General Q's on DCC and layout building

[RobinofLoxley]

Just spent an hour composing some questions, only to be told I was timed out by the site when I clicked submit, thereby losing the whole lot. It’s never the same second time around, but here goes....

 

I am rebuilding my loft layout and thinking of going to DCC control as I build so I want to install what I need as I build, assuming that I am working towards PC control in the end. So no retrofitting stuff later.

 

First question though is about points – I understand there can be some problems with shorts and I have 40 Insulfrog points that I don’t want to throw away or spend a lot of time modifying. Is shorting a serious problem? I have a mixture of steam era Bachmann and Hornby (tender drive) locos that I plan to modify as appropriate.

 

Matching the 40 points I have 40 PM2’s that I plan to install and wire in to command modules later. I can stand to run points manually at the start, one reason I am rebuilding is to put everything in easy reach. Any issues with this?

 

I have a number of semaphore signals that are basically just layout decor. Can these or similar be integrated into the control process?

 

I would want to be able to select a route through the DCC involving setting a route via point control then running the train through it manually. To go to PC control I understand that it is based on a block system but that it’s possible to put sensors in the track to pick up a passing loco. Can that cater for automatic stopping at a terminus, if for example some of the onboard sensors were in a tender and others in the loco body depending on the model? I should add that my proposed layout is end to end with 2 termini and an intermediate station, able to continuous run at the same time (2 main lines).

 

Given the end to end I will have locos facing the buffers and have a reversing loop planned in the layout – I would put in 2 but cant find the room. I understand that there is a digital switching device available but my question is, does it maintain loco orientation when the polarity switches?

 

Any general comments gratefully received. I am also starting a thread in the layout and track design section.

[ITG]

Robin,

I cannot claim to have all the answers, as I consider myself a relative novice, having restarted modelling 18 months ago, and plunging straight into DCC. Overall, for me, it is a total game changer compared to how I recall my DC modelling. Answers below are from my own novice experience, which might be helpful as you’re just starting in DCC.

 

It seems I have a similar layout, with a continuous twin mainline, with a reversing loop, and a terminus on a branch.

I have a mix of insulfrog and live frog points, and as with DC, I’d say the key thing is accurate track laying - level, flat, and of course clean. For insulfrogs, I’ve added point clips, and for live frogs, I use frog juicers to auto change polarity. All powered by Gaugemaster PM10Ds. I have what I would say is generally trouble free running, but not without effort and perseverance with track laying quality. Nothing to do with what type of point, or DCC generally. 
 

As I understand it, pretty much any device (point motor, signal, lights, level crossing gates) can be controlled with DCC; you just need a decoder linked to an appropriate device (LED, servo, solenoid,etc)
I use a Gaugemaster reverse loop module, works perfectly. Trains run through it without any hiccup, and no change at all in communication between controller and loco. To some degree, you can set routes in some DCC control systems, (eg in my Prodigy) by grouping points together as a route. But this wouldn’t extend to automatically stopping the train. I have briefly dabbled with JMRI which may help with your PC automation ambitions. Others will tell more.

 

hope that’s useful as a starter for 10!

[WIMorrison]

Lots of questions and the answer would be that all is possible and feasible.

 

go the points you have at the moment work with your trains and locos? If they do then just reuse them, they will work just fine.

 

autoreverse units can be bought for a little as £25, but if you are considering automation later buy one that will work with the automation Ls Digital 5410 or Digikeijs DR5013 which are about £50 and £90 respectively. The control system retains direction so forward is always forward even after reverse section.

 

for automation the modern way is not to use discreet sensors (reed switches, IR switches, etc) but use current seasoning which will tell you when a loco is in the section being monitored because current is being consumed.  The automation systems will all stop a train within a few mm of the desired position. My preference is iTrain and we have been creating a new series of videos explaining iTrain from the installation to running the train through series of videos that you can find here 

 

https://www.rmweb.co.uk/community/index.php?/topic/155617-new-itrain-tutorial-series/

 

if you want any advice on automation drop me a line ;)

[peach james]

These questions have been answered before on here, use the search function to get more details:

 

First question though is about points – I understand there can be some problems with shorts and I have 40 Insulfrog points that I don’t want to throw away or spend a lot of time modifying. Is shorting a serious problem? I have a mixture of steam era Bachmann and Hornby (tender drive) locos that I plan to modify as appropriate.

 

Depends on system.  Ideal wiring vs what works are 2 different things

 

Matching the 40 points I have 40 PM2’s that I plan to install and wire in to command modules later. I can stand to run points manually at the start, one reason I am rebuilding is to put everything in easy reach. Any issues with this?

 

Nope, but if you wire them to panels now, then re-wire to decoders later, it will involve reworking.  Depending on how you are presently operating, and what wire you are using, it can be cheaper to put in the accessory decoders now, at least if you have reason to value your time.  (if it is all hobby time, then either way works...)

 

I have a number of semaphore signals that are basically just layout decor. Can these or similar be integrated into the control process?

 

 Yes, they can be intergrated (be driven via the control process).  The "how" depends on which DCC system & software you choose to use.

 

 

I would want to be able to select a route through the DCC involving setting a route via point control then running the train through it manually.  

 

With Digitrax and DS64's, you can do most of that without using a computer.  DS 64 route setting will allow quite a lot of function without tying a computer to it.  For NCE, the same thing is possible using the Mini-Panels to set the routes.

 

To go to PC control I understand that it is based on a block system but that it’s possible to put sensors in the track to pick up a passing loco. Can that cater for automatic stopping at a terminus, if for example some of the onboard sensors were in a tender and others in the loco body depending on the model? I should add that my proposed layout is end to end with 2 termini and an intermediate station, able to continuous run at the same time (2 main lines).

 

Yes, you need to block the layout somehow with PC control.  The way I have done it is with current detection blocks.  Be advised that if you are going to go that way, installing the wiring in the layout early is the way ahead.  Basically, you want enough blocks that you know where each train is, at all times.  So, my shortest blocks are around 12", and the majority of them are 60" long.  If you have end points (for example, stations...) where you want stopping on a specific point always, then that will take a shorter block to achieve accurate & repeatable results.   Using RR&Co 7.0 Gold, I have seen repeated 1cm accuracy from calculated block end points (not a seperate sensor, and not on my layout...).  But in general, the more accurate you desire, the more blocks you will have.  

 

Electrical blocks do NOT have to directly match the train block sizes needed- It is easier in RR&Co 5.0 to tie several short electrical blocks together for a train block, than the other way around...

 

It is also possible to use position sensors (either optical, or physical) to detect the end of a train, which may be easier for a terminus.  That avoids the error of the loco plowing on through the buffer stops because you have it facing the other way.

 

Given the end to end I will have locos facing the buffers and have a reversing loop planned in the layout – I would put in 2 but cant find the room. I understand that there is a digital switching device available but my question is, does it maintain loco orientation when the polarity switches?

 

Yes, the DCC auto reverser's do NOT affect the direction of travel of the loco, because they are changing the (AC Polarity) of the reversing section.  The loco's direction is set by the decoder, not by the polarity on the track at any instant.    I use the very cheap Digitrax AR-1 for some of mine, and the much more expensive PSR-1 for another- I have what amounts to 2 Y's on Long Marton, and then use the AR-1's on the Lego as well.  AR-1 doesn't require external power...which is useful with the Lego and very temporary (2 day) setups.  The more expensive reversing controllers work better though- Digitrax takes a ham-handed approach at times.

 

James

[34theletterbetweenB&D]

1 hour ago, RobinofLoxley said:

...about points – I understand there can be some problems with shorts and I have 40 Insulfrog points that I don’t want to throw away or spend a lot of time modifying. Is shorting a serious problem?...

It was for me, on the few remaining insulfrog points on my layout when I went DCC. Slow moving trains of all metal wheeled stock would trip the DCC system's short protection about once every 200 wheel passes. The relatively minor mod of rewiring them to work identically with switched live crossings is a simple solution.

[RobinofLoxley]

Thank you for the comments so far.

 

A lot of jargon unfortunately for me. One question, within any given block how does the system know where the loco is within it, in general terms? Given that the position can be known to a high degree of accuracy  according to posts above.

[WIMorrison]

The train speeds for the various speeds steps is recorded and it then becomes a simple time and distance calculation to accurately position the train in the block.

 

the feedback going high initiates the calculation and because the speed is known the program can calculate the position

[Suzie]

If using Insulfrog points with DCC you will need to treat them as if they are Electrofrog or other live frog points and insulate the frog rails. Insulfrog points specifically are optimised for DC operation and will cause shorts when metal wheels go over the frog otherwise.

[Pete the Elaner]

10 hours ago, RobinofLoxley said:

Thank you for the comments so far.

 

A lot of jargon unfortunately for me. One question, within any given block how does the system know where the loco is within it, in general terms? Given that the position can be known to a high degree of accuracy  according to posts above.

 

It doesn't.

The more blocks you have, the more accurate detection will be. If your trains are stopping at a terminus, you may want shorter blocks for this reason.

We at the club are wiring a layout this way. We may want to use block detection later (especially in the large engine shed where we can't see where the locos are) but not straight away. We will therefore isolate the rail sections from each other then connect them together below the board. If & when we want to add block detection, we just flip the boards up & wire each section through its own detection loop.

 

Peco do not make insulfrog code 75 points, so you must have code 100. This has coarser tolerances than code 75 so the chances of your stock shorting the frog out are less, but it could still happen. Modern stock with well adjusted back to backs is usually less trouble than older stock running on coarser wheels.

Insulfrog points are easy to modify so the frogs are powered from switches then you can also wire the switch rails to the stock rails (which gives much better reliability). This does not take very long. 2 of us did about a dozen points in 1 club evening, including working out together exactly how we needed to do it. As expected, we were got faster after the first few.

If you check the electrofrog wiring diagram then you can easily work out what to do.

 

It may sound like the long way round, but a micro-layout to try some new ideas on first my be very useful. You won't have ruined much with something which you cannot get to work & rolling the change back will be much easier.

[WIMorrison]

16 minutes ago, Pete the Elaner said:

 

It doesn't.

The more blocks you have, the more accurate detection will be. If your trains are stopping at a terminus, you may want shorter blocks for this reason.

We at the club are wiring a layout this way. We may want to use block detection later (especially in the large engine shed where we can't see where the locos are) but not straight away. We will therefore isolate the rail sections from each other then connect them together below the board. If & when we want to add block detection, we just flip the boards up & wire each section through its own detection loop.

 

...

 

I think you confuse blocks which are logical concepts and feedbacks which are physical attributes. A block in most automation systems can contain one of more feedbacks. I agree that theoretically more feedbacks equals a higher level of accuracy however there is a trade off between cost, efficiency and complexity which must be considered.

 

I have blocks with single feedbacks that are 2m long and I can stop a train to within +/-5mm within them. I have blocks that are 500mm long and I can stop a train within mm of accuracy, the key to making them work is consistency in train performance, accuracy pf physical measurement and  correct speed profiles for the locos.

[RobinofLoxley]

2 hours ago, WIMorrison said:

The train speeds for the various speeds steps is recorded and it then becomes a simple time and distance calculation to accurately position the train in the block.

 

the feedback going high initiates the calculation and because the speed is known the program can calculate the position

Thanks for that. This is the kind of simple clear informaton I need as a novice with no DCC experence. I am sure I can find out for myself, but how are individual blocks identified? I visualised them as in DC as just track elements separated by IRJ's but if they are all connected to the 'Bus' anyway they arent identified?? This is important as I want to do all the wiring onle once! So to be absolutely clear, when a loco enters a block the time and distance calculation starts at that point.

[WIMorrison]

when the loco enters a feedback the counter would start. A block may have several feedbacks and each time the loco entered a new feedback the counter is reset to zero.

 

The best way to do the wiring once is to plan where the feedbacks will be and then ensure that there is an insulated rail joiner on both rails at each end. You would then solder droppers to the rails and you can take them back to the bus or a common connection point for each rail. At a later stage when you want to install automation then you would simply put the droppers from one side of the track through feedback decoders which connect to the DCC bus and they would report the occupancy to you computer program through whatever protocol you have chosen to implement.

 

I will PM you a presentation that you may find useful

[Dungrange]

1 hour ago, WIMorrison said:

I think you confuse blocks which are logical concepts and feedbacks which are physical attributes. A block in most automation systems can contain one of more feedbacks. I agree that theoretically more feedbacks equals a higher level of accuracy however there is a trade off between cost, efficiency and complexity which must be considered.

 

Since I have similar questions to the OP I'll jump in here with a question of my own: can you clarify the distinction between 'blocks' and 'feedbacks', as I'm not sure I understand the distinction you're making.  Is the distinction only within the computer software?

 

I'm not sure whether I'll go as far as computer automation, but I may change my mind at a future date, so I've been giving future conversion some thought before I lay the track and start the wiring.  I was therefore thinking about fitting, what I thought were block occupancy detectors like the DR4088 from Digikeijs - https://www.digikeijs.com/en/dr4088ln-2r-16-channels-s88n-feedback-module-with-l-net-cs.html, to monitor each fiddle yard road.

 

My understanding of these is that there needs to be a single dropper wire from each track section to the DR4088, which is then connected to the DCC Power bus.  In my case, each fiddle yard road is around nine feet in length (some are a bit longer), split across three baseboards, so I was assuming that I'd connect up a single DR4088 on the centre board and then carry 24 dropper wires (16/0.2 wire) for the twelve tracks across each of the two baseboard joins so that I'd only have one feed to each section (ie several droppers from the track all connected together before connecting to the DR4088). That would also mean that some of my block sections would be around 11 foot in length.  However, your comments, along with those of @peach james, suggests that I would need shorter lengths of track connected to each DR4088 (ie I should use more than one DR4088), with each feedback section being shorter (ie possibly one DR4088 per board) so that many of my loops would have multiple feedback sections.  Is that correct?

 

Is the block the concept of the fiddle yard road, where the feedback is the process of detecting the time that the monitoring is triggered?

[RobinofLoxley]

1 hour ago, Dungrange said:

 

Since I have similar questions to the OP I'll jump in here with a question of my own: can you clarify the distinction between 'blocks' and 'feedbacks', as I'm not sure I understand the distinction you're making.  Is the distinction only within the computer software?

 

I'm not sure whether I'll go as far as computer automation, but I may change my mind at a future date, so I've been giving future conversion some thought before I lay the track and start the wiring.  I was therefore thinking about fitting, what I thought were block occupancy detectors like the DR4088 from Digikeijs - https://www.digikeijs.com/en/dr4088ln-2r-16-channels-s88n-feedback-module-with-l-net-cs.html, to monitor each fiddle yard road.

 

My understanding of these is that there needs to be a single dropper wire from each track section to the DR4088, which is then connected to the DCC Power bus.  In my case, each fiddle yard road is around nine feet in length (some are a bit longer), split across three baseboards, so I was assuming that I'd connect up a single DR4088 on the centre board and then carry 24 dropper wires (16/0.2 wire) for the twelve tracks across each of the two baseboard joins so that I'd only have one feed to each section (ie several droppers from the track all connected together before connecting to the DR4088). That would also mean that some of my block sections would be around 11 foot in length.  However, your comments, along with those of @peach james, suggests that I would need shorter lengths of track connected to each DR4088 (ie I should use more than one DR4088), with each feedback section being shorter (ie possibly one DR4088 per board) so that many of my loops would have multiple feedback sections.  Is that correct?

 

Is the block the concept of the fiddle yard road, where the feedback is the process of detecting the time that the monitoring is triggered?

Blimey. i read the info on the link attached and it mentioned a 10A transient peak load. Would blow everything in the house!

[Dungrange]

4 minutes ago, RobinofLoxley said:

Blimey. i read the info on the link attached and it mentioned a 10A transient peak load. Would blow everything in the house!

 

Okay, but it's only rated for 3A use.  I think the reference to 10A is simply that in a short circuit scenario, the DR4088 won't be damaged by 10A flowing through it very briefly before the DCC system shuts down.  In my case, I think that would be suitable, provided I use a Circuit Breaker such as the PSX (https://www.coastaldcc.co.uk/products/dccspecialties/psx-powershield-x-2-block) to limit the current to the fiddle yard, as my DCC Command Station is capable of putting out 5A, which is more than the 3A continuous rating.

[WIMorrison]

Please be aware that although the DR4088xx is popular (and it has a 3a total output limitation) there are many other alternatives such as the DR5088RC to name but one of the dozens available many of which offer more than the DR4088 such as Railcom feedback, higher current, fewer outputs, alternative protocols. You can also use reed switches, IR detectors, contact switches, the list is very extensive.

 

The point about feedbacks (however they are fed) is that they are the eyes of the computer program, they are the only connection from the railway to the program and having enough and in the right place is key to a successful automation. There is no absolute rule, but generally more is better - I even put feedbacks on the turnouts for occupancy, some people put them on turnouts for position also.

 

getting it right at the start is well worth the planning effort as it can be much harder to implement after you have built the layout.

Edited September 19, 2020 by WIMorrison

[grriff]

6 hours ago, Suzie said:

If using Insulfrog points with DCC you will need to treat them as if they are Electrofrog or other live frog points and insulate the frog rails. Insulfrog points specifically are optimised for DC operation and will cause shorts when metal wheels go over the frog otherwise.

Don't understand this comment, I and my neighbour have DCC systems and insulfrogs with absolutely no problems. If a point shorts in DCC it will short in DC. You do not need to insulate the frog rails. 

[Dungrange]

3 minutes ago, grriff said:

Don't understand this comment, I and my neighbour have DCC systems and insulfrogs with absolutely no problems. If a point shorts in DCC it will short in DC. You do not need to insulate the frog rails. 

 

I think this is to ensure that you can create a new feed to whatever is beyond the frog so that all track is permanently powered.  If you don't add insulated rail joiners, then whatever siding the turnout is not set for will go dead, which means that DCC Sound and lights will go off.

[Dungrange]

1 hour ago, WIMorrison said:

Please be aware that although the DR4088xx is popular (and it has a 3a total output limitation) there are many other alternatives such as the DR5088RC to name but one of the dozens available many of which offer more than the DR4088 such as Railcom feedback, higher current, fewer outputs, alternative protocols. You can also use reed switches, IR detectors, contact switches, the list is very extensive.

 

The point about feedbacks (however they are fed) is that they are the eyes of the computer program, they are the only connection from the railway to the program and having enough and in the right place is key to a successful automation. There is no absolute rule, but generally more is better - I even put feedbacks on the turnouts for occupancy, some people put them on turnouts for position also.

 

I was aware that there are other detection methods and the existence of the DR5088RC although hadn't noted the higher current rating of that device compared to the DR4088xx series of detectors.  I thought it was just the inclusion of Railcom that increased the price and I'm undecided as to whether or not I need that.

 

My reason for looking at block occupancy detection is primarily to provide an indicator on a control panel as to whether or not a particular fiddle yard road is occupied or not rather than to be the eyes of a computer, but of course once the layout is up and running (probably in a few years time), I may change my mind, so thinking about putting any necessary rail breaks and track feeds in at the start to limit having to completely rewire in the future is what I'm pondering at the moment.  I was also thinking about including feedback from the turnouts so that these can't be thrown when occupied.  What sort of feedback device provides the turnout position indication?

[WIMorrison]

A common way to get position feedback from a turnout is to use a micro switch at one end of the tie bar which makes when it is closed. If you then put a 2k resistor in series (assuming a DR4088) then when the turnout is thrown you will get an occupancy indication that you can use as conformation. Depending on the software it will take this input as position confirmation 

[RobinofLoxley]

1 hour ago, WIMorrison said:

A common way to get position feedback from a turnout is to use a micro switch at one end of the tie bar which makes when it is closed. If you then put a 2k resistor in series (assuming a DR4088) then when the turnout is thrown you will get an occupancy indication that you can use as conformation. Depending on the software it will take this input as position confirmation 

Don't know whether to comments as Im a bit of a duffer here. But the SEEP Pm's I have require a pulse to activate either coil which corresponds to a position of the frog - provided the coil actually fires the frog. Can that pulse be used? I know its a random burst of energy.....

[WIMorrison]

Personally I rely on the fact that my software has told the turnout to change that the turnout has in fact changed. I have never had an issue where the turnout doesn’t change when instructed to do so however some people want positive confirmation of the position and the only way this can be done is by the movement of the tie bar.

 

It is possible (though unusual) for a motor to throw but for the tie bar not to move due to broken linkages, it is also possible (especially with seep/peco) motors for them not to move, or not move far enough because the supply is weak, the motor is slightly misaligned, or there is dirt in the coils.

 

as I say, for most people not required but the question was how can positional feedback be performed :)

[Suzie]

3 hours ago, grriff said:

Don't understand this comment, I and my neighbour have DCC systems and insulfrogs with absolutely no problems. If a point shorts in DCC it will short in DC. You do not need to insulate the frog rails. 

 

In DC you don't normally power the track beyond the frog independently - in DCC you do. When using Insulfrog specifically as the wheel crosses the frog it shorts the two frog rails together.

 

In DC this is not a problem because one will be powered by the switch rail contacting the stock rail, and the other will be left floating.

 

In DCC if no insulated joiners are used on the frog one frog rail will be connected to one stock rail, and the other frog rail will be connected to the other stock rail. As the wheel passes the frog it shorts the two frog rails together and consequently shorts the track feed.

 

The perception of what constitutes 'no problems' can vary quite a lot. Many people using Insulfrogs think it is quite normal for trains to stall on the point when going slowly, some people think that the occasional shutdown of the command station is quite normal, and some people think that serious pitting of the wheels is normal.

 

Other dead frog points are not affected like this - it is just the Peco Insulfrog points due to their extreme DC optimisation.

[grriff]

1 hour ago, Suzie said:

 

In DC you don't normally power the track beyond the frog independently - in DCC you do. When using Insulfrog specifically as the wheel crosses the frog it shorts the two frog rails together.

 

In DC this is not a problem because one will be powered by the switch rail contacting the stock rail, and the other will be left floating.

 

In DCC if no insulated joiners are used on the frog one frog rail will be connected to one stock rail, and the other frog rail will be connected to the other stock rail. As the wheel passes the frog it shorts the two frog rails together and consequently shorts the track feed.

 

The perception of what constitutes 'no problems' can vary quite a lot. Many people using Insulfrogs think it is quite normal for trains to stall on the point when going slowly, some people think that the occasional shutdown of the command station is quite normal, and some people think that serious pitting of the wheels is normal.

 

Other dead frog points are not affected like this - it is just the Peco Insulfrog points due to their extreme DC optimisation.

Hi Suzie

While I understand what you are saying, my neighbour has an extensive DCC layout using Hornby point clips to take power beyond the point (Dungrange's post) and ensure sidings are permanently powered. He has no problems with stalling, shutdowns or wheel pitting.

[WIMorrison]

Fleischmann also use the point clips solution - works well.