Lamp electrodes have to tolerate electrical, thermal and chemical stress. They tend to be made of tungsten in all but the lowest power lamps. In order for the gas in a lamp to produce light a current must flow through the gas. The electrodes assist with electron emission. Their ability to release electrons is often referred to as the work function or emissivity of the electrode. Heating the electrode lowers the work function and encourages electrons to leave the surface. This can complicate the design of the lamp and control gear when starting from cold. Electrodes are often coated with materials that aid emission for starting from cold. Once running the electrode is heated directly by the bombardment of ionised gas atoms. These charged particles are attracted to the electrode due to the charge difference and they collide with the electrode, generating heat.
The lamp supply must be able to start a lamp, control it as it heats up and operate it at full output. However, the electrical requirements to start a lamp are very different from running it. This has to be catered for by an electrical control circuit. As a result there are few discharge lamps which connect directly to a mains supply. Ones that do will contain some kind of self ballasting device.
Initially, when the lamp is turned on there will be no circuit. The gas in the lamp will be an insulator and nothing will happen. A voltage needs to be applied high enough to cause the gas to breakdown. There are many ways to achieve this. Some mercury lamps use a small third electrode. This creates local ionisation which spreads through the lamp. This electrode can be seen in the top picture on the left. Wires or conductive coatings around the arc tube can assist with starting as shown in the second picture top left. Here wires attached to the outside of the arc tube in a now obsolete design of a low pressure sodium lamp. Without starting devices the breakdown voltage of the gas can be very high. In order to restart or re-strike a hot lamp, the voltage requires to be even higher due to the increase in gas pressure when hot. This is very much the case with the HID Xenon lamp used for automotive lights. Strike voltages as high as 23kV are required to restart a hot lamp, yet the arc length is only around 5mm.
Once an arc has been struck the ionised gas becomes a conductor with a resistance often lower than 100 Ω. The voltage across the lamp must be reduced to around 100V for many lamps, but some super high pressure lamps may have running voltages in the tens of volts. If the voltage is not controlled the current in the lamp will be excessive and the lamp will be rapidly destroyed. The current can be controlled by a resistor as in the ballast filament in blended mercury lamps, but this is very inefficient. Inductive ballasts have been the main choice of lamp control to date. However, electronic ballasts or inverters are becoming more popular due to efficiency and light stability improvements that they can offer.
The schematics below illustrate three different circuits for starting and running discharge lamps. These are particular to low pressure sodium lamps:
Choke ballast and electronic ignitor