Japanese interlocking systems

[kvp]

I've accidentally run into an interesting site that describes the operation of a home made interlocking system:

http://www.roy.hi-ho.ne.jp/s-kato/signal1.htm

If i understand right, the red switches are the signal/running direction selectors, while the white buttons select the routes and the black switches select between manual and automatic turnout operation. The system also has route memory to avoid setting crossing routes. Some more explanation would be welcome here as i think i didn't quite get everything, like the usage of the white switches. (my best guess is they might be power routing/cab selection overrides)

 

Then i looked up what was used on the real railways:

Quite similar looking, but i found it without any explanation...

and another one from the Izuhakone railway, with even less info:

 

another, different, but similar type:

 

Can someone point me towards more information (english or google translatable) or explain me how these systems work?

[katoftw]

They are called "control boards".  If you google that, you should find the info you are looking for.

 

I cannot give you a description for the boards you have exampled, but I'll give you a vague description of the boards I know.  Some switches control points and signals.  Most real world ones will have lights in the lines to show the locations of the trains.  And some real world boards have kill switches for overhead power.

[kvp]

I'm looking for these japanese type of interlocking systems. I've already built a few european siemens (straight manual interlocking) and integra domino (route selection automatic interlocking) boards and even a soviet type (slider based route selection, manual locking). But i would like to know how the japanese version works, which seems different than anything mentioned above. They seem to be route based interlocking systems with route selection based on 3 position switches and pushbuttons with both route (turnout) selection and occupancy based route release automated or semi automatic.

 

I can't find info on the operation of the white and green switches. (afaik red switches are used for direction selection and white buttons are used for target track selection, while black switches are manual turnout overrides)

 

Any info on the axact system above would be great! (even the official name or trademark name of the system would help)

[Ochanomizu]

Hello Mr katoftw,

 

Can you post an image of what you mean?

[mrp]

If it’s of any help, Foxworks have a bunch of apps on Google Play with station simulations using similar control boards.  For example, Mikomayama Station.

 

 

They posted a video on YouTube showing the simlation running.

 

 

[marknewton]

This one appears to use Union Switch and Signal Company equipment, or perhaps licensed-built copies:

 

 

USS switches are normally marked L & R for signals, N & R for turnouts. The photo isn't big enough or clear enough for me to read the markings. But if the colours used to identify switches follow UK practice - which I'm guessing they do - then the black switches control turnouts and the red switches control signals.

 

Cheers,

 

Mark.

[kvp]

Thanks! The foxworks webpage contains the signal diagram of many stations with in game screenshots, so it's identifiable what controls what. It looks like red is used for main signals, white or green for shunting and yellow for auxiliary signals. The black switches are for turnout manual control and white buttons are route selection.

 

I found three types:

-manual interlocking (like the us examples i found)

-automatic interlocking with route selection and manual cancel

-automatic interlocking with automatic route removal based on occupancy detection

 

So far it seems all are implementable for a model railroad and increasingly harder to do. The linked site above seem to describe the 2nd variant with automatic routing and route protection but no occupancy detection.

 

I'm still looking for the exact brand or manufacturer as the japanese systems seem to have several improvements over the original.

[kvp]

Mode of operation, deducted from the videos:

-select the direction on the signal switch (it may have one or two)

-push the pushbutton of the target block

-if there is no route collision with already set up routes, it will light up the route with white light

-if there is occupancy detection, the train position will be indicated with a red light

-if there is occupancy detection, clearing each block on the trailing end may clear it to allow new routes

 

Switches:

-black is turnout or turnout override (center: automatic, left or right: manual)

-red is a main signal

-white or green are shunting signals

-yellows are auxiliary signals

-other switches can be white or black

-all switches have 2 or 3 positions

-pushbuttons are momentary and mostly single color (usually white)

 

Interesting info:

-unlike Integra Domino boards, most of the time there are no fixed building blocks, but each section can be as long as the diagram dictates and usually have only a single light instead of a chain (or as many as there are separate detectors)

-many automatic routing systems have the switches and pushbuttons on the track diagram

-while many manual interlocking systems have the switches in blocks (both for turnouts and signals)

-the blocks are color coded on most diagrams, unlike the monochrome variants seen elsewhere

-the system seems to be usable with a single hand, instead of the same time push of two buttons required for Domino

-the switch gives better direction feedback on what goes where than buttons and allows manual cancellation at the same location (instead of another separate button or other mechanism)

 

A model railway implementation can use the fully manual locking method, setting turnouts and signals by hand and only having minimal safety checking or use diodes and relays for automatic route selection. This combined with capacitor based turnout control allows very simple circuits to implement complex route selections with both light and form signal control added. (essentially the switch selects the signals and the pushbutton enables it, along with setting the route, while it disables all other routes and the relay holding of the signal/route is dropped by cancelling the route) Train detection is possible but makes it much more complex than viable for home application. If powered from the main, any fully power routing turnouts (like Tomix ones) automatically take care of the cab selection in analog as only the selected route receives track power and only from the controller of the selected main. For a double track mainline station this allows up to 4 valid mainline routes and if there is a storage yard, then it can have it's separate controller on it's entry point and which can be selected by selecting a routing between the station and the yard. Logic can use a per block holding relay, which enables each target button only when there is no active relay between it and the originating signal, but once the target is selected, the block relays are energised to self holding, only disconnectable by cancelling the route at the switch or completing it automatically (in case of train detection).

 

Did anyone ever see a similar implementation to the one linked in the first post? (in Japan or by a fan of japanese railways elsewhere)

[katoftw]

Hello Mr katoftw,

 

Can you post an image of what you mean?

I'll take some photos when I'm at one of my work locations next week.  Is very similar to the second pic down that kvp posted.

[NXCALE]

This post is very interesting.

I also have the idea of making a signalling control board but I think I am still far away to actually start on it.

if I built one, I would prefer to have it some sort of route setting and interlocking.

Anyway, I imagine that these control boards (the Japanese ones) are designed to work with DC rather than DCC.

If someone has mor examples, please post them.

 

[NXCALE]

Hi. I uploaded a couple of pictures with notes about the control boards (my understanding of it…). If you don’t want too much reading so just have a look to the pictures.

 

Pictures taken and modified from the link provided:http://www.roy.hi-ho...ato/signal1.htm

 

I haven’t watched/analysed the videos yet, I will try to do that later.

 

Picture 1:

• Route from signal 8 to siding (let’s called it route 8(1M)) set and locked by pushing/rotating RED button/knob then WHITE button.
• Note that there are smalls arrows painted on the board indicated if the route is unidirectional or bidirectional
• As indicated in previous posts, the black knobs are for turnout operation. Normal – Automatic-Reverse, LEDs help to indicted current position of the turnout blades.
• Shunt signal are the white knobs.

Picture 2:

• Route from signal 10 (let’s called it route 10(1M)) set and locked by pushing/rotating RED button/knob then WHITE button.
• Route from signal 8 (let’s called it route 8(2M)) cannot be set as is a conflicting route with route 10(1M).

Actually there is more to talk about this marvelous piece of engineering:

• This type of control board can be categorized as Entrance/Exit Control Board – you push a button to indicate the ENTRANCE then another is pushed for the EXIT of the route.
• Override knobs for turnout are necessary like in real life. This because if there is a failure in the system, you probably still need to set up some routes to keep trains moving. So, you need to override the automatic locking and start manual operation. Also, this function helps if people are testing the turnouts.
• The circuitry looks quite complex but I think is because the designer decided to do it at the “old school” style. The same functionality could be achieved with a microcontroller, I believe.

What interested me most is that I believe that the control board has been designed for DC. Some months ago, I was testing/trying how to build parts of one of this panel boards for DC. It was nothing too serious, just testing with train detection (IR sensors), signals, microcontrollers and… I temporally stop for personal reasons (i.e. shortage of spare time).

[kvp]

 

What interested me most is that I believe that the control board has been designed for DC. 

Yes, most japanese layouts are DC and if you use Tomix fully power routing turnouts, then power routing is taken care of by the turnouts, even on a multi track line as the not selected tracks are fully isolated and the turnouts route the power to the selected destination track. For turnout control, if you add a capacitor between one wire of the turnout and ground, then by selecting between power/ground on the other wire, you can switch them. This allows the straight/automatic/diverging switch to be no more than a 3 to 1 selector and the output can drive the turnout and the two leds directly. Route selection can be done by using a diode matrix with the pushbuttons and activating a holder relay for the route (that also disables all conflicting routes). This holder relay can trigger the diode matrix for the turnouts, that activates some of the output relays which select between power and ground. Lots of diodes (best done on a raster board), two morse relays for each route (main one deactivated by turning the power at the signal knob off, auxilary break on power ones activated by conflicting routes, can also trigger an error indication if it's a double throw) and one double throw morse relay for each turnout (shared between routes though the diode matrix). Signals (even multi aspect ones) can be powered directly from route signals by sharing the turnout diode matrix with the turnout relays.

 

It's easier, much faster and slightly less fun to do the same with a microcontroller. An arduino can be set up for this task rather easily and only a bunch of driver ic-s are required for the output stages.

[katoftw]

I'll take some photos when I'm at one of my work locations next week.  Is very similar to the second pic down that kvp posted.

Oops completely forgot.  Will do this week.

[NXCALE]

Yes, most japanese layouts are DC and if you use Tomix fully power routing turnouts, then power routing is taken care of by the turnouts, even on a multi track line as the not selected tracks are fully isolated and the turnouts route the power to the selected destination track. For turnout control, if you add a capacitor between one wire of the turnout and ground, then by selecting between power/ground on the other wire, you can switch them. This allows the straight/automatic/diverging switch to be no more than a 3 to 1 selector and the output can drive the turnout and the two leds directly. Route selection can be done by using a diode matrix with the pushbuttons and activating a holder relay for the route (that also disables all conflicting routes). This holder relay can trigger the diode matrix for the turnouts, that activates some of the output relays which select between power and ground. Lots of diodes (best done on a raster board), two morse relays for each route (main one deactivated by turning the power at the signal knob off, auxilary break on power ones activated by conflicting routes, can also trigger an error indication if it's a double throw) and one double throw morse relay for each turnout (shared between routes though the diode matrix). Signals (even multi aspect ones) can be powered directly from route signals by sharing the turnout diode matrix with the turnout relays.

 

It's easier, much faster and slightly less fun to do the same with a microcontroller. An arduino can be set up for this task rather easily and only a bunch of driver ic-s are required for the output stages.

 

Good info, thanks.

 

Yes, it could be less fun doing it with an Arduino but still challenging.

 

One of the tests I tried (some time ago) was programming the microcontroller in a way that it recognises the button sequence operation. That is, the board did not response if the Exit button was pushed first and then the Entrance button.

 

Another test was to allow a conflicting route for a second train after proving that the first train has fully stop for at least 5 seconds and this (the first train) is not over any section of the second route to set.

 

Then it started to get more complicated and I run out of I/O ports… and time...

[NXCALE]

Hi all. I had the chance of talking with a person who has worked with control boards in the real world (a person who does not anything about railway models but has experience working with the big ones).

 

I showed the picture below  to this person and he said that this is actually (from his point of view) a Unit Lever control board. A Unit Lever control board is not as “automatic” as an Entrance/Exit control board in few words.

 

 

 

 

I bother myself to find a kind of dictionary definition for “Unit Lever Operation” which I think could help, see below:

 

Unit Level Operation: “A method of signalling control with a separate lever for each set of points, each release switch and each signal. In setting up a signal route, the signaller is required to operate the levers for required points and/or releases individually before operating the lever for the particular signal. This contrasts with route control systems, entrance/exit or one control switch. Source: ESG 007”

 

I explained my understanding about the operation of the control board above and this person mostly agreed but he pointed out that the switches will not get back automatically. These have to be always controlled by the signaller.

 

My second question was about the white lever/switches shown on the top of the control board (see quotation). He does not understand Japanese nor knows about Japanese signalling practices (he only knows the UK ones). So he could not tell if these were for overhead wiring or for telephones.

 

If someone knows what those white levers/switches do, share it with the pack. Guessing is also valid as we all are learning.

 

Finally, I was checking online and I found a picture (not a good one) where an Entrance Exit (NX) control board and a Unit Level control board can be seen together. The link of the blog that holds this and other pictures is the following one:  http://position-light.blogspot.com.au/2013/03/photos-metra-blue-island-tower.html

[kvp]

First, the mentioned picture has those white pushbuttons that are used for automatic route selection. The black turnout control switches have 3 positions: straight, automatic, diverging. This means that the board is an automatic one with manual overrides and manual route release. The automatic route selection seems to be a japanese improvement over the american system.

[NXCALE]

Now I remember that I suggested to the person I asked about control boards (see post #15 ) that these white switches (yellow circle in the attached picture) could be for automatic route selection. He mentioned that this could be correct but he was not sure because there is something like a siding shown on the diagram (red line). So, these white switches are shown over that siding.

 

 

 

About the control board below, does anyone know the name/code of the switches installed on it?

 

 

 

p.s. I just noticed that the control board above even it has a key for switching on/off... so nice...

 

[NXCALE]

Hi all,

 

Checking some videos in youtube, I found one of the control board in operation!!!

 

 

Cheers,

 

Nxcale

 

 

 

 

[utrainia]

Hi all,

 

Checking some videos in youtube, I found one of the control board in operation!!!

 

 

Cheers,

 

Nxcale

Wow, those signals are a work of art! I'm assuming it is HO scale, because making working signals like that in N would send you straight to the mad-house. I wonder where he got his rotary switches from? They are very neat looking.

[cteno4]

Holy crap batman! That is a lot of great work. Love to see the backside of the panel!

 

Jeff

[kvp]

Yes, i've been trying to source switches like that for some time but without much luck.

 

I think i can help with the backside. The top black switches select between two fixed values or pass the route selection through in the middle. Route selection is matrix based, row/column selects are the signal switch and the pushbutton. This triggers a single self holding relay circuit (representing a route), that blocks the other buttons and drives a turnout matrix with the route pattern. The signal output is also a self holding relay circuit set with the route, that is cleared either by powering off with the switch or (optionallly) by the block detection inputs. The route relays have to be cleared manually by powering it off. The basic blocks are the diode logic matrices (row/column select to single output and single input to pattern select), the self holding relay (set and clear inputs, built from two morse relays) and the 2 or 3 position switches and close on push buttons. Turnouts and signals get a high or low signal and decide their drive outputs from that. The thing is pretty near to the original and requires only a bunch or morse relays and lots of diodes to work.

 

ps: i'm pretty sure it's doable in N scale too as you could mod off the shelf Kato/Tomix signals to be remote controllable or get ready made 3rd party ones with plain signal input wires

[Madsing]

ps: i'm pretty sure it's doable in N scale too as you could mod off the shelf Kato/Tomix signals to be remote controllable or get ready made 3rd party ones with plain signal input wires

Any recommendation for third party "Japanese" signals? I need remote controllable signals. I have purchased a Tomix TCS 3-aspect signal and plan to modify it but have not tried yet.

Edited July 19, 2016 by Madsing

[NXCALE]

ps: i'm pretty sure it's doable in N scale too as you could mod off the shelf Kato/Tomix signals to be remote controllable or get ready made 3rd party ones with plain signal input wires

 

I agree with kvp. Getting the signals (and perhaps downsizing the train detection sensors), it would be necessary to make it in N scale. The interlocking circuit and control board would be the same.

 

Any recommendation for third party "Japanese" signals? I need remote controllable signals. I have purchased a Tomix TCS 3-aspect signal and plan to modify it but have not tried yet.

 

If you are planning to use "normal" signals (i.e. the most common type signals like the ones offered by Tomix), I think you could go for any signals offered in internet.

Last year, I got these from ebay :

 

 

 

http://www.ebay.com/itm/5-pcs-N-Scale-1-160-LEDs-Made-Railroad-Signals-3-Aspects-Green-Yellow-Red-N-/400248654947?hash=item5d30adcc63:m:mnXJWDxdZsFI6g6_ALFgs1g

 

About these, I could say:

Pros: prewired signals, price is way lower than Tomix /Kato signal.

Cons: acceptable level of detail, it probably requires some painting, masking, etc.

 

 

 

If you require other "Japanese" signals, you could try for instance (the signals below are only painted):

 

Greenmax

Repeaters: http://www.1999.co.jp/eng/10359342

Shunt Signals: http://www.1999.co.jp/eng/10359346

 

TGW

Repeaters: http://www.1999.co.jp/eng/10080403

 

Cheers,

 

Nxcale

 

[Madsing]

Thank you, nxcale. I will purchase these models on ebay and try to paint them.

Edited July 22, 2016 by Madsing

[NXCALE]

Thank you, nxcale. I will purchase these models on ebay and try to paint them.

 

Not a problem, that is only one option.

You could have a look what else is in the market and let us know if you find something interesting.

 

Yes, i've been trying to source switches like that for some time but without much luck.

 

I join the quest! I am also interested in getting those switches.

These must have a code, we just need to get it to find them.

It would be great if someone who saw the switches (in their clubs, local model shops, friends' houses or exhibitions) could ask for this information.

 

I think i can help with the backside. The top black switches select between two fixed values or pass the route selection through in the middle. Route selection is matrix based, row/column selects are the signal switch and the pushbutton. This triggers a single self holding relay circuit (representing a route), that blocks the other buttons and drives a turnout matrix with the route pattern. The signal output is also a self holding relay circuit set with the route, that is cleared either by powering off with the switch or (optionallly) by the block detection inputs. The route relays have to be cleared manually by powering it off. The basic blocks are the diode logic matrices (row/column select to single output and single input to pattern select), the self holding relay (set and clear inputs, built from two morse relays) and the 2 or 3 position switches and close on push buttons. Turnouts and signals get a high or low signal and decide their drive outputs from that. The thing is pretty near to the original and requires only a bunch or morse relays and lots of diodes to work.

 

Thanks for the explanation. Once I would like to give it a go ( in the far away future) and built one control board with relays and diodes.

I was making some drawings of the circuits (on paper, nothing serious) and I think it would be also possible to implement a sort of approach locking to go with the route relays (which would work as route locking). Other things features could be proving if the signals are light, placing the signals in red in case of failure among others. Actually more features could be included but these probably would be only "enjoyable" for the person who operates the control board while the train drivers and spectators would not notice much of it.

 

Checking the website:http://www.roy.hi-ho.ne.jp/s-kato/signal1.htm (link shared in post #1 of this topic)

I noticed that we can see a picture of the control board opened and read what the author said about the interlocking:

 

“Is the contents of the interlocking equipment

 

 

 

Is the time period of this unit took about six months,

Or making other models were so real 4 months or even

So didn't I know nothing about microcontrollers will be easier if you use such as microprocessors relay and diode matrix (home or not), design

Think you might notice the thumb lever is real but uses generic switch lever itself and push button

Control Panel uses with 1 mm aluminum plate with two real leverage is too heavy and did not have in the aluminum panels

Had first done carefully routing many units to be sloppy, hard even to follow the wire breaks in now, it is the

Have the same movement as the real behavior of the signal as a function of interlocking than pretty simplifications have been made,

For example, model so don't close chain lock and overrun chain lock,

Other various circuit is omitted,

Would be more than doubled number of relay throw these circuits also if circuit even though relay close to 70 units.”

 

By the way, the system also has ATS! (ATS info: https://en.wikipedia.org/wiki/Automatic_train_stop)

 

 

 

There is a video and other pictures int he website of the author: http://www.roy.hi-ho.ne.jp/s-kato/ats1.htm

 

Power Up,

 

Nxcale

 

p.s. I am aware that implementing a large interlocking would require a large number of relays... it would be nice built one nevertheless.