7  Out of control

This article is an extended edited version of an article that appeared in Narrow Lines as 'Out of control'.

The previous articles have examined the considerations underpinning the day to day operation of the typical NG continental railway and have investigated how this can be modelled. The infrastructure, the pattern of train services and simulation of traffic flows have all been examined as have the methods by which these can be reliably replicated on our models.

It might seem that every aspect of layout operation has been covered but these procedures assume that everything is running like clockwork without any operational glitches. Having spent some time working on the real thing I have an insight into how realistic a vision of the railways that this is! Incidents of an untoward nature are a daily occurrence and are dealt with by control rooms, which supervise individual problems and supposedly have a strategic overview of operations. This might seem irrelevant to a small NG railway which would not have a dedicated control facility but some-one somewhere, performed this role, even on a line as basic as one operated with just one engine in steam. Model railways usually ignore such problems as they are created as personalised ideal worlds where such imperfections are banished.

On the typical rural NG railway, characterised by the sparse services outlined in earlier articles, any problems would most likely be resolved by simply running the affected train at a later time. Few trains were scheduled and many lines were lavishly provided with crossing loops to provide operational flexibility. In such circumstances there would be few operational ramifications other than the usual crossing place being altered. For a simple NG layout based around a single station the majority of incidents would have little impact and could quite well be ignored. The AFK has moved away from this template in becoming a quasi-regional railway that runs a much more intensive service. Many trains are scheduled to cross other trains on the lower, busier half of the system thereby meriting some consideration of the factors which might affect the train service.

1--130 (Session 1 photo 130 in the Archived Operating Sessions section). A late running connection has made the afternoon mixto 15 minutes late. The farm girls will have to hurry to load the milk into the van to avoid further delays.

American modellers have generally made more attempt to simulate these problems than their British counterparts. B. Chubb in his book, “How to operate your layout”, referring to his well known Sunset Valley, examined the use of 'situation cards' to replicate the typical 'spanner in the works'. Each card outlined a specific problem which was presented to the layout operators at a point during an operating session. These cards were drawn before the session and placed into envelopes which were only handed to the relevant operator at a predetermined time. As a result the layout owner had foreknowledge of the impending problem whereas the operating crew reacted to the situation rather than anticipating it. For an owner-operator layout drawing from a pack of playing cards at 2 or 3 model ‘hour’ intervals might remove the crystal ball aspect and the system could be tailored so that a specified value (e.g. the ace of spades) indicated an incident. A second card could be tied to a table to outline details of the problem, but I have to be honest and state that I have never used this system.

An analysis of Chubb’s suggested ‘situations’ shows that they could mainly be broken down into three broad categories covering traffic problems, equipment malfunction and unexpected events outside the railway’s control, although before proceeding further I would clarify that I am only interested in everyday occurrences rather than the spectacular exceptions to the norm. It is quite common to find that individual line histories include anecdotes concerning unusual incidents but the majority of these stories are given a disproportionate amount of space when compared to the daily workings of the railway. Inevitably, the tale of the lion in the Southwold waiting room or the misdirected CDR train that demolished Killybegs engine shed, are rather more graphically remembered than the many days when all the trains ran on time and there was no significant problem.​​

Of the three categories, the seasonal surges built into the traffic generator remove many of the arbitrary scenarios outlined in Chubb’s list, most of which involved dealing with extra traffic. Generally speaking this could be anticipated and planned for. The vast majority of disruption on the railways is caused by trains running late, usually for mundane reasons. NG trains may have to wait for SG connections or a train may simply not be allocated enough booked time to carry out the work required. A quick glance through the archived sessions will illustrate the ramifications that such simple problems have upon a busy single line such as the AFK..

3-5 (Archived Session 3 photo 5). Insufficient time was allocated for this special train to complete its work at Boursson which had ramifications for other trains as was outlined in the following photos in the session's sequence.
A detailed examination of the activities of train 2108, the morning down local goods between Lacono and Relforka illustrates the problems facing goods trains which can cause late running.
1-5 Train 2108 arrives at Boursson where it crosses the Fenditavalat goods (train 2303). The train was assembled by the Lacono pilot with the last van, carrying empty tin cans, correctly placed to access the canning factory at the rear of the train. The three (incomplete raw styrene) vans are running through to the junction. This is possible today because the loadings from the KTT villages are not likely to fill out the train to its maximum 180 tons load. They are carrying canned products and cured hams for trans-shipment at Relforka. A flat wagon carrying vinegar in containers will be transferred from the up train to the down train. This is bound for Sojonno which can only be accessed by down trains.
The running of this train is closely monitored, as it is with all AFK trains. The timings between stations are mainly hypothetical on a layout  operated by one person (see article 4) but the time taken for shunting manoeuvres is recorded by "click clocks" which are simple counters dressed up as clocks. Model Railroader has published many shunting puzzle challenges over the years, most of which have been based upon the actual situations facing crews on well known model railroads. These usually stipulate the time that it takes to complete a move, which can vary between 2 or 3 minutes. In operation each move is defined as a change of direction by the loco, or a coupling or uncoupling. An uncoupling and change of direction counts as one 'turn'. These moves have been set at 3 minutes on the AFK. This might seem rather leisurely but shunting was a time consuming and dangerous activity, as noted in article 6.
The ornate clock at the left can be found in the gate tower at Fenditavalat whilst that at Urteno is part of the Art-Deco station building. Both are 'worked' by cog wheels behind the building frontage.
These are the portable clocks. The one on the left is about 2 inches square and has a face printed out from the computer. The much older 'digital' clock displays the ravages of time having been prised apart for repairs on a number of occasions. The principle is displayed by the indented cog wheel at the bottom left which is held by  a springy piece of metal to prevent it moving accidentally.
The local goods shunts the Syvrone cannery siding at Boursson, protected by the guard with a dirty red flag. He would have had to walk into the road to stop traffic before signalling the driver to cross, all time consuming activities.

The points at the intermediate stations are secured by the Sclußelwerk system (based on the use of individual keys unlocking each point) which effectively means that they are thrown by hand for shunting moves. The AFK rule book stipulates that these keys must be replaced in the locking frame at least 5 minutes before the arrival of a train so that the signals can be released. Accounting for the time taken to perform these activities whilst shunting, and avoiding other trains, is another aspect of operation that allows us to get closer to the real thing. Some-one in the central office had to be informed of late running if necessary and a decision had to be made about whether to hold the train or let it proceed and delay other trains. On a simple stand alone layout these might be largely theoretical considerations but these decisions can have far reaching consequences on a single line multi station system. Even on a one station layout such ‘clocks’ would prevent the commonly occurring situation where train A, which crosses with B, never ever arrives until B has finished its station work.

The train is assembled in the yard at Lacono during the early morning by the Lacono pilot ready for its 04.12 departure. The preferred order of the train is set out in the marshalling instructions and is as follows:

Crew van ~ this is mandatory as the brake must be on the bottom end of trains on the steeply graded line.
Wagons for Boursson goods shed
Wagons for Sojonno,
Wagons for BdeG Co-op
Through wagons for Relforka
Wagons for BdeG goods shed
Wagons for Boursson Co-op
Wagons for Boursson (Syvrone) cannery

2-13 The same train but a different day. Boursson is often the site of complex manoeuvres as the train is sorted out on the limited amount of track provided. The next two stops demand that wagons are correctly blocked. The train is on the main running lines and is protected by the flashing light under the cross-bucks.
The wagons for Relforka are only carried by this train if it appears that traffic will be light and that there will be spare capacity available. This involves assessing the likely loads from the three stations served by the down train and the set outs and pick ups which will be made rather than whether there is space on the train as it leaves Lacono. Two later through goods trains usually carry this traffic but often there are problems in removing all the loads offered at Lacono and empty wagons needing to return to Relforka are sometimes transferred by the local goods.

It is unlikely that traffic for all seven destinations would be offered on one train but it often leaves the starting point incorrectly blocked because the pilot has many other duties to attend to in addition to marshalling train 2108. Upon arrival at Boursson the train is allowed 203 minutes to drop off any wagons for the station. This might seem a rather excessive allocation for working such a small station but three other trains pass through which the goods must avoid delaying. The first train is goods 2203 which is timed to detach wagons for Sojonno if necessary. This is done by the loco of 2108 which removes them from the rear of the long distance train before re-attaching the brake to the bottom end of the train. Once this has departed the local can then set out and pick up wagons as required.

The wagons for the Syvrone cannery can be pushed past the siding onto the main before being placed on the siding once the outgoing wagons have been picked up. Herein lies another catch however. The class III loco allocated to the train is only rated to push 180 tons (18 axles excluding the brake) into the siding which is level. This load is reduced to 160 tons on the section of line below the station as the loco is forced to push the train uphill. If the loco gets onto the gradient with more than 160 tons, excluding the brake, it cannot return to Boursson. When the train is 'heavier' than this the excess wagons must be placed on another siding before the shunt move can take place. Wagons are allowed to stand in the sidings and on the mainline towards Lacono without the brake van attached but no wagons whatsoever are allowed to stand on the mainline towards Glissent.  This drops away so steeply that it was the scene of a spectacular pile up in the earliest days of the layout before the danger was fully appreciated. It is quite possible for a heavy free running wagon to roll all the way to Glissent, including crossing the hump backed bridge over the Ero, before potentially reaching the buffers at Relforka if it remains on the track!

Having run round the train to shunt the Co-op siding and picked up any outbound wagons the crew must remarshal the train to ensure that any Sojonno wagons are next to the van because the siding there is so steep that the loco is limited to a 60T maximum load. This is also a good time to ensure that the wagons for Breĉo de Glissent are in their correct positions given the various wagons that may have been 'lifted' from Boursson. In between times the train will by now have crossed down railcar 1106 (Urteno-Relforka), placed any wagons for up destinations into convenient positions for the up local goods to pick up and should be ready to depart once railcar 1441 (Eromarbordo-Urteno) arrives.
1-28 Train 1106 runs down the mainline (RT) whilst the goods stands on the loop to let it past. The loco is on the rear end of the train.

Having shunted Sojonno the train arrives at Breĉo de Glissent where it has to work within similar constraints whilst avoiding three other trains between shunting the Co-op from the front of the train and the goods shed from the rear. Glissent is unusual in that the goods shed is on a dead end siding rather than a loop due to differences in ground levels.
1-50 The train shunts the preserves factory at Sojonno. The vinegar containers were transferred from the early morning Fenditavalat goods at Boursson allowing a much quicker delivery than would otherwise have been  the case. The rest of train, on the right, is held by a track brake. The handle for a brake operating in the siding can be seen at the lower left of the photo.
1-69 The loco shunts a van from the rear of the train into Breĉo de Glissent's siding whilst keeping clear of passing trains.

It should be apparent from the foregoing analysis that it is easy for this train to lose time which would create problems for other trains on the tightly scheduled lower section of the mainline. On the other side of the coin the shunting times are based upon a larger than average number of wagons being presented by the load generator. Whilst it is not unknown for the full 36 minutes to be needed at Sojonno the train often leaves 20 minutes early and occasionally does not shunt at all. Although Sojonno is a block post it is not a passing point and such mistimings are inevitable given that traffic varies on a daily basis. These are often exploited as the AFK control uses its local knowledge to finesse the published timings. Goods train operation is flexible with schedules being regarded as being advisory rather than set in stone. Provided that they do not get in the way of passenger trains, especially those with mainline connections to make, freights can run early or late and have their journeys curtailed or extended to suit demand. These events occur naturally on a single line network but it might be difficult to simulate them at a single station without things becoming a little too contrived.
2-32. On this occasion train 2108 appears out of the mists covering the Kasatritikakamaparoj running 35 minutes late having stopped to pick up two vans of jam at Sojonno.
The AFK simulates late running SG connections by using tables and dice. Each of the 8 connecting SG trains (4 in each direction) has had its probability of lateness evaluated. The morning connection running overnight from Sarip (the capital) is the worst offender as it travels most of the length of Thalnia overnight. These late starts have repercussions for the AFK which become more pronounced in the further reaches of the railway as the delays upon the single line are compounded. Whilst there is a reasonable margin of recovery for the stock of a late running express to be turned around this is not the case for the afternoon 'mixto' between Relforka and Fenditavalat. As a contingency a spare set of stock is kept at Fenditavalat Rivabordo which can deputise for the rostered vehicles. This takes preference over the late running train in planning crossings and if necessary the up train is held to allow the down train to keep to its timings.
Many sessions took place before those recorded on the website. This is a recreation of a photo used in Narrow Lines. The express passes itself at Ospicio d' Helcaraxë. The express in the background was running so late that it was held for the 0-10-0T and the spare set to run on the times of the returning train. It was noted that the restaurant car was doing a roaring trade and that the passengers had plenty of time to observe the alpine landscape (whenever it gets painted!).
This was a common procedure on many NG lines where SG connections had to be made. As was observed in article 4 FT (the Thalnian SG system) trains will wait for late running AFK connections within reason (which is currently unspecified!) but they will not be held indefinitely. Given that the mixto makes the last connection of the day and that there is no settlement as such at Relforka, other than railway workers housing, the AFK could be faced with an expensive array of bills for an overnight stay in the FT's hotel if it did not adopt this policy. In the up direction the AFK is forced to react to the SG's late running  which again causes problems with the last connection to Fenditavalat which nominally departs Relforka at 18.00. The importance of the former provincial capital ensures that this train is always held for late connections but the AFK cannot come to a standstill whilst waiting for it. The train is notorious for late running across the Altingablecaŭtoj as it requires banking up the Vuklpafaŭkangulo and the relevant assisting loco has often been assigned to other essential duties during the interim period.
2-134 The last railcar of the day is banked up the Vulpafaŭkangulo. On this occasion it was running late due to fog and its arrival at its destination was later still because of the need for snow ploughs to clear the line. The tractor is testing whether the ice on the Vardenamero is strong enough for skating!
Having considered the problems engendered by late running because of operational problems affecting the published timings some consideration needs to be given to the provision of extra unscheduled trains. These usually run to accommodate surges in traffic which can be anticipated because they are incorporated into the load generator. Market and fair days are known in advance and special mixed trains are run to these. The grain traffic at Boursson examined in section 5 (illustrated in Table 5.3) is worked by two or three grain specials and similar arrangements are made for other agricultural peaks. Other traffic, such as the fish loadings from Eromarbordo, are another area of uncertainty depending upon the tide tables and the seasonal movement of the shoals. These loads are determined using the probability system with the dice being thrown at a specified time during the day (which varies each day to reflect the tide) and would be easy to incorporate into a simple layout’s operating system. Given the volatility of the loadings and the timings Eromarbordo is scheduled to receive three local goods trains per day even though one of them often does not run. The fish traffic simply takes priority on the most convenient service.
3-157 SG refrigerator vans are shunted during the late afternoon at Eromarbordo. High tide was around midday and the van in front of the chapel is loaded.
3-79 The late January session was characterised by the need to run special trains to pick up sugar beet loads at Boursson.
As has been observed, the timetable for the AFK is plotted as a graph and special conditional paths have been incorporated for use by seasonal trains, which additionally provide ‘windows’ for late running freights when not required by specials. A plastic covered copy shows the basic service onto which seasonal trains are transferred for each operating session with notes about their working made in the margins. In some cases these times have been plotted onto the 'master timetable' and can simply be copied onto the daily sheet in the knowledge that the timings have been calculated beforehand. These 'conditional paths' were a common feature of real working timetables and were indicated by a Q in the column above the timings indicating 'Runs as required'. Scheduled special trains need to have their timings and train numbers arranged in advance and any alterations in the pattern of normal services need to be noted. As far as possible the AFK tries to limit these alterations to the status quo to a minimum and to ensure that they only affect freight trains. An ‘office session’ before running trains generates these timings and train numbers which are also marked onto the timetable sheet and noted onto blank timetable cards for the relevant trains. This, I suppose, is the European equivalent of the dispatcher issuing instructions for an ‘extra’ on an  American system.
3-92 The AFK's base timetable is shown on a train graph protected by plastic, hence the reflection of a fluorescent strip on the photo. Regular passenger trains are shown in red and goods trains in blue. The extra trains are shown in a water soluble pen  which can be erased as necessary. These are the green lines here. The smudges at the bottom of the sheet show the hectic rescheduling of trains to accommodate the late running freight (arriving at Relforka at 12.35). The gap between the down railcar's arrival at 12.22 and the mixto's departure at 12.55 was originally the slot in which the beet extra would have run. This has hurriedly been retimed to reach Glissent (BDG) and leave at 12.15. The extra goods will leave Lacono at 12.33 and will take the afternoon local good's path to reach Relforka. The local will have to run later. The Tippex in the bottom left corner indicates that the freight has been retimed from its old path in any case and the green lines at the top indicate paths for trains to the ski lift at Varden which are in the master timetable but not included on the daily version unless required. They are included on the train's individual timetable in italics.
3-93. The relevant "documents" for the re-timed trains are written onto plastic covered post cards. From left to right these are; train 9146 the LAC-RFK extra; train 9323 the RFK-BSN extra sugar beet train; train 2719 the RFK-LAC local goods and train 2116 the FDV-RFK goods. For modellers who just like to run trains as and when this might seem to be an avoidable layer of administration. To me this is the whole point of modelling and replicates the 'scrambling' that would have taken place on the real thing to try to square things up. I remember once reading an article in Railway Magazine (back in the seventies) about how Newcastle control could write the schedule for a mineral lead diagram on the back of a fag packet and provide a day's work for a class 37, brake van and crew without any problem in about 2 minutes. The AFK has not quite reached that casual level of impromptu sophistication at the moment but it is trying!
The pre-session arrangements for extra trains usually cover anticipated problems in accommodating traffic surges, but this has increased over the years as new flows have been added to the generator. In the more recent sessions it has become the practice to station an additional spare engine at Lacono Aspargo shed to operate a "Von Ryan's Express" if necessary. This term refers to the film of that name and according to what I understood from my late father was the common term used in Leeds Division offices for an ad hoc engineer's train which ran daily but without a timetable. I understand that the term 'Martini' was also common for such trains because they ran 'Any time, Any place'. Such trains are used on the AFK to solve unforeseen but pressing traffic problems and are avoided as much as possible because they almost inevitably cause problems for other trains, including passengers, as they are scheduled 'on the hoof'. This might be over egging the pudding for many NG systems but such chimera were a common feature of railway operating life as fourth and fifth thoughts occurred as different people poked their fingers into the pie in attempt to resolve a problem, usually to their advantage and everyone else's inconvenience!
3-94 A Von Ryan's Express. This trip working up the Vulpafaŭkangulo was the cause of many problems, as was often the case on the prototype. Unfortunately what started out as a simple transfer of coal wagons to the mine at Ithilarak became complicated when additional demands were made of the loco and crew. Matters were not greatly helped by the running of a second impromptu train. More details are contained in the archived operating session.
To date we have only considered fairly routine events concerned with traffic working but there were other less predictable events that also affected operations. These incidents usually fall into two categories, namely, mechanical problems and factors beyond the control of the railway. The simulation of equipment failure is more complex than might at first sight appear. Important items of plant, such as locomotives, would be highly unlikely to suffer catastrophic failures such as boiler explosions or axle breakage. On a well maintained railway most lesser failures would be detected before the locomotive came off shed, rather than in traffic, thereby causing delays or the re-allocation of engines between diagrams. I have still to devise a probability chart for these incidents but I intend to do so at some point. AFK locos ‘receive’ scheduled maintenance, recorded in tables, and have to be worked to Relforka by certain dates for major work but they can be stopped at Lacono or Fenditavalat for boiler washouts. For our simple through station these considerations would be of little relevance, beyond questioning the routine assignment of a favourite engine to a particular train, but they would impact upon a terminus to fiddleyard layout where a shed was incorporated. With reference to loco problems, advance decisions are taken on the AFK as to whether engines will need preparing to deal with traffic demands because, unlike our electric model locos, it took time to raise steam to working pressure. This may be yet another perverse example of adopting an operational mindset but it does explain why many AFK reserve locos are diesels (although I have not gone so far as to roster spare loco crews!)
The morning express and local trains have usually been photographed in the recorded sessions and provide an example of how stock is rotated around the diagrams.
1-64. With Christmas approaching the postal traffic was building up, hence the heavily loaded morning local passenger train at the left. Relforka shed turned out No 313 the 2-6-6-0T for this duty and assigned No 222 the 1-Co-Co-1 to the lightly loaded express (train 11005).
2-31. The express today (in early January) is hauled by the 2-10-2T and the local by the Austrian 1-B-1 loco.. Both trains are 15 minutes late as the fog has affected the SG connections.
3-67. The local and the express line up for the off in late January. The express is hauled by the 2-6-6-0T and the local by a 2-6-2T. The express was slated for the Swedish deisel but the model had to receive attention so the steam engine was a late deputy.
The rough treatment meted out to goods stock, in contrast, renders them the prime candidates for failure. This reflected the nomadic existence of stock on the SG systems and, presumably, the parsimony of the smaller NG railways where the foibles of individual items of stock were known and informally monitored. The AFK is quite prosperous and therefore pre-emptive in preventing the failure of wagons in service. Taking into account the smaller wagon fleets of model railways the occurrence rate has been set at 1 in 1000. Every goods train is checked on arrival at each intermediate stop such as Boursson (i.e. dice are thrown). Hot boxes, pulled couplings, shifted loads and brake defects are simulated and the time needed to fix the fault is noted, which varies depending on the location, reflecting staffing and ease of access. One side effect of this is that a slightly increased roster is needed to cover for the defective equipment and to allow other stock to receive routine maintenance. It is hoped, as well, that operational reliability will eventually increase to a point where derailments become an occasional rarity (don’t hold your breath!) These obviously would not need simulating but, rather than simply replacing the offending wagon onto the track it could be left there for some time to represent the jacking up and slewing processes needed on the real thing, with perhaps a 30 minute delay per axle.
In another recreated picture a (posed) derailment is shown at Caladonno. As has frequently been noted in the operating sessions derailments are still too common on the AFK although they are reducing in frequency as remedial work is carried out.
The incidents outside the control of the railway, but which affect its operation, are the most difficult to simulate. The category can effectively be subdivided into ‘acts of God’, essentially the weather, and human actions which affected the railway but were not caused by it. The nebulous nature of these factors presents an almost unlimited list of potential events! One of the background exercises carried out in the earliest days of the extended layout, once it had been firmly established what could be fitted into the railway room, was the production of a ‘definitive’ map of the AFK. Topological features were included in addition to settlement and vegetation patterns, although I have to add, as a trained geographer, that the AFK runs in a very unlikely setting! The availability of such a map, in conjunction with a progression throughout the year with each operating session, made it possible to produce probability tables for different types of weather conditions and their effects upon the railway. Snow and fog conditions have so far been simulated, again using probability charts and dice throws, with the subsequent delays being taken into account when running trains. Whilst I am sure that it would be possible to devise a probability table for the human factors I have not gone beyond simulating late running connections with the FT, discussed earlier. I suppose areas which could be considered would be accidents with road traffic and escaped farm animals but I have no plans to implement these ideas at the moment.
2-1 Fog and snow affect AFK operations and are symbolically represented here by a piece of insulation and cotton wool. The snow will take 5 minutes to remove and the fog will also add time to the journey but will be replaced once the train has passed. The line crosses a high alpine pass, some of which is modelled in winter, and can justify the use of specialised equipment but for many models this is probably an irrelevance. The plough can be seen in a more complete state in session 3 although ironically it was not required then as it was a bright clear day.
Reading this article makes it sound as though the progress of a train across the AFK is dogged by dice throws and fraught with dire consequences at every turn. I must admit that the dice are used a lot during an operating session, but how long does it take to throw one or two? The values are weighted to make sure that there are very few problems and that when they do occur they are credible incidents. The upshot of these simulations and the variations in traffic mean that each operating session is very different from the one that preceded it. This freshness, rather than following the same routine in every session, provides me with the stimulus needed to ensure that construction continues towards a completed layout. I often wonder if the root cause of many layouts never being finished is that they are run in a desultory manner with a consequent wane in interest for their builder. The articles probably go much further along the road of operation than many readers thought possible, although I doubt that much of it would be too great a revelation for an American modeller. Hopefully some of the ideas aired will help someone somewhere to improve their operating practices.

The operating sessions section of the website contains details of the current session and the archived sessions stores those that have finished. Unfortunately it only became my practice to take pictures when I began to run the website and post on internet forums, although some aspects of a previous session were photographed for the Narrow Lines article upon which piece is based. There is no pictorial record for the many sessions preceding this period.
Return to operational articles index