GB2353150A - Fluorescent lamp driver unit - Google Patents

Abstract

A fluorescent lamp driver unit comprises an inductive and capacitive output stage 106 which is driven by a micro-controller 320 that controls the frequency of the signals applied to the output stage. The micro-controller uses a counter and a stable high frequency oscillator 324 to produce a very stable frequency signal for application to the output stage which results in closely controlled lamp operating parameters. Precise control of the operating frequency significantly improves the life of the lamp, operating conditions and produces a consistent output intensity and colour. The driver unit also allows the flexibility to change the lamp drive parameters.

Description

2353150 LAMP DRIVER UNIT AND LAMP The present invention relates to a lamp

and driver unit, and, more particularly, to a lamp and driver unit implemented using a microcontroller.

A fluorescent lamp driver unit as is well known within the art comprises electronic circuitry that is capable of converting electricity- from a power supply source into a waveform suitable for driving a fluorescent lamp. The power source may be either AC or DC.

The design of fluorescent lamp driver units is intended to take into account a plurality of design objectives. These design objectives include, for example, improving lamp life, improved mean time between failures, improved number of attainable lamp start up cycles without significant lamp deterioration, high operating efficiency, negligible audio noise, various safety features, reduced component count, reduced manufacturing c'osts and lighting environment enhancing features.

A fluorescent lamp driver unit is available from Excil Electronics Limited, Ripley Drive, Normanton, West Yorkshire, WF6 1QT. For example, Excil Electronics manufactures and markets a fluorescent lamp driver unit known as an "Excil 388 Series" lamp driver unit.

It is an objective of the present invention to provide an improved lamp driver unit that at least mitigates some of the problems of prior art lamp driver units.

Accordingly, a first aspect of the present invention provides a lamp driver unit comprising a microcontroller arranged to generate an output waveform having a predeterminable frequency; a lamp drive means comprising at least one inductive element and one capacitive element for 2 producing a drive signal for the lamp in response to the output waveform.

Preferably, an embodiment provides a lamp driver unit in which the microcontroller derives the output waveform from an oscillator waveform produced by an oscillator.

Further, an embodiment provides a lamp driver unit in which the microcontroller comprises means for dividing the oscillator waveform. by a predeterminable divisor.

Still further, an embodiment provides a lamp driver unit in which the means for dividing comprises a programmable counter arranged to vary the count thereof in response to the oscillator waveform and the output waveform is produced in response to an output signal of the programmable counter.

An embodiment provides a lamp driver unit in which the divisor is determined according to a predeterminable frequency.

A further embodiment provides a lamp driver unit in which the predeterminable frequency is determined according to lamp type.

A still further embodiment provides a lamp driver unit in which the predeterminable frequency is determined according to a stage of operation of a lamp.

Yet another embodiment provides a lamp driver unit in which the stage of operation is one of pre-heating, generating a strike voltage and maintaining an arc current waveform.

Preferably, an embodiment provides a lamp driver unit in which the predeterminable frequency is selected from a plurality of frequencies.

Preferably, an embodiment provides a lamp driver unit wherein the inductive and capacitive elements of the lamp drive means comprises an inductor connected in series with a 3 capacitor, and means for receiving the lamp such that, in use, the lamp can be connected in parallel with the capacitor.

It will be appreciated that a common problem with prior art designs is that the inner glass wall of the lamp adopts a blackened appearance close to the electrodes. The blackened appearance is caused by excessive stress to the electrodes particularly during the strike process resulting in electrode emissive material being jettisoned from the electrode surface onto the lamp inner glass wall. The appearance is undesirable as loss of emissive material from the electrodes results in lamp failure.

Accordingly, an embodiment of the present invention provides a lamp driver unit comprising means for heating the lamp electrodes prior to application of a strike voltage.

Preferably, the lamp electrodes are heated to the point of thermionic emission.

Advantageously, the means for pre-heating ensures a free flow of electrons from the electrodes prior to arc discharge (lamp striking). This, in turn, significantly reduces the amount of the electrode emissive material jettisoned from the electrodes onto the lamp inner glass wall during striking.

Preferably, the pre-heat period is fixed and is closely controlled. In a preferred embodiment, the pre-heat period is between 0.5 and 2 seconds.

It will be appreciated that one source of premature lamp failure is excessive electrode stress.

Accordingly, an embodiment of the present invention provides a lamp driver unit comprising means for ensuring the arc current waveform has a predeterminable crest factor. Still further, a preferred embodiment provides a crest factor of 1.4.

4 A further source of premature lamp failure results from electrode wear.

Therefore, an embodiment provides a lamp driver unit in which the drive signal comprises a substantially fully symmetrical arc drive current waveform. Preferably, the arc drive current waveform is a near-sine wave current waveform, that is, a substantially sine wave waveform.

Advantageously, such a substantially symmetrical drive waveform ensures equal gas conduction in both directions within a lamp and ensures substantially equal electrode selfheating during arc discharge and corresponding substantially equal electrode wear. This prevents mercury vapour migration within the lamp which is commonly manifested as light output from only one end of a lamp.

A problem often associated with fluorescent lamp operation is the generation of audio frequency noise. Any such noise pollution adversely affects the environment within which the lamp is situated.

Accordingly, an embodiment of the present invention provides a lamp driver unit operable at a predeterminable drive frequency. Preferably, the drive frequency is substantially between 30 kHz and 50 kHz according to lamp style.

Advantageously, the choice of drive f requency, in particular in conjunction with the near-sine arc current waveform, dramatically reduces audio noise generated during lamp operation. Furthermore, such a choice of frequency reduces RFI emitted from the lamp wiring.

It will be appreciated that lamp life is dependent upon the correct level of arc discharge current flow. It is the arc discharge current flow that generates self-heating within the electrodes during lamp operation when a pre-heating current has been removed or reduced.

Accordingly, an embodiment of the present invention provides a lamp drive unit providing means for maintaining a substantially stable lamp arc current. The substantially stable lamp arc current is maintained at a predeterminable value notwithstanding variations in power supply voltage.

It will be appreciated that the appropriate lamp arc current is dependent upon lamp type and lamp operating parameters. Preferably, an embodiment of the present invention provides a lamp drive unit in which the lamp arc current is held to within 6% of a preferred lamp arc current level indicated by corresponding lamp manufacturer data.

Advantageously, the self-heating process results in a free flow of electrons from the electrodes into the lamp gas due to thermionic emission and reduces electrode erosion.

Maintaining the lamp arc current to within predeterminable tolerances also advantageously ensures constant light output over a predeterminable input power supply range.

A source of premature lamp failure can often be attributed to excessive component stress within the product.

Accordingly, an embodiment of the present invention provides a lamp driver unit operable at predeterminable efficiencies. Preferably, the operating efficiencies are at least 84% and, more preferably, are between 84% and 90%. Advantageously, such high operating efficiencies result in low internal component temperature changes which, in turn, results in reduced component stresses and improved reliability of operation.

During a power failure, emergency lighting is often provided by a DC power source. The improved efficiency of 6 operation of the lamp driver unit according to an embodiment of the present invention enables, for a given DC power source, extended periods of emergency lighting or, for a given design period of emergency lighting, a battery rating reduction to be realised.

Maintenance and testing of installed conventional lamp driver units often carries a risk of electric shock.

Accordingly, an embodiment of the present invention provides a lamp driver unit comprising means arranged to provide a floating lamp drive output. Preferably, the means arranged to provide a floating lamp drive output comprises an output transformer wound for 2 kV primary to secondary isolation. The risk of electric shock is advantageously reduced.

is It will be appreciated that there are many ways in which a lamp and lamp driver unit may be inadvertently damaged. For example, the following may result in lamp driver electronics failure: misconnection of the lamp, DC reverse polarity, failure of the lamp, open circuit lamp, short circuit of the lamp arc for an indefinite period and short circuit of either lamp electrode for an indefinite period. Typically, conventional prior art lamp driver units rely upon fuses to provide protection against the above. It will be appreciated that the provision or use of such fuses increases component count and manufacturing and maintenance costs.

Accordingly, an embodiment of the present invention comprises means for electronically providing at least one protection feature.

An embodiment of the present invention provides a lamp driver unit comprising means for maintaining a substantially stable arc discharge current within the lamp.

7 Such a substantially stable arc discharge current provides relief from colour bias and assists in ensuring consistent illumination levels and colour between lamps.

It is often the case, in the event of a failed lamp, with some electronic ballast designs and conventional electromagnetic lamp driver systems, that multiple re-strikes will be attempted. Clearly such failed multiple re-strikes are undesirable in that they provide an unpleasant environment and unnecessarily expend power.

Accordingly, an embodiment of the present invention provides a lamp driver unit comprising means for preventing attempted re- strikes in the event of a failed lamp. Advantageously, long term continuously attempted re-strikes in the event of lamp failure are avoided.

A further embodiment provides a lamp driver unit comprising a cut-out means for inhibiting the use of a lamp upon detection of a falling supply voltage or a supply voltage that is below a predetermined level. Preferably, the cut-out means is provided with some hysteresis to assist in providing a "clean" cut-out, that is to say, to avoid flickering as a failing supply voltage varies.

It is often the case with conventional driver units that the lamp or driver electronics may be damaged as a consequence of surges and transients present on a power supply.

Accordingly, an embodiment of the present invention provides a lamp driver unit comprising transient and surge suppression means.

Embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings in which:

figure 1 illustrates a lamp driver unit according to an embodiment; 8 f igure 2 shows an L/C circuit used in a lamp driver stage of an embodiment; figure 3 depicts schematically a lamp driver unit according to an embodiment; f igure 4 shows a circuit diagram of an embodiment; and f igure 5 shows a f low chart of an embodiment.

An embodiment of a lamp driver unit, for example a fluorescent lamp driver unit, according to the present invention comprises two independent functional stages, as can be seen from figure 1. The first stage 102 is a power supply stage, that is, a step-up boost converter. The second stage 104 comprises the lamp driving circuitry. The step-up boost converter 102 increases an input supply voltage, Vi,pu,, to a much higher voltage and operates in a closed loop mode so that the output voltage is maintained substantially constant irrespective of variations in the input voltage, Vi,p,,t.

In the case of an AC power supply being connected to the boost converter 102, the boost converter 102 operates on a full wave rectified but unsmoothed version of the AC supply and is power fac.tor correcting. Such a power factor correction stage employs active correction techniques so that the current waveshape consumed from the AC power supply is substantially identical to the waveshape of the applied voltage, which is usually a sine wave. Such a power f actor 2S correction stage results in low power supply harmonic distortion and a high power factor, typically exceeding 0.95.

The high voltage output produced by the boost converter 102 is known as a link voltage, Vlink. The link voltage is, typically, in the region of between 250 V DC and 380 V DC.

The stable, regulated link voltage is output from the boost converter 102 and input to the lamp driver stage 104. operating the lamp driver stage 104 using the stable regulated 9 link voltage produced by the boost converter 102 assists in ensuring or maintaining a substantially constant light output over the full permissible input power supply voltage range. The lamp driving stage 104 produces a square wave drive waveform using a symmetrical power drive circuit configuration that is described hereafter.

The square wave drive waveform is applied either directly or indirectly via an isolation transformer to an L/C resonant circuit 106 which acts as the lamp drive interface.

The L/C resonant circuit 106 is used to produce the three operating conditions required by the lamp 108; namely, preheating current, strike voltage and lamp operating arc current. It will be appreciated that the nature of the resonant circuit results in the lamp arc current operating under near sine wave current drive conditions.

Referring to figure 2, there is shown schematically the L/C resonant circuit 106 together with the lamp 108. According to an embodiment of the present invention, the L/C resonant circuit 106 is driven using a variable frequency, fixed amplitude square wave 202. The capacitance of capacitor C is preferably fixed as is the inductance of the inductor L. The three stages of ignition, that is, electrode pre-heat, strike voltage generation and arc current regulation are realised by varying the drive frequency of the square wave 202. It will be appreciated that the resonant frequency of the L/C resonant circuit 106 is fixed at some predeterminable value given by.

Fr = 1/27c J(L.Q.

The frequency of the square wave 202 is arranged such that it is preferably always above the fixed resonant frequency of the L/C resonant circuit 106. The drive frequency can be varied under the control of a microcontroller as described hereafter.

During an electrode pre-heat period, the square wave frequency is arranged to be high and well above the resonant frequency of the L/C resonant circuit 106. For example, the square wave frequency, Fp, is initially 59.5 kHz for a 58 Watt, T8, 1500 mm lamp. It will be appreciated as a result of the high drive frequency, relative to the resonant frequency of the L/C circuit 106, that voltage magnification is low and the voltage generated across the lamp arc, Va, is also low and insufficient to cause gas breakdown and strike the lamp.

However, during this stage, at frequency, Fp, current flows via the lamp electrodes 204 which causes pre-heating of these electrodes. The preheating current and duration of the preheating period is arranged such that thermionic emission from the electrodes results. Both the requisite pre-heating current and pre-heating period can usually be obtained from or derived from electrode emission constants available from manufacturers' data sheets.

An embodiment is provided in which the pre-heat period has a duration of between 0.5 seconds and 2 seconds. An embodiment is also provided in which the pre-heat current may be predetermined to be between 0.2 and 1.1 amps according to a given lamp. It will be appreciated that for a given model of lamp the pre-heat current and/or pre-heat period are substantially constant.

After completion of the pre-heating period, the operating frequency of the square wave 202 is reduced gradually, under the control of the microcontroller, over a predeterminable period, typically 100 milliseconds, to a strike frequency, Fs.

The strike frequency is arranged to be significantly closer to the resonant frequency of the L/C resonant circuit 106 than the pre- heating frequency. Preferably, the strike frequency is arranged to be above the resonant frequency. It will be appreciated that initially the lamp arc impedance will be high and does not load the resonant circuit. Accordingly, at the strike frequency, significant voltage magnification occurs and the resultant voltage developed across the capacitor, C, is high and known as the strike voltage. The strike voltage is that voltage which is sufficient to cause gas breakdown and thereby strike the lamp. In a preferred embodiment, the strike voltage has a value of X V. However, it will be appreciated that the strike voltage will also vary according to the model of lamp utilised.

Once the lamp 108 has struck, the arc impedance drops considerably and an arc current flows which results in light output. The drop in impedance across the lamp causes severe damping of the L/C resonant circuit which is arranged to result in a reduction of the arc voltage to the normal operating voltage for the lamp. The arc current for the lamp is principally determined by the value of the inductor L and is the operating frequency.

Once the lamp 108 has struck, the operating frequency of the square wave 202 is changed to Fr,,,,, which is the frequency required for correct operating arc current. The operating frequency is either equal to or greater than the strike frequency. This provides further independent means of determining or controlling the arc current rather than relying solely upon the value of the inductor L. This is because arc current is determined by the impedance of L ie X,=27EFnmL. Therefore, both Fm and L can influence X,. Preferably, the running f requency, Fn,,,, will be identical to the strike frequency Fs, but not essentially so.

The above frequencies are derived via software division of a signal from a high accuracy frequency source such as, for example, a ceramic resonator or crystal. The high frequency accuracy results in extremely accurate lamp drive parameters, such as pre- heat current, strike voltage and arc current, which are all frequency dependent, being derived from the L/C resonant network 106. The accuracy of controlling these parameters assists in improving lamp life from the start up cycle through to the normal operating discharge.

Advantageously, implementing the lamp driver stage 104 using a microcontroller allows different lamps to be accommodated simply by making software changes.

Referring to figure 3, there is shown a block diagram of an embodiment of the present invention illustrating the functional elements of the embodiment. Typically, an embodiment of the present invention may be utilised to drive fluorescent lamps on, for example, a railway carriage.

However, it will be appreciated by those skilled in the art that the present invention is not limited to application within the railway industry and can equally well be used in any other environment.

The schematic diagram 300 illustrates a transient protection network 302 which receives as an input a supply voltage, Vi,p,t. The input voltage is typically derived from a vehicle supply voltage. The transient protection network 302 provides protection against transients and surges in the vehicle supply.

The transient protection network 302 is coupled to a conducted emission filter and reverse polarity protection circuit 304 which provides for protection against incorrect connection of the vehicle supply to a boost converter 306 and driver circuit in general. The conducted emission filter and reverse polarity protection circuit 304 also prevents switching noise from the boost stage and driver stage being passed to the input power supply wiring and assists in ensuring compliance with European EMC regulations.

The boost converter 306 is coupled to the conducted emission filter and reverse polarity protection circuit 304 and arranged to produce a regulated and stable high voltage output.

13 The boost output feeds a push/pull driver stage 307 consisting of high and low side power devices. The high and low side driver devices are under the control of a controller 326. optionally, an isolation transformer 308 receives the S output of the driver stage and couples it to the L/C resonant circuit 106 including the lamp 108.

A start-up circuit 310 is coupled to the output of the conducted emission filter 304. The start-up circuit is utilised to provide a low voltage power supply for powering the control electronics and is only used for a brief period at system power-up. Once the boost converter is operational, the low voltage power supply for the control electronics is provided by the tertiary power supply 312 which is derived from the boost converter.

It will be appreciated that Vcontrol is the power supply for the control electronics of an entire embodiment. Vcontrol is initially provided by the start-up circuit and then by the boost converter tertiary supply. The two diodes enable mixing of the start-up and tertiary power supplies.

Preferably, an embodiment of the present invention also comprises a low voltage detector 314 arranged to provide or detect insufficient vehicle supply power levels or a failing vehicle power supply.

Interface circuitry 316 is arranged to provide, via a lamp status feedback signal 318, an indication to the microcontroller 320 of the status of the lamp.

The microcontroller 320 also receives a low voltage feedback signal 322 from the low voltage detector 314.

As indicated above the L/C circuit and lamp are driven using an accurately controlled high frequency square wave signal 202. The square wave signal 202 is derived from a very accurate frequency source 324 and supplied to the L/C circuit 14 via the microcontroller 320 and a high/low side driver 326 with dead time.

Preferably, a power on reset circuit 328 is coupled to the microcontroller 320 and is arranged to reset the microcontroller 320 at power-up.

Referring to figure 4, there is shown in greater detail schematic circuitry of an embodiment of a lamp driver unit according to the present invention. The driver unit 400 comprises two independent stages. As described in relation to figure 1, the first stage is a step-up boost converter 102 which increases the input supply voltage, Vi,,put, to a regulated link voltage, Vlink, that is arranged to be substantially constant over the full range of permitted input supply voltage, Vi,,p,,.

The second stage is the lamp driver stage which operates using the stable link voltage produced by the step-up boost converter 102. By utilising a regulated link voltage the lamp light output is substantially constant over the full power supply variation of the input voltage, Vinput The lamp driver stage 104 comprises a switched mode square wave generating circuit 402. The square wave generating circuit consists of U4 and drive mosfets Q4 474 and Q5 476, which feed the lamp via the L/C resonant circuit 106.

The resonant circuit comprises an inductor L3 and a capacitor C17. Preferably, the inductor has an inductance of 0.985 mH.

Preferably, the capacitor is rated 15 nF/160OV.

DC power, Vi,,p,,t, is coupled to the lamp driver unit at junction J1 (404) and junction J2 (406). Junction J1 is the positive supply connection and junction J2 is the ground connection. The input voltage is applied to the transient protection and filter networks 302 and 304. The transient protection and filter network comprises a 2 amp, 100iH inductor Ll 408, a 630V, 1ON capacitor C1 410 and a voltage dependent resistor M1 412 or metal oxide varistor available from Harris-semiconductors, part no. V130LA20A. The protection/ filter network is coupled via a fuse F1 414, a common mode choke LB 416, a reverse polarity protection diode D23 418 and a differential mode choke L2 420 to the boost converter circuitry 102.

Preferably, the fuse 414 is a 2.5A anti-surge fuse, the common mode choke 416 is a RN 124 available from Schaffner Components, the reverse protection polarity protection diode 418 is a UF5406 available from General Instruments, and the differential mode choke L2 420 is a 100 uH, 0.91A inductor.

It will be appreciated that inductors LB and L2 reduce emissions conducted onto the DC supply wiring in compliance with European EMC standard EN50121. Inductor LB 416 is a dual winding device and is designed for common mode interference suppression. Inductor L2 420 is a single winding device and is designed for differential mode interference suppression. The capacitor C1 410 also serves to reduce differential mode interference.

It will be appreciated that the incoming power supply, Vi,pu,, may be contaminated with high voltage transients of both positive and negative polarity. Therefore, a fast recovery device is selected for the reverse polarity protection diode 418 to ensure adequate blocking of the negative transients during which the device will be reserve biased.

The boost converter 102 comprises an inductor L7 422 (windings pin 2 to pin 6), a mosfet Q3 424, diode D5 426 and a capacitor C9 428. Mosfet Q3 may be an IRF740 available from International Rectifier, diode D5 may is a BYT03-400 available from SGS Thomson.

The boost converter 102 operates in a closed loop mode with a regulated output under the control of U1 430. U1 is a TL4941D available from Texas Instruments. U1 428 operates in a fixed frequency, variable duty cycle mode. The variable 16 duty cycle output from U1 is derived from pins 9 and 10 and used to control the resultant output voltage by varying the ratio of the on to off time of the mosfet Q3 424. The duty cycle is adjusted in response to the feedback from the boost converter output voltage.

In the case of an AC power supply, the boost converter 102 operates on a rectified but unsmoothed version of the incoming power supply voltage, Vi, ,P,,t, and provides power factor correction. Such a system may operate in variable frequency mode.

It will be appreciated that energy is stored in inductor L7 422 when mosfet Q3 424 is switched on and released via diode D5 42G into capacitor C9 428 when mosfet Q3 424 is switched off. The boost converter 102 operates in this discontinuous mode for the majority of power supply means. Therefore, the energy stored within inductor L7 422 while mosfet Q3 424 is "on" is completely exhausted by means of transfer to capacitor C9 428 when mosfet Q3 424 is "off". The output voltage of the boost converter 102, that is to say, the voltage across capacitor C9 428, is maintained at a substantially constant level by closed control which is achieved by varying the duty cycle of the mosfet Q3 424 in response to feedback from the output of the boost converter 102.

It will be appreciated that when mosfet Q3 424 is on, pin 6 of inductor L7 422 is pulled low and the inductor L7 422 is effectively connected across the DC power supply with pin 2 at +Ve and pin 6 at -Ve (ground). Accordingly, the current of inductor L7 422 increases linearly with time and energy is stored within inductor L7 422. When mosfet Q3 switches off, the energy stored within inductor L7 422 must be released. As a result, pin 6 of inductor L7 422 swings positively in polarity with respect to pin 2. The positive swing continues until diode D5 426 is forward biased. The point at which 17 diode D5 426 is forward biased is determined by thevoltage across the capacitor C9 428, which is in turn set by the negative feedback, variable duty cycle control system, and is selected to be a value that is considerably above the DC input voltage, V ip,t. It can therefore be appreciated that voltage step up occurs with a substantially constant output voltage resulting across capacitor C9 428, which is referred to as the link voltage as described earlier.

U1 430 contains an amplifier and variable duty signal pulse width system. The device U1 430 is arranged to compare the voltage feedback signal of the link voltage with a fixed voltage reference and to adjust the drive duty cycle according. If the link voltage is low, the duty cycle is increased. If the link voltage is high, the duty cycle is decreased. A stable voltage reference is generated by Z2 which is compared within U1 to a proportion of the link voltage derived via R11, R12, R13 and R4. The control loop is satisfied when the reference voltage and the derived proportion of the link voltage are equal. The duty cycle is adjusted to attain such equality. It will be appreciated that since a proportion of the link voltage is compared to the reference, the system has voltage gain determined by:- Boost Gain = 1+ ( (R11 + R12 + R13) /R4) Hence, the link voltage is determined by the above resistor values.

The operating frequency of U1 is determined by R8 and C4 and given by the approximate relationship F= 1.1/(R8.C4) and is chosen, in an embodiment, as approximately 55 kHz.

Two internal uncommitted transistors form the output of U1 (pin 8/9 and pin 10/11) and in this embodiment are wired in parallel as emitter followers with collectors tied high to the low tension power supply line "Vcontrol". The variable duty cycle output present at pin 9/10 (uncommitted transistor emitters), therefore requires a pull-down which is 18 accomplished by R14. As the duty cycle output (pin 9/10) has high positive drive but only a passive pull down, the signal is buffered in the negative drive polarity by Q2 but simply passed on in the positive polarity by D4, to form a signal suitable for driving Q3.

During normal operation, the low tension power supply line "Vcontrol" is provided by a tertiary winding on the boost inductor (L7 pin 7/10/11). However, the tertiary winding voltage is not initially present at powerup and an alternative means of generation of "Vcontrol" is required until boost oscillation occurs. The alternative source is provided by start-up circuitry R69, Z1, Q1, R70 and D2. This configuration forms a simple emitter follower power supply with an output voltage set by a zener diode Z1 448. The diode is D2 452 forms a means of override such that the tertiary winding supply may assume or take over the provision of "Vcontrol" when the voltage exceeds that of the startup circuit 444.

The tertiary winding has two taps that are selected by inclusion of either of diodes D24 454 and D25 456. The selection of the taps allows a wide range of link voltages to be accommodated as differing link voltages will result in different voltages across the main winding (pins 2 and 6 of inductor L7 422) and hence across the tertiary winding.

Tertiary rectification is provided by diodes D24 4S4 or D25 456. Smoothing is provided by capacitor C45 458 which is rated 1 uF/35V. An emitter follower regulated power supply 460 is arranged to accommodate changes in the input power supply voltage, Vinput Y since the tertiary voltage is proportional to the main winding voltage which is, in turn, dependent upon the input power supply voltage, Vip,,. Schottky diode D26 462 allows mixing of the tertiary derived voltage with the start-up circuit voltage such that the highest voltage acts as the Wcontrol" supply. After start up, the "Vcontrol" supply is provided by the tertiary derived voltage.

19 It will be appreciated that although U1 430 does have an internal voltage reference, this internal voltage reference is insufficiently accurate to act as a boost converter closed loop control reference. However, the internal reference is made available f or external use at pin 14 and is used as a +5V power supply. It can be seen that a 50 volt, 100 nanofarad capacitor C35 464, a 10 Ohm, 0.1 Watt resistor R64 466 and a 50 volt, 100 nanofarad capacitor 468 provide filtering and de-coupling for the +5V supply.

The link voltage, that is the voltage across capacitor C9 428, is passed to the lamp driver stage 104 via a ELM 41-P7505 inductor L4 and a 1 kV, 1 nanofarad capacitor C38 472. Inductor L4 470 is a ferrite bead manufactured by Murata, part no. BLM41-P750S. The inductor 470 and capacitor 472 form a high frequency filter to prevent high frequency ringing, typically present at the boost converter 102 output, from progressing into the lamp driver stage and hence the lamp wiring.

The lamp drive power stage 402 comprises a pair of mosfets Q4 474 and Q5 476 configured in a half bridge, push/pull mode. The square wave output from the half bridge configuration, which can be observed at test point nine drives the lamp via the L/C resonant circuit 106 comprising a 0.985 mH inductor L3 478 and a 1,60OV, 15 nanofarad capacitor C17 480. In all modes of operation, that is preheat, lamp strike and arc current regulation, the half bridge output frequency is arranged to be above that of the fixed resonant frequency of the L/C circuit 106. This ensures that the half bridge load always appears inductive and results in zero-volt mosfet switching.

Zero-volt mosfet switching advantageously reduces device dissipation by ensuring that the drain/source voltage is zero when the device is switched on thereby reducing switching losses. The two output mosfets 474 and 476 are driven under the control of U4 482 which is a dedicated combined high and low mosfet driver device. In an embodiment U4 can be realised using an IR2104S. Mosfets Q4 474 and Q5 476 are operated in anti-phase under the control of U4 482. To prevent cross conduction of the output of mosfets Q4 474 and Q5 476, U4 incorporates a built in "dead time". The dead time period is a short period in every drive cycle when neither mosfet device is "on". Without 11dead time", there would be drive overlap, where, for a short duration, mosfets Q4 474 and Q5 476 would be in the "on" state. This would result in high transient current peaks consumed from V,i,,, and possible device failure.

In an embodiment the dead time is preferably 500 nsec.

In addition to preventing cross conduction, inclusion of dead time also allows mosfet commutation or "zero-volt is switching". Assuming, initially, that mosfet Q4 474 is conducting, the half bridge output would be in the high state.

When the drive to mosfet Q4 474 terminates, the inductive nature of the load swings the half bridge output negatively.

This occurs during the dead time. The negative swing continues until caught by the intrinsic body diode of Q5. This action is known as commutation as at the end of the dead time when mosfet Q5 476 switched "on" the drain/source voltage thereof is already zero. The reverse happens on the next cycle with mosfet Q4 474 switching with zero drain/source voltage.

A snubber network 484, comprising a 1 watt, 10 ohm resistor and two 2 kV, 1 nanofarad capacitors C13 and C14 arranged in a series, limits the slew rate of the half bridge output and thereby reduces RFI during commutation. The snubber network 484 is arranged to ensure that slewing of the output has ceased prior to the end of the dead time thereby ensuring zero-volt switching.

Device U4 462 is a dedicated high/low side drive device.

It accepts as an input at pin 2 a single square wave and f rom.

21 this generates the drive signals for the high and low side device mosfets Q4 474 and Q5 476. The U4 device 482 also arranges the dead time between the high and low mosfet conduction. It will be appreciated that since both devices of the half bridge configuration are 'IN" channel, the upper mosfet device Q4 474 requires a gate drive voltage that is higher than the voltage present at the drain of mosfet Q4 474 to achieve saturation. Without such a condition being satisfied, the device would operate in a follower mode.

Satisfying this condition is achieved using a charge pump comprising a diode D16 486 and a 63 volt, 100 nanofarad capacitor C26 488. Pin 8 of device U4 482 is the power supply pin for the upper 474 drive stage. When the half bridge output is low (Q5 476 on), capacitor C26 488 is charged to I'Vcontrolll via diode D16 486.

The voltage across capacitor C26 488 is used to provide pin 8 of device U4 482 with a voltage that is higher than that of the drain of mosfet Q4 474 when that mosfet is "on" as the voltage of the capacitor C26 is added to that of the half bridge output. Such an arrangement ensures sufficient drive voltage for the gate of mosfet Q4 474 and ensures full saturation. The drive of the lower mosfet Q5 476 is readily available from Vcontrol.

The device U4 482 comprises a shut down function at pin 3. If pin 3 is pulled to a low state, the drive of both output mosfets 474 and 476 is inhibited. This function is described in further detail hereafter.

The output of the half bridge stage is used to drive the lamp via the L/C network comprising inductor L3 478 and capacitor C17 480. The L3/C17 network is driven using a variable frequency, fixed amplitude square wave.

In an embodiment, the inductor L3 is realised using an industry standard IIRM1011 core/bobbin style inductor. A low loss ferrite material is preferably utilised such as, for 22 example, a Philips 3C85. It should be appreciated that care should be taken during design/calculation of this component to ensure freedom from saturation particularly during the generation of the strike voltage where current flow is high.

Freedom from saturation is achieved by selection of a core with an adequate gap and high volume. There are certain lamp types that require a high strike voltage, and an RM12 core may be required thereby giving greater core volume. During strike voltage generation, the 3C85 material may be operated to a peak flux density of substantially 300 mT as this will drop considerably when the strike voltage ceases and the arc current begins to flow. The drive frequency of the L3/C17 resonant network is governed by software running within the microcontroller U3 490 and a crystal device X1 492. The is software is arranged such that the driving frequency appears on U3 device line "PBO", that is, pin 6. The driving frequency is then passed directly to the half bridge driver device U4 482 at pin 2.

Preferably, the software is used to realise a multi- strike attempt function whereby if the lamp fails to strike on the first attempt, further attempts will be made. This function is realised by repeating the pre-heat/strike voltage cycle with appropriate pauses between attempts. Typically, a period of 1 second is allowed between strike attempts. The number of attempts is configurable and under software control.

Furthermore, if the lamp does not strike after a predetermined number of attempts, for example three attempts, the lamp driver unit enters a shut down mode and lamp driving attempts are halted. This arrangement ensures that annoying flickering of life expired lamps associated with conventional electro - magnetic ballast systems is eliminated. Furthermore, the output stage is also shut down to prevent continuous heat dissipation in the output stage which could lead to failure of the lamp driver unit.

23 Preferably, a lamp struck detector 494 provides an indication of a successful lamp strike. A measure of arc current feedback is provided at CT1/CT2 via a current transformer 496. The current transformer 496 comprises three independent, single turn windings, two of the windings are the primary circuit and are wired in series with the connections to one lamp electrode "Elect 211 and the remaining winding, the secondary circuit, used to provide feedback signal CT1/CT2.

It will be appreciated that passing both connections of one lamp electrode via the primary windings results in an output from the secondary winding that represents arc current, that is, the light producing current which flows between the two lamp electrodes. During preheat, the output of the secondary winding is zero since the two primary windings carry equal but opposite currents. The secondary output produced is a current which therefore requires a load resistor R22 to produce a useful voltage at CT1/CT2. In anembodiment, resistor R22 is realised using a 1/8W, 2R2 resistor.

Signal CT1/CT2 is applied via a low pass filter 498 to a comparator arrangement comprising an LM2903D, a 0.1W, 150R resistor R15 and 1 nF/50V capacitor C12. The output of the comparator, pin 7, is an open collector type and has a pull down ability only. Therefore a 0.1W, 10K resistor R21 is provided. The output, STK, of the comparator is converted to a steady DC level via the arrangement of resistors R21, R23 and capacitor C15. In an embodiment, the resistor R23 is a 0.1 Watt, 1KO resistor and the capacitor is rated 1 uF/25V.

Without the network of R21, R23 and C15, STK would normally be high in the absence of arc current and a square wave in the presence of an arc current. The resistor-capacitor network R23 and ClS maintains a logical high state in the absence of an arc current and a logical low state in the presence of an arc current.

24 STK is passed to the microcontroller port input PBS (pin 11) and can be read internally by software.

In a preferred embodiment, there is provided a low voltage detector 500 which is used to prevent attempted drive during low power supply voltage conditions. Additionally, the low voltage detector 500 is operable to produce a clean shut down of the lamp during a slowly falling supply voltage which prevents erratic behaviour exhibited by many prior art driver units.

The low voltage de.tector 500 comprises hysteresis. on a rising supply voltage, the low voltage detector will permit lamp operation when a higher threshold "pick up" voltage is reached. However, on a falling supply voltage, lamp operation is inhibited when the supply voltage falls below a lower threshold "drop out" voltage. The hysteresis also provides tolerance of power supply ripple. In an embodiment, f or a 110V DC nominal Railway power supply, the higher threshold "pick up" voltage is 67V and the lower threshold "drop out" voltage is GOV. The low voltage detector 500 comprises a U2A 502 and associated components. A portion of the DC supply voltage is derived and filtered by an RC network comprising resistors R29, R30 and capacitor C19 and applied to the comparator 502 for comparison with a fixed reference voltage, Vref - Hysteresis is realised by a resistor-diode network comprising a O.1W, 6K8 resistor R28, an LL4148 diode DIO and a O.1W, 18K resistor R32.

The comparator output, at pin 1, drives a buffer stage transistor Q6 504 via interface circuitry comprising a Zener diode Z3 506, a 0.4W 3k3 resistor R33 508 and a O.1W, 47K resistor R24 510. The purpose of the Zener diode Z3 506 is to prevent spurious operation of the transistor Q6 504 due to erroneous operation of the comparator 502 under low supply voltage conditions.

The low voltage detector 500 also comprises means for providing a unidirectional delay function 512 such that when the pick up voltage is exceeded a small delay follows before operation begins to allow board initialisation. The pick-up voltage is that voltage at or above which the low voltage detector circuitry permits circuit operation. However, f or a falling supply voltage, the operation is arranged to terminate substantially immediately if the supply voltage falls below the drop out threshold.

In a preferred embodiment, the means for providing the uni-directional delay function comprises a 0.1W 10K resistor R25, a 0.1W, 680K resistor R26, a capacitor C18 rated 1 uF/25V, a diode D1 and a mosfet Q7.

A signal is generated on line IIVDU" 514 by a transistor is Q8 516 which is passed to the microcontroller 490 and read internally by software. If the VDU line is low then normal operation of the board continues. However, if the VDU line is high, this provides an indication that the low voltage detector 500 has determined that the operation of the lamp driver unit should be terminated.

The microcontroller 490 comprises means to shut down the operation of the lamp driver stage under software control. Shutting down the lamp driver stage is accomplished via the shutdown pin of device U4 482 (pin 3) and the signal SD. The 3D function shuts down when the line is in a logical low state. Normal operation follows when the line is in a logical high state.

The TD line is controlled via microcontroller 490 port line PB4 (pin 10) and, due to the action of transistor Q9 518, a low state at port pin 10 results in a high state of the SD line only if the output, VDU, from the low voltage detector circuit is high.

26 It will be appreciated that the low voltage detector 500 is arranged to produce a logical high signal on line VDU 514 when an adverse input voltage, Vj,,p,, has been detected. it will be appreciated that this arrangement, based around transistor Q9 518, ensures that SD is low if "VDU'1 is low irrespective of the microcontroller instruction.

In a preferred embodiment, the microcontroller 490 is coupled to a power up reset circuit for supply a power reset signal to the microcontroller. In a preferred embodiment, the power up reset circuit comprises a 0.1W, 47K resistor R34 520, a 50 volt, 10ON capacitor C23 522 and a diode D11 524.

optionally, to reduce the risk of electric shock from the lamp driver unit, galvanic isolation is provided between the lamp driver stage and the lamp output. Galvanic isolation ensures that the lamp drive output is floating notwithstanding the status of the power supply input, Vi,p,,t. It will be appreciated that there is no DC electrical continuity between the lamp drive output and the input power supply. However, a small high frequency AC path may exist due to capacitive coupling.

In a preferred embodiment, galvanic isolation is achieved by the use of a 1:1 ratio isolation transformer inserted between the half bridge push/pull driver stage 402 and the output resonant circuit 106. The transformer primary winding should preferably be AC coupled which is achieved, in an embodiment, by use of two series connected capacitors arranged across "Vlink" and configured to create a "Vlink/211 AC coupled centre point reference. The primary winding of the isolation transformer is connected, at one end, to the half bridge output (TP9) and the other end of the primary winding is connected to the centre tap between the series connected capacitors. This arrangement ensures freedom from any DC component that is present at the primary winding. The snubber network 484 is, preferably, retained on the primary (driver 27 stage) side and connected directly across the primary winding.

The secondary winding output is connected, at one end, to inductor L3 478 and the remaining end is connected to the output circuit AC coupling capacitor C16 526., In an embodiment of the lamp driver unit the capacitor C16 is a 250V, 470n capacitor.

A suitable isolation transformer, in an embodiment of the lamp driver unit, could be a transformer based around an assembly such as, for example, an ETD29. A low loss core material should preferably be used such as a Philips 3C85 or Philips 3C90 core. If such cores are used, the primary winding turns quantity should be calculated such that the resultant flux does not exceed 200 mT to 250 mT, which is an upper limit of flux. Ensuring that the flux does not exceed 200 mT to 250 MT ensures freedom from saturation and results in a low loss core.

It will be appreciated that inclusion of an isolation transformer has a further advantage in that a small core gap may be provided which will serve to increase primary energy storage during open circuit lamp conditions. This will aid commutation of the output stage resulting in reduced dissipation of mosfets Q4 and QS under "no lamp" conditions until software shutdown occurs.

Another advantage of the inclusion of an isolation transformer is that the device may be wound to provide a step up or step down function. The former being utilised to combat low Wlink" conditions due to a lack of step up dynamic range within the boost function.

The microcontroller 490 is employed as a management means for the lamp driver unit. The microcontroller 490 produces or controls the output to the L3/C17 network at the appropriate operating frequencies and can thereby fully control the lamp drive parameters pre-heat, strike voltage generation and operating arc current.

28 Additionally, the microcontroller 490 is employed to perform complete sequencing of the lamp driver unit and to perform continuous monitoring tasks. A summary of the microcontroller functions is shown below:

S (a) control of pre-heat current by control of operating frequency; (b) control of strike voltage by control of operating frequency; (c) control of operating lamp arc current by control of 10 operating frequency; (d) control of pre-heat duration; (e) implementation of multi-strike attempt system; (f) continuous monitoring of arc current flow instigating re-strike if lamp extinguishes; (g) monitoring of supply voltage in conjunction with external hardware; and (h) ability to shut down the lamp driver stage in the event of strike failure.

Referring to figure 5, there is shown a flowchart which illustrates the flow of control of the microcontroller 490. It will be appreciated that the software for running the microcontroller 490 is stored and addressed internally.

The generation of the appropriate operating frequencies is accomplished by the use of an internal eight-bit timer/counter in conjunction with an interrupt routine. The counter is advanced by a clock signal which is internally derived by frequency division of an external 20 MHz crystal 492. Within the software, a predetermined preset count is loaded into the counter and the counter is then advanced via the internally derived clock signal.

29 The microcontroller 490 is configured such that when overflows occurs, an interrupt request is generated which causes an immediate jump to an interrupt service routine. The interrupt service routine is arranged to toggle the frequency output port line PBO (pin 6) each time the interrupt routine is called.

Additionally, the interrupt routine is arranged to reload the counter with the preset count. As is well known within the art, the state of the microcontroller registers are pushed onto and pulled off a stack of the microcontroller at the beginning and end of the interrupt routine respectively.

The counter/timer operates substantially independently and the preset count is used to determine the operating frequency produced. It will therefore be appreciated that the is generation of the different frequencies required for pre-heat, strike voltage and arc current regulation is attained by loading appropriate pre- set count values into the counter, that is, the generation of different frequencies is under software control and can be altered readily to accommodate different lamps.

The microcontroller device, in a preferred embodiment, is realised using a PIC16C620A and uses an eight bit counter with a clock derived from a division by four of the 20 MHz crystal signal. Although this provides an adequate resolution for 2S most applications, a greater resolution may be achieved by using a microcontroller comprising a counter having a greater number of bits and a device where the counter/timer clock is derived from a higher frequency source.

Referring to figure 5, the main routine begins when the external power up reset period, set by the RC time constant resulting from the combination of resistor R34 520 and capacitor C23 S22 expires. The system variables are initialised at step 500. The system variables fall into two categories, that is, device specific and application specific variables. The set up of the device specific variables includes configuration of the port lines PBO to PB7 of the microcontroller 490. The configuration is required to ensure the correct input/output combination. Further variables are configured such that the internal counter/timer triggers an interrupt in the event of an overflow. Further details relating to the configuration of the variables used by the microcontroller are shown in the code of Table 2.

The initialisation step 500 also configures the application specific variables which include disabling the lamp driver stage via port line "PB411 of microcontroller 490 and selection of an initial start up operating frequency which is above the pre-heating frequency. After initialisation there is preferably a brief delay introduced at step 502 to is allow the product hardware to stabilise. Upon expiry of the delay, the status of the low voltage detector is determined via port line "PB111 of microcontroller 490 at step 504. if the low voltage detector 500 indicates that the supply voltage, Vi,,P,,t, is above the pick up threshold, control proceeds to step 506. However, if the low voltage detector 500 indicates that the supply voltage is below the pick up threshold, the step of determining, that is, step 504, is repeated until the supply voltage, Vinput, exceeds the low voltage detector pick up voltage.

As indicated earlier for a vehicle system based on a nominal voltage of 110V DC, the pick up voltage would be selected as 67V and the drop-out voltage would be selected as 60V. The voltages are selected to prevent excessive discharge of the vehicle back-up battery in the event of a power failure such as a loss of overhead line voltage for an electric locomotive or loss of alternator charge voltage for a diesel locomotive. In the absence of this feature, the lights would be operated irrespective of battery discharge level resulting in lack of power for other essential train features and also possibly causing battery damage.

31 Once the low voltage detection criteria has been satisfied, the output frequency is set to that required for pre-heating, Fp, and the lamp driver stage is enabled which results in flow of a pre-heating current in the lamp electrodes, at step 508. In an embodiment, for a 58 Watt, T8, 1500 mm lamp and the resonant components of figure 4, the preheating frequency, Fpis 59.5 kHz. The pre-heating duration is under software control and is selected together with the pre-heat current to attain thermionic emission within the electrodes. Table 1 below illustrates pre-heat currents and durations for various lamps which all comply with EN60081.

Lamp Pre-heat Pre-heat Arc Strike current, duration, Current, Voltage, mA (RMS) sec. mA(RMS) V (RMS) Watt, T8 lamp, 980 1 470 465 krypton gas 58 Watt, T8 lamp, 850 1 455 335 krypton gas 36 Watt, T8 lamp, 717 1 320 330 krypton gas 18 Watt, T8 lamp, 620 1 290 280 krypton gas Watt, T8 lamp, 460 1 290 280 krypton gas TABLE 1

The lamp driver stage is enabled at step 508. The microcontroller 490 then waits until expiry of the pre-heat period at 510. The pre-heat period is measured using the internal counter.

32 After expiry of the pre-heat period, the frequency output by the lamp driver unit is progressively reduced using a predetermined decrement until the output frequency corresponds to that required for strike voltage generation, that is, F..

In a preferred embodiment, the magnitude of the frequency decrement is 1 kHz. Further, in an embodiment, for a 58 Watt, T8, 1500 mm lamp and the resonant components of figure 4, the strike voltage generation frequency, F,, is 50 kHz. It will be appreciated that as the output frequency is reduced, the lamp strike voltage will increase. The frequency is progressively reduced by steps 512 and 514. Due to prevailing ambient temperature conditions, the lamp may strike before the output frequency is reduced to the target strike voltage generation frequency, F,,.

For low temperature conditions, it is quite often the case that full strike voltage may be required. A determination is made at step 516 viathe STK signal present on port line PB5 of microcontroller 490 whether the lamp has struck. If the lamp fails to strike, a determination is made at step 518 as to whether or not the most recent strike attempt is the last attempt via a variable, which is decreased until it is zero, that reflects the number of remaining strike attempts. This variable is also initialised at step 500. In a preferred embodiment, three strike attempts are allowed before lamp fail shut down occurs.

If it is determined at step 518 that the maximum number of strike attempts has been made, lamp fail shut down is arranged at step 520.

If it is determined at step 518 that there are remaining strike attempts, a delay of a predetermined duration is incurred before making any further strike attempts to allow the electrodes to cool down ' that is ' to prevent overheating of the electrodes, at 522. In a preferred embodiment, the 33 cool down period has a duration of four seconds. However, the cool down period can be a little as one second.

After expiry of the cool down period, control returns to step 506 and the process of pre-heat and strike attempt is repeated.

If at step 516 it is determined that the lamp has struck, the output frequency is set to equal the running or operating frequency, Fr., at step 524. In an embodiment, for a 58 Watt, T8, 1500 mm lamp and the resonant components of figure 4 (resonant at 41.4 kHz), the operating frequency, F,,,, is 48.07 kHz.

The microcontroller, having set the operating frequency, enters a loop within which there are made repeated determinations as to whether or not the lamp is still struck, at step 526, and whether or not the supply voltage is above a predetermined minimum level at step 528. The determination whether or not the lamp is still struck at step 526 is made by monitoring the arc current using the lamp struck detector 494 and the lamp L/C output circuitry 496. If it is determined that the lamp has extinguished, control is transferred to step 518 were a determination is made as to whether or not further strike attempts are permissible. It will be appreciated that the variable governing the number of remaining strike attempts is decremented every time a strike attempt fails or the lamp extinguishes.

If it is determined at step 528 that the input voltage, Vinput, has f allen below the drop-out threshold, the output of the lamp is disabled at step 530 via a signal SD. Control is then passed to step 504 where a determination is made repeatedly as to whether or not the supply voltage is above the minimum threshold level.

There is also shown in f igure 5 the interrupt service routine. Upon entry into the interrupt service routine all 34 registers are pushed onto a stack at step 532. The timer value is set at step 534. At step 536, the clock pin status is inverted, that is, toggled. The register values that were previously pushed onto the stack at step 532 are pulled from the stack at step 538 whereupon control then returns to the main routine at step 540.

A further embodiment also provides a dimming function which is realised by varying the operating frequency, F,,n, and thus the arc current once the lamp has struck. The variation in arc current causes a variation in lamp power which in turn determines light output. In an embodiment, the light output can be reduced to 20% of maximum.

By precise control of the resonant L3/Cl7 network drive frequency using the microcontroller it is possible to control very accurately the pre-heat current, strike voltage and arc current. Such accurate control results in improved lamp life and improved and consistent light output levels and colour.

A further embodiment of the present invention provides a lamp drive unit comprising an EEPROM. The EEPROM is utilised to record the operational history of the lamp performance. This operational history may record events such as number of lamp strikes thus far, number of first attempt light strike failures and prevailing system conditions in the event of failure or erratic behaviour.

A further embodiment of the present invention can be realised to accommodate cold cathode lamps. To accommodate cold cathode lamps the steps associated with pre-heat are omitted and the step associated with the strike voltage is suitably modified so that an appropriate frequency is output to the L3/C17 resonant circuit.

It will be appreciated that an advantage of the present invention is that the light output intensity and colour can be closely controlled and matched. This is particularly important in situations where a plurality of lamps is used in a given environment. Furthermore, within the railway industry, where the lamps are used in rows within railway vehicle passenger saloons, the impression of a continuous band of interrupted light is desirable.

The code for the microcontroller shown in Table 2 below is arranged to control the lamp driver unit to operate correctly a 58 Watt, 1500 mm, T8 lamp.

36

Claims (36)

1 A lamp driver unit comprising a microcontroller arranged to generate an output waveform having a predeterminable frequency; a lamp drive means comprising at least one inductive element and one capacitive element for producing a drive signal for the lamp in response to the output waveform.

2. A lamp driver unit as claimed in claim 1 in which the microcontroller derives the output waveform from an oscillator waveform produced by an oscillator.

3. A lamp driver unit as claimed in claim 2 in which the microcontroller comprises means for dividing the oscillator waveform by a predeterminable divisor.

4. A lamp driver unit as claimed in claim 3 in which the is means for dividing comprises a programmable counter arranged to vary the count thereof in response to the oscillator waveform and the output waveform is produced in response to an output signal of the programmable counter.

5. A lamp driver unit as claimed in claim 4 in which the divisor is determined according to apredeterminable frequency.

6. A lamp driver unit as claimed in any preceding claim in which the predeterminable frequency is determined according to lamp type.

7. A lamp driver unit as claimed in any preceding claim in which the predeterminable frequency is determined according to a stage of operation of a lamp

8. A lamp driver unit as claimed in claim 7 in which the stage of operation is one of pre-heating, generating a strike voltage and maintaining an arc current waveform.

37

9. A lamp driver unit as claimed in any preceding claim in which the predeterminable frequency is selected from a plurality of frequencies.

10. A lamp driver unit as claimed in any preceding claim, 5 wherein the inductive and capacitive elements of the lamp drive means comprise an inductor connected in series with a capacitor, and means for receiving the lamp such that, in use, the lamp is connected in parallel with the capacitor.

11. A lamp driver unit as claimed in any preceding claim comprising means for heating the lamp electrodes prior to application of a strike voltage.

12. A lamp driver unit as claimed in claim 11 in which the means for heating heats the lamp electrodes to a point of thermionic emission.

13. A lamp driver unit as claimed in either of claims 11 and 12 in which the means for heating is arranged to heat the electrodes for a predetermined fixed period.

14. A lamp driver unit as claimed in claim 13 in which the predetermined fixed period is determined according to a pre-heat current for a lamp type.

15. A lamp driver unit as claimed in any preceding claim comprising means for ensuring an arc current waveform, of the lamp has a predeterminable crest factor.

16. A lamp driver unit as claimed in claim 15 in which the crest factor is substantially 1.4.

17. A lamp driver unit as claimed in any preceding claim in which the drive signal comprises a substantially fully symmetrical arc drive current waveform.

18. A lamp driver unit as claimed in claim 17 in which the arc drive current waveform is a near-sine wave current 38 waveform.

19. A lamp driver unit as claimed in any preceding claim in which at least one of the output waveform and the drive signal has a predeterminable drive frequency.

20.A lamp driver unit as claimed in any preceding claim in which the arc current is at least one of 470 mA, 455 m.A, 3 2 0 mA, 2 9 0 mA or 2 9 0 mA.

21. A lamp driver unit as claimed in any preceding claim comprising means for maintaining a substantially stable arc current waveform.

22. A lamp driver unit as claimed in claim 21 in which the means for maintaining a substantially stable arc current waveform comprises means for maintaining the arc current waveform to within 6% of a predeterminable arc current waveform.

23. A lamp driver unit as claimed in any preceding claim operable at predeterminable efficiencies.

24. A lamp driver unit as claimed in claim 23 in which the operating efficiencies are at least 84%.

25. A lamp driver unit as claimed in claim 24 in which the operating efficiencies are between 84% and 90%.

26. A lamp driver unit as claimed in any preceding claim comprising means arranged to provide a floating lamp drive output.

27. A lamp driver unit as claimed in either of claims 26 in which the means arranged to provide a floating lamp drive output comprises an output transformer wound for 2 kV primary to secondary isolation.

28. A lamp driver unit as claimed in any preceding claim comprising means for electronically providing at least 39 one protection feature.

29. A lamp driver unit as claimed in any preceding claim comprising means for preventing attempted re-strikes in the event of a failed lamp.

30. A lamp driver unit as claimed in any preceding claim comprising a cut-out means for inhibiting the use of a lamp upon detection of a falling supply voltage or a supply voltage that is below a predetermined level.

31. A lamp driver unit as claimed in claim 30 in which the cut-out means is provided with some hysteresis to avoid flickering as the supply voltage varies.

32. A lamp driver unit as claimed in any preceding claim comprising transient and surge suppression means.

33. A lamp driver unit substantially as described herein with reference to and/or as illustrated in the accompanying drawings.

34. A passenger vehicle comprising a lamp driver unit as claimed in any preceding claim.

35. A lamp and lamp driver unit as claimed in any preceding claim.

36. A computer program element for controlling a lamp driver unit substantially as described herein with reference to and/or as illustrated in the accompanying drawings.