Updated 09-IX-2004
Mercury Vapour
Mercury Pressure
Mercury Spectrum
Lamp Nomenclature
Timeline of Developments
Mercury Vapour
J.T. Way
Küch and Retschinsky
MA Medium Pressure
The first lamp
The first installation
Lamp developments
Striking the discharge
Operating characteristics
Glass technology
Electrode technology
Production methods
MB High Pressure
MC Low Pressure
MD Water-Cooled
ME Super Pressure
UHP Ultra High Pressure
Mercury Vapour
Fluorescent Coated Lamps
Tungsten Ballasted Lamps
Lamp Electrodes
Additives to the Arc
Electrodeless Designs
Future Developments
Mercury Vapour
High Pressure Circuits
Low Pressure Circuits
Electronic Operation

MA Lamp Developments

The early days of MA lamp production was a time of intensive research and development, with work concentrating in two separate fields - firstly, methods had to be found to de-skill the production of the lamps and bring their price down.  The principal developments on that topic are explained in the section on lamp manufacture.  Secondly, there was a tremendous drive to further increase the performance of the lamps in just about every aspect - these being the subject of this page.

Introduction of New Lamp Wattages

In view of the tremendous luminous efficacy of the first 400W MA lamp, customers wasted no time in requesting lamps of different wattages to extend the range of applications.  A 250W version soon followed and by 1939 a larger 650W model had also been brought into production.  But customers were not satisfied with these ratings alone, what they really wanted was a much lower wattage so that the mercury lamp could be introduced to replace ordinary tungsten lamps.

To meet this demand lampmakers created the 150W MA lamp in 1936, but the reason for the delay soon became obvious.  Its efficacy was considerably reduced for two reasons.  Firstly, when smaller lamps are made, the volt drop across the arc should really be allowed to decrease but this could not be tolerated because it made brought with it an increase in lamp current.  This made the ballast very inefficient due to increased Joule heating of its windings when run at such a high current.  The lamp volts had to be held inconveniently high and a reduced current was required to make the lower wattage possible, this decreasing the power loading in the arc.  Figure 23 shows how lamp efficacy falls with reduced loading.

Figure 23 - Lamp Efficacy vs Power Loading

Secondly, the electrical losses occurring just in front of the electrodes are of a similar magnitude whatever the lamp wattage, so although the wattage dissipated in the arc column was decreased, the electrode losses remained substantially the same, meaning that these losses become more significant for the smaller lamps.  This phenomenon is explained more fully in the section on lamp electrodes.  As a result luminous efficacy always becomes progressively worse for lower wattage lamps of any type, this rule being applicable to all discharge lamps.  The 150W MA lamp was not a commercial success, and enjoyed only a few years in production.  The introduction of lower wattage mercury products had to wait a few years until the development of the MB style lamp made it feasible with higher efficacy, as described in the next section.

Meanwhile manufacturers concentrated on developing the high wattage market, where efficacy of course becomes higher for larger lamps.  A 1000W lamp joined the range in 1939 (this was a bare arc tube type), and in the USA a very long high voltage 3000W lamp soon followed (type H-9).  It operated at 535V with a lamp current of 6.25A, the arc length being 1220mm and the tube diameter 30mm.  Essentially it can be thought of as eight 400W arc tubes connected together end-to-end, the wattage being slightly lower than 8 x 400W on account of the fact that there are only power losses from two electrodes, rather than from the sixteen electrodes that would be used in eight 400W lamps.  The voltage also was lower to account for the fact that the volt drop of only one rather than eight pairs of electrodes was present. Its rather unwieldy dimensions confined it to industrial lighting areas with very high ceilings, but its length was found to be an advantage in photo-printing shops.

In the UK attention was directed at high wattage lamps of more compact dimensions and 2500W types were employed in many trial installations, but were never marketed on a large scale.  These had a high efficacy of 52lm/W derived from a much more compact arc 400mm in length with a tube diameter of 65mm.  The lamp voltage was 145 Volts and the current 18.5 Amps.  Although the arc bows upwards about 1/2" above the tube axis when used horizontally, it is prevented from touching the discharge tube wall by the field produced in the return lead which is affixed to the top of the arc tube.  An experimental BTH 2500W lamp for industrial floodlighting is pictured in Figure 24.

Figure 24 - BTH Experimental 2.5kW MA/H Lamp

Some of the more common technical specifications for the range of British MA lamps is detailed in Table 1 below.  All figures relate to 1950 lamps designed for a 230 / 240 V A.C. mains electricity supply.  Source - Electric Lamp Manufacturers Association, Electric Discharge Lamp Sub-Committee, Internal Agreed Data for Group IX Lamps.

Lamp Type













Lamp current 1.25 2.15 3.25 5.5 7.5 18.5
Lamp voltage 120-140 135-182 140-165 150-170 140-165 140-150
Luminous flux 4,800 9,250 16,800 31,200 50,000 130,000
Efficacy 32 37 42 48 50 52
Arc length 88 135 165 203 292 390
Lamp length   290 330 430 335 580
Lamp diameter 38 48 48 65 43 65
Cap type E27s E40s E40s E40s P28s x 2 G38

Table 1 - Some Common MA Lamp Specifications

Operating Position

The first MA lamps were strictly classified as the type MA/V, the V implying that they were suitable only for burning vertically in the cap-up position.  A similar MA/D lamp was also made available, this type being suitable for vertical cap-down use.  The difference was quite simply that the arc tube was inverted, such that the tip-off where its exhaust tube is sealed off from the vacuum system would always be on the upper end of the arc tube.  The reason for this is because temperatures can be a little lower inside this tip-off area, and mercury can condense here with the effect that it is not all vaporised into the discharge so the the lamp might not run up to full power.  Because heat rises, it is prudent to locate this tip-off at the uppermost end of the arc tube where it is less likely to present a cold spot and interfere with lamp performance.

Vertical operation is rather inconvenient though on account of the complex luminaire designs that are required to redirect the light and send it down onto the horizontal plane, especially for streetlighting lanterns.  Horizontal operation is much more desirable, but could not be tolerated.  In the MA lamp the central cord of the arc is at a very high temperature while the outer periphery is much cooler, and this temperature differential sets up strong convection currents.  In the horizontal position these are sufficiently strong to cause the arc to be swept upwards at its centre, bringing it into direct contact with the wall of the discharge tube.  The elevated temperatures this would cause could not be tolerated, since the discharge tube glass would soften and blister outwards on account of the positive internal pressure.  Lamp failure would quickly ensue.  A photograph illustrating this upward bow is shown in Figure 25.

Figure 25 - Natural upward bow of the arc in a lamp not designed for horizontal use

Since the arc itself is in effect a flexible electric conductor, and all conductors are influenced by the presence of a magnetic field, a system was conceived by BTH whereby the arc could be forced back down near to the centre of the discharge tube by an external electromagnet.  By wiring this electromagnet in series with the lamp the alternating field it produces is of course automatically brought into phase with the discharge current, and by correctly positioning the deflection coil it is most effective at preventing the arc from touching the discharge tube wall.  This kind of magnetic deflection proved very popular in Great Britain and thereafter, the vast majority of streetlighting lanterns switched over to horizontally burning lamps.

Typical deflectors only consumed about 3 to 5 Watts but had to be precisely located to be effective and soon came to be seen as a cumbersome and heavy addition to many luminaires.  Lampmakers therefore directed their attention to making a new lamp which could burn horizontally without the deflector, this resulting in the MA/H type.  By reducing the quantity of mercury per unit of arc length, the speed of the convection currents in the arc tube are reduced and the effect of this is to minimise the upward bow of the arc.  It was a combination of this and an arc tube 2mm larger in diameter that led to the introduction of the MA/H lamp.  However the lower mercury vapour pressure in these designs and their lower volt drop meant that their luminous efficacy was about 10% less than for vertical burning types.  This sacrifice was felt to be worthwhile in view of the fact that simple low-cost luminaires could then be created, which made up for the loss of lamp performance through the ability to employ more efficient optical systems.

In the early 1950's a new kind of lamp was placed on the market, this being the MA/U designed for universal burning without any deflection coil.  It was made possible through the development of a new glass type having a softening temperature approximately 50°C higher than the original kind.  In addition the arc length was slightly shortened and this reduced the extent to which it would bow upwards.  Consequently it could be burned horizontally without fear of overheating.  The quantity of mercury per unit arc length was almost as much as in the first MA/V types so efficacy was high, and as a result it was capable of superseding both previous types of lamp.  Horizontal efficacy was still 10% lower than vertical due to convection losses, but it matched the performance of each of the dedicated MA/H and MA/V lamps and was therefore able to supersede them both.

Improved Colour Lamps

Since the inception of the mercury lamp, attempts were made by numerous methods to improve its colour rendering properties.  Right from the start the GEC had considered adding various other metals to the mercury arc which would provide some radiation at the red end of the spectrum, and a lamp containing cadmium and zinc in addition to mercury was trialled.  This increased the red ratio of the lamp from 1% to about 2%, effectively doubling the small amount of red light present.  Such lamps were never marketed though in the MA format - firstly, these additive metals radiated many lines in the UV and IR parts of the spectrum and the efficacy of the mercury discharge was reduced by some 15%.  In addition early lamps made with these additions suffered rapid blackening of the arc tube and a useful service life could not be attained.

Attention was therefore focussed on a fluorescent mercury lamp instead, in which the inside of the outer bulb was coated with a fluorescent substance.  Ideally the material should absorb ultraviolet wavelengths and re-radiate them in the red part of the spectrum to improve the colour rendering.  In 1937 a so-called MAF lamp was placed on the market, based on a fluorescent coating of zinc cadmium sulphide.  This lamp is described fully under the section relating to sulphide phosphors.

An alternative and highly innovative solution to the problem of colour correction was presented in 1935 by Dr. J.N. Aldington and Mr. W.H. Le Maréchal of the British Siemens Bros. company (of no connection to Siemens of Germany).  Thirty years earlier Peter Cooper-Hewitt had proven than ordinary incandescent lamps of an appropriate rating could be successfully employed as a ballast for the mercury discharge, and that their excess of red light perfectly complemented the bluish light of the mercury discharge.  But it was the aforementioned Siemens men who were first to combine arc tube and filament all within one outer bulb though, creating the first mercury lamp which did not require an external ballast and which also had excellent colour rendering properties.  Details on these so-called mercury blended lamps are located here, and an actual example may be seen by clicking this link.