Tomix New Operation Control System (TNOS)

[mrp]

I found a video on YouTube showing the newest demonstration/prototype of the planned "Tomix New Control System" at the Shizuoka Hobby Show.

 

 

The demo layout has:

- 20 x TCS sensors

- 12 x independent power feeds

- 8 x remote turnouts

- 2 x control units

 

It looks like each control unit is daisy-chained to a number of I/O boxes/expanders, with each I/O box having a number of inputs and outputs - maybe something like 8 TCS inputs, 4 track power outputs and 4 turnout control outputs?

 

 

https://www.dropbox.com/s/9rmxjgmz8zahlvh/vlcsnap-2017-05-29-15h09m42s858.png?dl=1

 

The thing that interests me most is the wiring. It looks like they're using a new type of TCS sensor with hidden wiring from below, and I can't see any type of connector for track power so perhaps that's also a new style with access from underneath.

 

 

https://www.dropbox.com/s/rgh0pps57j0rrob/vlcsnap-2017-05-29-15h11m38s515.png?dl=1

Edited May 29, 2017 by cteno4

[katoftw]

Not a new power feed. The concrete tie single tracks can use the slab power feed from underneath.

 

All in all, it look like a very useful track system.

[mrp]

Not a new power feed. The concrete tie single tracks can use the slab power feed from underneath.

 

Yep.  I guess I'd forgotten that.

 

 

And looking at a piece of track, I can see a cutout with some plastic clips either end that looks like with a bit of hacking you could fit the 5567 TCS Sensor.  I'll have to order a couple and try it out.

 

Although twenty of those would set you back.  I guess that's where Tomix will be making the *real* money...

[kvp]

It looks like each automation unit supports 4 blocks, 4 turnouts and 8 sensors, while also allowing bidirectional running. Single direction running needs only 4 sensors for 4 blocks, always positioned at slowdown distance from the stopping point. (this is less precise than the two step slow and stop balise automatic platform stop used irl) The boxes could be daisy chained without any apparent communction cables between them, so this might allow full manual and automatic districts to work together. (like old european model railway single direction automatic block systems, accepting and feeding trains from/to the human controlled station)

[Kabutoni]

I'm really looking forward to this system. Although it'll be a significant investment, it's a flexible system for temporary and fixed layouts. It'll be a nice gimmick to relax and see your trains operate by themselves :)

[sandiway]

I love it!

 

I have the current very limiting TCS system and four sensors. A bit finicky with cutting out bits of the Wide PC track.

 

I wonder if the new system requires new points?

[bill937ca]

Generally new Tomix designs retain old compatibility with old designs. Track is an example.

[kvp]

I've seen the TCS automation unit used with the old brown shuttle unit sensors, so yes they are backwards comptible.

 

In this case all turnouts on the video have double track isolators installed, fremo analog style, so this new system uses ancient parts. The power routing in the turnouts however is not needed as stopped trains still have CL power, since the new unit is using an automatic Z circuit to switch traction power around, probably with a crossbar matrix. The nice thing is that it could be made to work with old pre finetrack 3 wire turnouts too and imho with some creative hacking even with things like maerklin z tracks. (as the sensor tracks are compatible)

[mrp]

I'd like to see more detail on how it will handle signals.  Some of the slides from last year seem to say that signals will automatically show red if the following block is occupied, and that trains will stop *before* the red signal.  It's hard to imagine how that could be done with the current Tomix stand-alone timer-based signals.

 

 

Unless it's just a "hack" using the known time delay of the signals changing from red to green.  If the control system knows when the leading train has passed a signal (from the TCS sensor at the stopping point) then - given the signal's built-in red-to-green timing delay - it could calculate when to release the following train.  It would look nice - but wouldn't correspond to exactly when the next block was cleared.  And it wouldn't work for station entry where the leading train might be stopped for longer.  On the other hand it's a huge amount of work for Tomix to release individually controllable signals.

 

I haven't noticed signals running on any of the prototype demonstration videos yet.

 

[kvp]

There are manually controllable Tomix signals, but they are all form signals. For individual light signal control, you would only need a single dumb signal stand without track and a suitable control signal, similar to the 2 and 3 aspect form signals. The main controller box knows the status of each block, so it's possible to control the signals based on that. The trains do stop before each block boundary, so that part is correct. This would actually eliminate all logic from the signals, but signals are not shown on the demonstration however.

 

Anyway, i have ordered a bunch of Tomix signal masts from HS that are built for automatic and/or individual control to test if i can make a nice and simple control circuit for them.

[mrp]

There's a good photo on this blog of a prototype/mockup of the connection box.

 

http://jg3t-ymzk.asablo.jp/blog/img/2017/05/14/48329d.jpg

 

It looks like the plan is for a modular system with plug-in adapters allowing up to 4 turnout connections and up to 4 track power connections per unit.  I suspect the primary reason for a modular design is pricing.  I'm guessing the plug-in modules won't be cheap - say ¥2,000 a pop? - about ¥20,000 for a fully populated connection box?

 

It does seem like Tomytec is judging the market pretty well for these "high-end" 40th Anniversary releases.  The new 4275 Island Platform Set with Lighting sold out the first run in about 30 minutes even at ¥10,800 list.  Then there's the complementary 5721 Platform Sound Effects unit for ¥17,800.  And not to forget the the new E233 Camera Train out soon for ¥22,000.  Ouch...

 

[bill937ca]

The figure I saw was Y50,000 which if you add all the pieces is about what the current 5563 runs from the manufacturers llist price view point. Our retail is usually a bit cheaper. But you don't see these high end toys go on sale.

[kvp]

There's a good photo on this blog of a prototype/mockup of the connection box.

 

It looks like the plan is for a modular system with plug-in adapters allowing up to 4 turnout connections and up to 4 track power connections per unit.  I suspect the primary reason for a modular design is pricing.  I'm guessing the plug-in modules won't be cheap - say ¥2,000 a pop? - about ¥20,000 for a fully populated connection box?

It's actually a pretty braindead idea to have this adapter modules as extras, since each traction controller could be as simple as a H bridge and each turnout controller could be as simple as another H bridge. If you have a double H bridge chip, that means you can drive two turnouts from the same chip. Now how much an L293D costs? 2 euro? (actually this is the local EU price, from china it's 1.5 USD) This chip could drive 2 bipolar turnouts, so for 3 USD, you get 4 turnouts. The same chip is also good as a traction driver for up to 700 mA (1000 mA peak). This is around the same as a normal Tomix controller. This means the driver chips would only cost 6*1.5 = 9 USD for a 4 blocks, 4 turnouts box. For control, you can get an Arduino mega2560 for around 10 USD from china with a USB connection built in.

 

My current guess based on the images is that the central unit is something with a stronger cpu (like and Atmega256 or even stronger) and there is a control bus going from that to the block driver, turnout driver and sensor boxes. The drive logic is in the extension modules, so a separate microcontroller is used in all modules to drive its output and communicate with the central unit. This is good and easy to expand, but requires extra circuits, so essentially each turnout needs its own decoder, while grouping 4 or 8 turnout outputs into a single box would have been better price wise. (the extension box might actually have a central controller) Also a modern microcontroller can drive (form hardware) up to 12-16 pwm ports, so careful design would easily allow 4 or 8 blocks to be managed by the same controller, with dedicated drive chips for each output. This would cut the manufacturing price of the units way down. Of course Tomix could sell it for much more, especially if it comes with special software, something like JMRI, but for the Tomix system as the developement of this software could get really expensive.

 

On the other hand, if this control bus idea is true, adding signals with a signal driver module would be easy. Also the older preview slides show 3 control boxes, one with a throttle (which is shown irl on the photo linked above with an USB connection and an SD slot), a layout trackplan on a screen (probalby a route controller box) and another with an SD slot and a video feed from what looks like a camera car. The latter has been changed to use a user supplied mobile phone while i'm not sure what happened to the middle one.

[sandiway]

And not to forget the the new E233 Camera Train out soon for ¥22,000.  Ouch...

Maybe better to get a drone and fly that yourself around your layout as some people do...

[brill27mcb]

Any news on this new control system and when it is expected?

 

Rich K.

[mrp]

There's some updated information on the Tomix web site.

 

http://www.tomytec.co.jp/tomix/necst/5701tnos/

 

It looks like release is planned for October.

 

Tomix 5701 TNOS Controller Set (including one T-ND100) - ¥46,000

Tomix 5702 T-ND100 I/O Hub - ¥19,800

[railsquid]

Eeek, I think I'll stick with traditional cab control.

[kvp]

Eeek, I think I'll stick with traditional cab control.

But that will have to be controlled manually. This system has everything packaged in. (and hopefully soon some 3 color light signal drivers too)

 

The base unit for (46000 JPY) has the following included:

● Control unit T-CU001 × 1 unit 

● ND unit T-ND 100 × 1 unit 

● Point driver T-DVP 01 × 4 

● Feeder driver T-DVF 02 × 4 

● AC adapter × 1 

● Communication cable × 1 

● Instructions (simple connection version) × 1 book 

● Manual (detailed version) × 1 book 

● Warranty card × 1 part 

● Special Memory Card × 1 

● Gap Joiner × 20 

● Seal × 1 sheet

 

On the other hand, if someone wants the same level of automation from cheap parts, then getting a 10 usd arduino and doing it yourself is possible. It's possible to make it plug and play the same way as the Tomix system, but that's a lot of work. Imho you can get 8-10 blocks, 4-8 individual throttles and around 20 sensors and/or turnouts out of a single arduino mega2560, but once you add up all the costs and make it plug and play, the Tomix system would be very much around the same mark. (one T-ND100 i/o unit can output 8 bipolar signals and handle 8 two channel inputs, which means at least 8*8=64 bits out, 8*2=16 bits in, a theoretical signal driver could piggy back onto a single 8 bit output channel, allowing up to two 3 color signals per slot, so a single base box could handle 4 blocks, 2 turnouts and 4 three color signals)

[railsquid]

But that will have to be controlled manually.

 

Not a problem, as it happens...

[Ochanomizu]

For that price you may as well go digital.

[velotrain]

kvp said, "once you add up all the costs and make it plug and play, the Tomix system would be very much around the same mark."

 

I should think the same is true for DCC, which would also require conversion of all your engines, and likely additional costs and work to provide all the functions of the Tomix unit.  Not to mention more conversions for any future engines that you may acquire.

[Kabutoni]

Somehow I really feel like saving up for this system. It might be expensive, but it's only about the cost of a few train sets. If your focus is more on running trains, rather than collecting, this would be a good choice IMHO. You can run anything from your collection on it and vary the layout every time, or experiment with it for a more permanent layout almost endlessly. It's expensive, but I can see quite a few people buying this system to build layouts and enjoy them with a beer while watching the trains roll by.

Edited July 14, 2017 by Kabutoni

[inobu]

Kabutoni,

 

You bring up a good point. Initially my though automatically goes to "why not just go DCC" but if one collects first and runs second this becomes a viable solution.

Buy it, run it, display it.

 

I have a friend that has 350 engines in his collection and I know for a fact that installing decoders will not happen.

 

Collection and modeling, two different needs, two different objectives. Learn something everyday.

 

Inobu

[brill27mcb]

This looks very interesting from my perspective of running a parade of trolleys and trams on single-direction loop tracks with automatic control and collision prevention. It would be great for layouts built with T-trak and the EasyTrolley concept at public shows. Digital control is not the same as automatic operation, since it requires attentive operators who can see every part of the layout in order to avoid crashes. When people get distracted or talking at a show, crashes happen!

 

kvp, am I right that the points and feeder modules are the small black blocks that look like they slide into the open holes on the front edge of the black I/O box? I wonder if it is possible to have 8 feeder and 0 point modules (or some other combination) instead of 4 and 4 in each box. It also appears from the latest, 13-minute video (with the wildly-uniformed operator) that they have 2 sensors per feeder block in that layout, doing the slowing and absolute stop functions, similar to the older 5563 Automatic Operation Unit.

 

It's unclear to me how the "programming" is done. Is it really by entering instructions into the system from its keypad or otherwise, or simply in how the cabling is arranged? I don't see how the station stops shown in this video, made when the block ahead is clear, could be programmed just by cable arrangement.

 

There's a lot more to learn about this system, but it is intriguing!

 

Rich K.

Edited July 14, 2017 by brill27mcb

[kvp]

The funny thing is you would still need the sensors and the turnout drivers for DCC automation and you'll need at least one throttle, a booster and a central. These are the same components as in the Tomix system, except it comes with 4 drivers (analog boosters). If you add up the price for detectors, accessory decoders (without vehicle decoders), a central with automation support, etc. they quikly go above the Tomix price. And then there is a need for decoders too, that won't really be used beyond what analog cab control could do.

 

In fact, i think it would be possible to change the pwm-cl modulation of the Tomix unit to dcc with software modification only. Not that there is any real need for it.