Fullscreen HTML5Med Res (??MB)HTML4

Pt.3: Fluorescent Lamps Electric Lighting Along with incandescent lamps, fluorescent lights are amongst the most widely used of lamps. Where diffuse, general lighting is required in commercial and industrial applications, fluorescent tubes rule supreme. By JULIAN EDGAR While we see just a white tube emitting visible light, the fluorescent lamp is in fact a low pressure mercury discharge lamp. It produces light when the fluorescent powder coating on the inside of the glass is activated by ultraviolet (UV) energy. Fluorescent lamp history In 1710, Englishman Sir Francis Hawksbee produced a glow discharge 4  Silicon Chip inside a glass tube from which air had been evacuated and mercury added. He called the glow “electric light” and claimed that his experiment had proved that electricity could produce light. This experiment took place more than a century before the first primitive incandescent light. It wasn’t until 1852 that Sir George Stokes discovered the basic principle of transforming ultraviolet radiation into vis­ ible light. Specifically, he found that quinine sulphate solution glowed when irradiated by ultraviolet energy. In the period between this discovery and the development in the 1930s of the fluorescent lamp, much work was done on low and high pressure electric discharges in both mercury and sodium vapour. However, all of these devices were relatively inefficient at producing visible radiation. A major breakthrough occurred in the 1920s when it was discovered that a mixture of mercury vapour and an inert gas was about 60% efficient in converting electrical input power in-to a single (253.7nm) wavelength of light. By 1935, a General Electric team led by GE Inman produced a prototype green fluorescent lamp with an efficacy of 60lm/W. This efficacy is far better than even current incandescent lamps can achieve and must have been the cause of quite some excitement at the time. As a result of their work, several important characteris­ tics of fluorescent lamp behaviour were identified. It was re­alised that the discharge process is best started by electrically heating oxide-coated filaments positioned at either end of the tube. This causes the filaments to emit electrons which disperse along the length of the tube. When a high voltage is subsequently applied, an electric discharge occurs through the inert gas, exciting the gas atoms which then emit ultraviolet light. Very high efficiency is obtained if the excited atoms are of mercury vapour, which produces a single wavelength of ultra­violet light at 253.7nm. To produce visible light, phosphors with a peak sensitivity at 253.7nm are applied to the inside of the tube. The reason that the phosphors must be on the inside is that 253.7nm ultraviolet light does not pass through ordinary glass. By April 1938, the fluorescent tube was ready for market. Initially, it was released in white plus six other colours. The ballast choke in a fluorescent lamp fitting con­sists of a large number of turns of enamelled copper wire on a laminated iron core. Its primary functions are to limit current and to provide sufficient open-circuit voltage to initiate igni­tion. The fluorescent lamp A fluorescent tube consists of a soda-lime glass tube that has been doped with iron oxide to control the amount of shortwave transmission. The most common tube diameters are 16mm, 25mm and 38mm, while the most common lamp lengths are 600mm, 1200mm and 1500mm. The most important factors affecting the light characteris­tics of a fluorescent lamp are the type and composition of the applied phosphors. Phos­phors commonly used include calcium ha­lophosphate (for white light), magnesium fluoro-germanate (red) and calcium tungstate (blue). Colour temperature, colour rendering and to a large extent luminous efficacy, are all affected by the phosphors. Standard phosphors give a lamp with good efficacy but poor colour render­ing. Tri-phosphor lamps use special fluorescent powders contain­ing certain rare earths that give radiation peaks at three well-defined wavelengths (in blue, green and red) that are equally distributed over the A capacitor is used to provide power-factor compensation. Fig.1: a simplified view of what goes on in a fluorescent lamp. The glass tube is coated inside with fluorescent pow­ders that glow when excited by the ultraviolet energy of the discharge (diagram from the Philips Lighting Manual). January 1998  5 The starter allows the filaments to be pre-heated, increasing their emission of electrons. visible spectrum. These lamps give very good colour rendering together with high efficacies. Finally, the latest lamps use socalled multi-phosphors, which employ a mix of phosphors chosen to cover the entire vis­ ible spectrum. These give the highest colour rendering of all the fluorescent lamp types. The filament windings located at either end of the tube can be of either coiled-coil or straight coil types. They are similar to incandescent lamp filaments but are coated with barium or strontium oxide to aid electron emission. Most fluorescent tubes use a starter to preheat the fila­ments with an electric current just prior to lamp ignition. However, “rapid-start” tubes have continuously heated filaments while “cold-start” (or “instant start”) tubes Fig.2: the energy consumption of a 36W fluorescent lamp in still air at an ambient temperature of 25°C. 10W of visible radiation is produced. 6  Silicon Chip use no preheating of the filaments at all. The latter types do not use a separate starter but often employ an auxiliary electrode or a conductive strip on the outside of the tube to facilitate ignition. The gas in a fluorescent tube consists of a mixture of saturated mercury vapour and an inert buffer gas, commonly argon or krypton. Under normal operating conditions, mercury is present in the tube in both liquid and vapour forms. Fig.1 shows a simplified view of what occurs within a fluorescent lamp. The biggest change in fluorescent lamp technology in recent years has been the release of compact fluorescent lamps. Designed as plug-in replacements for incandescent lamps, they combine high efficacy and good colour characteristics with a life expectancy which is typically eight times that of an incandescent lamp. Lamp performance Fig.2 shows the total energy consumption of a 26mm diamet­er, 36 watt (36W) fluorescent lamp operated in still air with an ambient temperature of 25°C. Of the 45W input power, there is just 10W of visible radiation. Infrared radiation, convection and conduction make up 25.8W, with the remaining 0.2 watts lost as UV radiation. The reason that the ambient temperature needed to be speci­ fied in the above example can be seen in Fig.3. The luminous flux of a typical fluorescent lamp is very dependent on temperature. It is at its greatest at about 25°C, falling by 40% as the temper­ ature drops to 0°C. So when you go out to the shed on cold winter nights and flick on the fluoros, it’s not just your imagination that it all looks dim and cold! As temperatures rise above 25°C, the output of the lamp again falls, being over 30% down at 70°C. Not only does the luminous flux of the lamp drop rapidly at higher temperatures but so does the luminous efficacy. However, the power dissipated by the lamp also decreases rapidly with increased temperatures, so the luminous efficacy falls off less rapidly than the luminous flux. Operating a fluorescent lamp on a high-frequency supply improves luminous efficacy by about 10%, a major incentive for employing high frequency electronic ballasts. Using a high fre­quency ballast has the added advantage of reducing lamp blacken­ ing, a problem that occurs at the ends of the lamp due to the deposition of dispersed emitter material lost from the filaments. Another cause of a decrease in luminous flux over the life of the lamp is that the fluorescent powders slowly become less effective. When a mix of powders has been used, discolouration can also occur. After 8000 hours, the luminous flux of a typical fluorescent lamp will be between 70% and 90% of its original value. After starting, a fluorescent lamp takes two to three minutes before its luminous flux reaches its maximum. However, the initial flux is about 60% of its final value and so this is not normally noticed. The reason for the delay is that the mer­cury vapour needs a short period before it reaches its working pressure. Lamp circuits Every fluorescent tube requires a “ballast” of some sort and its purpose is twofold, as we shall see. At first switch-on, it is necessary to apply a much higher than normal voltage to the lamp to assist ionisation and thus to get the lamp to ignite. However, once the gas has begun to con­duct, its resistance rapidly falls, resulting in a current flow that would spiral out of control unless checked. In fact, as with all gas discharge devices, it has a negative resistance; ie, as the current rises the voltage drop across the tube is reduced. Ultimately, unless something is done to prevent it, the current will rise to such a high value that the tube will be destroyed. It is therefore necessary to use a current limiting device, a “ballast” to prevent current runaway. This ballast can take the form of a resistor, an incandescent lamp, an iron-cored choke or an electronic control circuit. Although relatively simple, a resistive (or incandescent lamp) ballast wastes energy, which is dissipated as heat. In fact, the power lost in the resistor is comparable to the power taken by the lamp! Resistive ballasts are therefore rare and are employed only in some fluorescent lamps operated from a DC sup­ply. An iron-cored choke (inductor) is the most widely used ballast in AC applications. It consists of a single coil with a large number of turns of enam- Fig.3: the luminous flux of fluorescent lamps varies a great deal at different ambient temperatures (diagram from the Philips Lighting Manual). elled copper wire on a laminated iron core. In addition to limiting current, the ballast also: (1) provides sufficient open-circuit voltage to initiate igni­tion; (2) regulates the lamp current against power supply voltage changes; (3) permits electrode heating in preheat and rapid-start lamps. To understand how the ballast provides all these functions, it is necessary to consider the circuit of a normal fluorescent lamp fitting which is shown in Fig.4. Fig.4 shows that the fluorescent tube has a filament (heater) winding at each end and these are connected in series with the ballast choke via the starter. The starter consists of a bimetallic strip mounted within a small argon or neon-filled bulb. When the supply is switched on, the bimetallic strip is cool and its contacts are open. The applied voltage causes the gas in the starter to ionise, allowing a small current flow. This heats the bimetallic strip, causing it to bend enough to close the internal switch. Current can then flow through the ballast and the two filaments, which are heated and start to emit elec­trons. The starter cools and the bimetallic strip opens, inter­rupting the current through the filaments. Since the inductor is also in series with the starter, the sudden switchoff causes it to produce a brief high voltage spike which appears across the ends of the lamp, causing it to ignite. The voltage required to ignite the tube depends on its length and diameter, its age and the temperature. The longest tubes are hardest to start and all tubes are much harder to start at low ambient temperatures. The voltage required to start the tube can be as high as 800V; ie, much higher than the normal peak voltage of Fig.4: the circuit of a conventional fluorescent lamp with a glow switch starter. The starter enables current to flow through the filaments and it opens after a short delay, causing the ballast choke to produce a high voltage spike which ignites the tube. January 1998  7 leading to burnout of the ballast. The way to avoid this is to replace both the tube and the starter immediately they start to give trouble. Starter capacitor Inside the bulb of the starter is a pair of contacts with the movable contact actually being a bimetallic strip. Visible behind the glass bulb is the small capacitor which shunts the starter and helps suppress electromagnetic interference. the 240VAC mains waveform. As the starter and the tube get older, starting becomes progressively harder until eventually the tube will not start at all and will only flash spasmodically. If left in this condition, the starter’s contacts may eventually weld shut, In the circuit of Fig.4 you will notice a capacitor con­nected across the starter bulb. The value and voltage rating of this capacitor is critical to the starter’s operation. Typically, the capacitor has a value of about 0.006µF and will typically have a voltage rating of 3kV if it is a ceramic disc and around 1kV or more if it is a wound plastic type. Clearly, the capacitor needs a high voltage rating if it is to withstand the spike voltage produced by the inductor when the starter contacts open. Second, the capacitance is critical as well. If the capacitor is too small in value or open circuit, the starter’s contacts will arc badly and quickly burn out. The capacitor effectively controls the rate of rise of the inductor voltage and if it is too large, the voltage will rise too slowly and the tube will fail to ignite. But there is another important function of the capacitor and that is to help suppress the very considerable electromagnet­ic interference produced by the tube when it is conducting and also when the starter contacts open. This interference is radiat­ed over a very wide spectrum, including the UHF bands. It is strongest and most apparent in the AM and shortwave radio bands. Even with the capacitor present, the interference is strong and for that reason, fluorescent lights and other forms of gas dis­charge lighting cannot be used in applications where low EMI is necessary. Electronic starters which replace the bimetallic strip design with an integrated circuit are now available (see SILICON CHIP, August 1996) but the adoption of an entirely new electronic control system does away for the need for a separate starter entirely. In addition to this, electronic systems have other major advantages. These include: (1) improved lamp and system efficacy; (2) no flicker or stroboscopic effects; (3) increased lamp life; (4) excellent light regulation possibilities; (5) reduced heating; (6) no need for power-factor correction; and (7) no hum. Ballast power loss is significant. As shown in Fig.2, a 36W lamp using a conventional ballast has an actual power consumption of 45W, with the ballast dissipating around 9W (20%) of the power drawn. Even a low-loss ballast dissipates 6W, compared with around 4.5W from an electronic ballast. Note that some compact fluorescent lamps have the ballast built-in and so, for these lamps only, the power rating includes ballast losses. Power factor Electronic starters are now available to replace the glow switch starters in fluorescent lamps fittings. They have a number of advantages, including the ability to disconnect the power and protect the ballast if the tube cannot be started. 8  Silicon Chip While the diagram of Fig.4 shows the most common fluores­cent lamp circuit as installed in most homes, the type installed in industrial and commercial installations typically has an addi­tional large capacitor connected directly across the 240VAC mains supply. The capacitor is included to provide power factor correc­tion. “Power factor” becomes a problem in any 50Hz mains circuit where the current waveform or phase is not identical with that of the 240VAC sine waveform. To explain further, in a resistive load connected across the 50HZ 240VAC mains supply, the current is exactly in phase with the voltage and it has the same shape; ie, a sinewave. In an inductive load, the current Are Fluorescent Lamps Mercury Hazards? If all this talk of the mercury vapour within a standard fluorescent tube makes you wonder about safety, you are not alone. Mercury – especially in the form of a vapour – is extreme­ly toxic. While the bulb remains unbroken there is little or no chance of ingesting the mercury. The problem comes, however, in the disposal of the used tube. While there is apparently little thought given to fluores­cent lamp disposal in Australia, a very different situation exists in the USA. There, the Environmental Protection Agency established in 1990 a Toxic Characteristics Leaching Procedure (TCLP) to assess the impact of substances that may be leached away from landfill dumps. Normal US-market fluorescent lamps generally fail the procedure! As a result of this and other pressures such as cost, fluorescent lamp producers have reduced the mercury content of lamps. In the US, the industry average for mercury in their standard 1.2 metre, 40 watt lamp has waveform still has a sinewave shape but it lags the voltage by up to a quarter of a cycle, ie, the phase lag can be up to 90 degrees. This presents a real problem because the power consumed by an inductive or ca­pacitive load is denoted by the following formula: P = VI.cos φ where phi is the phase angle between the voltage and current. Now the if the phase angle is 90 degrees, which will be the case in an ideal capacitor or inductor, then the value of cos phi will be zero. So in that case: P = VI.cos 90° = 0. In other words, while voltage is applied and current is flowing, the power being measured is zero! Now while the induc­tance in a fluorescent circuit is not perfect, there is still quite a lag between the voltage and current and so the power being measured (and paid for) by the customer is still quite low. This causes the energy authorities serious concerns because their distribution system still has to provide the current and take care of all the resistive losses between the generator and the been reduced from 48.2mg in 1985 to 22.8mg in 1994. However, lamps with 22.8 milligrams of mercury still do not pass the TCLP test! Philips has developed a new lamp which uses significantly less mercury. Mercury capsules are mounted in the lamp and are activated only after most lamp impurities are removed. The use of buffer gases further reduces mercury loss, meaning that less than 10mg of mercury is required to be used in their ALTO model lamps. In the US the green end-cap ALTO lamps have been available since 1995. It was expected that by the end of 1997 80 percent of all Philips fluorescent lamps sold in the US would feature low-mercury technology. The lamps feature the same life, colour rendering and efficacy as conventional fluorescent lamps. In Australia, as far as we can determine, fluorescent tubes also now have reduced mercury and the so-called buffer gases, argon and neon, have been increased. final (inductive) load. For this reason, commercial and industrial installations are generally required to have power factor correction capacitors installed in fluorescent light fittings. The term “power factor” comes from “cos φ” in the above equation. When the phase angle φ is zero, as for a resistive load, cos φ = 1. This is said to be a power factor of unity and is the ideal. To overcome the problem of lagging power factor, a capaci­tor is often placed across the mains supply to the lamp circuit. The capacitor draws current which “leads” the voltage waveform and so compensates for the “lagging” current drawn by the induc­tive portion of the circuit. This substantially improves the power factor, typically giving a ratio of 0.85, instead of around 0.7 for a fitting without power factor correction. Typically, a 4.2µF capacitor is fitted for a 36 or 40W lamp, and a 6.5µF capacitor for a 58W or 65W lamp. The capacitor also provides some smoothing of the current pulses drawn by the fluorescent tubes and thereby provides some reduction of the 50Hz harmonics which would otherwise be superimposed on the 240VAC mains supply. Mains control tones Typically though, correcting one problem causes another and so it is with power factor correction. The electricity supply authorities also superimpose control tones (typically around 1kHz) on the mains supply to switch hot water systems and control their distribution network. Unfortunately, power factor correc­ tion capacitors also cause the mains control tones to be reduced so in any large installation (ie, in factories and shops) a blocking inductor is connected in series with the lighting cir­cuits at the customer’s switchboard. This article has only covered the most common fluorescent lamp circuit using a glow switch starter. There are many other circuits, including rapid start, quick start and electronic ballasts which are beyond the scope of this article. Next month: high pressure mercury SC lamps SILICON CHIP This advertisment is out of date and has been removed to prevent confusion. January 1998  9