Annual Conference at Folkestone
September 6th-9th, 1937

Abstract: Descriptions of lanterns and equipment displayed by Philips Lamps, Ltd., The Brighton Lighting And Electrical Engineering Company, Ltd., Engineering And Lighting Equipment Co., Ltd., Holophane, Limited, Wardle Engineering Co., Ltd, REVO Electric Co., Ltd., The Electric Street Lighting Apparatus Co., British Foreign and Colonial Automatic Light Controlling Co. Ltd., The British Thomson-Houston Co. Ltd., Bromford Tube Co., Ltd., Gas Meter Company Ltd., The General Electric Co., Ltd, Foster And Pullen Ltd., The Horstmann Gear Co., Ltd., James Keith And Blackman Co., Ltd., C. H. Kempton and Co., Ltd., Lighting Trades, Ltd., Metropolitan Pipe And Pole Company Ltd., Metropolitan Vickers Electrical Co., Ltd., Walter Slingsby and Co., Ltd., William Sugg And Co., Ltd., Concrete Utilities Ltd., Edison Swan Electric Co. Ltd. and Siemens Electric Lamps And Supplies Ltd.

Adverts: The General Electric Co., Ltd, Bromford Tube Co., Ltd., REVO Electric Co., Ltd., Philips Lamps, Ltd., The Electric Street Lighting Apparatus Co., Holophane, Ltd., The Brighton Lighting And Electrical Engineering Company, Ltd., Whelmer Gas Mantle Co., Metropolitan Pipe And Pole Company Ltd., James A. Jobling And Co. Ltd., Gas Meter Company Ltd., Gowshall Ltd., Falk, Stadelmann and Co. Ltd., Foster And Pullen Ltd., Venner Time Switches Ltd., Metropolitan Vickers Electrical Co., Ltd., The Horstmann Gear Co., Ltd., James Keith And Blackman Co., Ltd., C. H. Kempton and Co., Ltd., >Lighting Trades Ltd. and the Welsbach Light Co., Ltd., Walter Slingsby and Co., Ltd., William Sugg And Co., Ltd., British Commercial Gas Association, Concrete Utilities Ltd., Edison Swan Electric Co. Ltd., Siemens Electric Lamps And Supplies Ltd., Automatic Light Controlling Co. Ltd., W. Parkinson and Co., Crompton Parkinson Ltd., Newport And South Wales Tube Company Limited, British Electrical Development Association, Mosers Ltd, W. T. Henley's Telegraph Works Co. Ltd., Gas Light And Coke Company and The British Thomson-Houston Co. Ltd.

C. I. Winstone
Thursday, September 9th, 1937

Discussion published in: Public Lighting, Vol. 2, No. 7. September 1937

Mr. G. H. Wilson B.Sc. (Eng.), A.M.I.E.E.
Research Laboratories of The General Electric Co., Ltd., Wembley, England

Keywords: Lighting: Colour, Lighting: Distribution, Lighting: Lamps, Lighting: Luminaires, Lighting: Theory

Thursday, September 9th, 1937

Discussion published in: Public Lighting, Vol. 2, No. 7. September 1937

Abstract: "In the first part of the paper results are given of recent investigatory work on the planning of installations. The production of natural appearance under artifical lighting is dicussed and the influence of post position and height on the appearance of an installation, as shown by experimental results, is described in detail. Constants are given for an optimum installation as deduced from the results of experiemental work. The second part of the paper deals with new designs of lighting unit for existing lamps and new developments in lamps and in fittings for use with them. Details are given of the new horizontal burning 250-watt and 400-wat H.P.M.V. lmaps, the 80-watt and 125-watt H.P.M.V. lamps and the 400-watt luminescent powder lamp. Particulars of the the design and performance of these new lamps in recently designed fittings are also given."

Part 1 - The Lighting Installation

1. "Natural" Artificial Lighting
Fundamental research is continuing into factors which affect visibility and factors which give it a pleasing and natural appearance. The paper is concerned with those characteristics that produce a pleasing and natural appearance.

2. The Influence Of Post Position, Mounting Height And Power Of Sources On The Appearance Of An Installation
J. M. Waldram devised a series of experiments (which have not been previously published): aim to determine to what extent one lighting unit reveals part of the highway and its surroundings in a natural form; then find the arrangement of lighting units in a complete installation. Two types of lighting unit investigated: non-cut-off asymmetric (250W H.P.M.V. lamp bowl refractor) and cut-off (400W H.P.M.V. 75% cut-off). Fittings with intermediate distributions are now being studied. "Standard Representation" pictures taken according to R. G. Hopkinson.

(a)Non-Cut-Off:

Experiments On The Effects Of Post Position: The bright area in the carriageway has the characteristic "T" shape associated with this lighting unit on a worn carriageway. Slides of distribution were projected on a screen, and units at different distances were superimposed on it to determine the relationship between the distance of the fitting from an observer with the effective angular width of the bright area formed.

Post Positions On Bends: This had led to new ideas for planning of post positions on bends i.e. at angles which allow the bright patches to merge together, and on the outside of the curve.

Planning On Straight Roads: This is comparatively easy for the light sources in one row; but there are other considerations: if the sides of the road are not to apepar dark, then the units must be mounted on or near the kerb light. Central suspension non-cut-off distributions on all but narrow roads of less than about 25' in width will tend to leave dark regions near the parts of the highway to the right and left of the observer. If the spacing is excessive, then bad shadows are formed by large vehicles; the narrow tail of one bright area on just touches the broad "T" head of the next bright area; and an unduly large dark triange is left between them and the kerb. It has been found by experience that if the spacing per row does not exceed 300' then the highway will have a satisfactory natural apperance.

Arrangement Of Sources: Economy will dictate a spacing between spacings as great as possible, but some consideration is necesary of the relative positions of the posts on two sides of the road. A staggered (or "diagonal") is best so that the maximum recommended spacing of 300' per row becomes 150' between fittings.

Width Between Rows Of Sources: By multiplying the tangent of the effective angular width of the patch and the viewing distance, the maximum width in feed is obtained between two sources opposite one another at the distance concerned. For distances of 100', 200' and 300', the maximum width between pairs of opposite sources is 25', 24' and 21'. As a guide for general planning, it is fairly safe to say that the width between rows of sources should not exceed 30'. Kerb mounting should be satisfactory and is desirable as it improves the kerb visibility by increasing the contrast of the dark vertical face of the kerb and its lighter top. For wider roads, an overhang is desirable so that the width between rows of sources of 30' is not exceeded. For roads exceeding 40' to 45', a third row of sources in the middle of the road is desirable.

Effect Of Change Of Height On Appearance Of A Given Installation: This has been investigated by Waldram by using the projected image method. Diagrams were prepared giving contours of brightness.

(i) Effect Of Height On A Single Bright Areaa: The effect of reducing the height causes a reduction of the size of the area rather less than in proportion to the reduction in height; and the central brightness has increased.

(ii) Effects Of Height On Straight Roads: Decreasing the mounting height increases the darkness of the darker areas of the road: this is caused by the shrinking of the bright areas and the increase of brightness of the maxima. Further experiments took place with multiple sources with mounting height decreased to 18', spacing decreased to 100' and width decreased to 30': all showed a decrease in brightness of the road surface.

(iii) Effects Of Height On Curves: The effect of lowering the sources to 18' while retaining the same curvature as was used at 25', is to decrease markedly the brightness in the dark regions to one half or one third of its value at 25'., and to increase markedly the patchiness. If the spacing is now reduced, the eveness of the brightness can nearly be restored, but the extent of the bright region is reduced. It is interesting to note that a reduction in spacing with the low sources becomes much more effective in curved than in straight roads.

(iii) Effects Of Mounting Height On Glare: The effect of reducing the height of the fitting is to increase any tendency towards glare.

(b)Cut-Off: Experiments took place with a unit in which the maximum intensity occurred at an angle of 70° to the downward vertical with a rapid fall off in intensity above this angle. A 400W H.P.M.V. lamp was used and Standard Representation photographs were again taken. With spacings greater than 120', a dark band was found to occur across the road (which suggests a dip or a rise in the surface which is actually not present.) The spacing and height are almost directly proportional and an increased spacing over 120' could therefore be secured by increasing the mounting height above 25' but this would introduce mechanical difficulties in many situations. By raising the angle of the cut-off of the units, some increase in spacing could again be secured, but then the units would lose much of their virtue as cut-off units.

Position Of Units In A Cut-Off Installation The permissible spacing of 120' for a mounting height of 25' is strictly limited in any row of fittings. For the most effective lighting, central suspension must be used. (This will work well for both dry and wet road surfaces.

(c) Relative Mertis Of Cut-Off And Non-Cut-Off Installations: For a given light output per unit length of road, a good cut-off installation will not give such high road surface brightness or, on the average, such high contrasts between objects and their backgrounds as a good example of the non-cut-off type. On the other hand, the low intenstiy of the fittings directed towards an observer at normal angles of view, providing the fitting is well designed and focussed at a low angle, is considered by many an attractive feature because the sources lighting the road are inconspicuous. The widespread distribution, the reduced brightness of the carriageway and the absence of bright sources cause the side-walls, building fronts and hedges to appear well illuminated by comparison with a non-cut-off installation. From the point of view of economical street lighting, the important point is that units with an effectively low cut-off must be spaced close together if light and dark cross-banding of the carriageway is to be avoided, and this spacing of 120' cannot be exceeded. This spacing is lower than that frequently employed and will necessitate therefore increased capital and running costs.

(d) Power Of Sources: There is now general opinion that installations using 500W GLS, 250W HPMV and 150W SO can give satisfactory results. There is less doubt about installations which employ 750-1000W GLS or 400W HPMV especially under wet conditions. Installations using 300W GLS will frequently not give satisfaction. If an average spacing of 150' is assumed, the initial light output of the lamps to give satisfactory results is 5,000 to 13,000 lumens per 100' of road.

(e) Practical Conclusions Experimental work has been carried out under dry conditions. Conclusions correspond in some particulars with the recommendations made in an earlier paper [4] and the Interim Report of the Departmental Committee on Street Lighting by the Ministry Of Transport:

(i) Arrangement of Light Units (Non-cut-off):

On straight roads spacing not to exceed 300' per row, i.e. 150' in a staggered installation.
Units to be arranged on both sides of the carriageway, unless the carriageway is less than about 25' in width, when central suspension can give satisfactory results.
On bends, the units to be situation on the outside of the curve, their positions to be determined by the "departure" angle between them.

(ii) Arrangement of Light Units (Cut-Off):

A spacing of 120' should not be exceeded. Central suspension will usually be required.

(iii) Width Between Rows Of Sources On Staight Roads:

Width between rows should not exceed about 30'. On roads of this width, or less, kerb mounting is desirable. On wider roads, an overhang so that the width between sources does not exceed 30', will provide the best lighting. On roads greater than 40-45', more than two rows of sources are required to give good uniformity. An occasional central source will frequently provide adequate brightness at the middle of the road.

(iv) Mounting Height:

Not to be less than 25'.

Power Of Sources:
The initial light output of the lamps to be between 5,000 and 13,000 lumens per 100' of road.

(1) Lanterns For Existing Lamps

Lanterns For 250W And 400W Vertical Burning HPMV Lamps

The single-piece bowl refactor unit has been widely used with the 250W and 400W HPMV lamp. (Bi-Way Directional Refractor Lantern). A rectangular lantern and its optical system are also described (unknown lantern).

Horizontal Burning Sodium Lamp Units

Both cut-off and non-cut-off units have been available for some time.

A refractor plate unit is described (Lineal Lantern). Another unit is described with a different refractor plate (with vertical prisms) which gives a non-axial distribution.

(2) New Lamps And Their Lighting Units

The Horizontally Burning HPMV Lamp

Traditionally a method using an electro-magnetic deflector in conjunction with the lamp was used. The power loss in the deflector was slight so that the overall efficiency of the lamp was not materially reduced, but the cost of the lantern was increased by the extra component.

Further research had enabeld glasses with higher softening points so that with only a slight reduction in mercury vapour pressure, special lamps can be operated horizontally without a magnetic deflector. Lamps of this kind in the 400W size have been used for floodlighting for some time, but it's only recently that they have been developed to a stage suitable for street lighting in 250W and 400W sizes.

The proximity of the arc to the wall of the inner bulb produces a cooling action with affects the lamp voltage; therefore to ensure that the lamp wattage does not vary, choke tappings should be changed or specially designed chokes should be used.

The operation of the lamp in a horizontal position results in an increased loss of energy from the discharge to the wall of the inner bulb and a consequently small lost of efficiency i.e. 33 lm/W for the 250W version and 40 lm/W for the 400W version.

Lanterns For The Horizontal Burning HPMV Lamps

Lanterns already designed to take the horizontal burning lamp can now have their magnetic deflector removed, although a slight lowering of the lamp may be necessary to bring the arc into the focal position.

(i) Cut-off Type Lanterns

Silvered glass trough reflectors are employed for redirecting the light. By using this form of optical system it is possible to provide a sharp cut-off in the distribution below the horizontal. A curved silvered glass reflector is fixed to each side of the interior of the trough housing; the lampholder is fixed to one end, and in some designs can be raised and lowered to adjust the spread of light and cut-off angle; as normally used the cut-off angle is between 70° and 80° to the downward vertical. The maximum intensity is directed up and down the street, but there is no sideways concentration, the asymmetry resulting from the natural distribution of a linear source lamp placed with its axis across the street. On this account, the distribution is wide and sidewalks, building fronts, fences and other surfaces bordering the carriageway are will illuminated.

(ii) Non-Cut-off Type Lanterns

Does not have a short cut-off and plate refractors are used with horizontal prisms for the redirection of the light. In one design, a heat resisting glass cylinder is used to protect the lamp from rain splash.

80W and 125W H.P.M.V Lamps

Two new ratings have been added to the H.P.M.V. range. It is in the form of the inner tube in which the arc burns that changes have been made. In the larger lamps this tube is made of glass and a lower limit to its size is set by the heat of the arc. Lamps of smaller wattage have smaller efficiencies. Thereore with smaller lamps, the efficiency can be raised if the arc is formed in a tube smaller than is possible with glass and the mercury vapour pressure is thereby increased. Use has been made of fused silica, and so it has been possible to rate the 125W and 80W lamps at 40 and 32 lumens per watt respectively. The arc is considerably smaller than in the higher wattage lamps and its brightness is consequently some six or seven times as high. Partly on this account the lamps are normally made with internally frosted (Pearl) outer bulbs to provide a certain amount of diffusion.

Increasing the mercury vapour pressure improves the colour of the light. The percentage of red light has increased from 1% to 2% and the proportion of blue light is greater.

The overall dimensions are similar to those of the 200W and 150W tungsten lamps and the middle of the arc in the lamp is at a position corresponding to that of the filament in the tungsten lamps. Glass temperatures are such that the lamps can be burned in the open without risk of cracking from rain splash. In order to avoid accidental insertion of the lamps into BC lampholders connected directly to the mains, a three-pin bayonet cap is used. Development has also taken place in the design of chokes where the whole assembly is enclosed in a wax filled case. The advantages of this type of choke over the exposed type are the greater protection from weather and mechanical damage and the greater convenience in installation.

Lanterns For 80W And 125W H.P.M.V. Lamps

The size of the light source of the new low wattage mercury lamp is of the same order of the low wattage filament lamp although the light distribution slightly differ. In view of this, the new lamps may be used successfully in many fittings designed for filament lamps. Three example lanterns are shown: (a) one with a double-dome prismatic refractor, (b) a one-piece combined reflector and refractor and (c) a combined dome and bowl prismatic arrangement.

300 Watt Dual Lamp

A 300W dual lamp (a combination of the H.P.M.V. lamp and tungsten filament lamp) has just been introduced. In construction it is similar to the earlier lamp, the inner tube of the H.P.M.V. being mounted about the tungsten filament but the dimensions are smaller. No choke is required. The initial efficiency of the lamp is 21 lumens per watt.

400 Watt Luminescent Powder Lamp

Two versions of the 400W lamp will shortly be available. The primary light source is the inner tube of a special 400W mercury-cadmium discharge lamp designed to operated with the normal 400W chock coil. The outer bulb is internally coated with the luminescent powder. The isothermal bulb is designed to give the optimum performance where the lighting fitting imposes no restriction on the bulb shape; the tubular bulb is designed that it will go into normal discharge fittings. The efficiency is 37 lumens per watt. The red-ratio is now 5%-6% as compared with 1% of the normal lamp. When used in a diffusing fitting, the lamp is comparable with the common H.P.M.V. lamp. When used in a fitting with a redirecting optical system, the presence of the secondary luminescent powder source reduces the maximum concentration of the fitting. It reamins to be seen seen how much application these lamps will have for street lighting as they were developed for industry.

Conclusion: It has been explained how the results of experimental investigations on the planning of installations can facilitate the production of satisfactory lighting results with the maximum of economy. The development of new lamps and lighting units has been shown to have kept pace with the knowledge of installation design. These developments together should assist in a practical way, the erection of street lighting installations giving a satisfactoy performance.

Luminaires: Holophane Bi-Way Directional Refractor Lantern, Holophane Lineal Lantern, Holophane Standard Type Duo-Dome Lantern, GEC Pla-Frator Lantern and GEC Small Oxford Lantern.

References:
[1] Photographic Representation of Street Lighting Installations, R. G. Hopkinson. Trans. Illum. Eng. Soc., Vol I., No. 2, Feb., 1936. pp.19-31
[2] Some Further Aspects Of The Street Lighting Problem - Kerb Line Visibility, F. C. Smith, K. F. Sawyer and B. G. Winslow, A.P.L.E., Cheltenham, 1936
[3] J. M. Waldram, Illuminating Engineer (London), 1934, Vol. 27, p305.
[4] Wilson, Damant and Waldram, Jour. Inst. Elec. Eng., Sept. 1936, Vol. 79, pp 241-264.
[5] J. N. Aldington, Public Lighting, Vol 1, Number 4, December 1936, p106.
[6] J. T. Randall, Journal Of The Royal Society Arts, Vol 85, No 4398, March 5, 1937, pp 353-381.
[7] H. Warren, B.T.H. Activities And Developments, Vol 13, March-April 1937, pp. 37-53.
[8] S. English, Illuminating Engineer, Vol. XXVII., November 1934, pp. 352-361.

S. English, D.Sc., F.I.C., F.Inst.P. and E. Stroud
Research Laboratories, Holophane, Ltd.

Keywords: Lighting: Colour, Lighting: Distribution, Lighting: Luminaires, Lighting: Theory

Thursday, September 9th, 1937

Discussion published in: Public Lighting, Vol. 2, No. 7. September 1937

What was good streetlighting? and how was such street lighting best obtained? The aim of street lighting was not to produce uniform illumination along the roadway; nor was it to produce high road surface brightness; nor was it to eliminate all glare; neither was it to produce a uniform glare: none of these features individually produces good street lighting, but a suitable combination of them and other points, does. When ideal street lighting is obtained it wil not be judged by reference to any of these factors, but by the ability to see clearly, distinctly and easily along the roadway.

Visibility
The real reason for street lighting is to enable road users of whatever class they belong to, to see clearly and easily whatever they need to see.

To assist with this, it is necessary to have some means of measuring the visibility itself. It is obvious that the actual visibility of obejcts on the lighted roadway must be determined, and the eye must be functioning under the conditions imposed by the lighting installation. For this reason it is not possible to use an instrument which shields the eye from any glare or which limits its field of view. Visibilities must be determined by the unaided eye.

When seeing an object on the roadway there are numerous factors involved: size, colour and movement are attributes of the object itself; its apparent brightness, the brightness of the road surface, the general surrounding illumination and any glare present are attributes of the lighting installation. The latter are considered by the lighting engineer and these have to be welded into a satisfactory lighting scheme. In towns, general illumination is more important; on main roads between towns, then silhouette vision becomes more important. For road brightness, the porportion of light specularly reflected increases as the angle of incidence of the light on the surface increases.

Road Surface Reflection And Glare
Some investigators have measured the reflection factor of specimens of road surface, but these are not of much assistance, as the normal angle of view is substantially constant and is between 0° and 3° to the horizontal.

The brightness of a dry macadam road surface viewed at 3° below the horizontal only increases quite slowly until the light is incident at angles exceeding 83° from the normal, and the surface brightness increases very quickly when light is incident at 85° to 88° from the normal. Light incident on a road surface at these angles materially helps in building up brightness contrasts, and would assist visibility, but the glare simultaneously produced by these high beams reduces the sensitivity of the eye to such an extent as to destroy any advantage gained.

Glare does not seriously impede vision until the angle between the source and the line of sight exceeds 75° from the vertical. Therefore glare becomes detrimental to vision at distinctly lower angles than road reflection. However, the idea has been disseminated recently that glare does not matter when the road surface has a high brightness value; therefore this was to be confirmed by using an experimental road model.

Model Road Tests
The model road was 100 feet long and by means of a large plate glass mirror at one end its apparent length was doubled - it corresponded to an actual road length of 333 yards. It was 7 feet wide (corresponding to 35 feet) and was flanked by two pathways of dark stone colour, each 1 foot wide (corresponding to 5 feet). The surface was dark grey asphalt having a reflection factor of 8%. Five small lanterns were carried at a filament height of 4'2½" and a spacing of 25' (corresponding to 21' and 125' respectively) and giving a spacing height ratio of 6:1. The lanterns were fitted with 40W ring-filament clear bulb lamps housed in non-axial asymmetric 7" dome refractors, arranged in staggered formation with an overhang of 1' (corresponding to 5' from the kerb). Everything was kept constant except the angle of the centre of the beams given by the refractors which gave beams at 70°, 75°, 80° and 85°. At each of these beam inclinations the following series of readings were made:

• Illumination readings were taken at 15 regularly spacing points between the third and fourth fittings from the end and also at the two B.S.S. test points. From these, diversity ratios were calculated and isolux diagrams were prepared.
• The visibility along the roadway was measured. A brass disc, 5" in diameter, was provided with a projection 1" square on a point on its periphery, which could be set in one of 12 positions (corresponding to a clock face). The disc and its projection were painted a dark matt grey with a reflection factor of 8% corresponding to the average reflection factor of the dark clothes made by most men. The disc was placed on each of the 15 regularly spaced spots and observed from 2' behind the first lamp and 1'9" away from the near kerb. The eyes were kept 12" from the ground corresponding to 5' on an ordinary road. The area covered by the objservations extended from 52' to 77' corresponding to 87 and 128 yards respectively. The measure of visibility was obtained by spinning the disc and if it was possible to see the projection. Ten observations were made at each of the 15 test points and the results summarised to indicate the visibilities along the near side of the road, along the middle and along the offside. A full set of measurements was made with the units exposed to view as in ordinary practice, and another set of readings was then taken with the light sources screened to eliminate all glare. (The shield cuts of the view of the units).

Discharge Lamps
There is very little real data on the comparison of these two sources between themselves and against filament lamps. "Before" and "After" photographs give an exaggerated impression of improvement as the discharge lamps have higher lumen outputs than filament lamps and older schemes tend to be replaced by schemes having a smaller space-height ratio.

A street in Westminster was used for the experimental work.

The street had an asphalt surface with a fair degree of polish. The experimental length embodied 5 lamp posts arranged in staggered formation at a uniform spacing of 153'; the width of the road was 38'6". Between the third and fourth lamps from one end, 15 regularly spaced spots were painted on the road coressponding to those used on the model road - B.S.S. test points were also marked. Similar tests were made to those on the model. In addition an attempt was made to measure the glare: a pair of photocells was used and the reading was taken of light incident on the cells from the test installation, comprising the light received direct from the fitts and that reflected from the road surface and the sides of the houses. The direct light from the lamps was then cut off by suitable shields and then a measure of the reflected light obtained. By difference, the intensity of the light received direct from the fittings was deduced; the ratio of the direct to reflected light intensities being called the glare index.

The standard lighting was provided by 500W filament lamps (8470 lumens) in dome refractors mounted at 23'6", in staggered formation, and overhung 1'9". Test point readings were .117 foot candles, diversity factor 21 i.e. a good Class E installation.

The 125W mercury lamps (5000 lumens) were tested in four types of fitting. Test point illuminations fell to 0.52 foot candles and 0.85 foot candles i.e. a class F installation. Diversity factor was between 7.5 and 11.

• In smooth exterior asymmetric dome refractors giving a maximum beam at 75° and a rather wide vertical light distribution.
• In prismatic axial asymmetric reflectors giving a beam maximum at 65° and a rather wide curve.
• In non-axial asymmetric silvered glass reflectors giving a cut-off of the source at 80° and a strong maximum beam at 70° with a somewhat narrower beam.
• In an enclsoed reflector-refractor lantern giving a beam maximum at 72.5° with a virtual cut-off at about 85°

The 150W sodium lamps (9600 lumens) were tested in two types of fitting:

• In an inverted white enamel trough giving a weak beam maximum at 55° and having a cut-off at 85°. Test point readings gave a class E installation with a diversity factor of 19. Visibility was low at 52%.
• In a refractor panel lantern giving wide beams vertically with maxima at 75° and very broad beams horizontally. Test point readings gave a class D installation and the diversity fell to 7 to 1. Compared with the 125W mercury lamps, the sodium lamps in the refractor panel fittins gave B.S. test point readings approximately three times as high with a rather better diversity factor. Visibility was better at 59%.

The 400W mecury lamps were tested in three types of fitting. The illumination was Class D. In the case of the vertical burning lamps, the appearance of the road was very impressive, and was such as would immediately appeal to the members of a lighting committee. A visibility of 80% was expected but only 47% was recorded.

• Vertical burning lamp in 160° asymmetric refractor bowl with beam maxima at 75°
• Vertical burning lamp in 180° (axial) asymmetric refractor bowl with beam maxima at 75° vertically and somewhat narrow ones in the 75° cone.
• Horizontally burning lamp in refractor panel lantern giving beam maxima at 75° and very broad beams horizontally. These gave the highest general illuminations, the test point figures being .32 foot candles and .34 foot candles.

Visibilities
Visibilities near the near-side and off-side kerbs was poor for all lamp types and all lantern types. This was due to (a) the 5' overhang, (b) the well polished surface of the road and (c) it's camber close to the kerbs. Reducing the overhang to 1' reduces the problem considerably. It is noted that the Interim Report of the Ministry of Transport COmmittee on Street Lighting is oging to adopt a 6' overhang. All the tests were done when it was dry; if it was wet, the visibility in this area would've been zero. In conclusion: a small or no overhang is much preferable to the recommended 6' overhang.

One would expect the higher illumination values of the larger discharge lamps to be accompanied by better visibility values. However these were unexpectedly low. This low figure was due to difficulties inherent in the test road and (a) the 160° light distribution combined with the 5' overhang which left the sides of the roadway somewhat dark and (b) the contrast beween the brightness of the surface of the disc and the reflected brightness of the road surface was very low down both sides of the road. The possible reason for the lack of visibilty is that the roadway and houses had such a large volume of light thrown on them that details were being seen by direct vision: this was excellent for town street lighting and the pedestrian but poor for the motorist who wanted to see obstructions in silhouette. As a result there were positions along the roadway when direct vision was just changing to silhouette vision and at such points objects became invisible.

To try to overcome the relatively dark lanes along the sides of the carriageway while still maintaining the 5' overhang, a refractor giving an axial asymmetric light distribution was tried. Although lack of contrast was still present, it was not so pronounced as with non-axial asymmetric fittings. It appears certain that even on wide roads, if a 5' or 6' overhang is adopted, better visibilty is obtained by using axial fittings than by using non-axial fittings, more particularly as this increased visibility is largely due to an improvement in the near-side lane.

150W sodium lamps and horizontally burning mercury lamps in axial fittings also give good results. These give a better spread of illumination across the roadway, higher test point figures and visibility figures than vertically burning lamps - this is probably due to better light control in the vertical plane and to the much wider beam spread horizontally.

Glare
The effect of glare was measured directly by taking visibility readings with the light sources exposed and with screens interposed between them and the eyes. (It must be remembered that the refractor fittings have attempted to limit the main beam to 75° with a virtual cut-off above this angle.In the case of opaque fittings, the amount of light varies enormously according to the position of observation relative to the nearest fitting. At the standard observation position, the direct light from the next fitting was very small, and consequently the glare index derived from it was low, but had the observation position been halfway between two lamps, the direct light would've been much more intense and the glare index would've been higher. The figures recorded for the glare index in the case of opaque reflectosr were very misleading.

The actual intensity received from opaque lanterns was lower than that received from refractor lanterns, but with these latter there was always a number in the field of view. They also produced a much brighter road surface, and the combination of these two formed a kind of light background against which the nearest unit stood out, but not so prominently as did the opaque reflector unit against a black background. The constantly varying nature of the glare was disconcerting. The eye therefore never had a chance of settling down but was being called upon to adapt itself to constantly changing conditions. From this point of view, the refractor installations with their more or less constant glare were definitely preferable to the installations using opaque fittings.

Motor Road Lighting
In the case of motor roads, pedestrian traffic can be neglected, and the light available used to give the best possible results from the point of view of motorists. The problem has been simplified by the adoption of dual carriageways; any light that is emitted by the lantern in the same direciton that the motorist is travelling is wasted from the point of view of assisting vision. It is worse than wasted since it acutally helps to make vision more difficult by lighting up the face of any obstruction that may be present thus destroying the contrast.

Therefore it appears that the fitting for lighting unidirectional motor tracks should direct as much light as possible towards the oncoming driver and as little as possible in the direction of travel - a scheme which is just the opposite of that which has been recommended on the Continent. It is obvious that in applying such a scheme as this the greatest care will have to be exercised in order to avoid dangerous glare.

An experimental unit was produced by cutting direcitonal dome refractors and directional street lighting reflectors in half and putting the halves together in such a way that they both project the light in the same direction.

Using the model roadway with space height ratio of 6 to 1 and central mounting, it was possible to get the surface practially uniformally bright when observed facing the directed light, while when looked at from the other end it was very patchy. Neglecting the readings for the offside of the road, the visibility of the near side and middle of the roadway was 81%, while for the middle of the road it was 96%. Relatively road uniform brightness can be achieved if the space/height ratio does not exceed 6. For the standard spacing of 150', it would require the height raising to 30', whilst for a spacing of 125', the height of 25' could be used.

The unidirectional scheme was tried on the test roadway in Westminster. The fitting was a refractor panle lantern with a horizontal burning mercury lamp, and in which the refractor panel on the side away from the direction of travel was replaced by a reflector panel. Viewed in the direction of the flow of traffic, the appearance of the road was very satsifactory - however, when viewed in the opposite direction the appearance was decidedly patchy. Visibility figures were better than any that had been obtained with previous test installations and in no test position did the disc merge into the background. The average visibility was 75% while the shielded visibility was of 77%.

By courtesey of the Brighton Lighting Company and the Lighting Engineer of Hove Borough Council, the scheme was tried on a two-track byepass road (Hangleton Road, Hove). The units were refractor panel lanterns but using 150W sodium lamps. The lamps were mounted in-line over each track and had an overhang of 6' from the central island kerb. The refractor panel away from the direction of traffic was replaced by a reflector giving a virtual cut-off at 12°. The appearance of the road in the direction of travel was very good indeed. Visibility readings were 100% down the near side, 100% down the middle, and 54% down the offside (this was poor due to a hedge growing down the centre island).

In all three tests, unidirectional lighting produced perfect visibility from the motorists' point of view.

Luminaires: Holophane Bi-Way Directional Refractor Lantern, Holophane Refractor Plate Panel Lantern, Holophane Refractor Plate Uni-Directional Panel Lantern.

References:
[1] Trans. I.E.S. (London). Vol II, No. 6, 1937, p. 68.
[2] Taylor, Proc. I.C.I., 1928, p. 53. (Paper on road reflectance).
[3] P. W. Cobb and F. K. Moss, I.C.I., 1928, Paper 9, p. 5. (Paper on reduction in visibility produced by glare). [4] Prov. Pat. No. 9621/37.

G. Keith

Keywords: Lighting: Installations, Lighting: Luminaires

Wednesday, September 8th, 1937

Inverted gas mantles operate on two principles:

• Air-gas mixture only contains a proportion of oxygen required for combustion, a secondary supply beign required to complete combustion of the gas on the surface of the mantle.
• Air-gas mixture contains the whole of the oxygen required for complete combustion.

It is practically impossible to produce a mixture of gas and air sufficiently rich in oxygen to burn completely within the mantle by means of the kinetic energy in the gas alone at the ordinary main pressure of 4" water column. Apart from the difficulty of entraining the necesary quantity of air, the mixture must have a pressure sufficient to overcome the resistance of the mantle, and the velocity of the flow must be high enough to prevent firing back.

Therefore some source of energy in addition to the low-pressure gas mains must be employed. High-pressure gas systems uses a compressor to increase the pressure of the gas to 80" water column. The high-pressure gas is distributed through independent pipes; this is a drawback, but there are also advantages: relatively small diameter pipes will carry relatively large volumes of gas, and any number of gas lamps may be simultaneously lighted or extinguished from a central point.

Other methods of obtaining high power lighting have been tried:

• The supply of air to the lamps through special pipes under pressure.
• Supply of carefully controlled mixture of air and gas in proportions below the explosive limit, known as the Selas system to light.
• In America, Humpreys produced a system in which a small electrically driven fan is applied to each unit. (Which has been trialed in Edinburgh).

Before the advent of the inverted gas mantle Lucas designed a lamp to give "intensified" light with the upright incandescence burner and for this purpose he used a long chimney to give a "pull" on the burner - this gave a more highly aerated combustible mixture which gave more light from the mantle. The upright Lucas lamp was very successful for use indoors but was sensitive to wind but was still used for public lighting in its day. The Lucas upright lamp was superseded by the inverted mantle and inverted gas burner.

Lucas proposed a lantern using the chimney-pull principle for the inverted gas mantle but it was not successful. Keith continued these experiments to discover if inverted gas lamps for outside use using the mitrailleuse burner head principle operated by chimney suction. All experiments failed outside as a slight wind upset the steadiness of the light. Success was only found when a different method of causing the secondary air to enter the spaces between the tubes was tried: it pushes the secondary air into the spaces instead of pulling it through.

Various diagrams of two different units are then displayed and details are given of their installation in parts of Folkestone. It is generally accepted that the light output of a good low-pressure gas, using gas at 500 B.T.U. is about 200 lumens per cubic foot, and the light output of a good high pressure lamp is 350 lumens per cubic foot. In independent tests, the new lamp was rated at 300 lumens per foot.

Luminaires: Keith & Blackman Standard Lamp, Keith & Blackman Column Lamp, Keith & Blackman City Column Lamp, Lucas Upright Lamp, Lucas Prototype Chimney Pull Suspension Lamp, Keith & Blackman Prototype Magnalux #1 and Keith & Blackman Prototype Magnalux #2.

The discussion after the paper led to a disagreement between Mr. W. J. Jones of the Electric Lamp Manufacturers' Association and Mr. K. F. Sawyer of the Gas Light and Coke Company. The resolution of their disagreement was published in Public Lighting #9.

Dean Chandler, M.Inst.Gas.E.
Chief Technical Officer, District, South Metropolitan Gas Company

Keywords: Lighting: Columns, Lighting: Control, Lighting: Distribution, Lighting: History and Lighting: Luminaires

Wednesday, September 8th, 1937

The gas burners and gas lamps which were in use a decade ago in many parts of thecountry were for the most part indifferently mde and badly maintained. This did much to discredit gas lighting and influence public lighting against it. In more recent times the position has been entirely changed, owing to the attention given to detail and to the skill and ingeniuty displayed by gas engineers an lamp makers. There does appear to be a tendency to look upon the gas light with disdain and as being icnompatible with modern street lighting requirements.

There is a widespread opinion that gas lighting is about to make its final bow and it will not be longer before it remembered only as "the light of other days." It has been stated recently that the passing of the gas light is a "dead certainty" but half a centure ago the death sentence of the gas light was solemnly pronounced (two cartoons are included). But threatened industries, like threatened men, live long.

A table shows the steady progress of gas lighting in the streets of South London over a period of years. A map also shows that most of the important main roads in South London are lighted with high pressure incandescence gas lamps.

The term "high pressure" is a comparative term only. With the introduction of the incandescence mantle nearly half a century ago gas pressure became at once a matter of great importance. At first 1½" of water column pressure was a suitable pressure for the Welsbach light, but it soon became evident that better results could be obtained with higher pressures. Conditions changed when the inverted burner and mantle were introducted - 3" of water column pressure was desired. The important stage in the development of high pressure (or "intensified") gas lighting was when the gas pressure was raised to about 8" of water column pressure by mechanical contrivances - the earliest being water-driven pumps - to increase the velocity, and therefore the knetic energy, of the stream of gas issuing from the gas nipple orifice into the mixing tube of the burner. The energy of the higher pressure gas jet entrained a greater volume of air than was possible with the lower gas pressure and "this resulted in a hotter flame and a greater intensity in the incandescence of the gas mantle."

In 1896 Somzee and Greyson of Belgium were the first to construct a water-driven compressor to increase the pressure of the gas to about 8" water column. In the following year Mr. Keith also introduced a water-driven compressor which raised the pressure to 8" water column. William Sugg used a gas-heated hot air engine to drive a gas compressor which worked for many years in the Tottenham district.

The Scott-Snell Lamp was an early form of a "high-pressure" gas lamp. Probably introduced in the late 1890s. It included a novel and ingenious gas pressure raising device. Provided in the tent of the gas lamp above the gas burner was a hot air motor, which was actuated by the hot products from the combustion of the gas. The gas on its way to the burner passed through a chamber which contained a "displacer," the operation of which raised the pressure of the gas to about 10" water column. At this pressure the gas was then conveyed by a pipe to the gas ejector provided at the base of the burner. The gas issued from the ejector at high velocity, entraining thereby a considerable proportion of the air required for its combustion. The gas pressure raising mechanism continues to operate so long as the gas continues to burn. The system was later modified so that the air required for the combustion of the gas was compressed to one pound per square inch, the gas being used at the normal supply pressure. The lamps were used in Parliament Street and the lighting of the old Vauxhall Bridge.

Similar systems by Lucas and Welsbach are also mentioned. The great step forward occurred when Mr. George Keith introduced the now famous Keith Inverted High Pressure Gas Lamp in 1908. This pre-heated the gas-air mixture before combustion. Keith also developed the Keith Rotary Gas Compressor to ensure full justice be done to the lamps.

In 1935, a new development occurred with the high pressure gas lamp. The light from a cylindrical high pressure gas mantle must be redirected; this redistribution could be effected by altering the shape of the mantle itself so that a greater proportion of the luminous output was radiated in the direction where the greatest intensity of illumination was required. This was adopted in the Supervia lamp and there are now 1,800 lamps lighting nearly 73 miles of road.

Lamps operating on a high pressure gas system can be turned on and off by the simple process of starting or stopping the gas compressor, which is of definite advantage and in certain circumstances of very great importance.

Can the lighting of the streets by high pressure gas be undertaken on a competitive basis? Recently two tenders for the lighting of a borough of considerable size were submitted to the appropriate authority for installing the same class of lighting. When the two tenders were opened they were within a few pounds of each other. Therefore the answer, by experience, is that high pressure gas is competitive.

Descriptions of the equipment for a high pressure gas installation are then given. This includes the lamp, brackets or central suspension system, and columns. Cleaning and maintenance work must be regularly carreid out and arrangemetns must be made for it to be done safely, sppedily and conveniently. The older method involved the use of a wheeled tower ladder, but this has now been abandoned due to the unwieldiness of the ladder, the interference with traffic and the element of danger. The lamps are cleaned on average once a week. Instead of the attendant going up to the lamp, the lamp is brought down to the attendant by a simple, ingenious and reliable mechanical arrangement. (This version was designed by Keith).

The high pressure gas system is arranged to give a gas pressure of 81" water column or 3 lbs. per square inch. The South London system was built up around the busy shopping centres, each local high pressure mains network being supplied from a small compressor station (this was due to the system being used for the outside lighting of shops only, before being extended for public lighting). All but three of the ten compressor stations are now interconnected. The new large Kennington compressor station, together with another large station, will eventually supersede the small local stations.

Appendix I gives details of the Keith Rotary Gas Compressor and Appendix II gives details of raising and lowering gear for suspension lamps.

Luminaires: Scott-Snell Scott-Snell Lamp, Lucas Lucas Lantern Lamp, Lucas Lucas Suspension Lamp, Welsbach Welsbach Lantern Lamp, Welsbach Welsbach Suspension Lamp and South Metropolitan Gas Company Metro Supervia Lamp.

References:
None

F. C. Smith, F.C.S., Assoc.M.Inst.Gas.E. and K. F. Sawyer, B.Sc

Keywords: Lighting: Luminaires, Lighting: Theory

Thursday, September 9th, 1937

The effectiveness of a street lighting installation is judged by the easy and precision with which it enables one to see. This is of primary importance; all other considerations are of secondary importance, although in practice the economic aspect cannot be ignored. It is seeing in the street and not seeing the lights that is the important matter.

There will be occasions when special emphasis will be required on one or more features of the installation such as occurs in thoroughfares carrying fairly dense traffic when vision is not always by direct silhouette. In these cases, each object wil be seen against a background of others, and each will then be distinguished by its surface details. In such circumstances much greater importance must be attached to vertical illumination.

It is clear that distribution, intensity, siting, mounting height, etc., all have their separate bearing upon the ease of seeing. Although some of these have a greater influence than others upon the final result, it is manifestly impossible to select any one and regard it as an adequate index to the effectiveness of the installation. The failure to appreciate this all-important fact has mainly been responsible for much of the adverse criticism of the present BSI Specification, and for the same reason many have come to regard the provision of a certain amount of illumination at stated points as the sole aim of a street lighting installation.

It must be remembered that specifications, as the bases of contract obligations, can only be concerned with those factors which are under the control of, or can be verified by, the contracting parties. Therefore it is not surprising that the BSI Specification makes no direct requirement in regard to such factors as brightness contrast.

• The primary function of a street lighting installation is to enable all users of the road to see clearly, quickly and comfortably.
• This can only be accomplished if the quantity of light falling upon the road and other essential backgrounds exceeds a certain minimum.
• As the utilisable light is increased above this minimum value the case of seeing increases, rapidly at first, then more slowly to a level which is adequate. Above this level the small increase in seeing ability resulting from an increase in the light used is no longer worth the additional cost involved. This economic level of seeing has been interpreted by the M.O.T. Departmental Committee in their Interim Report as that provided by the number and power of sources associated with a generous class "F".
• This latter statement must not be taken to mean that installations are necessarily of equal merit as far as seeing is concerned, if a similar quantity of light is provided per unit length or area.
• Glare must be reduced to a minimum since it is detrimental to seeing.

Maintenance Of Street Lighting Installations

All the gas lamps described in this paper have been designed so that maintenance can be carried out quickly and efficiently without disturbing the directive equipment.

A very high standard of maintenance can be achieved at a reasonable cost. Many gas comparies are now prepared to contract to maintain lighting installations within 20% of their initial values. Such good standards of maintenance do not involve excessive expenditure on replacements of mantles and glassware. The mantles have an average life of 2,000 lighting hours and replacement is due to mechanical damage rather than because of any marked fall in the light output. Similarly good heat-resisting glassware has an average life of two years.

Modern Gas Lanterns

There then follows descriptions of various gas lanterns and photometric data.

Luminaires:
Sugg London Lamp, Keith & Blackman High Candlepower Low Pressure Lamp, Kempton Majestic Lamp, Parkinson Maxill Lamp, Sugg H. V. Lamp and Foster & Pullen Arcturus Lamp.

References:
[1] Modern Views On Street Lighting And Their Relation To Visibility, F. C. Smith, F.C.S., A.M.Inst.GasE. and K. F. Sawyer B.Sc, A.P.L.E., London, 1935
[2] First Report Of The Joint Lighting Committee, Institution Of Gas Engineers, Publication 125, November 1935.

T. Catten, R. Maxted and G. S. C. Lucas

Keywords: Lighting: Theory

Thursday, September 9th, 1937

The choice of light souce and lighting unit, the mounting height, spacing, and the position of the units above the road surface, are all dependent on one another, and it is the adjustment of these various factors that is termed the planning of the lighting installation.

The primary object of the paper is to deal briefly with some of the more important factors governing the planning of street lighting.

The main body of the paper is given as two films.

The first film, "Planned Stret Lighting (Location Of Units) Parts 1 And 2" (25 minutes) deals with one or two fundamenal principles of street lighting and in particular the importance of the correct location of the lantern unit with respect to the road surface. Considerable attention is often paid to the choice of light source and to the lantern design, while the all-important question of the location of the lantern is ignored of left to chance.

The film deals with three important points:

• It emphasises the difference between indoor and road lighting in the quantities of light available in each case. Where the quantity of light is small, it is better to use the light to illuminate the backgrounds against which objects can be seen in silhouette.
• It shows the difference between the plan view of a lighted road and the appearance of the same road to an observer on the road surface.
• It shows how bright backgrounds are produced and the importance of the correct placing of the lanterns in achieving the desired results.

The second film, "Planned Street Lighting (Lantern Characteristics)" (12 minutes), shows how the characteristics of the lantern are linked with the road surface conditions and the layout of the installations.

References:
[1] An Analysis of Modern Views on Street Lighting and their Relation to Visibility, F. C. Smith, F.C.S., A.M.Inst.GasE, K. F. Sawyer B.Sc, A.P.L.E., London, 1935
[2] The Importance of Kinematical Factors in Roadway Lighting, L. J. Davies, M.A., B.Sc., R. Maxted, B.E.(Elect.), B.E.(Mech.), G. S. Lucas, M.I.E.E, A.P.L.E., London, 1935