183

Technical Meeting of the Institution

hefd at The Institution of Electrical Engineers

Wednesday, March 13th, 1968

The President (Mr. H. W. HADAWAY) in the Chair.

The Minutes of the Technical Meeting held on February 8th, 1968, were read and approved. The President introduced and welcomed to the meeting Messrs. R. A. Bagnall (Associate) ; W. A. Mellin (Technician Member) ; J. C. Fowler, C. L. Ackerman, R. B. Parker, R. P. Little and F. W. Harris (Students) who were present for the first time since their election to membership. The President then invited Mr. J. V. Goldsbrough to read his paper entitled" Overlaps (British Practice)." Overlaps (British Practice) by J. V. Goldsbrough• (Member)

1. INTRODUCTION Institution are often as valuable as the paper itself, aud it is hoped that many The primary object of this paper is to contributions will be forthcoming on this provide a brief summary of the subject for the use of our Student and Technician new aspect of the subject so that they can be duly recorded in the Institution Members. Whilst it is referred to in the Proceedings for future reference. As will various booklets published by the In­ be appreciated, once the basic Block stitution, the pros and cons have not Regulation requirements have been satis­ been collected together under one heading fied, the subject of overlaps is more a for reference purposes, and the author matter of opinion rather than regulation. feels that this should be a useful purpose It is these opinions which the author for the paper. wishes to be recorded, for they will need In addition, for those who may consider to be kept constantly in mind when we the subject to be rather well-worn, it is try to make a machine give a decision apparent that in the not too-distant future which will be based only on logical when train describer, automatic or com­ assessment of available information. puter route-setting comes into use, we shall have to consider other methods for the operation of functions now per­ 2. THE ABSOLUTE BLOCK formed by signalmen ; and selection of REGULATIONS overlaps by automatic means may well The Absolute Block Regulations issued, prove necessary. The discussions which and from time to time amended, by the follow the presentation of a paper to the various former Railway Companies, and

*AEI-General Signal Ltd, 184 OVERLAPS (BRITISH PRACTICE) later by the Regions of British Railways expected to be able to run to schedule and the British Railways Board, all and to stop his train within a few yards ..::ontain the following clause with identical of a fixed point despite ignorance of these intention, if not in the exact words :- peculiarities, quite apart from other " Except where instructions are issued variables related to weather, train loading to the contrary, the line must not be and braking characteristics. This is not considered clear, nor must a train be an easy task, and it is right that our allowed to approach from the box in rear signalling systems should make allowance in accordance with Regulation 1, unless for it. the line, or at a junction the line for Some investigation was carried out by which the facing points are set, is clear the Medical Research Council in 1962/63 for at least ¼ mile ahead of the home on the subject of passing signals at danger, signal, and all the necessary points within which can presumably also be related to this distance have been placed in their the question of misjudgement. In contrast proper position for the safety of the to commercial market research, ,vhich approaching train." might be said to try and establish how Notice the phrase : " Except where many people react in the same way, the instructions are issued to the contrary." report wisely recognised that human This means that the Regulation is pro­ beings are not alike, even although they vided with means of modifying the ¼mile may have reactions in common. A requirement to suit the many and differing tendency for some individuals to be physical conditions encountered in careless, or perhaps "accident prone," practice, e.g. permissible speed over the was suggested; but this is well knmvn, section of line, gradients, weather con­ as also are the effects of fatigue after ditions, spacing of signal boxes and signal prolonged concentration. sighting distance etc. From a signalling point of view there Thus, in the acceptance of a train by is one very iinportant conclusion which a signalman, a safety margin is made can be related to approach-control of available beyond the home signal to signals. If a driver is confronted with a allow for misjudgements in braking. signal aspect which differs from that which If this margin is not available it is per­ he expects as a result of previous experi­ missible, where authorised, to accept a ence of working the route, then there is train under the Warning Arrangement of an increased possibility of his making an Regulation 5. In this case, the train error. In certain respects, approach must be brought to a stand, or nearly to control is inevitable ; how else (other than a stand, at the entrance signal to the by complex signal aspects or speed section before the signal is lowered to indications) can a train at present be allow it to proceed. brought under control before taking a route which demands low speed ? If 3. THE HUMAN ELEMENT approach control did not have its inherent If the overlap is provided to allow for disadvantage of anticipated clearance, we misjudgement, then we must consider might have been able to consider abolish­ the man involved, i.e. the driver. Quite ing the complications of the overlap at apart from his personal skill as a driver, junctions and intersections by introducing there is another factor often overlooked approach control of the outer signal when­ which relates to the demands of economics. ever the inner signal is at danger. Such In the early days of railways, a locomotive a practice would be considered as unduly and its crew were almost inseparable ; restrictive, but it could otherwise effect if the crew was off duty, so was the considerable economies. No doubt, auto­ engine. Thus, not only was the driver matic speed control will ultimately solve familar with the route, but he knew this problem, but its universal application the performance and capabilities of his is still far off. own engine to a fine degree. This was an ideal, but uneconomic situation. Now, he has to take any engine rostered 4. THE OVERLAP IN PRACTICE for the duty without previous knowledge 4.1. Length of Overlap of its individual peculiarities, other than Various formulae have been used to th()se common to -its class. He i.5 still calculate the appropriate length of an OVERLAPS {BRITISH J'RACTJCE) 185 overlap, except that in low-speed, power­ laden passenger train of 11 coaches, with signalled interlocking areas they are worn brake blocks, and two brake commonly chosen by reference to con­ cylinders out of action, requires a service venient track circuit block joints, with, braking distance on the level of 2 230 yd of course, due regard to minimum dis­ from 100 m.p.h. to a standstill. Even a tances. Although the formulae may differ loose-coupled unbraked freight train needs in detajl, they usually all include factors approximately 880 yd to stop from 30 of maximum approach speed, average m.p.h. Even if these distances were braking capacity and efficiency, gradients, reduced to terms of maximum braking signal sighting distances where difficult effort, their use as overlaps would severely conditions occur, and allowance for restrict service intervals on main lines, and weather conditions where significant. In could not be contemplated economically. certain interlocking areas where low speed An analysis of 13 reported over-runs limits are in force, specially selected at colour light signals in 3- and 4-aspect overlaps have actually been eliminated, areas over a period of 12 months showed and it has improved layout flexibility the average to have been 112 yd. Whilst and simplified signalling controls without all statistics and averages must be care­ any apparent loss of security. fully analysed, these figures were not With the overlap as a braking margin, selected for this purpose and exclude it is interesting to consider with respect cases where signals were actually missed to our main lines what the overlap would or mis-read, and can be said to refer only have to be if it were necessary to extend to misjudgements. On this basis alone, it to provide for maximum braking effort the choice of overlap distances by British as on London Transport. The so-called Railways for standard practice as given ' W ' curve issued by British Railways below, is shown to be both appropriate shows that a locomotive-hauled, heavily and safe.

YARDS OVERLAP

500

I 400n V I ~- s ~ ~ ~ ~-- ~ I ./ --- I----- 300 ~ . .J,f;9 i---- I s ~si.:..-- I/ '2,• -, - -- -· ,,. I ' v-- I ~ ~ - 200 - 'I V ~s ~ _,,."" ~ -- ~ ~~- - -- ~- ~100

I . ' ' 0 . l•2 l·t l·O 0·9 0·8 0·7 O·b O·S 0·4 0·3 0·2 O•I O·I 0·2 0·3 0·4 O·S O·b 0·7 0·8 0·9 1·0 1·1 1·2 + PER CENT ) ) PER CENT l I IN 100 RISING I IN 2b0 I IN 2b0 FALLING I IN 100 GRADIENT GRADIENT Fig. I. B.R.B. overlap adjustment for gradient. 186 OVERLAPS (BRITISH PRACTICE)

OVERLAP OVERLAP I·

Fig. 2. Overlaps without separate track circuits, 4.2. The British Railways Board Standard (b) The basic overlap lengths should be:- Signalling Principles 300 yd in 2/3-aspect territory 200 yd in 4-aspect territory Because of the varied operating require­ adjusted as necessary on account of ments encountered on the Regions of gradient and/or other factors. British Railways, the British Railways Fig. I shows how the length is to be Board, with the agreement of the Ministry adjusted in respect of gradients. of Transport, has prepared a number of (c) Where restricted approach arrange­ Standard Signalling Practices to achieve ments are in operation, an overlap of a measure of uniformity in treatment. 50 yd should be provided where Whilst these are generally available to practicable. There will be special Railway Officers, it will not be out of place cases where a reduction of the distance here to summarise the requirements for will be necessary. overlaps in colour-light signalled areas :- (d) The restricted approach should be subject to proximity and speed con­ (a) No separate overlap track circuit is to siderations. be provided unless it is required for (e) Trailing points in the overlap, and other purposes. opposing signals applying to the

SHARED OVERLAPS i

LATFORM

fig. 3. Typical shared overlaps. OVERLAPS (BRITISH PRACTICE) 187

overlap, must be locked until the of clearing." This means that when the route is cancelled or the movement overlap is set, detection is initially has come to a stand. included to ensure that the points have (f) Facing points in the overlap may be set and locked as required, but is not moved if they have time to operate continuously proved afterwards in the and an alternative overlap is avail­ outer signal. In this system, if a change able, of overlap is made, both overlaps are proved clear during transit of the points. Requirement (a) is derived from When detection of the new overlap is electrified areas where impedance bond complete, continuous detection is relaxed. track circuits necessitate keeping the These practices are all associated with number of track circuits to a minimum. the problem of avoiding interruption of Here, the block joint is located at overlap the outer signal controls when a change distance from the stop signal as shown in overlap takes place. There is, however, in fig 2. Although this gives delayed an approved solution which is simple in aspect replacement, the arrangement is circuit form and satisfies all desirable now in general use and is accepted as precautions. Where there is a choice of both safe and economical. overlaps, the principle is continuously to Shared overlaps are not referred to as prove all track circuits in all of them, and it is difficult to define precise requirements. to detect only those facing points which Generally, they are acceptable in low­ divert the overlap away from any one speed station areas, and fig 3 is given as a which may be obstructed. typical example. All such arrangements This is illustrated in circuit form in must be determined by individual con­ fig. 4 and means that if all point detection sideration and by operating requirements. in an overlap fails, the outer signal will not replace if all tracks in all overlaps 4.3. Point Detection. remain clear. In these circumstances it is pertinent No general rule has been established to question the likelihood of potential regarding continuous detection of overlap failures. Points rarely lose detection points in the controls of the outer signal. completely once they are static after Continuous detection, but not the setting having been set and locked. The majority and locking, of trailing points is often of detection failures occur immediately disregarded on the argument that a after actual transit when the new position trailing run-through is unlikely to derail fails to establish detection due to ob­ a train. 'Whilst this is generally acceptable, struction, etc. Fine detection can" bob " the author inclines to the view that if the with the passage of a train over the points, points nearest to the protecting signal of or on adjacent lines, but the fact that an overlap are trailing and are within detection was properly proved in the 50 yd, they should be continuously controlling signal before the train entered detected as a form of trap protection the route, and the points are held con­ against irregular movements occurring ditionally locked, means that there is where they are most likely to foul short little chance of danger with an over--run over-runs. in such circumstances. With facing points, the best practice The coincidence of both a static detec­ to adopt is that if they give access to a tion failure and a braking misjudgement choice of overlaps, then, if one overlap at a stop signal may seem remote, but it is obstructed, either by track occupation is essential that every possibility should or use in another route, facing detection always be given proper consideration to diverting the overlap away from the maintain the traditional safety of railways. obstruction should always be continuously proved. If the obstructed overlap forms part of another route, that route should 5. LONDON TRANSPORT also include continuous proving of the 5.I. Principles overlap diversion, so that mutual pro­ Probably, few transport authorities tection is given. in the world carry such a responsibility One practice, which has much to for the safety of the travelling public as commend it, is " detection at the time the London Transport Board. The main- 188 OVERLAPS (BRITISH PRACTICE)

r-OVERLAPS

B ■---➔------+--- ■- •t?I KY" .. II 13'-..._,, I

C ~3 ■---➔------+---•· TRACK LAYOUT

A B C TPR TPR TPR " " " / ' 13 PART OF SIGNAL I NWKR " CONTROL CIRCUIT / ....__ II II NWKR RWKR " '-

Fig. 4. Overlap controls in outer signal. tenance of a very close headway on lines points are held by the outer signal which partly in tunnels demands exacting atten­ is backlocked to maintain the previously tion to safety measures, and the overlap set overlap until released by delta track has received its due attention, not only occupation and time element. In push­ in precise methods of calculation, but also button route control systems, separate in ingenious and very effective arrange­ route buttons are used for the outer signal ments of overlap adjustment for measured to select the overlap corresponding to the approach speeds. inner route to be subsequently selected. The calculated overlap on London The delta track or rail circuit is described Transport is based on the braking distance in the paper on " Automatic Junction required by a train at its maximum service Working and Route Setting by Pro­ speed when the brake line is exhausted, gramme " (R. Dell, 1958). A principle either by the train being " tripped," or which is given much attention in all the maximum braking effort being exerted. signalling problems on London Transport An allowance of 30 % is added to the is that of endeavouring to avoid any distance to cover contingencies such as necessity of bringing a train to a stand, brake maladjustment or heavy loading as this can have serious repercussions on etc., and a further adjustment is made headway. It is achieved whenever possible, to allow for the approach and overlap by exercising control over the approach gradients. At stations where all trains speed of a train to the signal. The methods stop, the overlap at the starting signal employed for this are fully described in is calculated for 20 m.p.h. with the same comprehensive papers, previously read allowances. The swinging or switched to the Institution, on " Speed Control overlap is not employed ; the overlap Signalling" (W. Owen, 1949) ; "L.T. OVERLAPS (BRITISH PRACTICE) 189

Methods for the Control and Locking of length (420 ft), but as the weight of the Junctions." (H. W. Hadaway, 1961); train is assumed to be concentrated at and " Victoria Line Signalling Principles " its centre, the overlap plus 210 ft is (V. H. Smith, 1966). The student is considered to be the available effective recommended to study these examples braking distance. Hence :- Gradient of how trains can be kept on the move Compensation = Average % Gradient = and yet observe the very necessary restrictions imposed. T Cl!alrise or fall __ X lOO Overlap + 210 feet A standard table is available, relating 5.2. Calculation of the Overlap braking efficiency with maximum Calculation of the overlap is based on approach speed; and from this, the the maximum service speed which a train overlap distance can be directly deter­ can reach at the signal. This is obtained mined. from a speed/distance curve supplied to It may be of interest to note that it is the Chief Signal Engineer's office by the common practice in rapid transit systems Chief Mechanical Engineer. If the train in other parts of the world to use a full is " tripped " at this speed it must be block as an overlap. \\thilst this gives able to come to a stand within the overlap ; adequate security, it may mean shorter and the factors, apart from speed, which block sections to achieve minimum head­ will decide this are the braking efficiency way and does not give the precise results and the gradient of the approach and of achieved by London Transport. These the overlap. The braking efficiency of results, however, are by individual cal­ L.T. rolling stock on the level is main­ culation of each overlap and subsequently tained to achieve 12% in tunnels, but is by precise location of block joints. reduced to 10 ~~ in the open air to allow Whether the apparent additional costs for adverse weather conditions (e.g. wet of this procedure can show an advantage rails). If the gradient of the approach or over the simplicity of the double-block overlap is falling it will reduce the braking system is debatable. Even in double­ efficiency, and increase it if rising. Accord­ block, detailed and individual work is ingly, the length of the overlap is subject still necessary to arrive at correct signal to a gradient compensation, and this can spacing to achieve the desired headway be done directly by calculating the and in allowing for rolling-stock braking percentage net gradient, rise or fall, and and acceleration, gradient and layout adcling it to or subtracting it from the characteristics. braking efficiency percentage in the formula:- 6. THE SWINGING OVERLAP Length of overlap in feet = 6.r. General 4-34 V 2 There is no doubt that this aspect of n %± overlap net giadient -% the subject of overlaps is probably the where V = Maximum speed of train at most interesting and controversial. The signal (from speed/distance curve) in extremists, on the one hand, hold to the milejh. mechanical interlocking practice estab­ n = Braking efficiency, IO~,~in open air, lished many years ago where an outer 12% in tunnels. signal locked the overlap points to ensure The constant 4.34 in the formula is that there was no possibility of the points derived from 3.34 (the factor relating being moved if an over-run occurred at V rnile/h. and n which is a percentage of the stop signal protecting the points. 2 To achieve this object to the full, it is g (acceleration- due to gravity in ft/sec ) 30% 4.34. necessary to backlock the outer signal + = until the train can be proved at a stand The gradient compensation for the at the protecting inner signal. Before approach side of the signal is introduced the advent of electrical safeguards, how­ only when the approach gradient is falling, ever, such backlocking was difficult to and then only if it exceeds the falling achieve, and it appears possible, then, overlap gradient. The approach gradient that this form of locking served only as a is taken as an average over the train reminder to a signalman that he was 190 OVERLAPS (BRITISH PRACTICE)

obliged to hold the overlap for the evident that few layouts in complicated approaching train. Without backlocking, locking areas can accommodate sufficient it needs only the replacement of the outer clearance in rear of the stop signal for signal to release the points at the very long trains. This fundamentally sound time when the locking is progressively practice is favoured by London Transport becoming more important. whenever feasible, but it too often happens On the other hand, there is no doubt that the layout or sighting distance that to make the maximum use of a precludes it, and the signal has to be sited layout and to keep its essential flexibility, at the points. the holding of a selected overlap from the To maintain flexibility with a swinging time that the outer signal is cleared, until overlap it is essential that freedom shall the train comes to a stand at the inner, remain until the last possible moment, constitutes an appreciable hindrance. e.g. by provision of a 100 yd approach However, condition (f) of the Standard track to the stop signal. This is contrary Signalling Practices gives the required to the British Railways Board principles relaxation of this hindrance, so that in keeping the number of track circuits freedom to change the overlap points to a minimum, but if circumstances can can be maintained safely, until completion justify the provision, this is the only way of their movement can no longer be to achieve the maximum benefit. In many assured. cases short approach track circuits have It is, of course, possible to site the stop other purposes and may be justified signal at overlap distance from the first accordingly. facing points, when the problem im­ If the release timing is commenced mediately disappears, but it will be on occupation of a full-length approach

TRACK LAYOUT ...t?I B C D

I C D NLR TPR TPR B--"---+---¥--+-----"--1,--~----,-N I USR

D TJR

I USR

D TPR B -~,~,--

Fig. 5. Overlap route stick relay and time release. OVERLAPS (BRITISH PRACTICE) 191 track circuit, it has to be related to the 6.2. Holding the Overlap type of traffic using the line, For freight The method generally adopted for trains with a low approach speed, the holding the overlap is by a normally­ time interval must be long to prove the operated route-stick relay, released when train at a stand. If the same time interval the outer route is set. Once released, it is used for stock with better braking cannot be re-operated until the outer efficiency, it will result in delay because signal has been replaced to norm~l ~d the overlap is held longer than necessary. the train has cleared all track circmts A study of B.R. Standard Principle between the outer and inner signals, or No. 6, relating to the approach control is proved at a stand at the inner signal of junction signals, shows a similar problem by time occupation of its approach track. and a practical solution, but the require­ The circuit of a typical route-stick relay ments are in certain respects contradictory; is shown in fig 5. Two very important for overlap release the train must be circuit requirements arise from this proved at a stand ; for the junction it arrangement ; firstly, the route-stick relay requires only that the speed shall be under must be proved to release before signal 1 control. is allowed to show a proceed aspect, thus These problems will be more easily proving holding of the overlap. Secondly, solved when fully-fitted stock becomes the time element relay D.TJR must be universal, and braking efficiencies for all proved normal in the controls of signal 1 classes of train approach a common factor. to guard against premature timing due to

OVERLAPS TRACK LAYOUT AT SIGNAL 3 j K

--- D }-rn3 E 2~22 ···--+--...;;_.;~i...~----;...----:--tf- ~-+1 ~:-11--• 20~0 ~212~23 .. 0 N OVERLAPS AT SIGNAL 3 E. (0. N OR 20). {F (G OR 22R) (K OR 22) OR 20R}

D E 0 F G K 20 TPR TPR TPR TPR TPR TPR REVERSE 22 TO RWKR NORMAL I USR POINTS CONTROL CIRCUITS I USR

20 NORMAL TO REVERSE Fig. 6. Points controls for swinging overlap. 192 OVERLAPS (BRITISH PRACTICE)

a circuit supply or other failure to D.TPR. individual points keys is generally Note that in the event of replacement of accepted, but the view is taken, in some s·gaal I without the train taking the quarters, that these points keys are for oute, the overlap will be held by the emergency use only, and should not be approach lock of I because INLR cannot used for such a purpose. This attitude re-operate after replacement until the is difficult to understand because properly­ approach lock conditions have been circuited points keys are virtually no satisfied. different from an individual lever in a Consider now the way in which the frame, and the same locking conditions route-stick relay is used to impose should apply for key-setting as for route­ additional locking and control conditions setting. If keys are provided for on the points in the overlap. This is shown emergency purposes with circuits which in fig. 6, the track layout of the alternative exclude any locking, then they should be overlaps available at signal 3. The circuit sealed to prevent irregular use. diagram shows part of the points circuit On panel control systems which use controls for 20 points. \\Then no over­ entrance and exit buttons for route­ lap is held, only track circuits E and setting, a system is often employed 0 are included, but immediately the whereby a route with a choice of overlaps outer signal route is set, 1. USR releases is set by three buttons : entrance-exit­ and the following conditions are added:- overlap limit. Swinging the overlap is N orrnal to Reverse : Track D or time effected by operation of the route exit release, Track N. button as an overlap entrance, and the Reverse to Normal: Track D or time button at the extremity of the ne,v release. Tracks F. (G or 22R). (Kor 22). overlap. Whilst this is simple enough to Notice the useful facility given by perform, control panels tend towards I. USR in respect to D track control in the having too many buttons, particularly points. This track is added to the points where, in addition to those for overlap controls only for a movement from the selection, alternative routes and aspect outer signal i.e. overlap swinging is selection buttons are also provided. If prevented as soon as D is occupied and panels are to be simplified for quicker until time-released. However, if there operation in large control areas, and few are shunt movements involving reversal will dispute that they should be, !he only at signal 3, these movements will not be answer is to provide automatic operation hindered by having to wait for time of as many facilities as possible. release before moving off again. These points circuits are shown in full 7. THE AUTOMATIC OVERLAP detail in I.R.S.E. Booklet No. 22, the 7.1. General small modification to the control tables in respect of track D having been made It is probably fair to say that a signal­ in the circuits illustrated to provide the man operating a large power-signalled shunting facility. interlocking does spend a measurable It is always desirable in sv,:inging over­ time in selecting and/or changing overlaps laps to enforce the sequence of points in complicated track layouts. Whilst this operation; i.e. when swinging, the first is necessary, and requires intelligent facing points in the overlap should not anticipation of subsequent traffic move­ be permitted to move until all the new ments, it is strange that few efforts have overlap beyond them has first been set been made in the development of signalling and proved. This can be arranged in the control to relieve the signalman of this circuits by the use of the route-stick relay obligation. The reason for this is, however, to impose the sequence conditions when apparent, for at most signal boxes there an overlap is held. is ample time available, except at certain periods of the day, to allow for overlap selection. Nevertheless, with the present­ 6.3. Panel Control of the Overlap day objective of concentrating larger The methods employed for swinging areas of control in one signal box, any an overlap are subject to many variations development which can assist a signalman of opinion. With most relay interlocking in speedier and more efficient traffic control panels, the facility of using the handling with safety, must be considered OVERLAPS (BRITISH PRACTICE) 193 an advantage. Operating Department clear; if more than one is available, representatives with whom the author then that corresponding to the has had discussions, are generally agreed existing position of the facing points that they can name one or two signal shall be selected. In certain circum­ boxes in their respective areas where stances it may be desirable for the automatic overlap selection would assist circuits to be loaded and to have the in traffic handling at peak periods. At facility of giving preference to selected the same time, they are naturally con­ overlaps, e.g. one which corresponds to cerned that their signalmen shall not the most commonly used inner route. become "button-pushing-morons." This Also, to avoid the selection of others same argument can be used against (except where no other overlap 1s driverless trains, programme machines available), which may obstruct further and computers, but it does not arrest movements if they have to be held for progress in these directions because the any length of time. ultimate results of such progress are to (b) After the outer route has been cleared, eliminate tedious employment and call the overlap shall remain free to be for fewer, but more skilled operators. changed, either by key-setting, or by We still need the local knowledge and route-setting, to a ne,v overlap proved skill of our signalmen, but we are going clear prior to the change. Facilities to use it in a different way. It will be for manually holding a particular needed to make essentially human decis­ overlap for any reason, must be ions ,vhich a machine cannot be expected provided. to do, because of the limited amount of (c) If another route is initiated which data at its disposal. requires to use part of a held overlap, There is a coming development in which then provided that an acceptable preliminary work is already proceeding, alternative is first proved clear, the and which will soon necessitate serious new route shall divert the overlap consideration being given to the automatic to secure a free path for itself. overlap. This is in respect of route-setting, Detection of the diversion must be initially by train describers and ultimately continuously proved in the new route. by computers. Control by train describers (d) Changes in overlap will be permitted will obviously come first, and because only until the approaching train reaches a very limited amount of information a position beyond which the com­ about future traffic movements is available pletion of movement of overlap from a describer, the selection of a suitable points cannot be effected before the overlap will have to be achieved by train could reach them. After this automatic means. When route setting by position, the existing overlap must computer is eventually adopted, will be held until time-released by occu­ its task not then be complex enough pation. Changes in overlap must without the additional burden of selecting always commence with the points overlaps ? For this reason it would seem furthest away from the protecting prudent to develop the automatic overlap signal, so that the first facing points while we are solving the problems of are the last to move, being subject to describer route-setting, and the instal­ approach-track control. lation will then be ready for the ultimate introduction of a computer, without concern for overlap selection. It may well be argued that if the distance between inner and outer signals 7.2. Requirements is comparatively short, there is no purpose in introducing an automatic overlap The requirements of automatic overlap because little time is available for change. selection and diversion are comparatively This, however, is not quite the purpose. straightforward and can be stated as The facility provided means that a route follows:~ can be cleared as soon as a train is offered, (a) Where more than one overlap is or routes are demanded by a train des­ available at a signal, route-setting of criber. In a multiple-aspect area this the outer signal shall automatically means that the outer aspects are kept to select an overlap which is proved a minimum restriction. There is no need 194 OVERLAPS (BRITISH PRACTICE) to hold a train back nntil a preferred and aspect class selection. overlap is clear before setting the outer An entrance signal is given proof that route. Provided that more than one a route from it has been set and is clear, overlap is available, the train can be but supplementary information has to kept on the move right up to the inner be provided, not only of the identity of signal with a minimum of interference the route as determined by the selected to other movements which can continue exit, but also the route class applicable. and divert the overlap if necessary. Where overlaps are concerned, these may So far as the author is aware, only be a number of possible ultimate exits one power-signalled installation for for the same route located at the ex­ passenger traffic in Great Britain having tremities of the respective overlaps. It these facilities (with the exception of (a) is clearly not possible to feed this inform­ above), is in existence. It has received ation over the normal geographical levels Ministry of Transport approval, and may back to the exit signal protecting the thus be judged to be based on reliable overlap because the information is only safety principles. Although it cannot be that an overlap, and not a complete route, said that the traffic density on this is clear from this point. Even if the particular installation demanded such difference between an overlap and a facilities, there are concentrations of complete route be proved safely, a traffic at certain times of the day when difficulty still arises when an inner route they have proved their value, and the comes to be set. There is the possibility Operating Department considers the tbat this can be cancelled, yet the outer additional circuit work to have been signal must still hold an overlap if it is justified. off-normal. The circuit requirements in terms of The treatment of these problems has extra relays and wiring to introduce the been handled by two basic methods. facilities are comparatively modest, and Overlaps are arranged either by free­ this leads to an obvious question. Should wiring tailored to the individual require­ the facilities be provided on all overlaps ments ; or the geographical principles on an installation so as to preserve are maintained by providing supplemen­ uniformity in operation or should they tary overlap levels between the route be provided only where particularly exit and the overlap limits in parallel useful? With describer or computer with the usual geographical levels. The route-setting it will have to be universal, overlap levels comprise Availability, but on manual route systems, the author Selection, Setting and Locking, and inclines to the view that uniformity in Proving ; and are directed through sup­ panel operation is always desirable, par­ plementary points sets which add the ticularly where relief signalmen are con­ extra locking and controls imposed on cerned. However, since economy is the points when they are in an overlap. equally an important factor, it should Where control panel buttons are used to not be difficult for it to be suitably select overlaps, the separate levels may indicated on the diagram whether par­ terminate in an overlap button set. In ticular overlaps are automatic or other­ this way the route and overlap are set wise. by separate geographical levels and added together to form the complete route. 8. OVERLAPS IN GEOGRAPHICAL The economics of this system compared SYSTEMS with free-wiring are difficult to assess ; The packaged circuits of geographical with free-wiring, only the necessary relays systems, with their many advantages, do and circuits are provided, but more time present a number of problems in respect has to be spent on circuit design. With of overlaps including points. The reason true geographical sets there will always is that an overlap lies outside the route be circuits and/or relays not used, but proper, a problem also encountered in provided to maintain standard arrange­ flank protection in such systems, and its ments and essential flexibility, but special extent is not determined by fixed geo­ circuit design is less often required. On graphical principles, but by physical large and complicated interlockings the layout considerations, traffic requirements use of separate sets attempts a uniform OVERLAPS (BRITISH PRACTICE) 195 treatment of all overlaps ; but then again, given to the route proper. Neverthe­ treatment is rarely uniform even in the less, before its integrity is imperilled, same interlocking area. Is it too much at least two co-incidental failures to ask for a consistent policy to be followed should occur. in overlap selection methods, even if a (b) In straight overlaps, the location of precise specification is not possible? the block joint at overlap distance The methods of panel control for from the stop signal is the most overlap selection are dealt with in Mr. economical arrangement, and is to be J. S. Hawkes' excellent paper on encouraged. Geographical Circuitry, given to the (c) In swinging overlaps, freedom to Institution in London in October, 1965, change should be maintained until and no further comment is necessary. the last possible moment, consistent It appears that the automatic overlap with safety, to allow maximum could be incorporated in a geographical flexibility. Continuous facing point system using separate overlap levels detection diverting an overlap away because Availabilitv information is fed from an obstructed alternative is from the overlap limits back to the route very desirable. exit, and proving of setting is also fed To the Student and Technician Members in the same direction, so that the first of the Institution, to whom this paper is facing points would be the last to move specifically addressed, the author wishes in a change of overlap. to point out that it is in no sense an authoritative guide to the subject. It 9. CONCLUSIONS is only an introduction to a very interest­ It is not easy to draw definite con­ ing aspect of railway signalling which at clusions or derive any absolute standards some time in their careers will most on the treatment of overlaps. The British probably demand study and attention Railways Board Standard Signalling on their part. Principles were developed only after extensive discussions, and it is encouraging IO. ACKNOWLEDGEMENTS to see that they are phrased to recognise The author's thanks are due to AEI-· that there are always special cases where General Signal Ltd. for permission to definite rules cannot cover all require­ write and present the paper ; to the ments. The omission of reference to point Ministry of Transport, the British detection in the overlap is a pity, because Railways Board, and to members of the it seems to the author that general rules Chief Signal Engineers' staffs on Regions are possible in this instance and their of British Railways and London Trans­ adoption could have led to simplification port for their valuable and encouraging of needlessly complicated circuits. assistance. Finally, to the author's So far as main line practice is concerned, colleagues who have given their con­ the important points to note are :- structive criticism in a way that has (a) The overlap is a margin, and does helped so much in the preparation of not demand the detailed protection this paper.

DISCUSSION

Mr. E. A. Rogers congratulated Mr. subject if they could and so get something Goldsbrough on choosing a subject which constructive on the records. Any comment should produce a most interesting dis­ he made was intended in that spirit. cussion. Overlaps was a subject on which The author had set down the standards many people held strong opinions and the for overlaps on both British Railways and paper should invite a ventilation of them. London Transport, but he had not made Together they would form a useful record it quite clear why there should be so much in the Proceedings of the Institution. difference between them. British Railways Mr. Goldsbrough need not worry about used multiple-aspect signalling, three and being criticised for expressing his personal four aspects, giving full braking distance views ; they would all like to clarify the between the first warning signal and the 196 OVERLAPS (BRITISH PRACTICE) stop signal ; and this was, in general, machine operating time. Only if the backed up by A.W.S. on the train. The signal was less than this distance from overlap, therefore, was regarded purely the points did one need to resort to as an overrun, an allowance for mis­ approach locking. judgment, which was necessary and would In sections 7.1 and 7.2 the author men­ continue to be so while the anachronism tioned the future prospects of automatic of the unbraked freight train continued route setting. The rules he quoted were to be accepted. When that had been generally reasonable but until one reached abolished they might be able to rethink the stage of using computers which would overlaps and, perhaps, reduce them to be able to predict several moves ahead rather less than they were today. London of a particular route being set, there Transport, on the other hand, used would be a risk that an overlap would be basically a two-aspect system with train established which would inhibit a later, stops at the red signals, and it was of but perhaps higher priority, movement. course essential that the overlap should Finally, in section 8 the author pleaded be of sufficient length to enable a train for a consistent policy in overlap selection which had been tripped to come to a methods, and the speaker had every stand without risk of collision. sympathy for this plea; however, a In considering the detection of points little lower down, under " Conclusions," in the overlap it seemed difficult to avoid he noted that the author was encouraged a measure of inconsistency. It was quite to find that B.R.B. standard principles proper that they should detect facing were phrased to recognise that definite points which led away from an obstructed rules could not cover all requirements. overlap, but if they permitted swinging Would the author not accept that the of the overlap by using the series selection second comment was the answer to the arrangement which was shown on one of first ? the figures, there was no detection proving Mr. Goldshrough said in the course of at all when both overlaps were clear. his reply that a computer would cater Thus, if the points moved to select an for, perhaps, 90¾ of what was required available alternative overlap, there was of it, and even that was a high estimate; no means of ensuring that they completed but it was the other 10°/4.iwhich was their stroke. This was accepted on the always the bother. Any automatic basis that, to create a dangerous situation, process, among which one might include an overrun must coincide with a need to programme machines, might possibly be change the overlap and the failure of the regarded as a fairly simple form of on-line points to complete their movement in computer. An alternative manual control making this change. was always available, and this was He was not convinced of the need to essential. He felt that selecting an overlap detect trailing points in the overlap since which might prove embarrassing or one would expect that the route-setting difficult because of later traffic movements circuitry would have called for the was one of the circumstances in which appropriate points to operate and thereby the signalman should step in and override locked out the possibility of making a the operation of an automatic overlap. movement which could foul the overlap. That was why he very strongly emphasised In section 6.1 reference was made to the that one of the conditions for an automatic need for approach locking of facing points overlap was that it should be capable ahead of a signal to ensure that they had of being overriden by manual control, time to complete their operation after the and must always have this manual locking track circuit had become occupied. control, no matter what form it took. The signal did not need to be full overlap distance back from the points to avoid Mr. V. H. Smith thanked Mr.Goldsbrough the need for an approach lock ; it needed for his interesting and informative paper only to be far enough back for the points and joined with Mr. Rogers in congrat­ to complete their movement before the ulating him on producing a paper which fastest train would reach them after would be used for reference in dealing with passing the signal. For example, 132 yd this subject. He also felt that the Exam­ for a 90 m.p.h. train with a 3-second ination Committee would look very closely OVERLAPS (BRITISH PRACTICE) 197 at the paper when the time came to the train had come to a rest, or virtually prepare future examination questions. to rest, at the inner signal. One of the problems of the Examination In section 4.3 Mr. Goldsbrough stated Committee in setting questions was ensur­ that where there was a choice of overlaps ing that adequate information was avail­ the principle continued to be to prove able on a subject for the student and for all track circuits and to detect only those him to have acquired sufficient knowledge facing points which diverted the overlap to answer the question. He thought the away from any one which might be paper did this very thing. obstructed. He illustrated this in fig. 4, Mr. Rogers had already made one but Mr. Smith did not think that the comment on the relationship between point detection contacts really detected British Railways and London Transport, the points as Mr. Goldsbrough had drawn and for his part he tended to regard them. He thought they were only there British Railways as not having overlaps to cut out the track circuits which might but only overruns. Referring to Mr. he legitimately occupied if a parallel move Goldsbrough's paper, Fig. 3 showed over­ were taking place. In section 7 Mr. laps being shared by two or more signals. Goldsbrough talked of the automatic This was something that ,vent against overlap adopted on London Transport his nature. Its safety was dependent on on programme machine installations, and the belief that trains moving in opposite he felt it should be included as a feature directions would not both overrun their to assist the signalman if one must have stop signals at the same time. On London the signalman. Having simplified the Transport this arrangement was never form of control for the operator to one adopted, and where trains could move entrance and one exit button per route, in opposite directions, approaching each it seemed to him that the signal engineer's other, the distance between the two stop equipment was best fitted to devise the signals was always the sum of the two most appropriate setting for any swinging overlaps required. overlap, bearing in mind all other train Mr. Goldsbrough, in quoting the movements taking place in the area. standard signalling practice, agreed with Experience had shown in the past that the British Railways Board " that the where the arrangement was left to the facing points in the overlap may be moved signalman, the optimum from the train if they have time to operate and an service point of view was not obtained. alternative overlap is available." He In fact, at a fairly recent investigation himself felt that permitting facing points it was discovered that, given similar in the overlap to move with the train circumstances, different signalmen at the approaching the signal was a dangerous same signal box chose a different order practice and should be prevented. In for trains which had to proceed over fact, Mr. Goldshrough had pointed out conflicting paths. Each signalman, he in section 4.3 that points rarely lost detec­ believed, did his job conscientiously and tion once they were static, and if one thought he was right. The system had interpreted "loss of detection" as meaning now been automated, and he did not that the switches actually moved away think the automatic arrangement agreed from the stop rail, and not a detection with any of the three signalmen who failure as such, this meant to say that the worked it. They did not arrive at the possibility of switches not being closed solution without a lot of thought and, on to the stock rail, and being in the in fact, some trial runs. position to derail the train, was likely Mr. Goldsbrough had said he thought to occur immediately after points had manual control was essential, and he been thrown. This was the very dangerous would quarrel with him very much on condition that he was setting up. Here that. He certainly would disagree with was the train running at the points, and providing manual control which enabled he was permitting the points to move, a man to intervene and think he was with the subsequent risk of not com­ doing better than a machine. The manual pleting their stroke. He thought it much control on London Transport was only better for the outer signal to lock the provided to cater for the emergency that points, and then only to move them after arose where the service was gravely 198 OVERLAPS (BRITISH PRACTICE) disordered. For instance, that very day to do everything. One could programme there was an accident when somebody it up to a point but one was bound to fell under a train and this caused a major write into it the facility that if it found upheaval on the line. In fact, he believed it had got a problem it couldn't solve the line was shut down for three-quarters it would have to come back to the pro­ of an hour. Obviously they had not got grammer, and he, with his skill, would that degree of disruption built in, and in have to solve it. He thought this was the those circumstances manual control was attitude one must take on all these necessary, but he did not believe it should matters. be there, nor should it be permitted to He knew that a machine would do be used to allow the man to do, as he better than a man m many cases, but thought, better than the machine. not in all. There would always be the situation in which the machine would Mr. Goldshrough agreed with Mr. Smith say : " You'd better do this because I that an " overrun" was perhaps a more can't." appropriate term. but he knew that the principles of main-line overlaps were Mr. D. C. Webb added his thanks and absolutely horrifying to London Trans­ appreciation to Mr. Goldsbrough for port. He sympathised with the way they producing a most interesting paper. The felt about it, hut the fact was that the previous speakers had made it clear that Ministry of Transport accepted them, there were obviously very different and he could quote no less an authority problems on London Transport and tl1an Mr. Fews as to how useful swinging British Railways. In the case of British overlaps were in the London Midland Railways the overlap was purely to Region electrification. Mr. Fews had increase safety in the event of a driver quoted in his paper the valuable minutes failing to stop his train at a red signal, saved by using them and he considered having already had a warning signal, that the extra circuit work to provide and he found it interesting to consider them was money well spent. how long the overlap should be. Mr. Mr. Smith had referred to Fig. 4, Goldsbrough had said in section 4.1 that showing the overlap controls, and did the average of 13 reported overruns in not understand the case in which detection one year was 112 yd but that was an was included. If track circuit B was average figure. He wondered if Mr. either failed or occupied, one would not Goldshrough could tell them what was get the outer signal unless one had the maximum overrun which of course continuous detection of 11 points normal. could have produced an accident, and also whether these overruns were with Mr. Smith interposed that he preferred A.W.S. in operation or not. He would to put it the other way round-if B track like to know whether there were any was up, one was detecting, and Mr. Goldsbrough concurred. further statistics available, which led him to wonder whether it would be very Mr. Goldshrough, continuing, dealt with long before they could reduce the length the question of uniformity with automatic of their overlaps by finding other methods operation ; in that respect he agreed of ensuring that the train stopped when wholeheartedly with Mr. Smith, but not it should. This was touched on by Mr. in the way he put the matter of manual Rogers. When they lost the loose-coupled intervention. This was because, no matter freight train they would certainly have an how clever a computer was, there would improvement and as their stock became always be things which it could not solve more reliable, he felt it would be reasonable and it would virtually have to appeal to to think that the overlaps could be the user and say : " please do something reduced. The ultimate object might well because I certainly can't." This was be to do away with overlaps altogether, quite likely to happen with complicated but no doubt they would have to wait some traffic movements and attempting to time for that. assessing what was going to happen next. One would always have this problem; Mr. Goldsbrough said he was not at one could not programme a computer liberty to state the source of his inform- OVERLAPS (BRITISH PRACTICE) 199

ation on reported overruns. But he could fitted stock. If we could ultimately say that from the list he was given there reduce the length of an overlap, it would were about 16. He eliminated three of considerably economise in our signaUing them because it \Vas quite obvious from circuits and be of great satisfaction the remarks put against them that the to everyone. driver had no intention of stopping~he had not even seen the signal. No, there Mr. N. S. Hurford wished to emphasise was no A.\V.S. The maximum in the 13 the point about the terms " overlap" and reported ones was 400 yd. The minimum "overrun." On London Transport the was only 15 yd but Mr. Webb would overlap was something to which particular appreciate that there must be a number care and attention were given. It was of overruns well under even 15 vd, that very carefully calculated, checked and never got reported at all so that It would cross-checked and had a very real signific­ be very difficult to find an average figure. ance ; they knew that if a driver on He imagined that quite a number of London Transport took a train past a overruns were fairly short, and so it signal, that train was going to stop seemed quite reasonable to him that within that distance. It was very nice the B.R.B. figures were well chosen. and comforting to know that, of course, One or two people were disappointed but British Railways' use of the term that he did not include consideration of was, as had been said in the paper and overlaps on the Continent and in the implied by previous speakers, was merely United States in his paper. He had not a margin for drivers' error. He wished done so because it was difficult to get that the term " overrun " was used much information, but one thing he had rather than " overlap " because he thought found out was very illuminating. He there was a clear difference in meaning wrote to the American Association of between those two terms. Railroads about overlaps because every­ He would also like to refer to fig. 3, body here said that they never had the so~called shared overlap, which to overlaps in the United States ; if they anyone of London Transport sounded did not have them, whv should we ? rather horrific. If, as he understood the The truth was that thev" did not have arrangement, two trains were allowed overlaps except on rapid~ transH systems to proceed together, approaching each with a full block overlap, which was a other against the two back-to-back signals different case. What they did was to simultaneously, he really thought that provide an excessive braking distance the term was not shared overlap-the from the warner to the inner signal. VVe overlap was halved, and again if that term could not do that in this country because were used it would make the point rather the geography of our railways was so clearer. closely knit that we had to pack our His next point concerned the vexed traffic in between junctions so close problem of the circuit in fig. 4. His together that there would be no room to training had been that in a layout like provide this excessive braking distance. this quite clearly 11 points reversed Of course, if there was such a braking would lock 13 points normal. If that was distance, it would be a pretty bad driver so, he did not think the contact 11 RWKR who could not bring his train to a standstill would be necessary. at a red signal when he was given more Mr. Hurford referred finally to the than adequate warning. calculation of the overlap in section 5.2 As was well known, jn the United States by way of amplification, because he very sophisticated methods of cab thought the stress was not quite right. signalling were employed, which went Mr. Goldsbrough referred to the braking to show that it was all very well to accept efficiency of 12¾ in tunnels reduced to that somebody did not have overlaps, 10% in the open. He thought the stress but when one looked behind it there was should rather be that it was normally a very good reason for this. He, too, 10%, and they relaxed it to 12% in tube agreed that one looked forward to a time tunnels because in those conditions they when the length of the overlap could be knew the track would be dry and trains' reduced because of universal brake- brake blocks would also be dry. In fact 200 OVERLAPS (BRITISH PRACTICE) it was only applied after the first " all remembered a curve showing braking trains to stop " station when a train distance against speed in which the entered a tube tunnel. In the sub-surface distance was given as minimum braking tunnels the 10~,~ figure was applied, so distance, whereas it appeared to him that was the more normal figure. to be the maximum braking distance. It was only after some while that he Mr. Goldsbrough replied regarding fig. 3 realised it was maximum braking distance that he rather regretted having used the for the C.M.E's department, but minimum word " typical " because it was not typical, distance for the Signal Engineer ; and for it was rather special. He selected it that reason he feared that he might be because it appealed to him as showing referring to Signal Engineers' terms which many different overlap conditions that might have a different sense to the C.M.E. were possible. He might say that he Nevertheless, he drew attention to the knew of this particular layout in three author's reference to the calculation of separate places, and certainly in one of overlaps in para. 4.1, where he said that them the speed limit that applied was although formulae might differ in detail 15 m.p.h. He thought that these overlaps, they usually included factors of a as they stood in those limited conditions, maximum approach speed and so on. were perfectly acceptable. If one con­ So far as he could see from the paper, none sidered the actual possibilities of two of those factors was used by British trains approaching the stop signals at Railways in arriving at an overlap the same time, it probably rarely occurred. distance. Only two overlaps were given, The facility needed was, of course, to 200 yd and 300 yd, and it would appear berth two trains at the platform and that except for gradient, none of the allow either one of them to pull out other factors-speed, average braking without involving movement of the other, capacity, efficiency, signal sighting dis­ so that it was rather a special case; but tance and so on-was used in arriving at he agreed that they were halved overlaps the distances concerned. So far as he rather than shared ones. knew, the only system, at any rate in Returning to fig. 4, with the point this country, which did take account of detection, he had completely disregarded the relevant factors was London point-to-point locking in this. He was Transport. simply putting forward one argument, The proposals for the use of the term namely that where one had an obstruction, "overrun" rather than "overlap" became one must continuously prove detection obvious. The French used a rather away from that obstruction. He did not grandiose word when they talked about relate this to point-to-point locking at all. their philosophy of signalling, philosophy On the question of braking efficiency; of overlaps and so on. It was possible that in present company it would be very British Rail might he persuaded to unwise of him to make any comment change their philosophy-their attitude­ about London Transport principles at all. towards overlaps, and the reason lay in He put them in the paper because quite the introduction of A.W.S. As the author obviously this information ought to be said, it was not reasonable to talk about included in any study of the sort. He a safety margin for misjudgements in reported what he was told. The inform­ braking. There was, for instance, no ation given was produced in the paper misjudgement in braking with A.W.S. and he might have seen some private It was no longer possible to quote the opinions but he certainly did not think the cancellation of the brake by the he should elaborate on them. However, driver and his subsequent control, when he thanked Mr. Hurford for correcting the fact that he might be dead was the misunderstanding, or rather his mis­ accepted and the A.W.S. was designed representation of the respective braking to cover this contingency. The driver's efficiency percentages. safety device was a further recognition of the possibility that a driver might die Mr. H. R. Broadbent said he was always at his post. The position of British Rail­ a bit chary about speaking when the ways was no longer tenable and the cosy subject was not his own discipline. He feeling that an overlap was required to OVERLAPS (BRITISH PRACTICE) 201

be no more than a cover for misjudgement Mr. Goldsbrough replied that he was in braking, could no longer be enjoyed. being put in a very difficult position in There was already an overlap in the being asked to answer both for the present system--the distance from the British Railways Board and for London warning signal, A.VV.S. control point, Transport. He did not think he should to the end of what ,vas termed, by British do this, but with regard to the questions Raihvays, an overlap- and it was this about approach speeds, sighting distances distance which should now be considered. and other factors ; if he understood it On the L.M.R., for instance, where it had correctly, the British Railways graph been the practice to cover for speeds up was not to be applied literally. It was to 70 mile/h. on the maximum (minimum) a general guide, and if a Region considered braking distance to the home signal, that the particular circumstances at a and beyond that speed to provide a particular signal merited more consider­ greater sighting distance, the introduction ation being given, they were free to of A.W.S. had undermined the policy. adjust the overlap as they considered Sufficient sighting distance did not allow necessary. It was a merit of British for A.W.S. control and he would suggest Railways principles that they were made to that from now omvards, or from the accommodate exceptions, and as they all moment A.\V.S. came into full control, knew, the number of exceptions which the position was not as stated in the occurred was considerable, so he thought paper and a rethinking Was essential. they must not be taken too precisely. They were as a general guide, and it His other point was really only a seemed to him that they were a very some,vhat minor one for London Trans­ useful general guide. port, but as the paper was being addressed to students it might he as well that the With regard to A.W.S., here again, fallacy of talking about braking efficiency perhaps, he should not make any observ­ as a percentage of g should be got rid of. ations, it should be done by British Braking efficiency could not be a Railways. One of the points about A.W.S. percentage of g, and a quotation from was that it drew the driver's attention a glossary of terms produced by 13.R.B. to the fact that he had passed a warning for braking should clarify the matter. aspect and invited him to take charge Under the definition of the term braking of the situation. Now, he did know how efficiency, it was noted that the braking many occasions there had been when a efficiency (or braking ratio as it was driver had dropped dead at his post or known on Western Region) was of no had passed out completely. He did not value where rates of braking were being think British Railways ever had the idea considered, a rate of braking heing a that A.W.S. in every case would bring change in velocity in a given time. If a a train to a standstill at the stop signal. percentage of g was being quoted, the All he thought they were seeking to do reference was to a rate of change of was to make quite sure that the train velocity in time ; but the derivation of did not run away out of control, and that the braking efficiency formula was based was the merit in A.W.S. It was not entirely on initial speed and distance. possible to provide a system with precise It had no relation, therefore, to time results when one had a number of trains and it could not be quoted in terms of a with so many different characteristics. change of speed in relation to time. London Transport were to be envied Braking efficiency was the ratio of the in that they could measure the perform­ average retarding force during a stop ance of their rolling stock very accurately to the equivalent braked weight, i.e. the and they had taken advantage of it, as static weight of the train with allowance well they should, and had been able to of the rotating parts. Its conversion produce, if such a thing were possible, into a percentage of g on an assumption almost a perfect system. They were to of constant retardation from brake initi­ be commended for this, but one must ation until the train had stoppPd was not forget that it was not possible to do misleading and ended very often as an these precise calculations on the main unwarranted figure of a demand on lines of British Railways and one must adhesion. accept certain limitations in the facilities 202 OVERLAPS (BRITISH PRACTICE) provided. He invited comment from Mr. what was happening, and why he had no Rogers. brakes, and had put his brake into full airbrake, he was, in fact, past the signal. Mr. Rogers said that on the argument Perhaps he was shooting himself down which had been put forward, if it was a little bit on overlaps, but he felt basically assumed that every driver dropped dead this was a bad design in the braking when he got an A.W.S. warning, they system, and this state of affairs should would not need an overlap ; but the not be allowed to happen. driver did not do so------invariably, if he Turning to other matters in the paper, ,vere conscious, he cancelled the warning, he remarked that in connection with he acknowledged it and took control of approach release of signals at junctions, the train. Having established that for speed control etc., one went to great position one must again allow for his lengths in preventing the signal from misjudgement. They did not expect him clearing in an attempt to get the speed to fly past the stop signal at a high speed of the train down to a safe speed for the -they expected him to stop at or about junction ahead. What was not prevented the signal-but it was much better that was the actual movement of the points he should take control, otherwise the to the lower speed route. This seemed train would often stop very, very short wrong to him because if the train was of the stop signal. The driver's job was going to run past the signal and be to get to the stop signal ; for that they basically out of control, then it was gave him the allowance of the overrun wrong for points actually to have been (if one liked to call it that) ; they still set for the lower speed route ; and putting called it the overlap. approach control on to the signal did not really alleviate the problem. Mr. V. B. Orchard thought the question of overlaps was a hypothetical one. He In fig. 2 Mr. Goldsbrough showed a had been told by an old locomotive man junction signal with no separate overlap. some years ago, that any distance he He was quite prepared to accept no cared to name up to the signal at red separate track circuit for overlap on an was his, One quarter of an inch beyond auto signal, but he was a little worried that signal was not his and he shouldn't when it came to a junction signal, be there. He thought that was a golden especially if there might be a station to rule which was sometimes forgotten, and the rear of this signal and a train had the question of prolonging the agony, stopped to do station work with the should one say, by providing overlaps signal off. The train could have come to of varying degrees, was evading the a standstill with the driver and his engine question. London Transport this evening beyond the signal. He had met such a had put their case against the use of case at a place which should be nameless. overlaps on British Railways ; even The signalman, realising the train was London Transport were failing in the doing station work, enquired of the station exercise as he ~aw it They provided a staff how long it would be, and the train stop at a signaL If the driver was station staff said 10 minutes. The signal­ not already stopping, that train was not man at that time had a light engine going to stop at the sjgnal. He'd now movement about. He put the signal back gone past it and they were equally as to danger, the approach locking timed bad as they were on British Railways, out quite correctly, and the conflicting That train should not be past the signal. move was made at about the same time as the guard, willy nilly, decided to give It was rather ironical that with a fairly the right of way to the train in the recent inovation in braking equipment, platform. He saw the signal at red and Le, the electro-pneumatic brake, they had said, " Oh well I've just passed about now in fact a case where some form of two or three dozen others coming down overlap control was necessary because on from London that have been red by the a number of occasions he had personally time I saw them anyway" and gave the witnessed a driver making the last brake right of way. The driver was none the application at a signal and the E.P. fuse wiser because he was already past the had blown, and before he could think signal, and the train started. At which OVERLAPS (BIUTISH PRACTICE) 203 point a light engine shot straight across two copies of his paper so that they did in front of him. It was a lucky case in not have a failure ln the lecture. One of that the movement was all very smooth the arguments against redundancy tech­ and no catastrophe ensued. He mentioned niques was the claim that the existing this point as he felt very strongly about signalling system was ahvays fail-safe. it indeed, for it also applied in their With a three-channel redundancy system recent question of track circuit joints and majority voting, the advantages and their positioning and fouling points offered over existing systems were as etc., in connection with liner train vehicles. followed: 1. The same order of safety. Mr. Goldshrough thought they must 2. Better reliability. all agree with the principle that if one It could he argued that redundancy was ¼ in beyond the signal, one was in techniques used more equipment, but the wrong place. No one could question the equipment itself could be made to that, but it was apparently quite a a very much less rigorous specification difficult job to bring a train exactly and therefore much cheaper. The advan­ to the right position and stationary at tages of better reliability were many a signal in varying conditions. Regarding both from a commercial point of view fig. 2, overlaps without separate track and also of safety. It had been pointed circuits, he had rather expected that out on several occasions that many one or two people would try to read more accidents occurred when the existing into the diagram than was actually there. signalling system had failed and emergency It was put in only to show that by siting working had been introduced. It ,vas the signal back from the points at overlap of interest, therefore, that in ~ection 4.1 distance, one was relieved of the necessity of the paper it was stated that the British of considering the points in the overlap. Rail standard overlap, ,vhich he would He agreed that in dealing with the emphasise was not an absolute overlap, situation in practice, the circuitry and was justified from statistical experience. techniques for introducing point locking Would the author consider applying the and protection would be quite complex. same reasoning to redundancy techniques? Regarding engines standing beyond the signal, this was quite a common practice Mr. Goldsbrough asked to be excused and, in fact, rear indications were very commenting on the fail-safe and redun­ often given on signals to allow an engine dancy aspect, which was the President's to stand forward ; but in the cases he particular subject. Regarding the point had come across quite extensive pre­ in section 4.1, he admitted that he had cautions were taken to prevent any accepted that the figures he had been irregular movements. He did not think given justified British Railways' choice this particular difficulty was really related of the distance of overlaps. He hoped to overlaps at all; it was just normal more would not be read into it than was practice where a train was a bit longer really there. All that happened was that than normal and where special provision he had asked a particular Region if they was made for allowing the engine to stand would be kind enough to give him some ahead of the signal in a station. In any figures. and he was delighted to find, case, the train was at a standstill at the when he looked into them, that thev did time the starting signal was given. actually show the figures came inside the Therefore the difficulties of problems overlaps recommended by British Rail­ normally associated with the overlap ways. He was not in a position to make just did not exist. a statistical investigation of this problem ; this was a purely general and casual Mr. J. W. Birkby said that for the last observation, and no accurate statistics few years he had the honour of attending were involved, or any research whatsoever. the Institution's meetings as a visitor. He had been interested recently in The President (Mr. Hadaway) interposed listening to the arguments for and against to say he wondered if Mr. Birkby was redundancy techniques as opposed to the proposing that in the interests of re­ existing signalling systems, and was glad dundancy there should be three overlaps therefore to see that the author brought for every one ! 204 OVERLAPS (BRITISH PRACTICE)

Mr. D. Hotchkiss recalled that Mr. necessary flexibility ; one was not tied Rogers said they did not need to detect to using the same conditions in the trailing points in overlaps, but in geo­ overlap as one had for the route ; the graphical systems the aspect lines would two were quite separate. He agreed that have to pass through the trailing point overlap buttons on the panel were a units in the overlap, and be routed over good thing and he rather liked Mr. contacts according to whether the points Hotchkiss's term "mini-route." His were normal or reverse. He submitted that criticism about the number of buttons these contacts might just as well be on the panel merely came back to the detection contacts as contacts of any other automatic overlap, and he felt that point function, so one could say that it whilst they had buttons to do this and did not cost anything to detect trailing buttons to do that, if there was something points in the overlap. On the other hand, they could do automatically, why not one could say that if trailing points in do it, and get quite a lot of benefit from the overlap were detected, and a detection it at the same time. failure occurred, then one was holding the outer signal at red as well as the inner Mr. B. H. Grose said Mr. Goldsbrough signal. This could cause more delay to perhaps found it better to stand aside traffic. from the vexed question of redundancy techniques, and he would like to step Secondly. he wished to put the case into the breach. Mr. Birkby had come up for overlap buttons. Everyone seemed with the two old chestnuts which seemed to have agreed that it was much preferable to permeate all Research Department to swing the overlap by means of thinking. These were the reliability and independent point keys, but as a circuit the cheapness which could be offered by designer, he felt that, again with a redundancy systems in competition with geographical system, the method of over­ conventional fail-safe equipment. He lap swinging employing overlap buttons had pointed out some time ago, at another enabled one to swing an overlap using meeting, that adequate safety could only the existing route-setting facilities. In be assured with inferior equipment by other words, one treated it as a " rnini­ combining such items into voting circuits, route," using the button of the inner and as a consequence there was much signal as entrance and the required overlap more equipment involved and therefore button as destination. This was quite the reliability must suffer. simple, but if the same thing was now attempted with an independent point key, All that could be achieved by using one first of all had to seek out the new inferior equipment in redundancy circuits route and lock it. Then one had to swing was to prevent one wrong-side fault from the facing points themselves and thirdly causing a catastrophe. This improvement one had to release the remainder of the had to be bought at the cost of much old overlap. All with one point key. heavier maintenance. If, for example, So this point key suddenly stopped being one-out-of-three logic was used then there an independent point key and became must be three times as much equipment a sort of route-setting key. The circuits as would be employed by conventional one had to " hang " on to do this were, circuits, on top of which must be added he would say, at least twice as complex the extra equipment needed for the voting as those for a commensurate case of circuits. It was obvious that the overall overlap swinging by overlap buttons. fault rate of this equipment must be It all added to the cost of the job. much worse than that of the good equip­ ment it replaced, simply because there Mr. Goldsbrough said he thought he was more of it and because the basic had pointed out in the paper that in reason for its being there was that it geographical circuitry, and in the system was of inferior quality. It was therefore he was familiar with, the overlap levels simply not true that economical advan­ were quite distinct from the normal tages were to be obtained if the Signal route-setting levels, and therefore one Engineer would only drop his apparently could include trailing point detection antiquated methods in favour of the or not as one wished. That gave the latest ideas. OVERLAPS (BRITISH PRACTICE) 205

Mr. A. E. H. Jeffries asked how Mr. signalling and the assistance of the Goldsbrough felt about level crossings Automatic Warning System, the driver situated in the overlap a few feet away should be required to stop at the signal in from a signal. A substantial obstruction the same way as he was required to stop might exist at the same time as a train at the buffer stops at the terminal station. was approaching a signal at red under The new systems of signalling, which the Regulation 4. Research Engineers were at present developing, would certainly preclude over­ Mr. Goldsbrough replied that he thought leps and would simplify and cheapen it would be very unwise indeed in the the cost of present-day multiple-aspect circumstances of the moment for him to signalling if the Traffic Department would make any comment at all. The position agree to relax its requirements. Mr. as he understood it was that so far as Goldsbrough had said that his remarks an overlap was concerned, a level crossing were directed particularly to the new did not exist. He thought the reason for generation of Signal Engineers and he this was that since the line was well hoped that they would succeed in this, know, all the drivers are well aware that as the removal of overlap requirements there was a level crossing in the overlap, would considerably reduce signalling in­ and they adjusted their driving and stallation costs and at the same time exercised due care for that purpose. would not reduce the safety of the signal­ Signal Engineers were not, after all, the ling in terms of business risk. people who decided these things ; that was the position, but he would rather not The President (Mr. Hadaway) then make any further comment at this brought the discussion to a close, remarking juncture. that the value and interest of Mr. Goldsbrough's paper made him feel keenly Mr. J. S. S. Davis said he had listened the restrictions of presiding that prevented to the discussion with interest because him from taking part in the debate. He at one time he did some research to find did, however, wish to make one comment, out how the practice of providing an which was on the philosophy of the overlap originated and what had deter­ signalman stepping in to set an overlap mined the distance of 440 yd. He was when automatic means had gone as far not able to find any written information as they could and asked to be extricated on the subject, but was told reliably at from a dilemma. He thought perhaps it the time that the overlap derived from would be better if the design philosophy the emergency vacuum braking distances of the machine was that, having gone for express trains. This emergency as far as it could without producing a braking distance was stated to be 640 yd positive answer, it would still go on to and, therefore, if a driver was suddenly give an answer of some sort, because it confronted with a stop signal at danger, would do this better at short notice than it was calculated that he would, under the signalman, who did not know nearly normal conditions, be able to sight such as much as the machine. Therefore he a signal at 200 yd and accordingly would thought there was still a case for saying not be able to stop until some 440 yd past that the machine could do better than the the signal. Mr. Goldsbrough had stated signalman. in his paper that the overlaps varied The Signal Engineer was not able to according to gradient and other factors do all he would like in the matter of safety, and he could confirm that this was the and this situation would be with him case within his experience ; and that until the civil engineers and operators according to the type, class and speed designed and operated the perfect railway. of traffic, as well as the gradient, the Some of the problems they had been overlap could be varied. discussing arose from situations which He was in agreement with previous would not exist in an ideal environment. speakers that the time had come for The President then proposed a vote of revised thinking on overlaps and that thanks to Mr. Goldsbrough, which was with the introduction of multiple-aspect carried with acclamation. 206 Provincial Meeting of the Institution of Railway Signal Engineers held at Glasgow

Wednesday, March 20th, 1968

The President (Mr. H. W. HADAWAY) in the chair

Mr. J. V. Goldsbrough (AEI-General Signal Ltd.) read his paper on "Overlaps (British Practice)."

DISCliSSION

Mr. Baldwin opened the discussion and but this did not mean that they did not complimented Mr. Goldsbrough on his occur. In one Region, in a year, the paper, saying it would be of value to other maximum reported overrun was 400 yd. members of the Institution quite apart and the minimum 15 yd. from student and technician members. Lines fitted with A.W.S. equipment He remarked that the Ministry of Trans­ should be provided with overlaps until port Inspecting Officer had spent an it became possible for all trains to possess appreciabJe period during the inspection similar braking characteristics, when the of the Perth installation in the relay room, overlap length might be reconsidered. studying the circuits associated with the It should be noted that overlaps had not automatic overlap before he accepted the been provided on many low-speed lines principles involved. The principle of in the new Leeds installation. eliminating the separate overlap track circuit had been agreed so as to reduce Mr. Dean said it would be interesting signalling costs on all Regions, and not to know how the figure of ¼-mile for a primarily to reduce costs in areas where block acceptance overlap had come to be d.c. traction had been adopted. accepted initially and had been perpetu­ It was difficult to measure the value ated for such a long time. The first record of providing overlaps in actual practice. of this dated back to !875, when ¼-mile Reports dealing with cases of signals was agreed after the Newark trials, and passed at danger on two power installa­ whether it had any relationship to braking tions in the Scottish Region last year distances was doubtful ; but it was revealed that they were mainly cases of certainly accepted that under present misreading of signals as against cases of conditions the overlap was provided to slide-past or overrunning. He asked cover possible overrunning of a stop whether there was a case for eliminating signal. overlaps altogether, with a real reduction The provision of such overlaps did in in signalling costs, where A.W.S. was some cases complicate signalling controls; provided and in areas where speeds but its necessity was proved on several were very low ? occasions by the number of overruns even in modern colour-light areas. On Mr. J. V. Goldsbrough replied that this point, did the author think the elimination of the separate overlap did present practice of eliminating an overlap prevent an early replacement of the track circuit at outer homes, etc. was a signal, but this was considered acceptable. good thing, as this had deprived the It must be remembered that overruns signalman of a means of identifying of 20-30 ft. often were not reported, overruns. OVERLAPS (BRITISH PRACTICE) 207

With regard to swinging overlaps ; it position, the points would have completed was felt that a little too much importance their movement before the overrun was attached to these, as in his experience reached the points. it was not normally necessary to change an overlap once this had been decided Mr. J. V. Goldshrough agreed, but added upon, because simultaneous train move­ that if the points were included in the ments were normally known. overlap, the additional restrictions on In regard to the question of variable their operation must still be imposed. lengths of overlaps ; did the author think it was necessary to extend a normal Mr. R. James asked if the author would overlap during foggy weather, bearing comment on the positioning of runaway in mind the possible existence of A.\V.S. catch points within-or outwith the overlap equipment? on plain line remote from interlocking areas ? Mr, J. V. Goldshrough replied that some Regions considered that the provision Such points were generally placed beyond of swinging overlaps enabled valuable (outwith) the overlap and whilst this minutes to be saved. Mr. J. H. Fews' was generally desirable, complications paper, " Planning Principles Underlying might arise where long trains had to be the L.M. Region Main Line Electrification accommodated between the catch points Scheme", given to the I.R.S.E. in London and the next signal ahead. The position on October 8, 1963, stated : " This v,,ras aggravated by the usual closer facility (swinging overlaps) has proved spacing of signals on rising gradients, to particularly valuable during periods of maintain headway conditions. heavy traffic at such places as Manchester Where standage space was limited, (Piccadilly) and has enabled valuable would it not be preferable to consider minutes to be saved there." the overlap as a shared overlap between The Eastern Region, in their many the movement approaching the running relay interlockings, had always provided signal, and a possible breakaway approach­ this facility and its value was now well ing from the other direction, thereby established. Subsequent train movements permitting the catch points to be placed might be anticipated, but not accurately within the overlap and nearer the first forecast, and the facility of a last-minute signal ? change was very useful. Whether one liked it or not, the automatic overlap Mr. J. V. Goldshrough replied that this would have to come to allow describer question should be answered by raihvay and computer route-setting. There was operators. Nevertheless, the principle a reference to this last aspect under of shared overlaps was accepted and in section 7.2. of the paper. consequence it could be applied to this The extension of overlaps in foggy condition. The likelihood of a breakaway weather was strictly an operating matter. and a run-past occurring ahnost simul­ It would be expensive from a circuiting taneously was extremely remote, Even angle. In non-A.W.S. territory it \vas where shared overlaps were provided usual for speeds to be severely limited within low speed interlocking areas, as in fog and although overruns were more illustrated in fig. 5, the likelihood of two likely they would probably not exceed movements approaching signals at either the overlap distance provided for normal end of the overlap at precisely the same approach speeds. If a signal was actually time was fairly remote. missed, no overlap other than a full block would allow for this. Mr. Henderson, remarking that the testing of the automatic overlaps at Mr. R. C. Nelson said there was no Perth was comparatively easy and proved need to provide approach-locking of facing to be very interesting, commented that points in an overlap which was in excess the USR relay was proved in the released of 50 yd. ahead of the signal, because position before the route signal was if an overrun occurred immediately after permitted to clear. Was this control a the points had been called to a new matter of convenience or design ? 208 OVERLAPS (BRITISH PRACTICE)

Mr. J. V. Goldsbrough replied that it Since Mr. Go!dsbrough's paper would was always desirable to prove vital relays be used for reference, it was considered in the released position as far as possible, that the method of deriving the overlap and the USR certainly had a vital function. formula in paragraph 5.2. should be The specifications for present-day relays provided in greater detail. This might were less critical than formerly, e.g. be done by adding an appendix to the absence of jewelled bearings. There was paper. a case for adopting lower standard relays for all functions which could be fully Mr. J. V. Goldsbrough said in reply that proved, but insisting on high standards it was considered desirable that trailing for those which could not, e.g. TR and points in an overlap within SO yd. of the TPR. protecting signal should be detected, especially if the points were double Mr. D. A. Dickens thought the detection ended, to ensure that an effective trap of trailing points positioned within 50 yd. in the form of flank protection was avail­ ahead of the signal at the exit to a route able if an overrun occurred. did not seem necessary, since two overruns Changes in overlap must be obtained would have to occur simultaneouslv before by points moving in the correct sequence. a dangerous situation arose. A~ similar This did not imply that the other points situation was already accepted with shared must move in turn. It meant that the overlaps. rest of the new overlap must be proved With automatic overlaps, changes in before the first facing points were moved overlaps must always commence with to it. If there were several other sets, points furthest away from the protecting they could move simultaneously if points signal so that the first facing points were interlocking permitted. the last to move. Therefore, an appreciable Failure of the first facing points in time lag might occur before an alternative the overlap was an agreed hazard in the overlap became available in concentrated swinging overlap, but again it required point and crossing locations. When the both an overrun and a points failure to alternative overlap finally became avail­ be a potential source of danger. It was able, the key facing points might not accepted both by the M. of T. and by throw fully to the new position due to an the operating staff, and with these obstruction between the toe of the switch assurances the signal engineer should and the stock rail. This situation might accept it and use it to obtain the maximum be considered undesirable if a change operating flexibility. So far as he had in overlap were made whilst a train was been able to ascertain, there was no record approaching the signal at the exit of the of any fatality directly caused by this route. possibility.