Scale speed - Couplers - Setting up the game

Once you have built a shunting puzzle layout with sound baseboard foundations and track laid down in a trackplan arrangement which allows for the rules of your chosen shunting puzzle and provides smooth and reliable running of your stock, you will want to operate it - after all, you have set up the playing board, now you want to have fun actually playing the game - for which you may find a few useful or interesting hints here.

  The actual rules of various shunting puzzles are covered elsewhere (e.g. here for Inglenook Sidings and here for the Timesaver) - this page is all about making things move on a shunting puzzle layout.

Scale speed

A shunting puzzle requires a model locomotive which runs reliably at slow speeds. In doing so, it replicates the prototype which doesn't allow excessive crashing and banging simply because real railway shunting operations try to minimize damage inflicted on rolling stock (and the goods it carries) as well as accidents.

Class 08 (UK) - max speed 27mph

ALCO S-1 (USA) - max speed 60mph

Ee 3/3 (Switzerland) - max speed 25mph

  Back in the 1980s and early 1990s this was still a problem as many good looking model locomotives only ran reliably at speeds which were way too fast. Since then, the quality of the motors and gearing used even in low cost models has improved markedly and flywheels are now a common feature. If anything, it is the operator today who needs to be reminded to slow down and run his shunting puzzle moves at near to scale prototypical speeds. 

Obviously speed ranges of shunting locomotives vary according to location and era, but shunting moves in yards and on industrial tracks are usually very slow. Sometimes this is reflected in the maximum speed range of the prototype (at 27mph, the British Cl 08 shunter can't do anything else than trundle along even at top speed, and the Swiss Ee 3/3 couldn't even quite keep up with that, rated at a maximum speed of 25mph), but even an Alco S1 switcher would never do anything close to its potential 60mph when actually performing switching duties. "On the job", locomotives sorting out and dropping off freight stock (or even passenger stock) will usually be working at speeds of 5-10mph.

The best way to track the speed of your models is to measure a straight and level stretch of track and then note the time the locomotive takes to cover it. In 00/H0, a model travelling at a scale speed of 5mph (7,5 km/h) will take 12 seconds to cover a distance of 1 foot (30 cm). If you're travelling along at 15mph (23 km/h), it's still 4 seconds. In N scale, a model travels 0,55 inches (1,3 cm) per second at a scale speed of 5mph (7,5 km/h).

Scale speeds usually don't come for free, they require good equipment (both locomotives and track) which is well maintained. As far as track is concerned, good electrical conductivity is a must - if it can be achieved by regular cleaning, that's fine, if not, all electrical contacts need to be improved and maybe additional track feeds will be necessary - and maybe upgrading to a better control unit will be inevitable at some point.

If a locomotive still refuses to run slowly even under near perfect track and power input conditions, it will simply not qulaify as shunting puzzle motive power. Re-motoring or re-gearing used to be the playing field of experienced modellers prior to the 1990s, but today that hardly seems viable - model shunters are mostly geared to prevent extreme non-scale speeds and come with flywheels to ensure smooth running even when the locomotive is crawling.



As the word coupler implies, its primary function is to join up individual pieces of rolling stock and ensure that they stay coupled, be it whilst in motion or standing still. This effectively results in the necessary control to form and run a train without unwanted "runaway" incidents.

The same holds true for the model - couplers are designed to attach one piece of rolling stock to another as securely as possible. One specific aspect of shunting puzzle layouts, however, is the fact that uncoupling is just as important as coupling.


Buffer and chain coupling

(Chris McKenna)

Knuckle coupler
(Roy Smith)

  The variety in coupler types found on railway models is, in fact, mirrored by the prototype, where several completely different coupling systems existed and still exist today.

In Europe, the standard device is the buffer and chain coupling, first used by the pioneering Liverpool and Manchester Railway in 1830. Rolling stock is coupled by hand using what in British railway terminology is called a screw coupling, consisting of a hook and chain links which are secured through an integrated turnbuckle that draws the vehicles together.

This coupling requires and has brought about the typical European feature of buffers on the ends of rolling stock an essentially provides an arrangement which limits the slack in trains and, with sprung buffers, absorbs shock impact.

In North America, today's standard AAR (Association of American Railroads) knuckle coupler essentially goes back to Eli H. Janney who secured a patent for his newly designed coupling device in 1873. Thus also called a Janney coupler (or, in the UK, where this type of coupling is used on certain multiple units, a buckeye coupler), it is a semi-automatic coupler which - like the buffer and chain coupling - needs to be unlocked manually by using a cut lever but which - unlike the standard European coupling - couples and locks automatically as the knuckles of two couplers are pushed together.

This type of coupler is also used in a number of South American, Asian (noteably Japan and China) and African countries as well as in Australia and New Zealand.

The conclusion to be drawn from the prototype is that there is automatic coupling, but (with very few exceptions requiring special installations) no automatic uncoupling. The same, incidentally, goes for model couplings.

Ever since model trains were produced commercially the major concern with regard to couplings was to make sure that individual vehicules remained coupled. This certainly makes sense from the continuous run perspective, and therefore the vast majority of model coupling systems supplied with ready-to-run models for a very long time seemed to assume that once you had coupled two items of rolling stock together you would not want them to part again.

Standard UK tension lock coupler

  No matter whether you're looking at the (still current) tension lock coupler (00 scale, UK), the (now obsolete) hook and horn coupler (HO scale, US) , the (still predominant) drop loop coupler (HO scale, Europe), the (standard in Europe, fading in the US) Arnold Rapido coupler (N scale) or the (standard for Europe) Marklin horn coupler (Z scale) - they all have one thing in common:

None of them is designed to be uncoupled without either the famous "big hand from the sky" or some rather unsightly uncoupling devices which take the form of oversize ramps in the best of cases.

The problem from a shunting puzzle perspective is obvious: the fun to be gained from operation is in direct proportion to the amount of both coupling and uncoupling involved, and if this is a tedious process in itself, then the fun factor immediately drops radically.

In order to prevent this, there is the option on a DCC controlled layout of having rolling stock equipped with digitally controllable couplers - if these are available or can be retro-fitted, there are no uncoupling devices to be installed, and yet stock can be uncoupled anywhere on the layout. On a conventional DC layout, (electro-)magnetic uncoupling devices have proven their worth in terms of appearance and function over a long period of time.
The pioneer maker of magnetic uncoupling devices - Kadee (R) - was specifically founded in 1947 by twin brothers Dale and Keith Edwards with the intention of producing a coupler for railroad models which looked as much as the real thing as possible and allowed for automatic coupling and hands-off uncoupling.

Today, a fair number of producers have their own version on offer, and the principle underlying the operation of "magnematic" couplers (as the famous Kadee branding goes) is well known.

Fashioned on the knuckle coupler, the pin running down towards the track (and thus obviously specific to the model coupling) will be forced sideways when running over a magnet. If there is no pulling force on the coupling, i.e. if the rolling stock is standing still, the knuckle will swing sideways with the pin and open - effectively uncoupling itself.

As the pin is still forced outward by the magnet, pushing back into the (open) knuckle allows for delayed uncoupling: a piece of rolling stock can be moved to any point and dropped off after it is uncoupled without recoupling as long as the loco is pushing it.
  Today, most HO, N and Z scale US prototype models are sold with Kadee/Micro-Trains couplers or variations thereof, some of which were born out of necessity because Kadee/Micro-Trains would not license their couplers to some manufacturers. Whilst they all look pretty much the same at first sight - as knuckle couplers would - some do not allow for magnetically induced uncoupling.

Rokuhan, Japanese producer of Z scale track and models has published a video on youtube advertising their uncoupling track which nicely shows how "magnematic" uncoupling works, illustrating as a side-effect how well suited this type of coupling/uncoupling is for a shunting puzzle layout.

However, using Kadee couplers with European or British outline stock is not as straightforward as the existence of the NEM-362 coupler pocket would suggest, even though Kadee has a range of different couplers suitable for specific models. Converting these models to this type of couplers can still be a task requiring skill and time.

Just as important as the couplers are, of course, the uncoupling devices. They, like the couplings, should work reliably. Most layouts running stock which is equipped with "magnematic" couplers use uncoupling magnets which are either installed between (i.e. visible) or under the running rails (i.e. out of sight).

The magnets don't look all too prototypical, of course, but there are a number of ways to make them less conspicuous (e.g. using them as part of a yard pathway crossing the tracks). The magnets which are installed below the track usually produce a stronger magnetic field; apart from the fact that they require a lot of advance planing (they can't be installed once the track is fixed down and ballasted) they also have a strong tendency to interfere with the running of trains, causing unwanted side-effects such as metal axles and loco metal underframes being "locked on".

Some manufacturers offer set pieces of "uncoupling tracks" which have a magnet ready installed between the rails (e.g. Kato Unitrack) or out of sight (e.g. Rokuhan Z scale track which hides the magnet in the preformed roadbed).


A Micro-Trains N scale uncoupling magnet in place on an Inglenook Sidings layout

The positioning of uncoupling devices is also very crucial for a successful shunting puzzle layout - there should at least be enough of them to allow for all the necessary uncoupling moves.

If necessary or desired, an Inglenook Sidings layout can be operated with one uncoupling device only.
It would, however, be more advisable to use three uncoupling devices (one each at the throat of each siding) if possible.

A Timesaver layout requires four uncoupling devices located in specific positions.
It won't work with less, and it doesn't need more if couplings are used which allow for delayed uncoupling.

Some may want to reduce the number of uncoupling devices to a bare minimum and will accept having to sometimes make lengthy and not very prototypical shunting moves such as pulling the entire string of rolling stock from its siding in order to get at one single freight car, while others may want to spread a larger than needed number of uncoupling devices around the tracks. In any case, the trackplan and operating rules of a shunting puzzle need to be studied carefully in order to know what the minimum requirements for uncoupling devices are. The rest is, as so often, a matter of personal taste.

Personal taste also relates to as to whether or not to use magnetic uncoupling. Some modellers - and there seems to have been something of a revival of late - seem to prefer manual uncoupling. Their argument is that this has the advantage of being able to uncouple wherever you need or want to without any restrictions, rather than having to pass over a magnet first. Manual uncoupling - often using some device which can best be described as a "stick" (for N scale, a toothpick will indeed do the job) - also eliminates problems with unreliable uncoupling and the necessity to fine tune couplers. In the end, the important thing is that uncoupling is easy and reliable - possibly the most important point of all when operating a shunting puzzle layout.


Setting up the game

Finally, when all is well on the layout and running smoothly, it's time to start playing the game. With other games, this would mean perhaps throwing a dice or picking a card - but how do you start a game on a shunting puzzle layout?

With both classics, Inglenook Sidings and Timesaver, you need to know which freight car (8 of them in a standard game of Inglenook, 5 in John Allen's original way of switching the Timesaver) is required to go where before you can start. Obviously, there are almost endless ways of doing this, ranging from old tech to high-tech, but it will always involve handling tokens which represent the individual cars and where they are to be moved.

One good idea to start with is to have distinctive looking items of rolling stock which are easily distinguishable (unless the shunting puzzle replicates a prototype that used only one type of rolling stock, of course).

As for the selection process of which piece of rolling stock needs to go where - Alan Wright used his famous "tiddlywink computer" (a mug which held a tiddlywink (i.e. a token) for every piece of freight stock on the layout) from which the required number of 5 would be drawn, and the order in which this happened at the same time determined the order in which the cars were required to end up. You really don't need more than that, but I have found that using a system heavily influenced by the classic American way of creating switching orders, the card and waybill system, will make things unmistakeably clear even to someone having their first go ever. The following illustrations apply for an Inglenook Sidings layout, but they're just as valid for a Timesaver layout (which due to its US roots will probably feature car cards anyway).

The system is quite simple. On a classic Inglenook Sidings you have 8 freight cars, each of which needs its own card as a means of referring to this specific car. This can be as simple or as elaborate as you care to make it - perhaps a slightly more elaborate form could look something like this:

For this US prototype Inglenook Sidings layout, every car not only has its own card but cars have also been selected so that the colour of the car (there's only one yellow, one green, etc.) allows for immediate spotting of the car. All in all, there's a total of 8 cards of identical size which are shuffled and then the five uppermost are put down. The order in which this takes place also serves to determine the order in which the cars should end up.

Here, another two cards need to be drawn, but so far you can tell that the car which is required to be the closest to the switcher is the the brown D&H boxcar, then the yellow ACY boxcar, followed by the green Vermont Railway boxcar.

It takes a bit of work to make all the cards you'll need (though it can be done in 5 minutes if you don't feel like having any fancy special effects), but it's worth it, and it adds quite a bit of atmosphere to operating a shunting puzzle layout. Naturally, the same can be done e.g. for a UK prototype layout.

Naturally, you could also have a computer do the shuffling for you - which is precisely what the Inglenook Random Wagon Selector by William Pearson will do.
  This neat little piece of software dates from 2003 (it will still run fine on most hardware running Windows OS), and the screenshot shows an example list of cars on the layout and how the Selector produces a random list of cars in the order in which they are to be shunted (courteousy of and with the kind permission of Mark Kendrick, you can still download it from this site as a zip-file).

Or, if you prefer, you can input your rolling stock into an online random list generator and have your shunting order made up in your web browser by simply ignoring the last three items in the list (one example is the List Randomizer from random.org).



Text, pictures and illustrations not labelled otherwise are © Adrian Wymann
Page created: 11/OCT/2002
Last revised: 02/JUNE/2023