Automating a 4 siding, 3 train shuttle

[Sun VI]

Morning All,

Do any of you good folk have knowledge / experience of automating layouts and may be able to give advice, guidance or sources of information that might be of help? I have very little knowledge of electrics and view the following as a means of improving this. I have in mind a very simple OO layout on which I can experiment with automation. Basically, I want to shuttle 3 trains between 4 sidings. There would be a small station or halt on a single track between the two fiddle yards. Each fiddle yard will contain one point that will be motor operated and part of the automated process. Board sizes and the choice of rolling stock are not critical, although trains must be capable of running realistically in either direction without the need for a loco to run around. I’ve tried to analyse exactly what I require:

System should recognise individual trains

System should recognise if each siding is empty or occupied

System should select next train to move and ensure correct direction

System should recognise origin and destination and set points accordingly

System should provide auto delay between each sequence

System should provide auto delay for trains stopped at halt

Trains stopping at the halt should stop and start smoothly

Class 128, parcels, shall not stop at halt

Class 121, bubble car, shall always stop at halt

Class 108 will sometimes stop at halt

I’m undecided or open minded on the following:

DC or DCC?

Computer control?

Directional lighting?

Sound?

Signals?

 

Track plan

NB, I realise a 24” fiddle yard is too short for a point plus sidings for a 2 car class 108.

 

Schematic of start position

and so on …

 

Sequence of automated actions

a i Set point Y to R and point Z to R

ii Move train 121 from siding 1 left to right

iii Stop at halt, wait n seconds then continue

iv Go into empty siding 4 and stop at end

 

b i Wait n seconds

ii Set point Z to L and leave point Y at R

iii Move train 108 from siding 3 right to left

iv Stop at halt, wait n seconds then continue

v Go into empty siding 1 and stop at end

 

c i Wait n seconds

ii Set point Y to L and set point Z to L

iii Move train 128 from siding 2 left to right

iv Do not stop at halt

v Go into empty siding 3 and stop at end

 

d i Wait n seconds

ii Leave point Y at L and set point Z to R

iii Move train 121 from siding 4 right to left

iv Stop at halt, wait n seconds then continue

v Go into empty siding 2 and stop at end

 

e i Wait n seconds

ii Set point Y to R and leave point Z at R

iii Move train 108 from siding 1 left to right

iv Stop at halt, wait n seconds then continue

v Go into empty siding 4 and stop at end

 

f i Wait n seconds

ii Set point Z to L and leave point Y at R

iii Move train 128 from siding 3 right to left

iv Do not stop at halt

v Go into empty siding 1 and stop at end

and continue sequence until stopped.

 

Now, while the track plan may be simple, to actually automate this as I’ve envisaged appears to be anything but. Is it possible to set up a dc system to meet the requirements or must I look at dcc, possibly with computer control too? I’ve been given an Arduino starter kit which will support five devices such as servo motors: could this be of use in the layout?

I’ve had a look at the Heathcote Electronics and the MERG sites. I can’t see that they will do more than a shuttle between two points with an intermediate stop, but I may well be mistaken. I also have copies of Roger Amos’s ‘Practical Electronics for Railway Modellers’ volumes 1 and 2, but these don’t deal with the selection of which train to move, the operation of the points as part of a sequence, the change of direction and the recognition of which train is moving and whether it should stop at the halt or not.

If you’re still reading this and have not lost the will to live, any ideas?

[JohnR]

i should think that considering you want different trains to do different things, the easiest solution is DCC and computer control.

[Suzie]

If you want to keep it ultra simple and DC, look at it as two solutions:-

 

1. A simple shuttle with station stop which just moves a train from left to right and back again stopping half way.

 

2. A siding selector at each end which changes the point as soon as a train is sufficiently in the siding.

 

If you want to do any more than that it will probably be easiest to go DCC and use software like RR&Co, Rocrail or JMRI to do the automation for you.

[Gordon H]

I’ve had a look at the Heathcote Electronics and the MERG sites. I can’t see that they will do more than a shuttle between two points with an intermediate stop, but I may well be mistaken. ?

 

Can't speak for Heathcote, but relating to MERG, the ATC System could be set up to do this quite easily. The set-up used on the MERG exhibition stand test track actually incorporates your track plan, with a few extras thrown in. 

 

As it comes, the standard ATC kit has a built-in track configuration similar to yours, just without the central station stop, known as 'Config C'. It is two Ends to two Ends. The station stop would be added as a fifth 'End' in ATC terms.

To create your set-up you would employ the 'User Defined' option, which uses a straightforward spreadsheet where you enter the sequence you want the trains to follow, and which direction each one is to travel. If you need the acceleration, deceleration and dwell times to be individually adjusted for each train that can be accommodated too by adding an extra module. These can also be defined via the spreadsheet.

Once all the information is entered, a button is pressed and an uploadable file is automatically created. This is then sent to the ATC using a USB adaptor, which is another MERG kit. From the information you have provided already, I could create the required configuration file within minutes.

It might all sound a bit complicated, but it is actually quicker to do it than to write about it !

I should also mention there are no sensors required out on the tracks, only wires from the kit modules to the rails. Detection is performed within the ATC. Point controls are derived from the modules which switch the individual tracks on and off.

The diagram below shows the exhibition track and its track connections, though point motor drivers are not shown to maintain clarity. As you can see, your layout could be taken as 'End1, End2, End4 (Station), End5, End6.

 

[Gordon H]

To give an idea what the spreadsheet would look like for your layout, I have taken the suggested sequence and entered the information as shown below.

It amounts to a sequence of ten individual movements, which once complete repeats ad nauseum.

End 5 is the station stop, so each movement that has a station stop is effectively split into two separate movements.

Non-stop movements simply declare End 5 as a 'Via', meaning the track is powered, but not detected for stopping.

Direction convention here is (E)ast/(W)est rather than Left/Right, but it amounts to the same thing.

There are more features available which do not need to be filled in unless you need them.

The comments fields are very useful too, keeping track of where you think you are when entering the data.

 

[Gordon H]

Once all the information is entered, a button is pressed and an uploadable file is automatically created. This is then sent to the ATC using a USB adaptor, which is another MERG kit. From the information

 

Thinking about it, I should also have pointed out that you only ever need to do this the once. The ATC stores the uploaded sequence internally, so thereafter you only need to switch the system on and off at the supply.

[Sun VI]

Dutch Master, thanks for the link, unfortunately it won't open but it was kind of you to reply.

 

Gordon H, that looks to be just what I'm after so MERG's membership is about to increase by one. I assume that with the ATC the 'power in' is coming from a transformer but not a conventional dc controller such as Gaugemaster. The spread sheet looks straightforward, it appears that you can even set a different speed for each trip if so desired, and vary the waiting interval at the halt. That would make the layout appear less mechanical and monotonous. Many thanks for your time and effort.

[Dungrange]

The hyperlink provided by Dutch Master seems to have an extra space at the end.  Delete that and it should work fine, or try http://www.umelec.ch/hompa10.htm

[Gordon H]

Gordon H, that looks to be just what I'm after so MERG's membership is about to increase by one. I assume that with the ATC the 'power in' is coming from a transformer but not a conventional dc controller such as Gaugemaster. The spread sheet looks straightforward, it appears that you can even set a different speed for each trip if so desired, and vary the waiting interval at the halt. That would make the layout appear less mechanical and monotonous.

 

The system is designed to be powered from a standard DC regulated supply, similar to one you might have for a small laptop. Typically this needs to provide 15V DC at something like 3A - not that you actually need that much current, but that type of unit is readily available.

To take advantage of the individual control options, you need to include another module in the system which expands the number of control pots. This is simply added to the 'daisy-chain' module connection, as the first module in the chain. Then you just have to select the correct spreadsheet tab for that combination of modules and enter the numbers.

Although your current sequence only involves ten movements, there is enough storage is for up to 160 movements in the User defined sequence.

For your layout with this number of moves available, you could easily add to the range, apparently occasionally, including (say), going from 1 to 5, then back from 5 to 1, i.e. only going half way. As it happens, the MERG demo track shown above uses a sequence of 27 movements for three locos in its User table before repeating.

[bnb]

I have done a similar thing using DC and an auto reversing unit.  Micro switches can be fitted to the point motors so that the power is supplied to the appropriate siding so that the power is routed to the opposite siding when the auto unit switches over.  Point motor control was achieved by using magnets under the last carriage and reed switches  laid on the track that switched self latching relays.  The logic to get the trains to work in the right sequence should be simple but getting different trains not to stop at the station requires a little bit more circuitry.  Trickey but not impossible. (extra reed switches and latching relays).  While possible in and cheaper in DC , computer control may be the best solution. I would use tortoise point motors and take advantage of the switching circuit that comes with these. 

[Gordon H]

I have done a similar thing using DC and an auto reversing unit.  Micro switches can be fitted to the point motors so that the power is supplied to the appropriate siding so that the power is routed to the opposite siding when the auto unit switches over.  Point motor control was achieved by using magnets under the last carriage and reed switches  laid on the track that switched self latching relays.  The logic to get the trains to work in the right sequence should be simple but getting different trains not to stop at the station requires a little bit more circuitry.  Trickey but not impossible. (extra reed switches and latching relays).  While possible in and cheaper in DC , computer control may be the best solution. I would use tortoise point motors and take advantage of the switching circuit that comes with these. 

 

There have been a number of similar hard-wired solutions discussed here and elsewhere in the past. I have worked on such systems myself in the distant past (late 70's). Whilst no-one doubts that they can be made to work initially, their complexity tends to rise enormously when incorporating apparently simple additional features such as occasional station stops or different sequences of movements. The system I described above avoids these issues by design. The operating sequence can be added to or altered in seconds, and stopping or not is just different numbers in a table, no extra detection necessary.

[rodshaw]

It may be worth considering the offerings from CTI Electronics. I've used them myself and they are very flexible, but there is a fair bit of programming to do. Also you have to get them from the US, which involves import duty and VAT.

[Sun VI]

Well, you're certainly giving me a lot to think about.  The CTI system looks interesting but the lack of a UK distributor and cost puts me off somewhat.  I've looked at several threads on Arduino forums regarding use for model railway control, but the answers that are forthcoming leave me quite baffled.  So for the time being, it still looks as though the MERG options are the most appealing in terms of ease of use, availability and cost.  The only facility that the MERG system doesn't seem to provide is the individual train recognition which CTI appears to offer via RFID tags - if I've understood it correctly.  However, I can think of a way round this by using magnetic detection only on trains that are scheduled to stop at the halt but this wouldn't provide for one tagged to train to stop occasionally.  Still, for me to get a system up and running would be an achievement. 

 

Again, thanks to all of you for your contributions. 

[Gordon H]

Individual train recognition is not included in the MERG ATC system, simply because it isn't required for the method of operation it uses.

However, if you really want to include an RFID facility, there are MERG kits for that too - but it would be entirely separate from the track control system.

Not sure what you might do with the RFID data in this context other than use it to display details of the loco. That is what we use it for on the MERG demo stand, via a Powerpoint presentation whose individual slides are triggered by messages received via RFID.

[Sun VI]

Of course, having seen that what I would like to do is feasible, I've been having a host of other thoughts, such as: what about a level crossing with automatic gates (not barriers)and what about the Dapol signals and ... No, no, no - one step at a time!

[Sun VI]

Individual train recognition is not included in the MERG ATC system, simply because it isn't required for the method of operation it uses.

However, if you really want to include an RFID facility, there are MERG kits for that too - but it would be entirely separate from the track control system.

Not sure what you might do with the RFID data in this context other than use it to display details of the loco. That is what we use it for on the MERG demo stand, via a Powerpoint presentation whose individual slides are triggered by messages received via RFID.

I think that what I'm looking for here is an automated system that reacts to which ever train is passing, rather than following a tight script that would not be able to adjust if a train were changed during the sequence. What I would like the system to do is go through a decision process along the lines of:

is train x or y or z?

If x, train stops for 10 seconds.

If y, train does not stop.

If z, go to random selection of i) stop, or ii) continue. If i) select random delay before restart of whole seconds between 10 and 20.

 

But I'm well aware that I'm getting beyond what can be quickly and easily achieved, and I may well be getting into the realms of computer control. I really do need to get something running in the first instance that I can then use as a test bed for further improvements.

 

It's a shame that I hadn't thought of this project before the end of April. I went to the Bristol show and looked at the MERG stand but didn't pay a great deal of attention as I had no idea then of what I want to know now.

[Gordon H]

I think that what I'm looking for here is an automated system that reacts to which ever train is passing, rather than following a tight script that would not be able to adjust if a train were changed during the sequence.

 

 

However, the system should already know which train is going to pass because the system set it off running in the first place.

Only if you have two independent systems operating does one have to learn something from the other.

[Dungrange]

However, the system should already know which train is going to pass because the system set it off running in the first place.

Only if you have two independent systems operating does one have to learn something from the other.

 

You are correct, but in this case I think Sun VI wants two independent systems: one is himself (who is independently selecting stock for the operating session and startup positions) and the other is the automation process. He is therefore looking for the automation system to detect what units are on the track, so it can 'learn' what he wants it to do, without Sun VI having to reprogram the automation spreadsheet to accommodate a slightly different startup / stock.

 

I have no idea what the best solution would be, but I'd suggest that some form of computer control or something that can be coded with lots of IF statements would be the way forward.

[Gordon H]

You are correct, but in this case I think Sun VI wants two independent systems: one is himself (who is independently selecting stock for the operating session and startup positions) and the other is the automation process. He is therefore looking for the automation system to detect what units are on the track, so it can 'learn' what he wants it to do, without Sun VI having to reprogram the automation spreadsheet to accommodate a slightly different startup / stock.

 

Well maybe, but the original request was for a certain combination of capabilities and a particular operating sequence. To create a full-on self-learning layout operating system would be a rather daunting prospect if starting with little or no previous experience of such matters, though no doubt it is possible given sufficient hardware and software resources. As one of the stated aims is to learn a bit about electrics/electronics, it would probably be better to start out with something a bit simpler which meets the majority of the requirements, and work upwards from there.

[Sun VI]

Good morning All, Mrs and I have been away to stay with our younger son for a few days, so my apologies for not keeping up with the conversation.  I'll try to address the points that you've raised

 

I have done a similar thing using DC and an auto reversing unit.  Micro switches can be fitted to the point motors so that the power is supplied to the appropriate siding so that the power is routed to the opposite siding when the auto unit switches over.  Point motor control was achieved by using magnets under the last carriage and reed switches  laid on the track that switched self latching relays.  The logic to get the trains to work in the right sequence should be simple but getting different trains not to stop at the station requires a little bit more circuitry.  Trickey but not impossible. (extra reed switches and latching relays).  While possible in and cheaper in DC , computer control may be the best solution. I would use tortoise point motors and take advantage of the switching circuit that comes with these. 

Congratulations on your first post bnb and thanks for plunging into the fray.  I take it that the self latching relays that you've mentioned switch the power and the point at the same time.  Does this mean that the returning train starts immediately, or is there a timer incorporated somewhere to give a pause and is this adjustable? 

 

 

You are correct, but in this case I think Sun VI wants two independent systems: one is himself (who is independently selecting stock for the operating session and startup positions) and the other is the automation process. He is therefore looking for the automation system to detect what units are on the track, so it can 'learn' what he wants it to do, without Sun VI having to reprogram the automation spreadsheet to accommodate a slightly different startup / stock.

 

I have no idea what the best solution would be, but I'd suggest that some form of computer control or something that can be coded with lots of IF statements would be the way forward.

 

Hello Dungrange, your analysis is pretty much on target.  I would say that I am '... looking for the automation system to detect what units are on the track'  and direct them within the parameters that I have set out (i.e. whether to stop or not, the duration of the stop and at the end of the individual run the length of delay before the next train starts).  I'm aiming to introduce an element of variety where even I will not know exactly how the next train will perform unless it is a non-stop parcels unit. 

 

 

Well maybe, but the original request was for a certain combination of capabilities and a particular operating sequence. To create a full-on self-learning layout operating system would be a rather daunting prospect if starting with little or no previous experience of such matters, though no doubt it is possible given sufficient hardware and software resources. As one of the stated aims is to learn a bit about electrics/electronics, it would probably be better to start out with something a bit simpler which meets the majority of the requirements, and work upwards from there.

 

Hello again Gordon, 'To create a full-on self learning layout ...'  how right you are.  I do agree with starting out a bit simpler and working upwards.  I think that, unusually for me, I can see what I would like to achieve as an end result and I'd like to gallop there rather than walk step by step.  On numerous occasions I've watched shuttle layouts in model shops' windows, my problem is that they are too predictable, working with clockwork accuracy as they run to-and-fro'.  My feeling is that I want a degree of variety rather than just one train, and I would like a degree of uncertainty - think of it as British Railways rather than Swiss Railways.

 

Anyway, two things came out of this weekend that I hadn't anticipated.  First,  I outlined my thoughts regarding the four siding, three train shuttle problem to my son who is an electronics engineer and systems developer.  I hadn't run this past him previously as he's enough commitments of his own.  However, he was interested and said that the Arduino would be ideal  and suggested that we approach this as a joint project with me building the layout and him coaching me on the programming side.  He asked if I would give him a note of what I'm after, so I immediately directed him to this thread as the best introduction that I could think of.  The second development was visiting B and Q in Reading and seeing their sheet cutting machine: I couldn't resist it and one sheet of 4'x2' 6mm ply was promptly sliced and together with sufficient strip wood for battens plus timber for the sides at home in the garage, I now have the base board materials.  Mrs got in the car and merely said 'you bought some of that last week': 'yes' I replied, 'but it's already spoken for'.  What I need now are more brass screws and some pattern makers' dowels plus two short electro-frog Y points, I think a quick visit to Pecorama may be on the cards today.  Things are moving.