WO2010121968A1

WO2010121968A1 - Amplifier for electrodeless high intensity discharge (ehid) lamps with multiple stages, variable supply voltage, biasing and tuning, ehid-system and method for operating an ehid-system

Info

Publication number: WO2010121968A1
Application number: PCT/EP2010/055050
Authority: WO
Inventors: Bernhard Siessegger
Original assignee: Osram Gesellschaft mit beschränkter Haftung
Priority date: 2009-04-24
Filing date: 2010-04-16
Publication date: 2010-10-28

Abstract

The application relates to an RF amplifier for Electrodeless High Intensity Discharge Lamps (EHID). The RF amplifier comprises a plurality of amplifier stages (PA1-PAn) that are coupled between an input signal power divider (PD) and an output signal power combiner (PC). While the amplifier stages receive, amplify and output in parallel the RF signal, their DC supply terminals are connected in series between a DC supply voltage terminal (VCC) and ground. Thus, each amplifier stage (PA1-PAn) is operated at a DC voltage that is lower than DC supply voltage (VCC).

Description

Amplifier for Electrodeless High Intensity Discharge (EHID) lamps with multiple stages, variable supply voltage, bias- ing and tuning, EHID-System and method for operating an EHID-System.

Technical field

The invention relates to a RF amplifier, comprising at least one amplifier stage with variable supply voltage, biasing and tuning. The invention also relates to an EHIS System and a method for operating an EHID System.

Prior art

The invention concerns amplifiers for Electrodeless High Intensity Discharge (EHID) lamps.

Electrodeless High Intensity Discharge (EHID) lamps come with the advantage of superior maintenance and the possibility of having even higher colour rendering indices and higher efficacy than electroded HID lamps. Instead of using electrodes to couple the energy into the lighting plasma

EHID lamps use couplers - sometimes referred to as applicators - to couple microwave power into the plasma. The mi^¬ crowave power is provided by a power amplifier or power oscillator circuit.

Modern EHID systems use solid state power amplifiers. Such amplifiers use transistors like LDMOS transistors, BJTs, HEMTs, etc. and all commonly used amplifier/transistor technologies require supply voltages of a few ten volts (e.g. typically 28V for LDMOS) . Most lighting equipment is powered by the mains (an AC voltage of typically 110V or 230V) . Therefore in case of EHID ^λlow' voltage has to be supplied by a power supply (PS) to operate the power ampli- fier (Amp) which then supplies RF to the lamp-applicator- unit. Due to the fact that the electric power is processed twice the efficiencies of both stages should be high to achieve good overall system efficiency. The larger the difference between input and output voltage of a power supply gets the lower its efficiency becomes. Therefore an EHID system' s efficiency could be increased if the power supply would be allowed to put out higher voltages.

Figure 1 depicts an EHID system showing DC power paths indicated by solid lines and RF power paths indicated by dashed/dotted lines. The system comprises an oscillator (Osc) and a power amplifier (Amp) . A bidirectional (power) meter (Mtr) measures the reflected power or forward and reflected power or the standing wave ratio (SWR) . The fil- ter (Fi) is used to match the impedance of the amplifier or the amplifier-meter-combination to the cable (Cab) . The cable delivers the power to the coupler or applicator (App) which powers the discharge vessel or lamp (La) . Lamp and coupler may be combined in a lamp-coupler-unit (LCU) . Whereas the other parts of the system except for the cable are referred to as the electronic driver (ED) . The system is commanded by the control unit (Ctrl) which may have communications to the Outside world' (not shown) via a communications bus like DALI or DMX (or a 1-lOV interface, etc.) . The command paths from and to the control unit are not shown in the figure.

The power supply (PS) is usually a two-stage arrangement using a fist stage for power factor correction stage (PFC) and a DC-to-DC converter (DC2DC) as a second stage. Both stages are linked with a high voltage DC bus e.g. a 400V DC bus. A boost converter is typically used in the PFC stage which is fed by the mains (M) coming from the on-off switch (S) . Most of the known EHID Systems don't have any dimming capabilities at all. Figure 2 shows one of the known EHID systems with dimming capability, the system having a control input (DL) by which the user is able to command the dimming level of the lamp. The dimming level may be generated by external circuitry (not shown) processing video data e.g. in applications like video projection or displays. This might be an external 1- 10V signal or the angle/position of a potentiometer inside the system's housing. This command is converted by an interface (Infa) into a brightness command signal (Bcmd) which is fed into the control unit. The light sensor

(BrMea) feeds a signal (Br) corresponding to the actual brightness into the control unit. Inside the control unit a controller circuitry uses Bcmd as set-value which is compared to the actual value (Br) . The controller's output is the power command signal (Pcmd) by which the amplifier amplification of the power amplifier (Amp) can be regulated (between 0% and 300% of the nominal value; the nominal value gives full or nominal brightness in steady state operation of the system) . A reduced run-up time can be achieved due to the high power level the amplifier is capable of putting out. After the run-up and hence in steady state operation the signal (Pcmd) is limited inside the control unit (Ctrl) to 100% on average to prevent overpowering of the lamp. The instant value of the signal (Pcmd) may be limited to 250% for enhancement of the dynamic performance of the system (e.g. for applications like video projection, stage lighting or other specialty effects) .

The commanded RF power from the RF amplifier of an EHID system may be determined from the brightness level (which might be a commanded brightness level by a control knob on the electronic driver or by a command via a communications interface e.g. DMX or DALI or 1-lOV interface), measured or predicted temperatures of the system (which can lead to power reduction) and the actual operating state of the LCU (e.g. lamp being in run-up or steady state). The efficiency of an RF amplifier stage depends on the load, on the supply voltage, on the biasing (the biasing changes e.g. stand-by power, amplification and the power output capability of the stage) and on input and output impedance matching achieved by input and output tuning. The power amplifier of a (modern) EHID system (not using a power oscillator) shows the same dependencies because it is made up of at least one power amplifier stage (usually several stages are employed) . A change in the commanded power will result in a changed behaviour of the lamp- applicator-unit (App-La) which gives a different loading of the amplifier. Therefore the efficiency of the amplifier changes e.g. by dimming due to the change in commanded power. Usually an EHID system is constructed in such a way that its power amplifier has the best efficiency at nominal operating conditions. At other conditions the operation is sub-optimal regarding the amplifier efficiency.

Object

It is the object of the invention, to specify an Electrode- less High Intensity Discharge System that has a better efficiency over a wider range, and is capable of dimming the Electrodeless High Intensity Discharge Lamp.

Description of the invention

According to the invention, the object is achieved by an RF amplifier, comprising:

- at least one amplifier stage with variable supply voltage, biasing and tuning,

- the supply voltage, biasing and tuning depends on a commanded RF power, - each time the commanded power is changed, the supply voltage and/or biasing and/or tuning are changed for supplying RF power to the Electrodeless High Intensity Discharge (EHID) lamp. This EHID system has the following advantages: increased efficiency especially m EHID systems with dimming feature and hence it can be applied to EHID systems with multi-stage power amplifiers.

It can be applied to compact EHID systems. reduced operating cost.

A modular concept can be employed for the (identical) power stages which will reduce complexity and cost and gives an easy to maintain RF power amplifier.

Due to the fact that smaller voltage conversion ratios of the power supply are sufficient, transformer-less solutions are possible (e.g. a simple buck or boost converter just using a choke instead of a transformer) , which leads high efficiencies and might cause lower cost.

Some of the electronics of an EHID system (especially the RF power amplifier) is split into several units which are connected in series and are operated from a high voltage DC bus. The DC bus is powered by a power supply (PS) . The power supply achieves high efficiency because the voltage differences of mains voltage compared to the output voltage is low. A modular concept for the RF power amplifier can be applied using (identical) power amplifier stages. Each stage may provide means for clamping its input DC voltage.

The power amplifier (or one of the stages of a cascaded power amplifier) for the EHID system consists of several identical stages for high power output: RF outputs are in parallel, RF inputs are paralleled, too; see figure 3 for illustration. Appropriate power splitting and combining networks are applied - such networks are described in chapter 12-7 'Power Combiners and Splitters' on pages 379f in Herbert L. Krauss, Frederick H. Raab, Charles W. Bostian: Solid State Radio Engineering, John Wiley & Sons, Inc, 1980. In contrast to the paralleled RF path the dc supply paths are connected in series, leading to a high voltage supply (e.g. 6x28V=168V in case of 6 stages) even though each individual stage might only require a low voltage (e.g. 28V) .

Not all power amplifier stages have to be connected in parallel on the RF side: For example in case of 6 identical stages the first stage might be used as pre-amplifier to drive the remaining 5 stages which would be parallel con^¬ nected concerning their RF path (see figure 4 for illustration) . Other units of the EHID system might be included. As an example a 7th unit may be added to the system just described. This additional unit is connected in series to the arrangement of the 6 stages already mentioned. The 7th unit may include the control, the oscillator and a pre-pre- amplifier. The 7th unit is designed in such a way, that it consumes the same amount of current as the 6 stages mentioned afore (e.g. IA) . The dc paths of the functional blocks inside the 7th stage might be in parallel (e.g. oscillator and control may consume 1/3 of the overall current, whereas the pre-amp consumes 2/3 of the current) . The 7th stage might be supplied with 18 volts. Therefore the power supply has to put out 168V+18V=186V at a current of IA.

Power architecture may be used in which several of such strings (each consisting of several of in series connected units) are connected in parallel (see figure 5) . The voltage VCC delivered by the PS is chosen in such a way that it always stays below the minimal sum of the clamping voltages (see next paragraph) of the clamping devices in the individual strings.

Each stage or unit may provide means for clamping its input dc voltage (e.g. a Zener diode across the dc input of the stage) . In addition each stage may provide means to bypass energy (e.g. a power transistor across the dc input of the stage) or to distribute energy (e.g. a small dc/dc- converter connected to the high voltage bus to feed back energy not consumed by the individual amplifier state; instead of connecting the small dc/dc-converter to the high voltage bus it might be connected to the input of another stage, meaning that no voltage step-up would be needed) to lead to the desired voltage distribution across the unit (even though tolerances in the individual units are present) .

Each stage or unit may provide a voltage controller for regulating its input voltage and thereby bypassing energy or distributing energy to other stages (see figure 6) . The set-voltage for the voltage controller is chosen below the clamping voltage of the individual stage. Therefore only the voltage controller is active during normal operation. The clamping circuit acts only in case of a transient over- voltage being present across the stage or unit.

The voltage control may be incorporated locally in each stage. Alternatively an overall master unit commanding all stages may be used. This overall master may control the energy processed in the individual stages as well as the energy flow between the individual stages.

The power supply supplying the RF amplifier may be controlled in such a way that its output current is controlled (e.g. to IA in the example mentioned above) in contrast to state of the art power supplies used in such applications which keep its output voltage (e.g. 28V) constant. The voltage at the output is generated automatically by the summation of the individual stages of the series connected elements each fixing or controlling its own voltage.

In addition to the voltage regulator or instead of a volt- age regulator a shift in the bias supply of the power amplifier stage can be applied. In case of a rising voltage across the stage the bias is increased leading to (a) more RF output power and hence more power consumption of the power amplifier stage and (b) more stand-by power which also increases the consumption of the stage. Both effects lead to a lowering of equivalent resistance of the stage and hence to an action tending to reduce the voltage across the stage.

An RF amplifier consisting of one ore more amplifier stages with variable supply voltage, biasing and tuning depending on the commanded RF power is described. To achieve high efficiencies each time the commanded power is changed supply voltage and/or biasing and/or tuning are changed as well.

A voltage control for each individual stage may be employed for a multi-stage power amplifier using variable set-values for the voltage controller instead of just having a single set-value corresponding to the nominal operating point of the EHID system.

As described the efficiency of an RF amplifier stage de^¬ pends on the load, on the supply voltage, on the biasing and on the input and output tuning. Therefore to achieve high efficiencies each time the commanded power is changed (and thereby the loading of the stage) one or more of the other parameters affecting the efficiency (supply voltage, biasing and tuning) are changed as well. Therefore the power amplifier may always be operated at highest efficiency .

The supply voltage (s) produced by one or more DC-to-DC converters may be changed. Some of those converters are fed by a bus voltage coming from a PFC stage (The PFC stage is a part of the box ^λConv' shown in figure 2.) . The bus voltage may also be changed upon the commanded RF power. There- fore the PFC stage will receive commands from the control unit of the EHID system. Preferred topologies are buck, SEPIC and flyback for the PFC stage. In case of a multi-stage power amplifier based upon the commanded power one or more parameters affecting the efficiency (supply voltage, biasing and tuning) in one or more of the power amplifier stages are changed.

The voltage control may still be incorporated locally in each stage as shown on figure 6 but the set-value fed to the voltage controller (the reference voltage Vref in this case) is changed depending upon the commanded RF output power of the stage. A communication between the Ctrl and the individual stage is established commanding the set- value .

Alternatively an overall master unit commanding all stages may be used. This overall master may control the energy processed in the individual stages as well as the energy flow between the individual stages. Therefore this master unit may command different supply voltages to the individual stages depending on the commanded RF power.

Dimming information is used in a feed-forward control type to improve the system's dynamics. Depending upon the com- manded power the power converters are influenced. This can be used to increase the dynamic performance of the EHID system which is especially desired for applications requiring fast dimming. The bus voltages and/or the power amplifier supply voltages may not drop or over-shoot due to a dimming action because the commanded voltages are changed in advance and therefore the changed power consumption is not negatively affecting the supply voltages.

Brief description of the drawings

Further advantages, features and details of the invention are obtained by means of the subsequent description of exemplary embodiments and by means of the drawings, in which identical or functionally identical elements are provided with identical reference symbols and in which:

Fig. 1 shows a block diagram of an EHID system with electronic driver and lamp-coupler-unit relating to a first aspect of the invention,

Fig. 2 shows a block diagram of an EHID system that has a control input by which the user is able to command the dimming level of the lamp relating to a second aspect of the invention,

Fig. 3 shows a first embodiment of a RF Power Amplifier consisting of n power amplification stages, Fig. 4 shows a second embodiment of a RF Power Amplifier consisting of n power amplification stages, Fig. 5 shows a third embodiment of a RF Power Amplifier consisting of three stages,

Fig. 6 shows a single stage x of a power amplifier arrangement according to figures 3 or 4,

Fig. 7 shows a Block diagram of a fourth embodiment of an EHID system with a power command signal relating to a second aspect of the invention,

Fig. 8 shows a fifth embodiment of an EHID system which contains a sub-unit which determines the power command signal relating to a second aspect of the invention,

Fig. 9 shows a first exemplary amplifier stage PAx of the embodiment of figure 8, Fig. 10 shows a sixth embodiment using a multi-stage power amplifier and a chain of optocouplers to communicate the set-values of the voltage control to the individual stages by the use of a pulse width modulation signal relating to a second aspect of the invention,

Fig. 11 shows a block diagram of a seventh embodiment of an EHID system with a power command signal relat^¬ ing to a second aspect of the invention, Fig. 12 shows a second exemplary amplifier stage PAx of the embodiment of figure 8.

Preferred embodiment of the invention

There are two aspects of the invention concerning amplifiers for Electrodeless High Intensity discharge (EHID) lamps. The first aspect relates to maximizing efficiency by splitting one big amplifier in many smaller amplifiers in several stages. The second aspect relates to driving these amplifiers in a way that they are able to dim the EHID lamp. Figure 3 shows a first embodiment of a first aspect of the invention in which a RF power amplifier consists of n identical power amplification stages (PAl ... PAn) . A power splitter or power divider (PD) is used to split the input RF power (RFi) and distribute it to the individual stages. After amplification in each stage the outputs of those stages are combined to the output RF power (RFo) by means of the power combiner (PC) . The DC power supply provides a voltage VCC which is n times the supply voltage of a single stage. The input voltage of every stage is stabilized by a capacitor (Cl ... Cn) and clamped by a Zener-diode (Dl ... Dn) .

Each power stage has two DC blocking capacitors one at its input and one at its output (as adumbrated in stage PAl of figure 3) . Therefore only the RF power is seen by PD and PC. The DC blocking capacitors are used in state of the art amplifiers to decouple the biasing of the transistor from the RF signal; therefore no additional capacitors are needed. The only difference to a state of the art arrangement is the fact that the capacitors have to be rated for the full supply voltage VCC and not only for VCC/n as in case of a state of the art amplifier. But this is no strong constraint on the capacitors because most low-loss capaci^¬ tors used in those applications have high DC breakdown voltages anyhow.

As an example a power amplifier consisting of 15 stages is considered. Each stage consumes 28V yielding to a power supply output voltage of 420V. This voltage may be provided by a PFC stage using a boost-converter having no galvanic isolation to the mains.

As a second example a power amplifier consisting of 3 stages is considered. Each stage consumes 28V yielding to a power supply output voltage of 84V. This voltage may be provided by a PFC stage using a buck converter having no galvanic isolation to the mains. For European line voltage of 230V this arrangement is still sufficient to fulfil the regulations for lighting regarding line harmonics (rule of thumb: o.k. as long as the output voltage of the buck is not sufficiently above 80V) even though the buck converter is not operating all the time (If the line voltage is below the output voltage no power is transferred from the line to the output hence the buck converter is inactive and the power used by the amplifier is coming from the capacitors at the output of the power supply) . Those capacitors may be designed as ^Λhold-up capacitors' storing enough energy to operate the system even if some half waves of the line voltage are totally missing (zero volts) . The buck converter comes with the advantage of superior efficiency compared to other converter topologies.

As a third example a power amplifier is proposed which consists of that many stages that it can be operated di- rectly from the mains just using a simple (bridge) recti^¬ fier charging an electrolytic capacitor which provides the high voltage DC bus and hence powers the system. A passive PFC circuit might be used (e.g. a joke between rectifier and electrolytic capacitor) . The ripple on the bus voltage depends on the size of the used capacitor. Some ripple is allowed leading to a 100 or 120Hz modulation in the RF power which doesn't matter in this application.

Preferably the n power stages are identical yielding to identical power consumption and therefore voltage balancing of the series connection. This can be achieved by using power transistors or integrated power devices from the same production run, preferably from the same wafer. An on-chip integration of all power devices needed for the n stages may be considered. A further integration of all n stages including passive components into a Monolithic microwave integrated circuit (MMIC) might be considered (e.g. by applying LDMOS technology on a silicon substrate) .

Figure 4 shows a second embodiment of the first aspect of the invention in which the RF power amplifier (Amp) con- sists of three cascaded power amplifiers. The oscillator (Osc) delivers the RF power RFa to Amp which is amplified in the pre-pre-power-amplifier (PPA) yielding to signal RFb. RFb is then amplified by PAO (which is identical to PAl . PAn) putting out RFi. RFi is the input signal of the final amplifier consisting of n identical power amplifica^¬ tion stages (PAl ... PAn) and is identical to the amplifier structure shown in figure 3.

Figure 5 shows a third preferred embodiment relating to the first aspect of the invention. The Power Amplifier (Amp) consists of three stages. The pre-amplifier stage made of PAO amplifies the RF input signal RFi. PAO puts out the signal RFl which is split into three equal parts which are fed into the stages PAl ... PA3 which act as intermediate amplifier stage. The final amplifier stage is made of the 8 stages P4 ... PAIl delivering the output signal RFo. Consid^¬ ering the DC power the 12 identical amplifier units are arranged in 2 strings with 6 units in series connection, respectively.

Figure 6 illustrates a single stage x of a power amplifier arrangement according to figures 3 to 5. In parallel to the power amplifier (PAx) a voltage controller (VCtrl) is con^¬ nected. VCtrl is regulating the input voltage of that stage and thereby bypassing energy by the use of transistor Q. A Pi-controller employing an operational amplifier (OpA) is used as controller. The set-voltage of the controller is determined by the reference voltage Vref which is determined by the Zener-diode DxI. The set-voltage is chosen below the clamping voltage of the Zener-Diode Dx which acts as transient over-voltage protection. Therefore only the voltage controller is active during normal operation.

More than one string each containing one or more units in series connection may be paralleled.

The supply voltage may be changed according to the dimming signal as per particulars given below. A fourth embodiment based on the system according to figure 2 is shown in figure 7. This figure shows a Block diagram of a fourth embodiment of an EHID system with a power com^¬ mand signal (Pcmd) relating to a second aspect of the invention. Depending upon the power command signal (Pcmd) which is primarily used to modify the amplification of the power amplifier (Amp) the DC voltage supplied by the DC-to- DC converter (DC2DC) changes its output voltage (DC2) supplying the amplifier with DC power.

In addition the power command signal (Pcmd) is also pro- vided to the PFC stage. The power command signal is used to change the DC voltage (DC) supplied by the PFC stage (not shown) analogously to the output voltage (DC2) of the DC- to-DC converter (DC2DC) .

Figure 11 shows a Block diagram of a seventh embodiment of an EHID system with a power command signal relating to a second aspect of the invention. In contrast to figure 7 power command signal is also used to change the DC voltage (DCl) supplied by the PFC stage to the DC-to-DC converter (DC2DC) . The PFC stage has therefore two outputs DC and DCl. Only DCl is influenced by the power command signal. In this embodiment there is a series connection of two con^¬ verters, the PFC stage and the DC-to-DC converter, which both are changing their output voltage according to the power command signal. This is a very energy efficient and economical way of adjusting the voltage DC2 fed to the power amplifier as both converters need only to change their conversion ratio in a relatively small range as the two voltage conversion ratios are multiplied. Such converters exhibit better energy efficiencies, less expensive components and have a better dynamical response compared to converters having a relatively large adjustable range of the conversion ratio. The voltage conversion ratio is defined by the average value of the output voltage divided by the average value of the input voltage. Figure 8 shows a fifth embodiment of an EHID system which contains a sub-unit which determines the power command signal relating to a second aspect of the invention. The control unit (Ctrl) contains a sub-unit (CtrlP) which determines the power command signal (Pcmd) . By the use of lookup-tables the Pcmd signal is converted to set-value signals for supply voltage (SPS), for biasing (SB), for input tuning (STi) and for output tuning (STo) . A change in commanded power leads to a change in the mentioned set- values. The set-values change simultaneously. Within the amplifier (Amp) only a single stage (PAx) is shown which receives 3 of the 4 set-value signals. The signal SPS is used as reference voltage for the pulse width modulation employed for generating the switching commands in the buck or SEPIC converter used in the DC-to-DC converter (DC2DC) . The amplifier stage PAx may be constructed as illustrated in figure 9. The amplifier stage consists of a bias control part (BCtrl) and a class A or C power amplifier (as described i.e. in Mihai Albulet: Rf Power Amplifiers, Noble Publishing, 2001 and Herbert L. Krauss, Frederick H. Raab, Charles W. Bostian: Solid State Radio Engineering, John

Wiley & Sons, Inc, 1980) marked as PA. RF input and output are named RFi and RFo, respectively. The impedance tuning units for the input and output are shown as boxes ITi and ITo and my include PiN-diodes or varactor diodes which are used for tuning depending on the signals STi and STo. The tuning boxes have a galvanic connection (not shown) in horizontal direction (between their left RF input and right RF output) .

The higher the set-value signal for biasing (SB) becomes the more conducting Ql and Q2 become and thereby increasing the current flow through R6. This leads to a higher (DC or low freguency components of the) gate to source voltage and thereby biasing the LDMOS FET Q3 more into the conducting state. R7 sets the minimal gate voltage of Q3. R7 might be connected to a constant voltage source instead of VCC to accomplish a minimal gate voltage regardless of the actual supply voltage VCC. In addition this constant voltage source might be varied depending upon the (junction) temperature of Q3 to compensate for temperature influences of the biasing (current) of Q3. Figure 12 shows a second exemplary amplifier stage PAx of the embodiment of figure 8. This amplifier stage includes an internal generation of the signal SB in contrast to the previously described. The signal SB is generated in the block SBgen based on the supply voltage delivered to the amplifier stage. The signal SB shown in figure 8 is not used as it is internally generated in the amplifier stage. The resistors R18 and R19 form a voltage divider generating the biasing signal Depending on the individual design of the amplifier stage the resistors R18 and/or R19 may be chosen as NTC or PTC thermistor or a combination of them. The thermistors are thermally coupled to other semiconductors of the amplifier stage like Ql, Q2 or Q3. To accomplish a nonlinear dependency of SB from the supply voltage a combination of zener diodes (D16, D17) having different Zener voltages and resistors (R16, R17) are used. Other circuits showing non-linear dependencies may be used alter- natively. As any high frequency modulation of the supply voltage may be coupled via the bias generating or the bias signal generating circuit (SBgen) into the Rf portion of the amplifier stage and thereby leading to undesired oscil- lation of the output signal additional low pass filtering is applied consisting of capacitor Cb and the combination of inductor Lb and capacitor Cb2.

In case of a multi-stage power amplifier regarding to one of the figures 3 to 5 a voltage control (VCtrl) for each individual stage may be employed using variable set-values for the voltage controller instead of just having a single set-value. In case of a single set-value this value corresponds to the nominal operating point of the EHID system.

Figure 10 shows a sixth embodiment using a multi-stage power amplifier and a chain of optocouplers to communicate the set-values of the voltage control to the individual stages by the use of a pulse width modulation signal relating to a second aspect of the invention. This signal is generated from the SPS signal (e.g. an analog 2-4V signal) coming from the control unit by the use of a pulse width modulator (PWM) . The duty ratio of the optocouplers (which are used in binary communication mode) are determined by the duty ratio of the signal PWSPS. The duty ratio modulates the reference voltage (e.g. Vrefl in the nth stage) inside each stage yielding to a set-value (e.g. Vref2) after low-pass filtering (e.g. done by R12 and C2) . This set-value is used inside the voltage control to control the operating voltage of the individual stage in a closed-loop arrangement. Due to the fact that not all reference volt- ages (e.g. Vrefl, etc.) need to have the same value it is possible to set different set-values and therefore operating voltages for the individual stages with a single chain of optocouplers and just one PWM unit. All of the operating voltages change synchronously according to the SPS signal. This synchronous change is preferred because usually all stages of the amplifier need to behave in the same way. The PWSPS signal might be generated directly (without the use of a discrete PWM unit) from a microcontroller or DSP in a modified embodiment (not shown) .

List of reference designations

Amp RF power amplifier

App coupler or applicator

App-La applicator-lamp-unit

Bcmd brightness command signal

BCtrl bias control part of amplifier stage

Br brightness signal from light sensor

BrMea light sensor

Cab cable

Crtl control unit

CtrlP sub-unit of the control unit (Ctrl) which determines the power command signal (Pcmd)

DCl output voltage of PFC stage fed to DC-to-DC converter

DC2 output voltage of DC-to-DC converter

DC2DC DC-to-DC converter

DL control input

ED electronic driver

Fi impedance tuning circuit or filter

Infa interface

ITi impedance tuning units for the input

ITo impedance tuning units for the output

La discharge vessel or lamp

LCU lamp-coupler-unit

M mains

Mtr power meter

Osc oscillator

PAx stage of power amplifier or power amplification stage

PC power combiner

Pcmd power command signal

PD power splitter or power divider

PFC power factor correction stage

PS power supply

PWM pulse width modulator

PWSPS duty ratio modulated signal from PWM

RFi input RF power

RFo output RF power

S on-off switch

SB set-value signal for biasing

SBgen building block of amplifier stage generating the signal SB based on the supply volt^¬ age of the amplifier stage

SPS set-value signal for supply voltage

STi set-value signal for input tuning STo set-value signal for output tuning VCC supply voltage

Claims

Patent Claims

1. An RF amplifier, comprising:

- at least one amplifier stage (PAx) with variable supply voltage and/or biasing and/or tuning, - the supply voltage and/or biasing and/or tuning depends on a commanded RF power,

- each time the commanded power is changed, the supply voltage and/or biasing and/or tuning are changed for supplying RF power to at least one Electrodeless High Intensity Discharge (EHID) lamp.

2. The RF amplifier as claimed m claim 1, characterized in that the RF amplifier includes more than one amplifier stage.

3. The RF amplifier as claimed m claim 2, characterized m that the dc supply paths of at least two of the amplifier stages are connected in series.

4. The RF amplifier as claimed m claim 2 or 3, characterized m that a voltage control for each individual stage is employed using variable set-values for the voltage controller corresponding to the nominal operating point of the EHID system.

5. The RF amplifier as claimed m claim 1, characterized m that each time the commanded power and thereby the load of the stage is changed one or more of the supply voltage, the biasing or the tuning are changed.

6. The RF amplifier as claimed m claim 2 or 3, characterized m that the voltage control is incorporated locally m each amplifier stage.

7. The RF amplifier as claimed m claim 6, characterized m that the set-value for the voltage is changed depending upon the commanded RF output power of the stage .

8. The RF amplifier as claimed in claim 4, characterized in that a communication between the voltage controller and the individual stage is established commanding the set-value .

9. The RF amplifier as claimed in claim 3, characterized in that the amplifier includes an overall master unit commanding all stages.

10. The RF amplifier as claimed in claim 9, characterized in that the overall master controls the energy proc- essed in the individual stages as well as the energy flow between the individual stages.

11. The RF amplifier as claimed in claim 8 or 9, charac^¬ terized in that the master unit commands different supply voltages to the individual stages depending on the commanded RF power.

12. An EHID system including an amplifier as claimed in claim 1, characterized in that the EHID system also includes one or more DC-to-DC converters and/or PFC stages .

13. The EHID system as claimed in claim 12, characterized in that the supply voltages produced by the DC-to-DC converters and/or PFC stages are changed.

14. The EHID system as claimed in claim 13, characterized in that some of those converters are fed by a bus voltage coming from a PFC stage.

15. The EHID system as claimed in claim 14, characterized in that the bus voltage may also be changed upon the commanded RF power.

16. Method for operating an EHID system, characterized in that the commanded RF power simultaneously influences the set-values of the supply voltage and/or the bias^¬ ing and/or tuning of the amplifier stages.

17. Method for operating an EHID system as claimed in claim 16, characterized in that the commanded RF power is generated from a light dimming signal fed to or generated in the EHID system.