A new technology to increase enjoyment of the hobby?

Very interesting

I wrote a program to allow my grandsons in Kansas to operate my trains in California.  Skype was used for the visual interface.  I tested the program using my home network and I could control the trains wirelessly over my home network.  I abandoned attempts to use the program over the Internet because I had to place a hole in my firewall and I have no idea how vulnerable that would make my computer to hackers or viruses.  

 

Earl

       


Skype uses 256-bit AES encryption (although some security concerns still exist.)  If you want to make security more robust, the best and reasonably uncomplicated approach to this is to use a separate computer as the host computer for the application with nothing else installed (except antivirus) and as we're talking about a home network, you would want to set up port forwarding on your router/firewall for that computer.  That way it is accessible from the Internet with its own public IP address but your private network is kept hidden.  If you want to get fancier you could beef up the security on that PC more by making it a web server, designing your own web interface and secure it with SSL so it's encrypted.  There are other ways to segregate a home computer to be shared on the Web but that warrants its own topic.

Both DCS and Legacy are outdated technologies.  Why can't we have a Wifi receiver/transmitter in each engine, network the entire layout and run our trains from the computer, Ipad/tablet, Smart phone, touch, pod what have you.  Change sound sets on the fly, program them with the features we want.  No more wiring issues, ground plan issues. 

Why can't we have a Wifi receiver/transmitter in each engine, network the entire layout and run our trains from the computer, Ipad/tablet, Smart phone, touch, pod what have you.

You would rapidly run out of IP addresses and/or bandwidth.  People already have issues managing their WiFi nets at home with only a handful of devices.  Imagine trying to run a net with 50-100 and that's for a mid sized home layout.  A club layout could require more complex sub netting and expertise in network administration that would make ground plane/star wiring look like child's play.

An internal IP4 class C /24 subnet is designed for up to 254 available IP addresses.  On the average only a few at most would be eaten up/reserved by pesonal home computer(s), router, WAP and/or switches.  Maybe several factoring in other wifi devices.  So for the average hobbiest that may have at most a dozen or so engines the available scope of IP addresses and resorting to subnetting wouldn't really be the issue.  But I do agree with the potential bandwidth issues, not to also mention the inherit problems, especially with signal reliability, that's associated with wireless.  Hard-wired will always be more reliable (not to mention more secure) than wifi.

I think the bandwidth requirements would be pretty low.  You're not constantly sending detailed information to every loco on the layout.  A couple of packets at best if you're using a DCS-like protocol.  Wireless networks span entire enterprises - you just have to be creative.  I have seen 30 clients streaming video from a single, enterprise class access point. I'm pretty sure someone could get 30 trains to blow their horns at the same time

 

From an addressing perspective, a flat Class C subnet could hold a maximum of 250 hosts.  I'm sure there are some here who have that many locos but most of us don't.  Plus, if you wrote the spec right, DHCP could be used to assign addresses to the hosts like it does now.  Unless you have ALL 256 locos on the track and fired up at the same time, that should be enough for just about anyone.  If you want to extend it to other layout devices or switching, think about it in terms of power blocks.  We deal with power blocks now on layouts to effectively distribute power to different areas of the layout.  I assume you could use address blocks (or subnets) to define different groups of trains.  Bear in mind that DHCP is a pretty deep protocol - the engine discovery and pre-configuration options are virtually endless.  For example; a loco could come on line and based on it's configuration data, could be configured on the fly with specific features or limitations based on user-defined parameters.  

 

This stuff is child's play to those who have worked with it a lot/grown up with it, and I'd be willing to guess that these days you'd have an easier time finding WiFi gurus than you would layout/electrical gurus.  Thats what is scary, but yet intriguing - it might open up the hobby to a generation that has grown up "untethered".  And once the trains are "on the net", then things get real interesting.  Remote control via a browser interface from anywhere on the planet?  Done.   Engine configuration via a simple  tool or browser-based interface?  Done.  Simple command line troubleshooting/debugging?  Done.  Think about putting a small camera in your engine and then opening up a browser session to issue commands to the loco ad "see" out the front window as the train moves down the track.  You can send this view to someone else, and they can cruise down your layout in real time.  

 

<sigh>

 

But I suppose DCS will have to do for now....

John, you'd be surprised how fast the addresses go.  With laptops, tablets, smart phones, wireless printers, etc the actual computer net in the house could easily run up to 20+ addresses.  I'm guessing the typical train nut that has command control has over 20 loco's.  When you throw in accessories, switches, controllable power supplies etc the numbers start running up pretty fast.  

 

Managing them is also an issue.  Unless you name everything and set up a local DNS server you would have to set up static addresses.  You then have to come up with some way manage the static addresses.  Managing an IP net isn't much fun and it isn't that easy.

Actually Chuck, I wouldn't be surprised at all.  I'm an IT Manager/Systems Engineer by trade, I work with this type of stuff every day.

To clarify, my reply above was geared more towards the average home layout which isn't going to have nowhere near the same amount of nodes, including trains and accessories as would a mid or large size personal or club layout.  Also factoring in the average age demographics and the fact that most in this hobby are empty nesters, the amount of wifis in the average hobbyist's home consuming IP addresses would probably be less than you believe, IMO.

May want to look into the newer tablets coming to the market. Almost all are wireless - and almost all have forward and rear firing cams. So far, the rear cam is a bit grainy - but I am sure they will beef it up shortly.

 

I've set up one commercial instance for a training environment using McTivia between a tablet and a 70" panel. The tablet is wireless so you have the mobility+ the touch screen...and the cams allow you to send images directly to the TV.  The forward firing cam looks great on the 70 incher! The lag time is maybe a quarter to half a second.

 

There is some great technology out there...coming to store near you soon!

 

As for wirelessly controlling trains...I'll bet 10 bucks to a doughnut that soon we will hear....

 

"They got an app for that..."

To the question, Yes I use Skype, mostly for work when dealing with colleagues, customers and others around the world.  Audio quality is better than the phone.  I have used it for video, and it typically works fine, I think it could be interesting interacting with others.

 

As to the question on the use of wifi for train control, I have to agree.  Given that we are in a relatively small market, we have to leverage widely used technologies to make things cost effective.  I don't see an issue with the capacity of having a wifi approach to a typically train layout.  Would you separate it from your home network, probably, we shouldn't confuse the technology selection with the application, by that I mean just because the physical layer is wifi, that doesn't necessarily mean it is on the same network we use for our computer, printer and so forth in the home.  The cost of adding a wifi device to a circuit board is cheap, cents, and a wireless router is under $50 at most stores.  With respect to bandwidth, we are sending close to real time control, very low bandwidth, this is not a serious issue.

 

In resent years, the industry I am in, Radio and TV broadcasting, has adopted IP technology with relatively good success. Is there still some dedicated hardware, sure there are just some jobs that need to be done with dedicated hardware, and off the shelf is not applicable. Managing digital content in baseband audio and video yes, high power radio and tv transmitter, no.  Chances are very good that the TV show you are watching, or the radio you listen to is going thru a IP system at some point, and live audio and video require a lot more bandwidth (especially HD) than our control protocol for trains.

 

 

The hardware is the easy part, it is the software that is where the investment is.

I would be happy with an app for iPad and iPhone to run my MTH trains and another one for Lionel trains.  IMHO this is definitely the way to go.  Want to appeal to the younger generation - do apps!

Harold

If you want the remote technology to become wide spread you would want receivers in each device (e.g. new FastTrack switches) instead of just the loco's.  This causes the number of nodes to jump quite a bit.  Each switch, each accessory, each operating car needs it's own address.

Here is what I don't understand. How do the R/C plane, car, and boats do it? They seem way ahead of Legacy or DCS.

 

Ray Marion

Here is what I don't understand. How do the R/C plane, car, and boats do it? They seem way ahead of Legacy or DCS.

 

It isn't just DCS/Legacy, DCC has similar issues.  The main difference between cars/planes/boats, etc is that there is usually a limited number of them and these don't require the backdrop of a "layout" to operate them.

Sorry to be a wet rag about this, but give me the feel of a ZW or Z4000 handle in my hand anyday... 

 

Rusty

Thanks, but no thanks.  I'm already getting more than enough enjoyment from this hobby.  For me, it's all about the trains, railroading and the history of railroads, and creative modeling; not some fancy gadgetry. 

This is so simple a 13 year-old kid could understand it. Go out and get me a 13 year-old kid, I can't make heads or tails of it! (apologies to Groucho Marx).

I'll stick to Legacy and DCS!

Why use a class C lol.

 

10.0.0.0 would work perfectly well and seriously if you have more engines then a class A....   well..  then you have a problem I'll never have.

 

I dont know much about this subject, but there was a Marklin train video posted on this forum that used an I-phone for control. I know it just interfaced (may not of said that right) with current control systems, but I thought it was pretty impressive. Why not not something like that? Apparently that App is already out there.

Technology is already too invasive to suit me. Trains are my escape from it. But you go ahead and have fun anyway.

 

Pete

Apps/Software to control existing systems like DCC and TMCC and some of the earlier control systems have been around for a while.  Marklin introduced in 1984 a digital control system for their AC based stud rail trains and it too supported  a computer interface.  This technology allows you to use a smart phone/tablet/computer to replace the remote and potentially run the layout from anywhere.   You still have to do all of the work to set up the existing system.  The phone/pad/laptop is an adjunct device.  

 

The original poster's thought was to toss all of that out and have everything on the layout have wifi web enabled controllers that you would talk to directly from a web interface.  No DCC/TMCC/DCS.

I've no doubt that you could do some amazing things using a computer to run a layout through a wireless network, but the last thing I personally want to look at when I'm in my train room is a computer screen!

 

It would be nice to see a port on our modern locomotives for diagnosis like the OBDII on cars, which would greatly assist repairs and upgrades, but for me the fun of running trains is doing so manually.

 

It is the same for me in the "real" world. When I go to a spot to watch real trains, I'll bring a scanner, watch the signals and listen for the distant sound of a train coming. But I won't be hunched over a laptop watching ATCS and knowing exactly when the trains will be passing through. I enjoy the anticipation as well as watching the trains go by. However, I can certainly understand that younger folks might find using technology enhances their real and model train experiences and we all benefit from technological advances.

 

So, I encourage the technology folks to continue to expand what can be done with our trains, but a computer will never "run" my train layout.

Just because something can be done doesn't mean it should be done.

The modern 2.4GHz DSSS (Direct Sequence Spread Spectrum) used for R/C airplanes has put to rest the old problem of frequency interference and someone ‘shooting down’ your airplane by inadvertently turning on a radio on the same frequency. It is also practically bullet proof against other types of interference.

 

Think about this. In the mid 1970’s I worked in the development of barcode reading technology for a variety of applications in addition to actual barcode reading. Back then, when attempting to read the series of small parallel lines that made a barcode, said barcode had to be a precise distance from the reader, had to be exactly a certain size, and it required a number of additional stipulations for it to work.

 

If anyone would told me back then, that in the near future a barcode could be printed on the side of a loaf of bread flimsy, flexible plastic wrapper, and could be read regardless of orientation at a distance of a foot, I what have deemed you delirious.

 

Yes, all that you wish for our trains, technology-wise, will happen, and sooner than we think.

 

Alex

Why is it then when I am in the self-checkout aisle in the supermarket I need to rotate to bread a dozen times before the reader knows it's there?

Just curious, why would such as system use a private class C network with the standard subnet mask (255.255.255.0) and only get 254 addresses?
Why not use 255.255.0.0, and get 65,536 addresses?

Still not enough, the mask could go smaller, or how about IPV6?

If it was a train-centric system, you could address it however you want.  I think most here are thinking from the perspective of a home network perspective.  In those cases, most are set up as simple class C, but in most cases that could be changed.  

 

Was thinking about this more today in the car.  All of the features and functionality of IP baked into a train system would be pretty cool.  Multicast (all PRR engines do 'this'), unicast (PRR E-7 5859 do 'that'), broadcast (everyone do 'this').... 

 

I'm guessing we may see it someday... and a possibly a whole lot more.  

I still think the R/C hobbies have it partially figured out. There is a R/C car race course near where I live. They race a dozen or more cars at any given time on several different tracks. How can this be less complex than even a club layout?

 

Ray Marion

CW Burfle you are spot on!

 

Also DHCP!  You  don't need  dedicated addresses for each engine its not DCS or TMCC.

 

I am pretty sure this would work.  

 

 

 

quote:

I still think the R/C hobbies have it partially figured out. There is a R/C car race course near where I live. They race a dozen or more cars at any given time on several different tracks. How can this be less complex than even a club layout?

 

You have 12 guys each controlling ONE car/vehicle.  The controller and vehicle's radio's are paired to make sure that they don't control someone else's vehicle.  On a layout you might have twelve guys each controlling any one of the 12 or more locomotives, switches, accessories on the layout at the same time.

 

quote:

Also DHCP!  You  don't need  dedicated addresses for each engine its not DCS or TMCC.

 

For this to work you would have to set up a RESERVED DHCP addressing scheme which usually means a data base with MAC addresses and "names".  You don't want a piece of equipment that will pick up a new and different address each time its powered up?  Changing the subnet mask will allow subnet spanning (aka larger than 253 addresses) but it's not going to make the set up easier.

 

There is also the issue of how you would actually try to control the trains?  Someone has to write the software to actually control the trains instead of just programming their set up?  

Just my two cents.  I have two grown boys and two grandsons.  They are unimpressed with Legacy and DCS.  Their generations are living on text messaging and iPhone applications that were inconceivable only a few years ago.  Whoever comes up with a way to integrate these new technologies with our model railroads has a chance to win and hold the devotion of the younger generations, and, maybe, just maybe, will sustain the hobby for yet another century. While there are many young people involved in the hobby (check Youtube!), I'm not certain that will be enough to sustain the hobby and the manufacturers supporting the hobby.  I hope someone steps up, the next Neil Young, to create the next generation of interactivity.  Thanks for your time.

 

Class C was only used as an example in my case because that's the most common subnet mask used and 254 IP addresses would actually go a long way for the average home layout.  A class A/16 subnet mask you used as an example, well, you could go further than that and use a class A 255.128.0.0/9 mask which would give you a whopping 838,8606 hosts and 2 subnets.

 

In regards to IPV6, not all networkable nodes are compatible with it, and it's actually more complicated on average to set up compared to IPv4.  IPv6 was really developed due to public IPv4 addresses getting close to being maxed out.  It really has no signficance in regards to addresses constraints on a private home network, let alone a small to mid-size corporate network.

If you want to have better consistency and predictability when it comes to programming or troubleshooting, Static IP would make more sense.  Also, DHCP when it comes to IP networking requires bi-directional communication to work which TMCC doesn't have as Chuck above basically alluded to.  When a networked node is first fired up, it broadcasts out a DHCP request by providing its MAC (hardware) address to the listening DHCP "server."  So both the DHCP server and the node its providing a DHCP lease to need to be able to both "talk" and "listen" to each other (the server needs to know a node is needing an IP address, and the node needs to tell a DHCP server it needs one).  On the flip side, if you have a huge amount of locomotives, switches or accessories that you would be leveraging wireless IP technology with, managing a static IP scope can be a bit of a headache. 

It's interesting how a topic takes on a life of it's own. My original post was more about communication technology, but the train control posts have been a fun read. Thanks Rich883 for addressing the original topic, and John, could you please elaborate about the security concerns? Thanks.

Probably a Rockwell Retro Encabulator.  I think the cardinal gram meters and spurving bearings are a big enhancement to our hobby.

 

How about a small internal combustion engine,it could run on smoke fluid and provide power for lights during power failure.

That Rockwell presentation is worthy of an SNL skit!

 

TOO funny!