.

[NXCALE]

.

Edited January 29, 2017 by nxcale

[kvp]

I will be checking more information but perhaps to save time, what sort of train detection do they use?

I notice some sensors (maybe from level crossing sets), I still need to make some research.

I think it's just current detection as the layout is divided into blocks. This is the cleanest solution but requires at least some minimal current consumption from each end car to work reliably (head/tail lights, internal lights, resistor wheelsets or at least a ballast resistor) and the blocks should be large enough to have in them at least one car with active current consumption while the train passes over them. The flickering of the block occupancy on one scene also suggest the use of this. (and some very small buffer caps) This detection could be done with cheap and simple series diode/parallel optocoupler detectors, but i don't see any circuits, so i can't tell.

 

The other option for unitrack would be Kato crossing gate sensors (both the optical and the electric kind), but that either requires too many of them or a complex entry/exit logic. And the 3rd option is Kato signal sensors, as they are essentially directional wheel sensors, very much like the Tomix TCS types, but integrated with the automatic signals. I don't see these at all on the video, so i assume they don't use them.

 

ps: Some stations seem to be manually signalled, while some are automatic or semi automatic. This means there are multiple technologies in use.

[NXCALE]

.

Edited January 29, 2017 by nxcale

[kvp]

Those current sensors seem to lack one imporant component, the high current series diodes that the traction current is going through. For flickering, the easiest method is to have a pull up resistor/pull down optocoupler based system with a large enough capacitor that provides a timing delay. Tomix and Kato both use this system for their threadle sensors, but it works for occupancy detection too, you just connect the detector parallel to the series diodes and with a lower resistor instead of between the spring loaded detector rail and the other rail. The effect is that once the block is signalled occupied, the capacitor is discharged (=signalled as red), then gets charged slowly to half (=yellow) and then gets fully charged (=green). This works for block occupancy and you can add external logic (next block occupied=red, next block free+next next block occupied=yellow, both blocks free=green). Very very simple and cheap and requires around 5 components (diode bridge used as a series 2x2 diode group, optocoupler input series resistor, optocoupler, optocoupler output pullup resistor, output parallel noise filter capacitor) Personally it seems easier than wiring up every piece of track with a cartload of optical sensors and it works with turnouts too.

 

On the other hand, you don't need resistors wheelsets, you just have to size your blocks to be large enough that at least one current collecting part of every train (at least the front and back cars) are in them. This means that each block must be at least one train length or you have to connect multiple blocks. (a central pullup per input allows a wired low active OR and diode logic allows A+B and B+C grouping forming an AB and a BC block with 3 small physical blocks driving 2 logical ones, this is usually in use for short turnout districts)

 

Power routing: The default output of the detection system is not occupied, but adding a small continous power to each block is possible and CL power is great for this. Also assuming all nonselected routes to be occupied by default is an easy way to wire everything. This means if you select a station track and it has a train on it, then the signals will drop to red. Thrid option is to monitor track power with a parallel power detector and monitor the occupancy with a series detector and only clear an occupied block when the track power is present. This means a block could only get occupied while under power, but it remains so until you power it and the detector could actually check if is still occupied or not.

 

 

 

They also have small control panels and it can be seen how they are trying to follow a train graph (i.e. follow the schedule).

The idea behind the european Fremo system is exactly the same. It even suggest getting rid of most automation and always signal manually, so each station must be manned in one way or another (manual, key operated, electronic or even ctc is allowed). Drivers must follow the signals, written or audio directions from traffic control and the timetable if possible. My club tries to do this on most exhibitions if the layout is large enough and we have enough people to control it.

[NXCALE]

.

Edited January 29, 2017 by nxcale

[kvp]

That sounds interesting and cheap but  there is also the cost of the sensors (Tomix/Kato).

The Tomix and Kato sensors are threadle type, meaning they detect the metal wheels passing over them. It's better to add the aforementioned diode/optocoupler detector to the traction feed of every block as this makes a really simple and really cheap system.

 

This is one example, the optocoupler drives a comparator module for signal amplification:

http://home.cogeco.ca/~rpaisley4/OptoBOD.html

(on the top diagram R1 could be left out and the bidirectional optoisolator replaced with two single ones, while the comparator is only for driving higher loads than what is viable with TTL signals)

 

Fremo: https://www.fremo-net.eu/en/home/

(most info is in german though, but google translate works nicely)

Edited December 14, 2016 by kvp

[medusa]

nxscale,

 

Hall sensors work fine in model trains. Some time ago I built a speed measure with one (Hall sensor glued into bogie, one axis got a really tiny supermagnet glued to it, pulsed signal fed into motorbike tachometer). Can give a description in another thread if someone is interested.

 

~D.

[NXCALE]

.

Edited January 29, 2017 by nxcale

[kvp]

 

 

Just a query of the above. So, does the capacitor work as a timer? In case a train stays in a track for long.

More or less. The idea is that if a train keeps the contacts active (and the sense line pulled to ground), the timer is reset. This happens with every passing wheel, so actually the last axle (re)starts the countdown. If you stop across the sensor or right behind it, then the timer will not be reset and this means the signal will eventually turn to green, regardless of a train standing across/behind it. A quick fix is to add a manual switch between the reverse direction sense line and the circuit ground which could keep the signal red indefinetly. (good for exit signals as this could be interlocked with the turnouts) Afaik Kato has already made this upgrade available for their H0 range.

 

You should make a video to see the operations in your club.

There should be some videos of our layout at the Temofeszt exhibition in the Railway Historical Park (old Nyugati station roundhouse) from this fall. We don't have the space to assemble all the modules in the club room (it's over 100 meters long), so usually only place at most two smaller stations along the wall for playing. I'll ask everyone who made a video as i only took a few pictures with my phone. The schedules are usually written by IST, who is also here on this forum.

 

Hall sensors: While these tend to work, i found it hard to equip some locomotives with magnets, while some has quite large, but changing fields under them due to the motors being placed low. A more easier way is to use infrared or mechanical threadle type sensors for speed measurement.

[medusa]

Thanks for your comment.

Please share any pictures, videos, diagrams, etc you have, I believe that all of us benefit when there are more project/examples available

 

 

Done. Please see

http://www.jnsforum.com/community/topic/12300-speed-measure-car/?p=147116

[NXCALE]

.

Edited January 29, 2017 by nxcale

[NXCALE]

.

Edited January 29, 2017 by nxcale

[NXCALE]

.

Edited January 29, 2017 by nxcale

[cteno4]

Beautiful! Where is that?

 

Looks like you are having too much fun! Trains at the beach, that's a new idea!

 

Jeff

[JR 500系]

Wow I cant help but ask this... Your office looks so high up! Must be great sitting outside on the balcony taking in the sights!  :)

[NXCALE]

.

Edited January 29, 2017 by nxcale

[medusa]

Good idea.

 

It reminds me to other forums where always a number of contests are going on. Could become here the contest of "Model Railway in the most unusual place".

So? :)

Edited December 24, 2016 by medusa

[NXCALE]

.

Edited January 29, 2017 by nxcale

[velotrain]

It reminds me to other forums where always a number of contests are going on.

 

 

Not a contest, but it seems there used to be regular "project parties" - who can tell us why they stopped?

 

 

http://www.jnsforum.com/community/forum/49-summer-2011-project-party/

[cteno4]

I think it was a mix of having time to set them up and enough folks having time to participate. we have talked now and then or getting them going again. if anyone would be interested in participating and or helping organize and run one, pm an admin and we can see if we could get them going again.

 

cheers

 

jeff