in the switches due to the current limitation on the Tortoise's switch contacts.  From the installation manual:  

"These auxiliary switches can switch a maximum of one (1) amp of current"

You may want to use the switch contacts to control relays in that application.