Facebook Page
Coupling, Handing and UNDMs

Underground-Symbol-Small.gif (1234 bytes)Tubeprune - The Tube Professionals' Rumour Network
Tubeprune is an unofficial web site for professional railway people working for London Underground and for those interested in the London Underground railway system.  To the Tubeprune Home Page

Links to pages on this site:-

Home Page

What is the Underground?

Underground Map




Life Underground


Signalling Pages

Photo Page



Jubilee Line Photos

Victoria Line ATO

The Central Line at White City

FAQs - General

Driver Training


FAQs on the PPP

SATS and Bats

Train Operator's Notebook

Headcodes and Marker Lights

Life Underground Overseas


C Stock

D Stock

1967 Tube Stock Photo List

1973 Tube Stock

Track & Traction Current

Rolling Stock

Links to other LU sites

Contact Tubeprune here

Coupling, Handing and UNDMs

A question in the newsgroup about the use of NDMs (Non Driving Motor cars) and shunting control cabinets instead of driving cabs set off a whole series of messages asking various questions about London Underground train formations and types of cars.  This article tries to answer those questions and it also covers a lot of details about London Underground train operating practices.


Background - The Central London Railway - The Wood Lane Loop - Control Trailers - Pairing and Lettering - LER Tube Lines - Main and Train Lines - Developments In the 1920s - Numbering - Loops and Triangles - Uncoupling in Service - Four plus Three - Sub Surface Trains - Metropolitan Complications - District Formations - The District in the 1920s - The F Stock - District Uncoupling - Unit Stock - UNDMs - Uncoupling - 8-car trains with 4 cabs - Single-ended units and double-ended units - 1972 Tube Stock - Universal Couplers - The C Stock - 1973 Tube Stock and D Stock - The 1990s - Automatic Couplers - Coupler Controls - Manual Control


London Underground has been running electric multiple unit trains since 1901 and was the first operator of such trains in the UK.   Since that time, a unique range of train formations and coupling strategies have been tried, developed and improved.  The reasons for the choices made and the development of the strategies are covered in this article, which describes the progress of the Underground's train formation policy over a 100 years of operation.

The Central London Railway

CLR 03 TS gate.jpg (42426 bytes)The first line in London to be provided with multiple unit trains was the Central Line (then called the Central London Railway) in 1903, following some experiments in 1901.  Trains were formed of a set of four or five carriages with a powered carriage at each end.  This photo shows a motor car with a trailer at the leading end of a train.

Click on the image for the full size view and detailed description.

Following American practice the carriages became known as cars and the powered ones at the ends, motor cars.  Intermediate, non-powered cars were called trailer cars or "trailers".

Central London trains were restricted for space because of the diameter of the tube tunnels through which they ran.  The car floors were about 1 foot (300 mm) lower than those of conventional railway carriages, so the clearance between the floor and the track was rather tight.  The clearance was further reduced by the location of the current rail in the centre of the track.   These restrictions led to a special design for the couplings between the cars.   The standard UK screw coupling would have fouled the central current rail used on the Central London so the coupling was a simple link and pin arrangement.  A central, sprung buffer plate was provided instead of the traditional British side buffers.   The hoses used to connect the brake pipe and main reservoirs were also arranged differently so that, instead of being hung under the headstock, they were hung from standpipes at car waist level.

Hanging the hoses at waist level produced another problem.  The cars ran in tube tunnels where the only means of escape from a stalled train was through the ends of the train.  Therefore, there had to be a walkway through all cars to allow everyone to get out.  Between cars, the gangway had to be kept free of couplings and hoses of course, so the brake hose was placed on one side of the gangway and the main reservoir hose on the other. 

LER gate ends.jpg (41650 bytes)With the space on either side of the gangway occupied by hoses, the multi-core control cable and the lighting supply cable had to be hung at roof level, as shown in this photo of a pair of LER gate stock cars coupled. 

Click on the image for the full size view and description.

The cables usually got in the way of the head of the gateman employed to stand on the end platforms to open and close the gates, so straps were used to tie them up tightly.   This sometimes caused the coupler heads to pull out of the sockets as the cars went round sharp curves.  It was a problem which lasted as long as gate stock was in use on the Underground, up to 1930.

To the Top of this Page

The Wood Lane Loop

Until 1908, Central London trains ran between Shepherds Bush and Bank.  At each end of the line, trains were reversed when the driver changed ends and drove off in the opposite direction.  The simple CLR shuttle service meant that the train's brake pipe was always on the south side of the train and the main reservoir pipe on the north side.  In fact, the driver's position always stayed on one side too, the north side, regardless of which direction he was driving, because the driving cabs were built for left hand drive when going eastbound and right hand side driving when going westbound.  This was a legacy from the days of locomotive operation when the driver’s position was on the north side of the locomotive.  Signal sighting was originally designed for this position of driving.

In 1908, the Central London was extended from Shepherds Bush to Wood Lane and a terminal loop was built.  As a result, at the west end of the line, the trains ran from Shepherds Bush to Wood Lane and back to Shepherds Bush without the driver changing ends.   This meant that the whole train was turned round and the cab which had faced west, now faced east.  At the other end of the line (Bank) the driver changed ends to reverse the train in the usual way.  For more information about the loop at Wood Lane, click here.

As a result of this new method of operation, a complication was introduced into depot operations.  Trains could arrive back in the depot (at Wood Lane) either way round.  If a car was removed from a train for maintenance or repair, it had to be replaced by another car which was facing the same way round, otherwise the hoses wouldn’t couple.  Sometimes, a suitable car wasn’t available and, rather than cancel a train, short trains occasionally appeared in service.  The depot management soon found this to be a recurring problem and they built a turntable in the yard late in 1908 to turn individual cars when necessary. 

To the Top of this Page

Control Trailers

1908 was a busy year on the Central London because, apart from the Wood Lane extension, they started that year to introduce short trains into service during off-peak periods.  This was partly a response to the car shortage mentioned above and partly a mileage and energy saving exercise.  As the trains were usually of 6-cars in a fixed DM - T - T - T - T - DM formation, running shorter trains required either a train to be taken into the depot for intermediate trailer cars removed or for additional driving positions to be provided on the middle pair of trailers. The latter was the sensible solution, especially since it saved a lot of complicated depot manoeuvres and there was a bonus in that a 3-car set with one motor car had the same performance as a 6-car set with two.

So, additional driving cabs were provided on the trailers in the middle of a number of trains (but not all) and the train formation became DM - T - CT + CT - T - DM on these trains.  The + indicates the normal uncoupling point.  The CT was the designation for a trailer car which had been provided with a driving cab.  It was called a Control Trailer.  Some railways, which had similar vehicles (including the Metropolitan Railway), referred to these cars as Driving Trailers.   Curiously, in view of the arrangements of the motor car driving positions, the Central London control trailers all had the driving position on the left hand side. Trains could therefore have left hand or right hand driving, depending on whether you were driving from the DM car end or the CT end and which way round the train was facing.   No, Tubeprune, hasn’t tried to work out the possible variations.   However, all the variations were limited by the need to ensure that each pair of 3-car units which had to couple together in the afternoon for the evening peak were facing in the same direction.  This was ensured by timetabling them to run one following the other after uncoupling through the off-peak service and then re-coupling the same pair for the peak service.  Trains which didn't uncouple were stabled during the off-peak period.

To the Top of this Page

Pairing and Lettering

When multiple unit traction was introduced on the Central London, it adopted a means of identifying trains by "set" numbers.   The "set" was allocated a number corresponding to the road where the train was stabled in the depot.  An identification letter was added to each motor car in the set, "A" or "B", apparently as a way of denoting that the original four powered axles of the locomotives had now been split into two sets of two at either end of the train (see photo).  The two motor cars of train No 7 (normally stabled on 7 road at Wood Lane) were therefore lettered 7A and 7B.  When control trailers were introduced, they were given the same letter as the motor car with which they would run when operating as a 3-car unit

In a complete train, the lettering would look like this, reading from west to east: ‘B’ DM - T -  ‘B’ CT + ‘A’ CT - T - ‘A’ DM, see Fig. 3 below.  Half the train was the “A” half and the other was the “B” half.   Put another way, there was an “A” unit and a “B” unit.   Either unit could run as a 3-car train.  When being driven from the Control Trailer end, power was provided by the motor car at the rear.

Gate Stock train formations.gif (8888 bytes)Typical train formations of gate stock used on tube lines. 

Click on the image for the full size view and description.

The Central London formation was unique in the way a 3-car unit had the same letters at each end.  The orientation was also unique in that "A" units normally faced east and "B" units faced west.  The Bakerloo and Hampstead formations were similar to the Piccadilly but orientated north to south.  The same 6-car formation was used for Standard Tube Stock until the introduction of 7-car trains in the late 1920s.  

To the Top of this Page

LER Tube Lines

The other tube lines, the Bakerloo, Piccadilly and Hampstead Lines (the Hampstead was the Charing Cross, Euston and Hampstead railway, to give it its full title, eventually became part of the Northern Line) which were collectively known as the London Electric Railway (LER), all had basically similar arrangements to the Central London.  They used gate stock, which was similar in design to the Central London’s, except bodies were of steel instead of wood and that control trailers were provided from the outset, not converted from trailer cars.

Services over the three LER lines were introduced between 1905 and 1907.  This was boom time for tube railways in London.  LER trains were equipped with the brake pipe along the north side (taking the east-west running Piccadilly Line as the base) and the main reservoir pipe along the south side.   Control and lighting cables were hung at roof level.  Automatic mechanical couplers, designed by two Americans named Stearn and Ward in Chicago in 1902 for the North Western Elevated Railway, were fitted between car with central, sprung buffers. Ward came to London with Yerkes to assist with the electrification of the District.  The coupler appeared on all District and tube cars from 1905.  Stearn and Ward were paid 1 for each one, as it was their patent.  It was only superseded by the Wedgelock automatic coupler adopted on the 1936 Tube Stock onwards.

The Ward couplers a mechanical coupler which works automatically when two cars are pushed together.  A spring-loaded cam locks the couplers together.  To uncouple, a "coupling pole" with a hook at the end is used to pull back the cam and allow the couplers to be separated as one car is pulled away from the other.  Hoses and jumpers had to be connected and disconnected separately.  There are still a few in use on some LUL engineers vehicles.

On the Bakerloo and Hampstead Lines, the train orientation was originally arranged so that the brake pipe was on the east side and the main reservoir pipe on the west side.   One can imagine that the delivery arrangements had to be very precise, since each car was delivered individually to the depot (by road in the case of the Bakerloo and Hampstead Lines) and they each had to be the right way round to allow them to couple to other cars.

The LER didn’t use letter identification at first but soon adopted it when cars began to be swapped between lines from 1912.   Their system was slightly different from the Central London in that they lettered all west facing cars “A” (i.e. cars with the driving cabs facing west, including control trailers) and east facing cars “B”.  Some lines had little enamelled plates fixed to the cab ends showing the orientation letter.  A 6-car train of LER Gate Stock was formed like this: ‘A’ DM - T - ‘B’ CT + ‘A’ CT - T - ‘B’ DM (see diagram of Piccadilly Line train in 1908).   This arrangement did not allow units to be referred to a “A” units or “B” units so they became know as “West end units” or “East end units” (or north or south) as appropriate.

The LER lettering was in exact opposition to the system adopted by the Central London, in that the "A" cars normally faced east on the CLR instead of west as on the LER.  This peculiarity persisted until the old Central London stock was replaced in 1938-9.

Hamp Gate Stock at GG.jpg (49698 bytes)A pair of LER gate stock cars coupled.

Click on the image for the full size view and description.

To the Top of this Page

Main and Train Lines

These are American terms.  The London Underground was originally electrified using money obtained though American sources and was designed by American engineers.  Many of the LER cars were built in the US.  The electric traction technology was largely imported from the US.  Much of the language used was American and much of it has stuck to this day. That’s why the carriages are still called cars and bogies are still called trucks on London Underground.  Such well-known phrases as “southbound” and “eastbound” are also relics of the American heritage. Even drivers were still called motormen up to the introduction of one person operation in the early 1980s.

As much of the technology was American in origin, many train equipment names were American too.  The brake pipe was called the train line and the main reservoir pipe was called the main line.  These names have stuck to this day.  Train lines, which are the name given to electrical cables on rolling stock in the UK, are called control cables and the multi-core control cables between cars are called “jumpers”.

To the Top of this Page

Developments In the 1920s

In the 1920s, there were significant technical developments on the Underground. Throughout the decade, many new tube trains (the Standard Stock) and some older ones (the Central London stock) were provided with powered sliding doors instead of entrance gates and, towards the end of the period, electro-pneumatic brakes were fitted to all new and many recently built cars.  As a result, there were a number of changes to the coupling arrangements.

The most significant changes for coupling were that all car ends were enclosed and that more jumpers were needed between cars.   Main line and train line hoses remained unchanged in terms of location but jumpers were mounted on the body ends just above the headstock, instead of at roof level.  There were three jumpers, reading from left to right when looking at an “A” end cab, Auxiliary, Brake and Control (A, B and C).   Auxiliary (originally just lighting but now including door controls) and control jumpers were provided on all cars but the brake jumper was added from 1929 on those cars equipped with electro-pneumatic brakes .   The jumpers were all 10-core cables with similar heads and it became important for shunters to ensure that the right connections were made when coupling - they had to know their “ABC”.

26 TS DM at GG.jpg (55947 bytes)1926  Standard Stock "B" end driving motor (DM) car at Golders Green

Click on the image for the full size view and description of coupling.

Another jumper was installed at roof level when emergency lighting was introduced (about 1914) and this was retained at roof level on the new Standard Stock trains of the 1920s.  The problem was that, during uncoupling, it was sometimes forgotten and units were parted with this jumper still connected.   Damage to them remained a problem, although many cars were modified later with the offending jumpers moved down to waist level. 

Another feature of Underground cars throughout the pre-1938 era was “side chains”.  They were attached on either side of the headstock of all vehicles to act as emergency coupling links if the Ward coupler broke.  They were supposed to be coupled at all times but often, they were left loose and sometimes caused problems when they touched the positive current rail.  They were not provided on cars built after 1936.

29 TS A DM front.jpg (35426 bytes)Fig 6. Front of 1929  Standard Stock "A" end driving motor

Click on the image for the full size view and description of coupling.


To the Top of this Page


In 1932, a car renumbering scheme was introduced across the Underground which included a new car lettering system.  The lettering scheme was based on the use of the letters A, B, C and D to identify the four axles under a car.  Cars with the cab nearest the “A” axle became “A” cars, those with the cab nearest the “D” axle became “D” cars.  Of course, axle lettering took account of the existing car orientation so that existing “A” cars remained “A” cars while the “B” cars all became “D” cars.  The “B” identification plates on the cab fronts were all changed to “D” plates.  At the same time, the north side of the train became the No 1 side and the south side the No 2 side. Diagram and more details here.  The system remains in use to this day.

All cars were renumbered.  “A” cars were given even numbers and “D” cars odd numbers.  This continued the practice for “A” and “B” cars first started on the Hampstead Line about 1914 and gradually adopted on other lines after World War I.  Cars were renumbered in four-figure groups; 3xxx for Driving Motor cars, 5xxx for Control Trailer cars and 7xxx for Trailer cars.  The remaining 4-figure numbers were reserved for sub surface line cars.

To the Top of this Page

Loops and Triangles

The same problems experienced on the Central London following the introduction of the loop at Wood Lane in 1908 were introduced on the Hampstead Line in 1914 with the opening of a loop at the new Charing Cross terminus.   All the Hampstead stock now became turned on every trip.  The Central London train make up problems were repeated on the Hampstead and the same solution, the building of a turntable, was repeated at Golders Green Depot.

The Bakerloo was a north to south running line which was gradually extended from its northern terminus at Baker Street in 1906 to a country terminus at Watford in 1917.  In the process it became connected to the District via connections at Willesden Junction and Kensington (Olympia).  When a central rolling stock overhaul works was opened at Acton Town in 1922, Bakerloo cars could be transferred to and from the works via this link but they became turned in the process so that the “A” cars faced east instead of west at Acton.  To ensure all cars passed through the works with the correct orientation, the whole Bakerloo fleet was turned (allegedly over one weekend) using the triangle of lines available at Croxley Depot, near Watford.  This meant that, from then on, Bakerloo cars ran “wrong way round” on their own line (“A” cars face south and “D” cars face north), a situation which has continued to this day, even though overhauls at Acton have ceased.

To the Top of this Page

Uncoupling in Service

All Underground lines adopted a policy of changing train lengths during the traffic day.  The procedure usually took place whilst the trains were in service and normally in a station platform.  Gaps in the current rails were often provided at the coupling point to allow staff to work safely between cars whilst coupling jumpers, side chains and hoses.  As mentioned above in the description of the Central London system, careful timetabling was needed to ensure that the correct halves of trains uncoupled after the morning peak came together in the right place for coupling prior to the evening peak.  This became quite complicated.   Some lines introduced short trains in the early morning, lengthened them for the morning peak, reduced them during the midday peak, lengthened them for the evening peak and reduced them afterwards only to lengthen them a third time for the late evening “theatre” peak. Such complex operations required tight train management discipline and rapid recovery techniques, which were developed to a fine art in the period up to the Second World War.  Since that time however, the cost of labour, the reductions in social discipline, changing travel patterns and the difficulties in managing regular disruptions to services, has meant that uncoupling is no longer considered a viable option.

To the Top of this Page

Four plus Three

During the late 1920s, increases in traffic forced some lines to increase train lengths, the Northern and Piccadilly Lines in particular.  The original, standard 6-car train formation was increased to 7-cars by the insertion of an additional DM car.  Trains were now formed: ‘A’ DM - T - T - ‘D’ DM + ‘A’ CT - T - ‘D’ DM or ‘D’ DM - T - T - ‘A’ DM + ‘D’ CT - T - ‘A’ DM, depending on the line of use and orientation.  In addition, it became the practice after World War Two, not to allow trains to operate in service with less than two compressors.  This policy required that all Standard Stock trains had to run with two motor cars, since these were the only cars with compressors.  Uncoupling was therefore re-arranged so that the 4-car portion always remained in service and the 3-car was uncoupled and sent to a depot or sidings.

7 Car Std Stock train formation.gif (5076 bytes)Typical train formation for 7-car Standard Stock. 

Click on the image for the full size view and description.

Trains were still operated in 6-car and shorter off-peak formations and, when the Central Line extensions were opened after the Second World War, some 8-car Standard Stock trains were operated.

To the Top of this Page

Sub Surface Trains

Compared with the tube lines, the coupling of sub surface line trains (District and Metropolitan Railways) was a simple business to begin with.  On both lines, which first introduced electric services on some of their main routes from 1905, automatic mechanical couplers with centre buffers were provided.   Side buffer pads were provided on the Metropolitan cars so that they could be pushed by a steam locomotive if necessary in emergency. 

The District used the Ward coupler while the Metropolitan, after an initial flirt with link and pin couplings, adopted the Buckeye coupler. Hoses were coupled below the coupler (there was plenty of room under these cars) and jumpers were duplicated on either side of the headstock, so it didn’t matter which way round cars were coupled.

To the Top of this Page

Metropolitan Complications

The Metropolitan’s basic electric train formation was 6-cars comprising: DM - T - DT + DT - T - DM.  The DT was a driving trailer, the same sort of vehicle as the Control Trailer used on the tube lines. However, the Metropolitan soon introduced complications to its electric stock fleet by using two different equipment suppliers - Westinghouse and BTH (British Thompson-Houston) - which used different control jumpers.  In order that trailer cars could run with motor cars fitted with either equipment, many were fitted with one firm’s jumpers on one side and the other’s on the other side.  Some of the earliest electric vehicles had Westinghouse jumpers on both sides.   From 1925, cars got turned on the triangle between Watford, Rickmansworth and Moor Park and things became so complicated that special letter codes were added to car ends to help shunters when they were making up trains.    Tubeprune doesn't know what these code were.  Does anyone out there know?  Please contact Tubeprune if you can help.

A further complication was introduced on the Metropolitan when they introduced, from 1908, mixed trains of saloon-bodied motor cars and compartment stock trailers and sets of steam stock converted to electric operation.   Over the years up to the take-over of the Metropolitan by London Transport in 1934, an increasing variety of train types and coupling variations appeared, too numerous to mention here. 

To the Top of this Page

District Formations

The District started electric operations with a 7-car standard train comprising: DM - T - T - MM - T - T - DM.  The DM was a driving motor (some of which were equipped with luggage compartments) with a driver's cab at one end, while the MM was a driving motor (Middle Motor) with a driving cab at both ends - otherwise referred to as a double-ended motor car. The double-ended cars were used on the branch line services to South Acton, Hounslow, South Harrow and Uxbridge and, sometimes, down to Putney Bridge.

From 1908, some trains were arranged to split in service at Acton Town so that four cars went on to Ealing and three to South Harrow or Uxbridge. To allow the 3-car portion to operate as a train, some trailers were converted to Control Trailers. From 1910 there was a gradual increase in the District fleet which allowed the standard train formation to be increased to 8-cars. Splitting trains were then divided into 3-car and 5-car sets.

To the Top of this Page

The District in the 1920s

By the end of the first world war, the original District Line fleet had become very unreliable and some cars had been stored out of service.  Much of the trouble was due to the poor condition of the wooden bodies.   There was then a purchase of a fleet of new cars (the F Stock) which were designed to be formed into the standard 3 + 5 formation.  When the fleet arrived, they were found to have a number of design problems including, most importantly, excessive power requirements.  In a subsequent re-evaluation of the District Line fleet, it was decided to divide the rolling stock into three groups.  These were “local” wooden stock, old cars patched up to work lightly used branches, F Stock - the new cars re-organised to reduce the power and increase operational flexibility and main line “steel” stock, newer steel cars plus older wooden motor cars rebuilt as trailers to allow further reliable service.

C Stock DR at Oly.jpg (34388 bytes)Fig 8. District Railway C Stock motor car at Olympia

Click on the image for the full size view and description of coupling.


There was a lot of car rebuilding and conversion work to do and it was the end of the 1920s before the reorganisation was complete.  By then, all three groups were operationally incompatible.  The Local Stock rarely worked as more than 2-car (DM-CT) sets, the F Stock ran in the 3+5-car formation while the Steel Stock was reorganised into a 5+3 formation but eventually, this was developed into a 4+2+2 formation.

The 4+2+2 arrangement looked like this: ‘A’ DM - T - T - ‘D’ DM + T - ‘D’ DM + T - ‘D’ DM.  It was developed in the late 1920s and survived until the early 1970s.  It suffered from a significant limitation in that there was no driving position at the west end of the 2-car units so all uncoupling was limited in method and location as described later below.

District Train formations.gif (9706 bytes)Typical train formations of District Line stock. 

Click on the image for the full size view and description.

The Steel Stock had its coupling arrangements altered during the 1920s reorganisation so that the air hoses were on either side of the end doorways instead of under the coupler as they had been on older cars.  At the west end of cars the hoses were close to the doorway, whilst at the east end they were mounted on the outside corners of the body.  Electrical jumpers were hung below the headstock as on earlier cars.  Busline jumpers were on one side and control jumpers on the other.  Eventually, the wooden cars all disappeared and were replaced by Q Stock - a mixture of old and new steel bodies built at various times between 1923 and 1938. These trains survived until the early 1970s.

G Stock car end.jpg (40759 bytes)Trailing end of District Railway G Stock motor car

Click on the image for the full size view and description of coupling.


To the Top of this Page

The F Stock

It is worth mentioning here that the District's F Stock was not compatible with the main District fleet for two reasons.  First the motor cars were more powerful than other District cars and would have led to uneven acceleration in a mixed train.  Secondly, the F Stock had non-automatic acceleration, a feature which existed on all District cars up to the F Stock's delivery but which was superseded by automatic acceleration on the next batch of cars, the G Stock of 1923.  It was not possible to couple the two types, for obvious reasons.

F Stock FWI.jpg (38964 bytes)Photo of F Stock

Click on the image for the full size view and description

It was possible to couple the F Stock physically to other cars.  Photos of the first cars delivered show them coupled to older cars for test purposes but, interestingly, the new cars are only trailers and the older cars are providing the power.

The F Stock went to the Metropolitan after the second world war but was retained in a block fleet - always incompatible with other stocks - and was scrapped in the early 1960s, when it was replaced by the A Stock. 

To the Top of this Page

District Uncoupling

Up to the late 1920s, District trains tended to be uncoupled in service in 3+5 car split formations. However, from 1925, there was a gradual shift to a 4+2+2 arrangement was developed and by the Second World War, this was the usual set up.   Trains were normally run as 4- or 6-cars in the off peak and 8 cars in the peak.   The 2-car portions were always located at the east end of the train, which required a special uncoupling procedure.

Taking Ealing Broadway as an example, the 8-car train to be uncoupled would arrive and detrain passengers.  The two cars would be uncoupled from the east end and driven to the storage sidings at the east end of the station.  The remaining six cars would then depart on their next timetable trip.   To recouple, the 6-car would arrive in the terminus and detrain passengers.   The whole 6-car train would then be driven into the sidings to meet the 2-car unit parked there, where it was recoupled and then driven back to the terminus to pick up passengers for its next working.

To the Top of this Page

Unit Stock

In 1936 London Underground introduced a new type of tube train which was a complete departure from anything which had gone before. Units were formed into blocks, with cars coupled internally by bar couplers which could only be disconnected in a workshop environment.  Units were coupled together by fully automatic “Wedgelock” couplers, operated by a push button in the driver’s cab.  All mechanical, electrical and pneumatic connections were made through this coupler. The new stock was called 1936 Tube Stock and was the prototype for over 1100 cars of 1938 Tube Stock.

To differentiate between the new coupling system and the old, the new trains were referred to as "unit stock" while older trains became known as "car stock", the latter because individual cars could be uncoupled from a train at any time without the need to get the train over a pit in a depot.

A surface line version of the new tube stock appeared at the same time on the Metropolitan and District Lines as the O and P Stocks.   More appeared on the District in 1947-59 as the R Stock and the tube lines got an aluminium bodied version of the 1938 Tube Stock known as 1950 and 1962 Tube Stocks, which replaced the Standard Stock.   All LU stocks have been designed to the same basic coupling principles ever since.  

One new feature of the 1938 Tube Stock was the non driving motor car.  This was a motor car with a full set of traction equipment but no cab.  It became known as an NDM and appeared in the normal train formation thus: DM-T-DM + DM-T-NDM-DM.  Cars would be classified “A” or “D” as appropriate.  The “A” and “D” orientation prevailed as it had done on the earlier stocks so “A” ends could still only couple to “D” ends. The story of the variety of train formations and coupling strategies used on the 1938 Tube Stock is complicated and is not covered here.   "The 1938 Tube Stock" by Piers Connor (Capital Transport, 1989) is a good source of additional information.  However, one feature of the 1938 TS which is important to this article is the UNDM, which is covered next.

38 Stock train formations.gif (8405 bytes)This diagram shows two versions of the 1938 Tube Stock train formation as used between 1938 and 1988.  The lower version with the UNDM was seen regularly from the early 1950s.  The 1972 Tube Stock currently on the Bakerloo has the same formation with the "A cars facing south and the "D" cars facing north.

Click on the image for the full size view and description.

To the Top of this Page


A new type of car was introduced to the Underground in 1949 when a re-organisation of the 1938 Tube stock allowed an opportunity to develop a middle uncoupling point without a driving cab.  The Underground was suffering from capacity problems at this time and any way of increasing passenger space was important.  Getting way of a little used cab was a way of doing this. 

The new type of car was called an Uncoupling Non Driving Motor car or UNDM.   It quickly became known as an “undum”.   The UNDM was a non driving motor car with an automatic coupler at the outer end and a shunting control panel instead of a full width, enclosed driver’s cab.  The shunting control panel allowed driving at restricted speed for coupling purposes only.  The idea was to allow units to be uncoupled and driven into a depot but, at the same time, to provide more passenger space where the driver’s cab would have been. The usual train formation was: DM-T-UNDM + DM-T-NDM-DM.

38 TS UNDM with DM.jpg (36790 bytes)1949 Tube Stock UNDM coupled to DM car plus another photo of the interior of an UNDM car

Click on the image for the full size view and description.

To the Top of this Page


Uncoupling remained a standard operating feature of the Underground until it was eliminated on the tube lines around 1960 and on the District in 1971. The Metropolitan hung on until 1981. Now, all trains on a line stay the same length all day.  However all London Underground trains have retained the facility to be uncoupled in or near the middle point of the train.  These days this is solely for maintenance purposes plus, it does help if there is an emergency like re-railing a train or getting a body from under the train after a suicide.

It is worth mentioning here how uncoupling was carried out at on the tube lines in the 1940s and 50s.  The principle established in the early 1930s that trains should not run with less than two compressors led to a train formation for uncoupling trains of 4- + 3-cars with the 3-car unit being stabled while the 4-car set ran in off peak service.  On the Piccadilly Line, for example, the 3-car unit was normally at the east end of the train.  Uncoupling was always carried out so that the unit could run into a depot or siding and could only take place at Northfields (with WB trains uncoupling and EB trains coupling).

To the Top of this Page

8-car trains with 4 cabs

As we have seen, a train only requires a cab at each end for normal operation, giving a total of two cabs for any length of train.   If the train is to split during its duty, each portion to be separated will require a cab at each end.  If this is two portions, there would have to be four cabs on a full length train.  LU trains originally designed for uncoupling in passenger service had full width drivers cabs at each end of every unit - four on a full length train. The Metropolitan’s A Stock is a good example. Being designed for regular uncoupling in service, it has two 4-car units in an 8-car train.  The 1967 Tube Stock (Victoria Line) was also designed this way, although it was never intended to uncouple the trains in service.  It did however allow 4-car trains of the stock to be tested in service on the Central Line’s Woodford-Hainault service, where the original ATO trials were done.  The 67TS 4 + 4 formation gave a lot of flexibility when forming trains for service in the depot.  As long as a train had an operative cab at each outer end, it didn’t matter which end of the train a unit was positioned.  A defective unit could quickly be replaced by a good one since all units were the same, with operational cabs at both ends.

To the Top of this Page

Single ended units and double ended units

After a few years of operation, spares became scarce on the Victoria Line, (particularly ATO equipment) and some driving cars had equipment “borrowed” (cab seats, destination blinds, automatic driving boxes, etc.) which rendered them un-useable at the ends of service trains.  Eventually, because of the “robbed” cab at one end, many of the 1967 TS units became permanently restricted to one end of a train.  They then became known as south end units or north end units, depending which end had an operative cab.

To allow some flexibility in forming trains for service, a small block of ten units of 67TS were reserved as “double-ended units” which could replace a “single-ended unit” at either end of a train.   If you look at the middle coupling point of a 67TS train on the Victoria Line, you will see that, where the two middle cab fronts come together, there is no red paint on the end, unless it is a double-ended unit.  Apart from their red paint, the double-ended unit cab ends can be identified by the ugly blisters fitted recently for mounting the inter-car safety barriers, in case they are coupled in the middle of a train.

The A Stock is configured as 4 + 4 cars in a similar way to the 67TS. It has to be since it needs a number of 4-car sets to operate on the East London Line.  Identification of the cabs of double-ended units is similar to the 67TS.

To the Top of this Page

1972 Tube Stock (Photo)

The 1972 Tube Stock operating on the Bakerloo (and which also worked on the Northern until 3 years ago) was derived from the 67TS but it was designed to allow trains to be formed from a 3-car unit and a 4-car unit because that’s all the Northern Line platforms could take.  The 4-car unit was given a cab at each end like the 67TS but the 3-car only had a full cab at the outer end. The inner end had an UNDM car with a shunting control position fitted behind a door on the left hand end wall of the car like the 1938 Tube Stock. The shunting controls was taken from scrapped 38TS UNDMs. 

Unlike the Victoria Line, the Northern Line has a loop at Kennington.  This replaced the loop at Charing Cross when the line was extended to Morden in 1926.  Quite why another loop was built after all the problems of the two at Charing Cross and Wood Lane escapes me.  I digress.  The loop means that trains get turned.  For many years, the line’s operations were plagued by the need to ensure that units faced the right way round otherwise they couldn’t be coupled.  In order to be able to couple trains both units have to face the right way round.  The 72TS on the Northern was designed so that the couplers on the UNDM could couple to either end of the 4-car unit.  As long as the UNDM was coupled to a DM car, a 7-car train could be formed.

To the Top of this Page

Universal Couplers

The coupling arrangements for the 72TS were derived from a development first introduced on the A Stock built in 1960-63.  This stock was designed to operate over the Metropolitan’s routes to Uxbridge, Watford and Amersham.  Trains on these routes are regularly operated over the triangle of lines between Watford, Moor Park and Rickmansworth and therefore get turned.  In order to allow units to be coupled either way round, the electrical connections were duplicated on either side of the coupler.  This removed the need to ensure that only “A” ends coupled to “D” ends and produced what the Underground called the “universal” coupler. It allowed units to be coupled “A” end to “A” end as well as “A” end to “D” end.

To the Top of this Page

The C Stock (Photo)

This stock was introduced to replace the old COP Stock working the Hammersmith and Circle Lines.  One of the features of Circle Line operation is that wheel wear is uneven.  Although trains ran round and round the Circle all day, they never got turned in respect of the way the wheels wore.   One side (the outside) of the wheelset always wore more than the other, so cars had to be lifted every six months to swap the wheels from one side to the other to balance the wear.  In order to eliminate this expensive and time consuming process, universal couplers were provided on the C Stock so that it could operate with units either way round. To ensure that they were turned regularly, two passenger trips per day are arranged for Circle trains to run from Tower Hill to Whitechapel, where they reverse and then go to Hammersmith.

Another unusual feature of the C Stock is that it is designed as a 2+2+2 train formed DM-T + DM-T + T-DM.  This was because it was intended that it would be the forerunner of a new fleet for the sub surface lines which would run as 4-, 6- or 8-car trains on the District or Metropolitan Lines.  It was not to be.  In any event, the 2-car units were provided with automatic couplers at each end. 

Originally, the C Stock did not have driving facilities at the trailing end either but they were able to be uncoupled in a depot yard. The trailers were actually “uncoupling trailers” but they were never officially described as such, probably because you couldn’t drive from the trailer.   Without driving facilities at the trailing end, coupling, which took place only in depots, was a cautious affair.  A man had to ride on the trailing end directing the driver who was driving in reverse from the cab at the other end of the unit.  This situation was rectified with the refurbishment of the fleet in the early 1990s, when all trailers were fitted with shunting controls.

To the Top of this Page

1973 Tube Stock (photo) and D Stock (photo)

With a logical progression of the design process, the operationally developed "single- and double-ended unit" Victoria Line train formation was applied to the 1973 Tube Stock (73TS) built for the Piccadilly Line. Trains were formed of single-ended 3-car units coupled back to back to form 6-car service trains. The 3-car single-ended units were provided with UNDMs equipped with automatic couplers and shunting control panels.  A batch of 20 double-ended 3-car sets were kept for flexibility and to allow a unit to work the (now closed) Aldwych branch.

The D Stock, built from 1978-81 for the District Line, was simply a surface line version of the 73TS.  The main fleet was made up of single-ended units and some double ended units were included to provide flexibility. 

73 & D TS train formations.gif (10071 bytes)Fig. 13 showing the 6-car formation of  both 1973 Tube Stock  (Piccadilly Line) and D Stock (District Line).  It is the same for both stocks. 
Click on the image for the full size view and description

The upper formation shows a double-ended unit in the place of an "A" end single-ended unit, while the lower diagram shows "A" single-ended unit coupled to a "D" single-ended unit.   Note that both ends of the double-ended unit have automatic couplers but the single-ended units have couplers on the UNDMs only.  The outer (cab) end has an emergency mechanical coupler.

On both 1972 and 1973 Tube Stocks and the D Stock, not all cab ends are equipped with automatic couplers.  Only the cab ends where train formations require the coupler, have it.  The outer ends of any unit with an UNDM do not need it and these positions are fitted with an emergency mechanical coupler.

To the Top of this Page 

The 1990s

Three sets of tube stocks have appeared during the 1990s - the 1992 Tube Stock for the Central Line, the 1995 Tube Stock for the Northern and the 1996 Tube Stock for the Jubilee. Looking at the later stocks first, both 95 and 96 stocks have the 3 + 3 formation with cabs at the outer ends and UNDMs in the middle. There are no double-ended units.  It is thought that better maintenance regimes will remove the need to have spare flexibility.  We'll see.

96 TS train formation.gif (5377 bytes)Fig. 14.  Diagram of 6-car formation of   1996 Tube Stock for Jubilee Line. 

Click on the image for the full size view and description

The 1995 Tube Stock on the Northern Line is similar except that the "universal" couplers allow "A" units to couple to other "A" units and likewise with "D" units. 

The only difference in their coupling arrangements is that the 95TS on the Northern gets turned on the Kennington loop and therefore universal couplers are used to allow “A” ends to couple to other “A” ends as well as “D” ends.  The 96TS doesn’t get turned and is “handed” so only “A” ends will couple to “D” ends.   Curiously, both stocks are equipped with automatic couplers at the cabs ends.   There seems to be little logic in this, since the ends cannot be coupled to anything for normal service use and, if emergency coupling is needed, electrical connections are notoriously unreliable under failure conditions and may end up transmitting the fault from the defective train to the assisting train.

The 92TS built for the Central Line is based on a 2+2+2+2+2-car formation.  The 2-car unit is self-contained with cabs provided at one end of some units, while the rest have no cabs. Each end of every unit has a fully automatic universal coupler, necessary because the stock gets turned regularly on the Hainault loop. As there are more units with cabs, it is possible to see trains formed with cabs at coupling points but this is rare.

To the Top of this Page

Automatic Couplers

Wedgelock Dwg.gif (51489 bytes)

Fig 15. Diagram of current design of LUL auto-coupler, showing main parts. 

The automatic coupler used by London Underground is basically the same design which first appeared in 1936. The mechanical portion consists of a cast steel coupler head with a tongue and throat which mate with the tongue and throat of the other coupler. Once the tongues are in the throats, they are locked into place with wedges driven by air pressure operated pistons.

When the two couplers come together, the air connections between the units are completed by matching holes lining up. Rubber washers prevent air escaping at the joint. Electrical connections are housed in two terminal blocks mounted on either side of the coupler head. The connections are made through spring loaded butt contacts provided for each circuit. A universal coupler can have up to 64 electrical connections across the coupler faces.

In order to protect the electrical contacts while the units are uncoupled, the contact blocks are covered automatically as the two coupler heads part. The covers are curved so that they swing over the face as the uncoupling takes place. They are referred to by LU as "Dutch ovens". There is a lug at the base of each contact block which pushes the "oven door" open as the two couplers meet. The doors are spring loaded so they will close automatically as the units part.

A Stock coupler.JPG (40760 bytes)Fig 16.  Universal automatic coupler on A Stock driving motor car.

Click on the image for the full size view and description of coupling.


To the Top of this Page

Coupler Controls

The original design of auto coupler had automatic controls. The driver wanting to uncouple two units would release the parking brakes of the unit he was to operate from, enter the cab of that unit at the coupling point, "open up" the controls to release the brakes and push the "Uncouple" button in the cab.

As the uncouple button was pushed, an electro-pneumatic "coupler valve" on each unit operated a "coupler engine", a pneumatically driven camshaft which drove the valves controlling the couplers. Each coupling engine went through a process of releasing the wedges securing the coupler tongues, isolating the main and train line connections from the coupler face and resetting the tripcock, which was isolated while the units were coupled. These operations were achieved through a series of five cam-operated valves on the camshaft.

As the coupler wedges released the tongues, the spring buffer provided between the units pushed the operating unit away from the static one and the two units were automatically uncoupled. Of course, the whole system depended on the two e-p valves operating at the same moment and on both coupling engines turning together to release the wedges. The difficulty of achieving this on couplers which had not been used for a long period often led to failures. If this occurred on a service train due to uncouple after the morning peak, the train would remain coupled until it could be changed over (if enough spares were available) or until the evening peak.

Coupling was a simpler operation. One unit was driven up to the other so that the two couplers engaged and a "couple" button was pressed to turn the coupling engines. The coupling engines operated the valves to complete the main and train line connections, lock the wedges into place and isolate the tripcocks.

Coupler failures, usually during uncoupling led to some bizarre procedures to get units to uncouple. The first step usually involved a man in each car, manually lifting the uncouple valves on the count of "one, two, three". If this didn’t work, the air was manually bled off the units and a pin and long lever was used to manually push back the wedges through a slot in the top of the coupler head. If this still didn’t work, the air connections to the couple and uncouple valves were reversed and air re-applied to the train.

To the Top of this Page

Manual Control

Frustration with the push button control led to a new design for the 1967 Tube Stock built for the Victoria Line. The coupling process was redesigned to allow manual operation on each unit sequentially instead of requiring parallel and instantaneous operation. The coupling engine was moved from its former remote position under the offside cab floor to a cabinet inside the driver’s cab.

The camshaft is mounted vertically and has three operating positions, "Coupled", "Wedges Back" and "Uncoupled". When a train is to be uncoupled, the control is moved to "Wedges Back" in both cabs, the units are driven apart and then the control moved to "Uncoupled". Although it takes longer than the push button control, this is a more reliable system and time is not a crucial as before since uncoupling no longer takes place in service.


To the Top of this Page  or to the Home Page                                                                 This page upgraded 24 August 2002

Copyright Tubeprune 2001, 2002.   If you have comments or if you would like to use any part of this site for publishing or commercial reasons, please e-mail me