© Henk Funk, 2005
Introduction to digital model railroading
It is my intention with this page to get you started in the wonderful world
of digital model railroading. Have written this page as a result of my own
adventures in trying to figure out what all the digital components do and
how they fit together. Do not expect ready-made solutions, the page is more
aimed at providing an overview but also as a counterweight to the stories
the manufacturers tell you. Have tried to use a step-by-step approach and
stay as impartial as possible towards the various systems as the underlying
principles are all the same anyway. It is strongly recommended to read the
story as a whole as only then the relations and interdependencies will get
a bit clearer.
FYI, I kept the pictures in Dutch, don't think
that will be much of an issue for you. If it is, just let me know.
Good luck!
Contents:
• The Basics
• Standardisation
• Booster and Control-unit
• Blocks
• Sections
• Occupancy detection
• Brands and Systems
• Variations on Digital
• Digitising Locomotives
On an analogue track the locs are controlled by changing the voltage, however,
on a digital track this works completely different. There is a continuously
available single voltage AC current on the track but hidden in the AC current
are signals that can be picked up by so called decoders connected to the
track. The smart thing is that each instruction hidden in the AC current
is preceded by an address and only the decoder with the same address will
act upon the instruction, all the other decoders will ignore the instruction.
Using that principle you can control multiple trains on the same track all
using the same circuit.
Instructions for a loc are entered in a control-unit which converts the
instructions into the appropriate digital signal and using a booster the
digital signal will be put on the track where the various loc-decoders can
read the signal via the rails. A rough overview:
The above basically works the same for
2- and 3-rail systems, however, to avoid
duplication I have restrict myself to 2-rail
only but mention differences where
appropriate.
OK, but what is it that these components do?
The transformer is easy, it converts the 110 or 230V main power to 16-18V
AC that the trains can use. The control-unit is used to enter instructions
and it will convert those instructions into the appropriate digital signals,
the booster will then boost this signal, mix it into the 16V AC and put
it on the track. Officially we should split the control-unit into two parts;
an input-device and a digital central that converts the input into digital
signals. However, in most situations the two functions are combined in the
same piece of equipment we then call the control-unit.
The explanation above is not very scientific but more is not really necessary.
For those who want to know more, a search on internet will provide with
all the in-depth technical data.
For larger lay-outs multiple boosters will be
used as the overall demand for current
would be too much for a single booster.
Typically a booster can supply 3-4 Ampere.
Do note that each booster has its own circuit
and the rails are isolated between the circuits.
Anyway, we’ll leave this topic as we’re still beginners.
Switches
Just an oval for driving your trains will become boring very quickly so a number of switches will be needed. These switches used to be controlled by a row of buttons or sometimes even a big switch-panel showing the lay-out of the track. In the digital world you can still do that but if you’re considering using a PC in the future you will need switch-decoders. These decoders are electrically connected to he track, will pick up he instructions from the control-unit and convert that into a pulse for the switch-engine.
Important: there are basically three types of switch engines; some requiring a brief moment of current (a pulse) to set a magnetic switch motor, some requiring a few seconds of current and finally a type that requires continuous power. Each type has its own type of decoder, be careful t select the one appropriate for you when buying switch-decoders.
It is strongly to be recommended to always select decoders that have the
possibility of using an external power source. You can then use a cheap
transformer to set your switches and thereby safe costly digital current
for driving the trains. See schematic:
Connection ‘A’ is only needed to get the digital
signal to the decoder. Usually a switch-decoder
can set 4 or 6 switches.
Be careful with pulse-decoders as some
switch-motors need a large amount of current
(sometimes even 2-3 Ampere) and many
pulse decoders cannot provide that.
Types of decoders, a summary:
Locomotive: decoder built-in in the loc and controls the driving and lights
of the locomotive.
Pulse: decoder used for switching magnetic drives
or relays that require a pulse
Motor: decoder capable of providing current for
a set amount of time, usually 1-10 seconds. Used for setting switches that
use an electric motor.
Relay: decoder with on-board relays, used for
switch-motors that require continuous current. Do note that connecting a
relay to a pulse-decoder gives exactly the same effect and often a lot cheaper.
Function: decoders used in carriages to control
lights or other extra functions.
Silly question, of course not!! The various manufacturers have each their
own systems and each have their own, sometimes fanatic, fan club. Fortunately
it isn’t as bad as the above might suggest because there are several
multi-system solutions possible.
The following protocols (languages) are currently around:
Each manufacturer have the full range of equipment for sale and they all
claim their products are the easiest to use and install. Up to a point they
are a little bit right as the equipment of a specific manufacturer is designed
to be used in combination with each other. However, manufacturer specific
solutions are horrendously expensive. It is perfectly possible to shop around
or even build quite a lot yourself. Next to that there are also control-units
that can speak various ‘languages’ especially the Intellibox
(IB) made by Uhlenbrock is worth mentioning here.
Building stuff yourself is very doable and you don’t need much more
than some elementary logic and DIY-skills. Without those model-railroading
will become extremely difficult anyway!!
For those well versed in electronics there is a huge amount of options available
on the internet, for the rest, who need clear instructions, there is a bit
less but still enough around. Check out the DIY menu on this site for some
examples.
Power supply
An essential part of your lay-out is the power supply, transformers are available in al sorts of sizes and prices.
The capacity of a transformer is expressed in VA (volt-ampere), this is
the voltage multiplied with the number of available amperes. A 64VA transformer
can provide 4 ampere when the voltage is 16 or 3.5 when the voltage is 18.
You could also build your own power supply for your track,
do be careful though as you’re working with high voltages.
Based on an example from Huib Maaskant you can
find a suggestion for your own power supply on
the DIY pages. This power supply uses a high
capacity toroidal transformer and splits the output
into six separate circuits each safeguarded with a
dedicated 3A fuse.
Computer
Many, if not most, people who are thinking about a digital model railroad
will also be thinking about automating their lay-out using a computer. This
certainly provides an extra dimension to your hobby.
Dutch speaking, or more importantly Dutch-reading, people are lucky as there
is a brilliant free-ware application available called ‘Koploper’.
Koploper is at least a match and often surpasses commercially available
applications with respect to flexibility and capabilities and it doesn’t
ship for 200 Euros!!
Back to the basic lay-out:
The computer controls the Control-unit that
generates the digital signals which the
booster supplies to the track.
There’s a lot more to say about this but first we
need to make a very important sidestep.
Booster and Control-unit
Boosters and control-units are available in all sorts of types and capacities.
It is also possible to build your own booster, on the net a variety of choices
is available.
Limiting factor is basically the level of electronics skills you have. Be
careful to select a DIY booster that can speak the protocols you need, almost
can handle DCC, most can handle Motorola and few can handle Selectrix.
For size N or H0 choose a booster with a capacity of
max 3-5 ampere, some of the DIY boosters go as far
as 10 ampere but that’s way too much for N or H0.
Always make sure you connect your track using a sufficient
number of connections and use thick wire to bring the
current to the connection point. There are varying stories
on closing the loop or not, I do not know for sure so I
didn't to avoid possible interference problems.
Next to DIY boosters there are also alternatives to the
control-units as offered by the main manufacturers. Two varieties worth
mentioning:
• Using a dedicated PC as a control-unit,
several versions are around on the internet,
some better than others. However, usually
they require quite some fiddling and most
are not very intuitive in usage.
• Buy a control-unit from an independent manufacturer.
Preferably a multi-protocol version so you are
fully flexible. Especially the Intellibox (in jargon
the ‘IB’) from Uhlenbrock is worth mentioning
here. This machine combines a control-unit and
booster in a single box and can speak DCC,
Motorola and Selectrix at he same time on the
same track.
Of the two options mentioned the IB is clearly the more expensive one but,
in my eyes, it is well worth it. A easy to use control-unit, a 3 ampere
booster and a truck-load of interfaces of which the PC and occupancy detection
are the most important ones.
Starting Sets
Of course do the big manufacturers also provide a wide range of boosters
and control-units, however, in general they are only able to use a single
protocol. Sometimes you’re even stuck to that manufacturer for all
other equipment as well, so be careful!
All manufacturers also have relatively cheap starter kits in their product
range, usually containing 1 or 2 digital locs, some track, a transformer
and a contol-unit. Do be careful because very often the transformer and
control-unit have very little capacity (sometimes just 1 ampere) or the
control-unit misses important interface capabilities like a occupancy detectors
or even a computer interface.
For the rest of the story it doesn’t matter very much as all the basic principles remain the same.
Disadvantages of computer control.
Are there any? Sure, a computer is just a stupid thing you need to tell
it everything.
Imagine driving a train from A to B on your track without a PC. As a human
you will do lots of things without even being consciously aware that you’re
doing it. You see where the loc is, how to set the switches, you see the
speed, you know you’ve arrived in B and when to stop.
To let a PC do all these things you need to tell it everything, what does
the track look like, how switches are positioned, where the train is, where
it should go and when it should stop.
All the above needs to be made clear to a computer, in the next sections
we will focus therefore on items such as Blocks, Sections, Occupancy Detectors
and Occupancy Reporters.
Blocks
Let’s take the following reasonably simple
lay-out (FI, this used to be my test-track).
Controlling this lay-out manually and
running two locs simultaneously is still
doable. You could drive around without
much trouble in loosing control. However,
a PC doesn’t understand this at all, maybe
without being aware of it you would never
send a loc into a section of track that is
already occupied. Let’s call such a piece of
track a Block and can have the status of ‘free’ or ‘occupied’.
Take this thought one step further and you could divide the lay-out into
7 separate blocks, like this:
The blocks have been numbered I to VII,
the small red lines indicate the block-
boundaries. A train will always travel from
one block to another.
Two important remarks:
1. In most applications a switch is NOT part
of a block as this creates confusion on the
required position of switches when going
from one block to another.
2. A block should, in general, be as long as the longest
train. Remember that a PC cannot see the train! Sometimes applications have
the capabilities of handling a long train in a short block, but certainly
not all application can do this.
As far as the computer in concerned the lay-out above would look like this:
The computer sees a lay-out as the combination of the
various relations between bocks. For our human
understanding I put the blocks in roughly the right position
but the computer couldn’t care less. Lets look at the logic
behind the schematic.
The relations show that a train can never go directly from
block VII to block I but always has to travel through
block IV or VI to be able to reach block I.
Of course there are many more properties involved in describing the relations
between blocks but for no the above is sufficient.
OK, you’re curious, other variables could be one-way traffic or two-way,
direction of traffic, priorities, maximum or minimum speed, is train allowed
to stop etceteras.
Sections
After defining the blocks you’re not finished as a block always consists
out of 3 sections. Using these sections you can indicate where a train should
start to brake and where it should stop. Maybe a bit fuzzy now but examples
below is an attempt to clarify this.
Situation 1, a through route (e.g. block IV in earlier
lay-out);
Actually not much happening in this block other than that the train should
stop just before the switch whenever the switch or next block is occupied.
To achieve this you need to define a piece of the block as the stop-section
(in this case even two as the block can be travelled in both directions).
If the train in block IV cannot continue to block I because the switch is occupied by traffic going from block VI to VII you want the train to stop in S3 and not already in S1 or S2. Going in the other direction is similar, stopping in S1 and not S2 or S3.
Situation 2, a station (e.g. block II);
If a long train needs to stop at the platform S2 or S5 (depending on direction)
would clearly be the preferred stopping positions, while a short train should
stop at S3 or S4. A passing train that needs to stop because the switch
is occupied needs to stop at S1 or S6.
The way you can actually arrange these things in an application depends very much on the application itself, but the principle of blocks and sections will remain.
Occupancy detection
So far we’ve seen how you can make the PC understand the lay-out
of your track, but how does the computer know where a locomotive is?
This is achieved through occupancy detection, in general every section of
your lay-out will be equipped with an occupancy detector and if these detect
an occupancy (=train arriving in a particular section) this will be reported
back to the control-unit and subsequently to the PC. If you have told the
PC at the start where a particular loc is situated on the lay-out, the PC
can then follow because of the occupancy detectors where the train is going.
Occupancy detection can be done in several ways but will always require
additional hardware (and costs). It is therefore useful to take a critical
look if occupancy detection is really necessary in every section. Take for
example situation 1 as described above, the through route. Almost all applications
do not really require section 2 to be detected at all, it suffices for them
to have section 1 and 3 detected. This is already a 30% reduction in occupancy
detectors! Beware that such a block still contains 3 sections! It’s
just that he middle one is not detected.
However, there is also quite a significant group of digital model-railroaders
who do detect the middle section with an eye to security as accidentally
uncoupled cars can still be detected in the fully detected situation.
Anyway, tThere are good arguments pro and con but the choice is yours.
Back to methods of detection. There are 5 ways in which occupancy detection can be realised:
Terminology
There is quite often confusion in terminology of the occupancy detection.
Things like Occupancy detector, Occupancy reporter and current detector
are often used as if they are interchangeable.
Better is to call the thingy that reports back to the control-unit the ‘Occupancy
reporter’ and the thing that detects occupancy the ‘Occupancy
detector’ and if the method of current detection is used you
would have a ‘Current detector’ acting
as an ‘Occupancy detector’.
To make life even more confusing, there’s also equipment out there
that combines current detection and reporting back to the control-unit.
However, in general the pieces of equipment are a lot more expensive per
section that others.
Occupancy Reporter
The current and Reed-methods are the most widely used methods. Each method has its pro’s and cons but there are also some similarities. The most obvious one is that both methods require an Occupancy Reporter to get the signals back to the control-unit. Usually this is done through one or more S88-units.
There are also other systems around, Selectrix system has
one for example, but the S88-unit is widely used
and almost
the common standard. It can be used in combination with
most control-units that have the option for occupancy
detection (not all do!!). Additional advantage is that you
have a wide choice in manufacturers.
For example: Märklin, Viessmann, LDT, Uhlenbrock, Lenz and others.
Lenz LR100/101
If the Lenz Digital Plus system together with the occupancy detectors LR100/101
has been chosen you cannot combine this with other occupancy detectors.
You will have to stick with Lenz.
Selectrix occupancy detection
With the Selectrix system the current-detection is build into the units
and they can therefore be connected to the rails directly. However, it is
very expensive and not to be recommended!
Reed contacts
Reed contacts are small glass tubes containing a very small switch that
is sensitive to a magnet. So whenever a magnet passes such a Reed contact
the circuit is closed and a signal can be generated. The contacts are usually
buried immediately below the sleepers (tie).
A disadvantage of using Reed contact is that in general the reliability
is slightly lower than using the current detection method. Two reasons;
positioning is to be done extremely careful and precise, secondly, by definition
the detection is a pulse and will therefore rely heavily on an flawless
functioning of the rest of the signal chain.
Current detection
As the name already suggests current detection works on the principle that a detection takes place whenever current is used in a particular rail section. This can be a locomotive but also a lighted carriage or even a boxcar where the wheels have been made a little bit conductive using resistor-paint or a plain resistor.
2-rail tracks have the disadvantage that they require some extra electronics
to make current detection in combination with a S88 work.
3-rail systems do not have this problem. Fortunately that extra piece of
kit can easily be made yourself and is fairly cheap. The required extra
steps certainly outweigh, in my eyes, the ugliness of 3-rail track.
As you can see things become a little more complicated. In effect is each
section isolated from its neighbours and separately supplied with current.
Whenever a loc in a particular section is using current, that section is
reported back as ‘Occupied’ to the control-unit and PC.
For those who are capable of translating electronics schematics into a print
lay-out there is a wide variety of current detectors available o the web.
For us, normal mortals, I have a found an easy to build unit that can detect
4 rail sections. You will need 4 of these units and 1 S88 to be able to
detect 16 rail sections. Out of experience I can assure you it works well,
currently I have 14 of these current-detection units active on my lay-out.
Terminus loop issue with 2-rail
Also in a digital set-up the terminus loop problem remains with 2-rail systems. However, this can very well be resolved using the Ultimate Terminus Loop project as described on the DIY pages.
Brand and Systems overview
Below a list of the main digital systems and their supplier. Do note the comments are just my personal views.
Märklin Digital
Web: www.maerklin.nl/digital.html
Mainly aimed at 3-rail H0 track. Not interesting for me for scale and ugliness
but even apart from that I always have this feeling they are acting to monopolistic
with associated prices.
Lenz Digital Plus
Web: http://digital-plus.de/
DCC system with a solid reputation, mainly 2-rail. Too expensive for my
taste.
Digitrax
Web: www.digitrax.com
Hardly used in Europe but big in America. No personal opinion on it.
Selectrix
Web: http://www.trix.nl/selectrix.html
Beautiful solid digital system with excellent driving characteristics. Terribly
expensive and as the smallest of the 3 protocols becoming a bit marginal.
Mainly used for N-scale. However, for those with N-track a very useful protocol
to have available on their controller.
IntelliBox
Web: www.uhlenbrock.de
Multi-protocol control-unit. It can address DCC, Motorola and Selectrix
decoders simultaneously on the same track. The machine contains two controllers,
a booster, PC interface, a possibility for programming loc-decoders and
a lot more. As far as I’m concerned the best value for money for ready
made control-units. Especially as you can easily build your track using
the cheapest or DIY-components without having compatibility problems. See
also the DIY-section on this site.
Twin Centre
Web: www.fleischmann.de
Basically the same machine as the Intellibox with one major difference.
The TC can handle DCC, Selectrix and FMZ but NOT Motorola.
FMZ is an ancient protocol introduced by Fleischmann in the eighties, not
used anymore but Fleischmann wanted to stay downward compatible. Go for
an Intellibox unless you bump into an extremely good offer.
Variations on digital
Next to pure analogue or digital there are numerous mixed systems available.
However, these systems usually have a fairly small audience and require
a significant amount of electronics skills.
If you are skilled in electronics do Google a bit and you will find trainloads
full of options on the web.
Digitising Locomotives
By far the majority of reasonably modern locs (since approx. 1990-1995) are equipped with a so-called NEM-plug. That means that converting a loc from analogue to digital is childs play.
1. Open the loc.
2. Pull the little AC or DC print from the plug.
3. Stick the decoder into the plug (observe the orientation!).
4. Program the decoder.
5. Make a test-drive.
6. Close the loc again.
Older locs without a NEM-plug can also be digitised but that requires some
more work.
In general follow these steps:
1. Open the loc.
2. Check for available space to put the decoder. Nowadays not much of a problem anymore as decoders become smaller and smaller.
3. Isolate the engine from the frame!!
4. Remove the analogue “rubbish”.
5. Solder wires from decoder at appropriate places.
6. Check and look again for mistakes.
7. Make sure isolations are made at correct places.
8. Check and look again for mistakes.
9. Program the decoder.
10. Make a test-drive.
11. Close the loc again.
For some graphical explanations see:
2-rail locs: http://people.zeelandnet.nl/rosoft/decoderinb.html
3-rail locs: http://people.zeelandnet.nl/rosoft/DecoderinbM.html
Now go do it!!
By now you’ve had quite a lot of information to absorb but please
do realise that this is just the beginning. Don’t be discouraged by
the above info, thousands of colleague railroaders have gone before you
and by far the majority have become active and enthusiastic digital model
railroaders.
However, start small. Digital railroading is complex at first and you will
go through a steep learning curve. As a starter you could consider building
the before mentioned test track.
Driving your locs digitally, manually or using a PC, is an adventure you
will undoubtedly enjoy for many hours to come and then I’m not even
mentioning the pleasure of building all sorts of equipment yourself, satisfaction
assured!!
Have lots of fun!
Henk