automatic movement of stacked trains

[chaz]

Hello,

 

This is a question relating to a possible future project. This may be a double track oval with only the front section visible. The two semi circular curves and the back connecting straight will be used for storing trains. 

Imagine the front train on one of the tracks is to run. As it comes on to the scenic section it leaves a space on the storage track at the front. The other trains (maybe two or three) will need to move up so that the empty section is at the back of the stack ready for the train that is running to run into.

I had this arrangement on a DC powered layout sometime ago, with relays and a BBC Master computer (remember those?) doing the work, moving one train at a time into the vacant space ahead of it so that the empty space on the loop was at the back, ready.

Is this possible with DCC? I know about detecting locos in sections (I have a number of MERG detector PCBs which will detect the current drawn by a decoder, even if the loco is stationary). Of course it will be easy enough for an operator to move the trains up but I am wondering if it can be done automatically.

 

Chaz

[Gordon A]

Ask Tim V, he has a system that works as you describe on his Clutton layout..

It does involve all the locos in a loop being given the same number and dead sections.

 

Gordon A

[Nigelcliffe]

Yes its possible.   As to how to do it, depends on a lot of factors.  Here's a few suggested approaches:

 

a)  crude power-off relays to drop the track power to sections of track, and restore when movement wanted.  But, a loco with a big stay-alive might just drive right through the section.  And stopping trains instantly isn't nice. 

 

b)  Asymmetric DCC braking sections.  Probably the simplest and cheapest mechanism to implement.  But, requires all loco decoders to support the Asymmetric Braking mechanism; which means essentially Lenz, Zimo, ESU (v4 and later), and a few others.   

 

c)  Relay switching of a "DCC all trains stop" signal into different sections.   Should work with any decoder, but the track isolation needed might be complex. 

 

d)  Full computer tracking of trains and their movement.  Not that hard if you have section detection AND can tell the computer the identity of a train as it enters the storage area.   The identification stage could be any of:  human tells computer (ie. a means whereby the human tells the computer "please park train XYZ entering south FiddleYard",  track identification method (eg. RailCom, but limits decoder makers to those supporting RailCom, and places limits on DCC system as well),  non-track identification method (eg. RFID tags on locos, and other methods). 

 

e)   Robot throttle device to look after trains.   Several around which could do it as software projects.  In the commercial area for LocoNet based systems the CML LocoShuttle which can definitely do the job (when connected to some external sensors),  there are ways of making its use quite slick.   The NCE "minipanel" can sort of do it, but would have quite a few limitations on how to deal with locomotive addresses; its not really intended for this job. 

 

I'm sure there are others as well.

 

 

[WIMorrison]

@Nigelcliffe for you option d) the computer program will track without any feedback of the loco id provided you identify  or confirm where the loco is at the start of the session. If you only ever use the computer to move the locos and trains then it will always know where they are.

 

The advantage for Railcom is that it will identify the postion of all Railcom enabled locos at switch on irrespective of whether the computer, or hand of God moved them to where they are - but it does require everything to support Railcom, not just the decoders

[chaz]

12 hours ago, Nigelcliffe said:

Yes its possible.   As to how to do it, depends on a lot of factors.  Here's a few suggested approaches:

 

a)  crude power-off relays to drop the track power to sections of track, and restore when movement wanted.  But, a loco with a big stay-alive might just drive right through the section.  And stopping trains instantly isn't nice. 

 

b)  Asymmetric DCC braking sections.  Probably the simplest and cheapest mechanism to implement.  But, requires all loco decoders to support the Asymmetric Braking mechanism; which means essentially Lenz, Zimo, ESU (v4 and later), and a few others.   

 

c)  Relay switching of a "DCC all trains stop" signal into different sections.   Should work with any decoder, but the track isolation needed might be complex. 

 

d)  Full computer tracking of trains and their movement.  Not that hard if you have section detection AND can tell the computer the identity of a train as it enters the storage area.   The identification stage could be any of:  human tells computer (ie. a means whereby the human tells the computer "please park train XYZ entering south FiddleYard",  track identification method (eg. RailCom, but limits decoder makers to those supporting RailCom, and places limits on DCC system as well),  non-track identification method (eg. RFID tags on locos, and other methods). 

 

e)   Robot throttle device to look after trains.   Several around which could do it as software projects.  In the commercial area for LocoNet based systems the CML LocoShuttle which can definitely do the job (when connected to some external sensors),  there are ways of making its use quite slick.   The NCE "minipanel" can sort of do it, but would have quite a few limitations on how to deal with locomotive addresses; its not really intended for this job. 

 

I'm sure there are others as well.

 

 

 

Thanks Nigel for the detailed answer.

 

I hope you don't mind if I respond to your points.

 

Option (a) is not available. All my locomotives have keep-alives fitted, and these would, I'm pretty sure, defeat the dead sections.

 

On option (b) all locos will have TCS WOW decoders. I don't know if these support the Asymmetric Braking mechanism. Are they on the list?

 

Does option (c) require each section to have its own booster? 

 

Option (d) and (e) both look like they will demand more investment in time (and expertise) than I think I want to commit.

 

On reflection automating this problem looks to involve too much time and effort*. I think the best way for me to go is to eschew an automatic solution and instead rely on an operator who will be able to identify loco addresses, measure the length of the gap in front of stored trains and control the movements with accurate stopping.

 

A human operator can do all this but a rota that changes their role before boredom sets in will be vital.

 

*If I am wrong about this and a simple method is possible please tell me!

 

Chaz

[chaz]

12 hours ago, WIMorrison said:

@Nigelcliffe for you option d) the computer program will track without any feedback of the loco id provided you identify  or confirm where the loco is at the start of the session. If you only ever use the computer to move the locos and trains then it will always know where they are.

 

The advantage for Railcom is that it will identify the postion of all Railcom enabled locos at switch on irrespective of whether the computer, or hand of God moved them to where they are - but it does require everything to support Railcom, not just the decoders

 

Thanks Iain. I see from the tag at the bottom of your posts that you have experience of electronic control.

 

Can you tell me please would the TCS WOW decoders work with Railcom? One of your posts elsewhere on forum suggests they do.

 

Another important factor for me is that I would not want a solution that would divert a large amount of time and effort away from the models and into electronics. The cost is not an issue, the time and work required might well be.

 

If you think I can solve the problem of managing the staging area (fiddle yard) of the layout with electronics, rather than using a "dedicated" human I'd be grateful if you could give me a pointer or two.

 

Chaz

[Nigelcliffe]

I think TCS have supported RailCom in their decoders for a few years, but I haven't used them for a while, so someone else needs to confirm things.     I don't think TCS support Asymmetric DCC braking. 

 

 

A completely different low-ish tech approach is a few cameras and some monitors, or cameras linked to a computer, tablet or smartphone.   From those, you can see the fiddle-yard and thus move the trains manually.   Requires that you (the human operators) can identify each train from the camera in order to select and drive it.  
One solution to the identification problem is to place "train fridge magnets" on a board showing the fiddle yard, and move them by hand as you drive trains up the fiddle yard.   All fairly low-tech (the cameras being the highest tech bits), and doesn't require any intervention on the track, DCC system, decoder types, etc.. 

 

 

 

- Nigel

 

[chaz]

1 hour ago, Nigelcliffe said:

I think TCS have supported RailCom in their decoders for a few years, but I haven't used them for a while, so someone else needs to confirm things.     I don't think TCS support Asymmetric DCC braking. 

 

 

A completely different low-ish tech approach is a few cameras and some monitors, or cameras linked to a computer, tablet or smartphone.   From those, you can see the fiddle-yard and thus move the trains manually.   Requires that you (the human operators) can identify each train from the camera in order to select and drive it.  
One solution to the identification problem is to place "train fridge magnets" on a board showing the fiddle yard, and move them by hand as you drive trains up the fiddle yard.   All fairly low-tech (the cameras being the highest tech bits), and doesn't require any intervention on the track, DCC system, decoder types, etc.. 

 

 

 

- Nigel

 

 

Thanks Nigel. You obviously understand my dilemma. The project is a model railway - it's not an electronic or computer project. if I can enlist electronics and/or a computer to automate what might be an irksome task I would - provide it doesn't divert too much of my time and effort away from my primary focus. 

 

The fridge magnets method might well have mileage - the only drawback might be that an operator might just forget to move them. He/she could recover the position by going round the back and looking. I'm not sure of the need for cameras but they too might have merit. 

 

I envision at least two operators on at a time. One is in front of the scenic part of the layout (probably sitting to one side) and doing the nice stuff - the other is behind managing the oval staging (storage area). It's the latter role (the short straw) that I was contemplating automating. 

 

A rota will be an essential - the mug who gets lumbered with the staging role will have to be promised a go "round the front". 

 

Chaz

[NIK]

2 hours ago, chaz said:

 

Thanks Iain. I see from the tag at the bottom of your posts that you have experience of electronic control.

 

Can you tell me please would the TCS WOW decoders work with Railcom? One of your posts elsewhere on forum suggests they do.

 

Another important factor for me is that I would not want a solution that would divert a large amount of time and effort away from the models and into electronics. The cost is not an issue, the time and work required might well be.

 

If you think I can solve the problem of managing the staging area (fiddle yard) of the layout with electronics, rather than using a "dedicated" human I'd be grateful if you could give me a pointer or two.

 

Chaz

Hi,

 

You asked if it was possible with DCC and it is.

 

If cost is no issue but time and work is, its possible Train Controller software package and Infra Red position detectors wired to feedback units (and a suitable DCC command station if yours is not Train Controller compatible) will do the job. You will have to initially configure the software on screen so that it knows the fiddle yard tracks are split into short blocks between the Infrared detectors but from then on the software should deal with the 'Auto Staging'.

 

Perhaps in any future questions you could give more specific information about your criteria and your model railway hardware to help those reading your questions.

 

Regards

 

Nick

[RFS]

I actually do this extensively in my storage yard using Traincontroller and it's not difficult to set up with software like this. My yard has 16 roads which have either 2, 3 or 4 blocks. When a train leaves the first block of a road, its occupancy detector going off triggers TC to run a schedule to move the train from block 2 to 1 (if 2 is occupied), and that in turn starts a schedule to move the train from 3 to 2 etc. Also TC schedules that send a train into the storage yard always have the last block of the road as their ending block, but have a successor schedule to go from that to the next block. Eg for a 3-block road the train will end in block 3, but if block 2 is free will continue straight through etc.

 

Previously I had a point ladder in the middle, but it took up a lot of space and moving to simple long roads divided into blocks greatly increased the capacity.  

[Alan Kettlewell]

 

Like RFS, I also use Train Controller to operate my hidden storage roads automatically, involving a 'Line up in a block' feature of the software.  A single line can hold as many trains that will fit in, as the leading train exits the block, aĺl the others move up by the same distance corresponding to the space now available.  This feature only requires a single sensor (I use a reed switch) at the entrance to the block, all the rest is done by the clever software and some built in formulae to determine the distance to move up. 

 

Please note that although you said money is not an issue - this is expensive software costing over £500 for the Gold version and is predominantly for those who wish to fully or partly automate their layouts.    If you are interested you can download it for a free trial from the RR&Co website.   Alternatively you could have a browse through my layout thread where I explain quite a bit about some of Train Controller's features.

 

Good luck with it and I hope you sort out a method.

 

Cheers.. Alan

 

 

 

[Nigelcliffe]

On the software front, whilst TrainController is perhaps the best known, I suspect several other packages could do it just as well - iTrain for one example. 

[Andymsa]

What would be interesting is the various methods that has been used in traincontroller for move up in the hidden sidings that have been referred to. Personally where I have two sections or blocks I use a simple flagman and a schedule. The line up in a block is a bit more tricky.

[RFS]

The problem I found with Traincontroller's line-up in a block is that trains are moved up only at the loco's threshold speed.  Far too slow for me, as the train that left is usually ready to re-enter at the other end long before the line-up has completed. Seems the process was designed for locos in a staging track rather than full-length trains.

 

With my process, a schedule is started to move the next train up as an action on the release of the first block when the train leaves.

[Andymsa]

10 minutes ago, RFS said:

The problem I found with Traincontroller's line-up in a block is that trains are moved up only at the loco's threshold speed.  Far too slow for me, as the train that left is usually ready to re-enter at the other end long before the line-up has completed. Seems the process was designed for locos in a staging track rather than full-length trains.

 

With my process, a schedule is started to move the next train up as an action on the release of the first block when the train leaves.

 

Yes I also found this with line up in block.

 

with your method what happens when there are no trains in the blocks. Will it go to the far end of the hidden siding 

[RFS]

27 minutes ago, Andymsa said:

 

with your method what happens when there are no trains in the blocks. Will it go to the far end of the hidden siding 

 

Yes it will. Let's say the train is entering a 3-block road. Block 3 will be the ending block for the schedule. However the schedule will have a successor that moves the train from block 3 to 2, with the condition on that schedule that block 2 is free, and also that TC will not keep retrying the schedule if the block isn't free. TC recognizes this scenario and the train will continue without slowing down into block 2 if block 2 is free. Likewise this 3->2 schedule has a successor to move the train from 2->1. So if 1 is also free the train will run all the way up to the front of 1 without stopping or slowing down in blocks 3 and 2.  Took a while to set up these schedules but it all works very well indeed. 

 

The only thing I had to watch is that, when a train leaves a block on a schedule that returns it after its scenic journey to the rear block of the same road, it spends long enough doing this (eg via a station stop) to give the other trains time to move up. Otherwise it gets signal-checked at the entrance to the storage road.  Non-stop trains either occupy 2-block roads or are slow-moving goods trains.

[Andymsa]

10 minutes ago, RFS said:

 

Yes it will. Let's say the train is entering a 3-block road. Block 3 will be the ending block for the schedule. However the schedule will have a successor that moves the train from block 3 to 2, with the condition on that schedule that block 2 is free, and also that TC will not keep retrying the schedule if the block isn't free. TC recognizes this scenario and the train will continue without slowing down into block 2 if block 2 is free. Likewise this 3->2 schedule has a successor to move the train from 2->1. So if 1 is also free the train will run all the way up to the front of 1 without stopping or slowing down in blocks 3 and 2.  Took a while to set up these schedules but it all works very well indeed. 

 

The only thing I had to watch is that, when a train leaves a block on a schedule that returns it after its scenic journey to the rear block of the same road, it spends long enough doing this (eg via a station stop) to give the other trains time to move up. Otherwise it gets signal-checked at the entrance to the storage road.  Non-stop trains either occupy 2-block roads or are slow-moving goods trains.

 

It always amazes me how users come up with different methods to achieve the same thing. 

 

So if the schedule is setup not to keep retrying what is the trigger in that case when the next block becomes free.

 

Just curious with your method can you set up a condition that it will only apply to certain trains. I ask this as one of my hidden sidings is a big loop. So if I want a train to go around the loop to face the other way your method might be wa way to go. 

 

Going to try try some tests in simulator mode tonight.

Edited April 24, 2019 by Andymsa

[RFS]

19 minutes ago, Andymsa said:

 

So if the schedule is setup not to keep retrying what is the trigger in that case when the next block becomes free.

 

 

The trigger is an action on the release of the block that comes free. This is what starts the move-up schedule. 

[Andymsa]

38 minutes ago, RFS said:

 

The trigger is an action on the release of the block that comes free. This is what starts the move-up schedule. 

 

 

Ah I should of realised that, thanks

[chaz]

8 hours ago, NIK said:

Hi,

 

You asked if it was possible with DCC and it is.

 

If cost is no issue but time and work is, its possible Train Controller software package and Infra Red position detectors wired to feedback units (and a suitable DCC command station if yours is not Train Controller compatible) will do the job. You will have to initially configure the software on screen so that it knows the fiddle yard tracks are split into short blocks between the Infrared detectors but from then on the software should deal with the 'Auto Staging'.

 

Perhaps in any future questions you could give more specific information about your criteria and your model railway hardware to help those reading your questions.

 

Regards

 

Nick

 

Nick, thanks for that useful and positive shove in the right direction!

 

You asked for more specific information so here is what I have planned so far.

 

The scale will be 1:48 and the gauge 16.5mm (so On30). Track will be Peco code 100 on the staging and hand-laid spiked code 83 on wood ties on the scenic area.

The staging is a two track oval, but the station is on a single track line. The single line will divide into two just out of sight of the viewed scenic section. The two turnouts will be lightly biased (either a spring or a counterweighted crank) so that trains leaving the scenic section always take the right hand road. Trains entering the scenic section will trail through the switch blades - the live crossings will have their polarity sorted with frog juicers.

 

All locomotives will have TCS WOW decoders with keep-alives. There is ample room in the locomotives for these. They have been purchased and I have made a start on fitting them (2 down - 10 to go!).

 

I haven't decided on a DCC control system yet. The choice may well be influenced by the need to work with any automation arrangement for the staging. I use Roco Multimaus handsets on my home railway and like these so a Z21 system looks a good choice - yes?

 

I have an Apple Notebook (laptop) which can be used with the layout if it is needed.

 

You mention IR detection. Could I use current amplifier detectors instead? I have lots of the MERG ones ready to go. 

 

Chaz

[chaz]

4 hours ago, RFS said:

The problem I found with Traincontroller's line-up in a block is that trains are moved up only at the loco's threshold speed.  Far too slow for me, as the train that left is usually ready to re-enter at the other end long before the line-up has completed. Seems the process was designed for locos in a staging track rather than full-length trains.

 

With my process, a schedule is started to move the next train up as an action on the release of the first block when the train leaves.

 

The low speed moves may not be a problem. It may well be possible to leave the first section empty so that a train leaving the last section can run through the scenic part of the layout without stopping and find an empty section waiting. In practice most traffic will stop in the station and on the narrow gauge nothing moves that fast anyway.

If the "move-up" process is very slow could it be triggered as soon as the train on the last section start to move? Little chance of the train moving up colliding with the departing train.

 

Chaz

[chaz]

Plenty of ideas to play around with in your responses to my initial question lads. I will not be starting this layout until the autumn so I am very much at the planning stage. 

 

I would be happy to see any further advice, particularly on the choice of DCC system and any other hardware that might be needed.

 

Chaz

[NIK]

15 minutes ago, chaz said:

 

Nick, thanks for that useful and positive shove in the right direction!

 

You asked for more specific information so here is what I have planned so far.

 

The scale will be 1:48 and the gauge 16.5mm (so On30). Track will be Peco code 100 on the staging and hand-laid spiked code 83 on wood ties on the scenic area.

The staging is a two track oval, but the station is on a single track line. The single line will divide into two just out of sight of the viewed scenic section. The two turnouts will be lightly biased (either a spring or a counterweighted crank) so that trains leaving the scenic section always take the right hand road. Trains entering the scenic section will trail through the switch blades - the live crossings will have their polarity sorted with frog juicers.

 

All locomotives will have TCS WOW decoders with keep-alives. There is ample room in the locomotives for these. They have been purchased and I have made a start on fitting them (2 down - 10 to go!).

 

I haven't decided on a DCC control system yet. The choice may well be influenced by the need to work with any automation arrangement for the staging. I use Roco Multimaus handsets on my home railway and like these so a Z21 system looks a good choice - yes?

 

I have an Apple Notebook (laptop) which can be used with the layout if it is needed.

 

You mention IR detection. Could I use current amplifier detectors instead? I have lots of the MERG ones ready to go. 

 

Chaz

Hi,

 

If you always have locos or a suitable vehicle that draws current at the front of your trains then you might be able to use current detectors such as MERG ToTis instead.

 

I have lots of multiple units some of which don't draw current on their trailing coaches so I immediately thought of Infrared detectors (which are also a bit more accurate as regards position).

 

Regards

 

Nick

[Nigelcliffe]

Just to add to the mix - Traincontroller isn't the only software package around which can do this "off the shelf".  I think iTrain can do it as well, and may be somewhat cheaper.  

 

Whatever software is chosen, and whatever detectors, you'll then need a way of putting the detector events into the computer or DCC system's bus network.  Quite a few options here, but pick something which works with both your chosen DCC system and the software choice.  (ie. not every hardware option plays well with all software packages).

 

- Nigel

[jpendle]

The Z21 is (imho) a very good system and you have the ability to use the Multimaus and/or a phone/tablet as the 'throttle'. One possible additional benefit is that loco's can be identified visually using a photo rather than needing to know the loco address.

 

Regards,

 

John P