Very soon after the invention of the MA lamp by Osram-GEC, Philips also took up the production of this kind of lamp in the Netherlands. Following this, Elenbaas made a landmark discovery in 1934 that once the mercury vapour pressure has exceeded a certain level, higher pressures do not have such a significant effect on increasing efficacy as power loading does (usually expressed in watts per cm of arc length). It was therefore surmised that once high pressure conditions had been achieved, it would be of crucial importance to reduce the arc length of the lamp in an effort to increase power loading, and hence build more efficient lamps. Luminous efficacy increases with power loading in the manner shown in Fig 36. |
Fig. 36 - Power Loading vs. Efficacy |
Low Wattage Lamp Requirements
Increasing the luminous efficacy was of significant importance for the simple reason that customers were crying out for mercury lamps in lower wattages than the 400W MA lamp which was first to become available in 1932. Smaller ratings of 250W and 150W were put on the market, but they had considerably lower luminous efficacies, to the point where the 150W version showed little advantage over a similarly sized incandescent lamp. The reasons behind this are explained below.
The first lamp was the 400W version and it had an arc voltage of approx. 130V, roughly half that of the 240V mains voltage in the UK. With this arrangement whereby half the mains voltage is dropped across the ballast, an efficient system can be made. Consider now making a 250W version of this lamp. It is clear that all other things being kept constant, the arc must become shorter to dissipate proportionately less power, and of course the arc voltage will also fall. We then run into difficulties that the ballast must drop a considerably higher voltage across itself, and this makes it very inefficient. So inefficient that such a system could not be marketed.
The volt drop could be restored to normal levels by reducing the tube diameter and increasing the mercury pressure, but this would have taken the arc tube temperature above what it could stand and posed a risk of melting. The aluminosilicate material was already running at its very limits in the 400W lamp. So to make lower wattage lamps, one had to be content with making the arc length much longer than it should have been, thus bringing its volt drop back up to similar levels as were found in the 400W lamp. With the voltage roughly the same, the discharge current then had to be reduced to make a 250W lamp, this causing a reduction in power loading, and a fall in efficacy roughly according to the variation shown in Figure 36. The situation was bad for the 250W lamps, and even worse in the 150W case.
Philips undertook many tests to try and greatly reduce the arc gap so that it would follow the curve in Figure 36 to higher efficacies. Spherical aluminosilicate bulbs were employed but to keep the wall temperature within its limits the bulb had to be quite large. Such large bulbs were physically weak and they could not endure the high mercury vapour pressures that were required to push the discharge into the zone where increasing power loading causes increased efficacy.
A New Quartz-to-Metal Seal
The project had to be shelved until 1935, when a technique was perfected at Eindhoven to reliably seal metal wires through quartz for the first time ever. Previously the only known method was the technique employed in the lamp of Küch & Retschinsky, which used precision-ground cones of invar that mated with a similarly ground socket in a quartz tube, this being a very labour intensive and highly unreliable method.
Philips' new seals centred around the development of a glass having a coefficient of thermal expansion that was intermediate between quartz and tungsten. Lamps could successfully be fabricated by this method, and because of the very high softening temperature of quartz, high power loadings could be tolerated in a lamp made from this material. But on account of the still very large mismatch in thermal expansions, the operating temperature of the seal was limited.
It did however enable Philips' engineer Cornelis Bol to create a small high-pressure mercury lamp in which the discharge took place in a very narrow capillary tube of quartz. The tube diameter had to be small so that capillary action caused a globule of excess mercury to be permanently lodged just behind each electrode, this acting as a heat shield to protect the end seals from high temperatures. Lamps would crack immediately after lighting if this mercury bead was absent. The tube diameter was so narrow that forced water cooling was required to prevent its immediate destruction when lighted. Nevertheless this lamp quickly found an application in high intensity cinema projection systems, and is described fully under the MD Water-Cooled lamp section.
It exhibited two crucial features - both a high mercury vapour pressure, and a high power density and it therefore allowed a lamp of rather high efficacy to be created. The challenge was to then adapt the technology to make lamps which did not require forced water cooling, and would be suitable for general lighting applications.
A year later in 1936 Philips was successful, and it led to the invention of the MB family of high pressure mercury lamps described on the next page. The MB lamps are classified as types operating with a power density of 10-100W/cm of arc length and at a mercury vapour pressure of roughly 5-20 atmospheres.
|
