WO1992011742A1

WO1992011742A1 - A method of and apparatus for operating an electric discharge lamp

Info

Publication number: WO1992011742A1
Application number: PCT/GB1991/002257
Authority: WO
Inventors: Philip John Rimmer
Original assignee: Tunewell Transformers Limited
Priority date: 1990-12-17
Filing date: 1991-12-17
Publication date: 1992-07-09
Also published as: AU9057891A; GB9027282D0

Abstract

An A.C. power source adapted to supply alternating current electrical power with a net flow of current in a predetermined direction over a plurality of cycles of the alternating power is supplied to a discharge tube. The direction of the net flow of current is reversed after a selected or predetermined period of time. Impurities in the tube are selectively occluded by the electrodes in the tube.

Description

"A METHOD OF AND APPARATUS FOR OPERATING AN ELECTRIC DISCHARGE LAMP"

THE PRESENT INVENTION relates to a method of and an app¬ aratus for operating an electric discharge lamp or lamps such as a luminous gas discharge tube, e.g. a neon tube.

A typical neon tube is provided with an elec¬ tric alternating current having a frequency of between 50 and 60 cycles per second. Thus the current undergoes a current reversal every 8 to 10 milliseconds. A quant¬ ity of gas in the tube is ionised within the discharge and is consequently moved as part of the current flow in the tube. The power supply manages to displace a sig¬ nificant quantity of gas leading to a stirring action. Any pockets of contaminating gas or vapour within the tube will be broken up by this stirring action and eventually the entire gas contents of the tube will have been brought into contact with the electrodes at either end of the tube. This gas stirring prevents uneveness in colour and luminosity in compound gas and in gas/mer¬ cury tubes.

Often when a tube is manufactured it will con¬ tain contaminating gases or vapours that can cause im¬ proper tube function, for example, discoloured areas, excessive dissipation, and patterning in the discharge. These contaminants are the result of poor manufacture or the result of incomplete "ageing" of the tube.

It is to be understood that manufacturers carry out the process of "ageing" on all tubes that they manufacture. To do this a higher than normal current of 50 or 60 cycles per second is passed through the new tube. The high level of current flow causes significant stirring of the gas and also generates a high level of ion bombardment of the electrodes which results in a higher than normal electrode temperature. Contaminants present in the tube will normally exist mainly in the metal of the electrodes. The contaminants will be adsorbed in the metal but not trapped or occluded. Heating of the electrodes to a high level helps to drive off the contaminants and release them to the atmosphere inside the tube. Once there, the contaminants become ionised, just like the other gases, and under the influence of the powerful electric fields near the more negative electrode they are accelerated towards that electrode with considerable force. Of all the ions bombarding the negative electrode, the contaminant ions are generally preferentially trapped or occluded when they impact on the electrode. This results in a reduction in the quantity of potentially free contaminants.

It is relatively costly to carry out the "ageing" operation, both with regard to the time taken to perform the function, the space occupied while the function is performed and the electric power consumed. Thus a typical manufacturer does not completely "age" tubes before they are sold. The "ageing" process is actually completed during the first hours of actual operation of the tube, from a conventional mains transformer. However, this further "ageing" is carried out only at an ordinary current level which means.that the gas is not stirred to the same extent that it was stirred during the "ageing" performed by the manufacturer and also the electrodes are not heated to a higher than normal temp¬ erature. Thus the "ageing" process is not continued in an ideal manner.

Some discharge tubes are operated at a high frequency, for example, greater than 1 kilohertz. It is to be appreciated that in utilising a high frequency for a gas discharge tube, the current displaces proportion¬ ally less gas during each cycle, and thus may fail to break-up any gas or vapour concentrations within the tube. Consequently, only a minimal stirring effect is achieved, which can lead to a problem that contaminants are not trapped or occluded by the electrodes, but remain free within the tube. A further problem that may be found in connection with a tube driven at high fre¬ quency, particularly an argon/mercury tube of a complex shape, is that temperature variations at the glass wall can cause a preferential condensation of mercury in the cooler areas. This can result in a progressive reduc¬ tion in luminosity of parts of the tube where mercury vapour is deficient.

Luminous gas discharge tubes that run at high frequency are responsible for high levels of radio-fre¬ quency interference. This is certainly true of units powered by high frequency inverters. However, some in¬ terference may also arise from tubes having conventional power supplies where the sudden "snap-on" conduction of the tube each half cycle can generate a very fast current and voltage transition, consequently giving rise to interference.

Tubes run from high frequency supplies may also suffer from power loss. The very high voltages in¬ volved, coupled with the high frequency, can cause con¬ siderable currents to flow from the tube and its wiring via stray capacitance into the environment. This may result in brightness variations along the length of the tube and can cause nearby metal work, if improperly earthed, to become live.

A further disadvantage that has been exper^¬ ienced with luminous gas discharge tubes utilising high frequency power supplies is the creation of standing wave patterns in the tube. The particular frequencⁱes used, coupled with the typical ion velocities, may lead to visible "bubbles" in the luminous discharge.

It has been proposed to try and overcome the problem of the standing wave patterns by utilising a current wave form to drive the tube which is asymmetrⁱc. Thus the current may be an alternating current, but the alternating current is super-imposed upon a constant "bias" current, so that there is an average current flow in one direction over a number of cycles of the wave form. This tends to move the standing wave, thus makⁱng it blurred or even invisible. One problem of using a power supply of this type is that the life of the tube may well be shortened due to the uneven degradatⁱon of the tube electrodes. It is to be understood that ⁱf there is an average current flow in one sense, one of the two electrodes would act to have principally "cath^¬ ode" properties and the other electrode will act to have principally "anode" properties. A problem that can arise in neon/argon tubes when driven with a hⁱghly asymmetric current is that, due to the process of cato- phoresis, the gas with the lowest ionic concentratⁱon will be moved to the more negative electrode, resultⁱng in a colour change along the length of the tube.

The present invention seeks to provide a new apparatus for operating a discharge tube in which the above described disadvantages are obviated or reduced. According to this invention there is provided an apparatus for operating a discharge tube or tubes, said apparatus comprising an AC power source adapted to supply alternating current electrical power and means to direct that alternating electrical power to the dis¬ charge tube, the alternating electrical power being such that there is a net flow of current along the discharge tube in a predetermined direction over a plurality of cycles of the alternating power, means being provided to reverse the direction of the net flow of current after a selected or predetermined period of time, said period of time having a duration of at least 200 milliseconds.

Preferably the predetermined period of time has a duration of at least one second.

Conveniently the predetermined period of time has a duration of at least one minute.

Preferably the predetermined period of time has a duration of at least 30 minutes.

In one embodiment the means for altering the direction of flow of current comprises manually operable means.

Preferably said manually operable means com¬ prise means adapted to switch the arrangement on and off.

In another embodiment said means comprise timer means adapted to reverse the flow of current at the expiry of the predetermined period of time.

In a further embodiment the means comprise means adapted to integrate a signal representative of the current, and to effect a reversal of the direction of flow of current when the integrated signal reaches the predetermined level.

Preferably the output of the integrator is con¬ nected to two comparators each provided with a respect¬ ive reference voltage, each comparator being adapted to reverse the direction of current flow when it provides a predetermined output signal.

Preferably the power source comprises means to develop an off-set alternating current.

Conveniently the said reversing means comprise ganged switches adapted to reverse the application of the off-set signal to the tube.

Advantageously the reversing means comprise means for generating an off-set voltage and means for adding that off-set voltage to the power from the power source.

Preferably the off-set voltage is developed upon impedance means, such as a capacitor.

Conveniently the voltage is applied to a cap¬ acitor from a DC power source, switching means being provided to reverse the DC potential applied to the cap¬ acitor by the power source.

Preferably two capacitors are provided con¬ nected in series, each capacitor being connected in parallel with a respective diode, the diodes being con¬ nected in the opposite sense, switching means being pro¬ vided adapted selectively to short either one or the other of said capacitors. Preferably means are provided to heat both the electrodes of the discharge tube.

In order that the invention may be more readily understood, and so that further features thereof may be appreciated, the invention will now be described, by way of example, with reference to the accompanying drawings in which

FIGURE 1 is a block diagram of one example of an apparatus in accordance with the invention,

FIGURE 2 is an illustration of one form of the control for use in Figure 1 ,

FIGURE 3 is a block diagram of an alternative form of the control for use in Figure 1 ,

FIGURE 4 is a diagrammatic view of a hot cath¬ ode tube and some associated circuitry for use with the apparatus of Figures 1 to 3_>

FIGURE 5 illustrates an alternative apparatus in accordance with the invention,

FIGURE 6 is an example of the control apparatus as used in Figure 5,

FIGURE 7 is an example of the wave form that may be obtained from the arrangement of Figure 5,

FIGURE 8 illustrates a further alternative em¬ bodiment of the invention,

FIGURE 9 illustrates the embodiment of Fig¬ ure 8 in greater detail, and FIGURE 10 illustrates component PCB 2 of Figure 9 in greater detail.

Referring initially to Figure 1 of the accompanying drawings, an asymmetric inverter 1 may be utilised as the power source. The output of the asymmetric inverter will be an asymmetric alternating current which has a greater current flow in one sense than in the opposite sense, over a number of cycles. Thus the inverter provides a net current flow. Many inexpensive inverters are actually asymmetric, and the fact that they are asymmetric may, in many circumstances, make them undesirable for use. In the present case, however, the asymmetric nature of the output of the inverter is very advantageous.

In an alternating current a number of coulombs of charge will flow in one sense in the positive cycle of the current, and a number of coulombs will flow in the opposite sense during the negative cycle of the current. In a symmetrical current the number of coulombs in the positive cycle is the same as the number in the negative cycle. In an asymmetric alternating current more coulombs flow in one cycle, than flow in the opposite cycle, so there is a net current flow in one sense. An asymmetric alternating current may have an asymmetric waveform, or may have a symmetric waveform with a D.C. bias.

The outputs 2,3 of the inverter 1 are fed to a three-way switch arrangement 4 which is actuated by means of a control 5. It can be seen that the switch arrangement 4 comprises two parallel ganged switches 7,8, which are connected to the two electrodes 9,10 respectively of a discharge lamp 11. When the switches are in the position illustrated in Figure 1 -9-

the net current flow through the tube 11 is in one sense, whereas when the switches are moved to another position (the lower-most position shown in Figure 1) the net current flow through the tube will be in the opposite sense. The switches may have an intermediate position in which no current flows through the tube 11 and the tube 11 is thus "off".

The control 5 may comprise a rotary switch of the type illustrated in Figure 2 as item 5A. The switch will only rotate in one sense, and the arrangement is such that as the switch 5A is rotated the parallel switches 7 and 8 move from their upper position as illustrated in Figure 1 to the intermediate position and then to the lower position and thence back to the inter¬ mediate position and finally again to the upper pos¬ ition. Thus, as the switch 5A illustrated in Figure 2 is operated successively to turn the tube on and off, net current will flow through the tube in opposite senses.

In an alternative arrangement the control 5 of Figure 1 may comprise a timer 12 and an associated sole¬ noid 13, the solenoid serving to operate the parallel ganged switches 7 and 8. Thus, in response to the timer 12 the switches 7 and 8 will be successively moved between their upper position and their lower position thus reversing the flow of DC current through the tube 11 after predetermined intervals of time.

It is envisaged that the timer will measure periods of time having a minimum duration of 200 milli¬ seconds, and preferably having a duration longer than one minute, most preferably having a duration in excess of thirty minutes.

The net current will flow in one sense for a - -

period of time either as determined by the timer 12 or for the entire duration of the illumination cycle of the tube 11 if the switching arrangement as illustrated in Figure 2 is used as the control. Thus, on one occasion that the tube is illuminated the net current will flow in one sense, and the current will continue to flow in that sense until the tube is switched off. The net current will subsequently, the next time the tube is illuminated, flow in the opposite sense.

When the net current flows through the tube 11 only in one sense, there will be a significant transfer of charge along the length of the tube, thus providing a substantial "stirring" effect. Contaminants will also be swept towards the more negative electrode (that is to say the electrode that presents negative properties when the current is actually in the direction of the net current), thus being trapped by that electrode.

It is to be appreciated that for the tube 11 to have an optimum life, the current flow through the tube should be symmetric which is to say that the total move¬ ment of current in one direction through the tube should be equal to the total direction of current in the oppo¬ site direction through the tube. The switching arrange¬ ment shown in Figure 2 will, it is believed, provide a sufficient degree of accuracy over a prolonged period of use. Thus, whilst the tube 11 may be switched on for periods of time of varying length, over a period of weeks and months, the periods of time during the switches 7 and 8 are in the upper-most position as illustrated in Figure 1 will be substantially equal to the periods of time in which the switches are in the lower-most position as illustrated in Figure 1. Thus the total current flow through the tube in each direc¬ tion will be substantially the same. The switches will only be used to switch the tube on for periods of greater than 200 milliseconds. Using an asymmetric inverter, when the switches 7 and 8 are in the upper position as illus¬ trated in Figure 1 an alternating current will flow through the tube 11 , but the nature of that alternating current is such that there is a net charge flow, during a period of time, along the tube 11, from one electrode, such as the electrode 14, to the other electrode, such as the electrode 15. This means that there is a stirring effect which is provided, and also any im¬ purities are moved consistently along the tube towards one of the electrodes, and the impurities can be trapped at that electrode. Because there is a net current flow along the tube the "ageing" process is completed rel¬ atively swiftly. Each electrode-will, for at least part of the time during each cycle of the alternating current, act as a cathode. Of course, both the elect¬ rodes will, during each cycle, act as the cathode and as the anode, during separate periods of time. In a hot cathode tube both the electrodes will thus be self-heat¬ ing, but a separate heating arrangement, as shown in Figure 4 may be necessary.

It is to be appreciated that, in some circum¬ stances (with a hot cathode tube), even though an alter¬ nating current is applied to the tube, the reduced ionic bombardment at the electrode which, during any specific period of time, tends to be the more positive electrode, may result in that electrode becoming cooler and poss¬ ibly failing to become "thermal". This condition, when insufficient electrons are liberated from the elect¬ rode, when that electrode is acting as a cathode, -~~Y cause power loss and considerable electrode damage. This damage is the result of the bombardment of the electrode by massively energetic ions as they attempt to make good the deficiency, by the liberation of the required electrons by secondary emission from the tiny electrodes. The heating deficiency in the electrodes may, in such a situation, be made good by the heating arrangement which is illustrated in Figure 4. Referring to Figure 4 it can be.seen that the electrodes 14 and 15 in a tube such as the tube 11 are each in the form of a heater element, and each heater element has an asso¬ ciated power supply 16,17. The power supply 16 would be activated to operate the heating element 14 shortly before the element 14 is due to become the "cathode" under the control of the timer 12, that is to say before the net current is due to flow towards the element 14. Once the electrode has been heated the electrode becomes the "cathode" and then no further power need be supplied from the power supply 16 since, when the electrode is operating as a cathode, it is self-heating. Of course, a similar power supply 17 has to be associated with the electrode 15 for use when the flow of net current again reverses.

Whilst, in the embodiment described above, reference has been made to the use of parallel ganged switches 7 and 8, it is to be appreciated that the switches may be formed as solid state switches con¬ trolled by appropriate logic.

Figure 5 illustrates another embodiment of the invention, which again incorporates a discharge tube 11. The discharge tube 11 is provided with an alternating current signal from the secondary coil 19 of a transformer 20, the primary coil 21 of which may be connected to the mains electricity supply or to the out¬ put of an appropriate inverter. However, included in the circuit between the secondary 19 of the trans¬ former 20 and one electrode of the tube 11 is a capac¬ itor 22 and a ganged switching arrangement 23 (which is similar to the switch arrangement 4 of Figure 1) is pro¬ vided to connect a DC power supply 24 across the cap¬ acitor. The switching arrangement is controlled by a control arrangement 25 which may again comprise a rotary switch as shown in Figure 2 or a timer and solenoid as shown in Figure 3 or any other appropriate arrangement. It is to be appreciated, therefore, that the DC power supply 24 applies a voltage to the capacitor 22 and that voltage is effectively added to the AC voltage generated in the secondary coil 19 of the transformer 20 to pro¬ vide a composite voltage which is applied to the tube 11. It is to be appreciated, therefore, that when the capacitor 22 is charged in a particular sense there will be a net current flow through the tube in one dir¬ ection, whereas when the capacitor 22 is charged in the opposite sense there will be a net current flow through the tube in the opposite direction. Effectively the combination of the solenoid and the capacitor provides an alternating voltage with a reversible off-set.

The control 25 may, as mentioned above, com¬ prise the rotary switch or the timer and the solenoid as described above, or any other appropriate arrangement, but one particular arrangement for the control 25 is illustrated in Figure 6. A current sensing resistor 27 is incorporated in the circuit with the capacitor 22 and the secondary coil 19 of the transformer 20. As current flows through the resistor 27 a voltage is developed across that resistor, and that voltage is applied to an integrator 28. The output of the integrator is fed to the input terminals of two respective comparators 29 and 30. The comparator 29 is provided with a positive ref¬ erence voltage 31 and the comparator 30 is provided with a negative corresponding reference voltage 32. When the integrator 28 has operated and produces a positive out¬ put voltage which is equivalent to the positive refer¬ ence voltage 31 _» the comparator 29 provides a signal which causes the switching arrangement 23 to change its condition. The net current sensed by the integrator is then a negative current, because the direction of current flow changes. When the integrator has inte- grated this to produce a negative output voltage which is as great as the negative reference voltage 32, the comparator 30 operates to again cause the condition of the switching arrangement 23 to reverse. This arrange¬ ment possesses the advantage that the control arrange¬ ment is responsive to the actual current flow in each direction, rather than being responsive simply to the elapse of a period of time. Thus if there is any un- desired or unintentional degree of asymmetry in the cir¬ cuit, this will be compensated for by the arrangement as shown in Figure 6. This control arrangement may be used in further embodiments of the invention.

Figure 7 illustrates a typical wave form as supplied to the tube 11 in the embodiment of Figure 5. As can be seen during an initial period T_ a positive shift is applied to the signal, and so the signal ex¬ tends between +6 and -3. However, during the period T, a negative shift is applied to the signal, and the signal then extends between +5 and -5. Each cycle of the signal is not symmetric within itself and the line 33 illustrates the net current flow during the period T 3. while the line 34 indicates the net current flow during the period 1* . It can be seen, therefore, that the net current flow during the period T_ is equal and opposite to the net current during the flow T..

Figure 8 illustrates another circuit arrange¬ ment that can be utilised to obtain a wave form such as that shown in Figure 7. In the arrangement of Figure 8 a transformer 35 has its primary 36 connected to an appropriate source of AC voltage or the output of an in¬ verter. The transformer has two secondaries 37,38 which are wound in the same sense, one secondary 37 is con¬ nected to one electrode present in a tube 11 and the other secondary 38 is connected to the other electrode of the tube 11. The two secondaries are inter-connected by a parallel connection consisting of two diodes 39,40 which are connected back-to-back, two capacitors 41,42 and two switches 42,43. The node between the two di¬ odes 39,40 is connected to the node between the two cap¬ acitors 41,42 which is also connected to the node between the two switches 42,43 which is also connected to earth 44. Under an appropriate control either the switch 42 is closed or the switch 43 is closed. When the switch 42 is closed, the capacitor 41 is shorted, but the capacitor 42 serves to complete the circuit in¬ corporating the two secondaries 37 and 38 so current can flow through the tube 11. However, because of the pres¬ ence of the diode 40 a voltage potential in one pre¬ determined sense appears across the plates of the cap¬ acitor 42 thus effectively adding an off-set voltage to the signal applied to the tube 11. When the switch 43 is closed and the switch 42 is opened the capacitor 42 is snorted, but the capacitor 41 is then open circuit and a voltage in the opposite sense appears across the capacitor 41.

Thus this circuit again provides an alternat¬ ing signal with an off-set, the off-set being switched. If one assumes that this arrangement operates at a high frequency at 30 m Amps Rms, but has an off-set current of 1 m Amps RMS that is reversed every 500 seconds, the charge moved during the 500 second period is 500 milli- coulombs. The charge moved in a similar prior art tube operated at 30 m Amps RMS at 50 cycles per second would only be 0.27 millicouloms during one-half cycle. Thus the invention provides a high degree of charge transfer compared with the prior art.

The bulk of the visible discharge in a tube comprises a region known as the "positive column" within which the majority of the current is carried not by ions but by electrons due to their much higher mobility. An electron is about 4,000 times more mobile than an ion. In a 15 millimetre diameter tube, containing argon and pumped to 10 torr the ionic concentration would be such that if the tube is driven by a conventional supply, all ions would be displaced by about 0.5 metres during the half-cycle of the current. On the other hand, if the arrangement in Figure 8 is used with the current level specified above and current reversal every 500 seconds, there is a likelihood that a large fraction of all the gas atoms will be displaced along the entire length of the tube (assuming the tube has a length of 2 metres or so) as ions are continuously being neutralised at the cathode and the tube walls and new ones created in ion¬ ising collisions.

Figures 9 and 10 illustrate an arrangement equivalent to that shown scheramatically in Figure 8 in greater detail.

It is to be appreciated that in using preferred embodiments of the invention mixed gas and gas/mercury tubes may be used to produce light having a uniform col¬ our and uniform luminosity. Also new tubes may be fully "aged" in a straightforward manner and the "ageing" pro¬ cess will be speeded up. However, the method does not lead to premature tube and electrode degradation due to asymmetric current flow since, over realistic periods of time, the current flow is substantially symmetric. Pat¬ terning in luminous gas discharge tubes will be elimin¬ ated without compromising tube life. By allowing the average current within a period of operation to be such as to increase the net ion flux to each electrode in turn, the peak electrode temperatures may be increased. Thus the electrode which, at any instant, is acting more as a cathode than as an anode will have a higher temper¬ ature than the other electrode. As ion occlusion is an exponential function of electrode temperature, the average rate of occlusion will thus be increased even though the average electrode temperature, taken over a number of cycles of operation, may remain unchanged.

It is important to note, however, that tube life may be maximised if the electrodes are treated in an entirely symmetrical manner over a relatively long period of time and, by using the procedures described above, this is achieved.

While various manual switches have been referred to above, it is to be understood that such switches may be replaced by automatic switches which respond to signals from appropriate sensors, if appropriate. A typical sensor might be an infra-red sensor to sense the presence of a person and then to switch on a light automatically.

The preferred embodiments of the invention, intended to operate at high frequency, use waveforms with sufficient asymmetry to prevent the formation of standing waves. In alternative embodiments, however, it may be preferable to use a power supply which only has a small degree of asymmetry in its output. Such a power supply may include a transformer with asymmetry produced on its primary side. A power supply of this type may be relatively inexpensive. If this type of power supply is used with a phosphor coated tube the standing wave problems will generally be of no consequence. However, if used with other types of tube, the standing wave problem may be eliminated by modulating the output frequency of the power supply or by mixing two or more frequencies simultaneously. In the case of some self oscillating inverters the output frequency varies with input voltage supply, and the desired effect may be achieved by having a high level of mains ripple on the D.C. supply rail. The two or more frequencies may be obtained using two inverter circuits or by using output circuitry that provides resonances.

Claims

- 18-CLAIMS :

1. An apparatus for operating a discharge tube or tubes, said apparatus comprising an AC power source ad¬ apted to supply alternating current electrical power and means to direct that alternating electrical power to the discharge tube, the alternating electrical power being such that there is a net flow of current along the dis¬ charge tube in a predetermined direction over a plur¬ ality of cycles of the alternating power, means being provided to reverse the direction of the net flow of current after a selected or predetermined period of time, said period of time having a duration of at least 200 milliseconds.

2. An apparatus according to Claim 1 wherein the predetermined period of time has a duration of at least one second.

3. An apparatus according to Claim 1 wherein the predetermined period of time has a duration of at least one minute.

4. An apparatus according to Claim 1 wherein the predetermined period of time has a duration of at least 30 minutes.

5. An apparatus according to Claim 1 wherein the means for altering the direction of flow of current com¬ prises manually operable means.

6. An apparatus according to Claim 5 wherein said manually operable means comprise means adapted to switch the arrangement on and off.

7. An apparatus according to any one of Claims 1 to 4 wherein the said means comprise timer means adapted to reverse the flow of current at the expiry of the pre_¬ determined period of time.

8. An apparatus according to any one of Claims 1 to 4 wherein the means comprise means adapted to inte_¬ grate a signal representative of the current, an_d to effect a reversal of the direction of flow of current when the integrated signal reaches the predetermined level.

9- An apparatus according to Claim 8 wherein the output of the integrator is connected to two comparators each provided with a respective reference voltage, each comparator being adapted to reverse the direction of current flow when it provides a predetermined output signal.

10. An apparatus according to any one of the pre_¬ ceding Claims wherein the power source comprises means to develop an off-set alternating current.

11. An apparatus according to Claim 10 wherein the said reversing means comprise ganged switches adapted to reverse the application of the off-set signal to the tube.

12. An apparatus according to Claim 10 wherein the reversing means comprise means for generating an off-set voltage and means for adding that off-set voltage to the power from the power source.

13. An apparatus according to Claim 12 wherein the off-set voltage is developed upon impedance means.

14. An apparatus according to Claim 13 wherein the i pedance means comprises capacitor means.

15. An apparatus according to Claim 14 wherein the voltage is applied to a capacitor from a D.C. power source, switching means being provided to reverse the D.C. potential applied to the capacitor by the power source.

16. An apparatus according to Claim 14 wherein two capacitors are provided connected in series, each capacitor being connected in parallel with a respective diode, the diodes being connected in the opposite sense, switching means being provided adapted selectively to short either one or the other of said capacitors.

17. An apparatus according to any one of the preceding Claims wherein means are provided to heat both the electrodes of the discharge tube.

18. An apparatus according to any one of the preceding Claims wherein the A.C. power source is provided with means which modulate the output frequency.

19. An apparatus according to any one of Claims 1 to 18 wherein the A.C. power source is provided with means to produce an output comprising two or more frequencies simultaneously.

20. A method of operating a discharge tube comprising the steps of directing alternating current electrical power to the tube, which alternating electrical current is such that there is a net flow of current along the tube in a predetermined direction over a plurality of cycles of the alternating power, and reversing the direction of the net flow of current after a selected or predetermined period of time, said period of time having a duration of at least 200 milliseconds.