Lamp-Coupler-Unit for Electrodeless High Intensity Discharge (EHID) lamps with a data memory and communication and an impedance-controlled feedthrough and Electrodeless High Intensity Discharge system with such lamp-coupler- unit .

Technical field

The invention relates to lamp-coupler-unit for Electrodeless High Intensity Discharge (EHID) lamps. The invention also relates to a High Intensity Discharge (HID) system with such lamp-coupler-unit.

Prior art

The invention concerns lamp-coupler-units for Electrodeless High Intensity Discharge (EHID) lamps and systems with such lamp-coupler-units .

Electrodeless High Intensity Discharge (EHID) lamps come with the advantage of superior maintenance and the possi- bility of having even higher color 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.

To achieve high efficiency in an RF system the impedance of the load (e.g. coupler) should be matched to the source (e.g. amplifier, cable, etc.) what means that these imped- ances are identical or are close to each other. In case of unmatched impedances a significant amount of RF power coming from the source (forward power) is reflected by the load back to the source (reflected power) . This degrades the efficiency because some of the reflected power is transformed into heat in the source.

This results in the necessity for impedance matching in EHID systems due to the need for high efficiencies. Impedance matching can be achieved by

A) constructing the source in such a way that its output impedance comes close to the input impedance of the load or by

B) using a tuning circuit between source and load. The impedance of the load is transformed by the tuning circuit in such a way that the resulting impedance seen by the source is better suited to the source.

Figure 1 depicts an example of an EHID system containing an oscillator (Osc) and power amplifier (Amp) . A bidirectional (power) meter (Mtr) measures the reflected power or the forward power and the reflected power or the standing wave ratio (SWR) . The SWR is defined as the voltage standing wave ratio sometimes referred to as VSWR which is the ratio of the (voltage) amplitude of a partial standing wave at an antinode (maximum) Vmax to the (voltage) amplitude at an adjacent node (minimum) Vmin in an electrical transmission line .

SWR = %≡- (1) mm

The filter (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) . The RF power path is illustrated by a dashed/dotted line. Some EHID systems, i.e. the system described in the US- Application US12/720, 769, which is incorporated herewith as reference, use an evacuated or gas-filled (e.g. N2) outer bulb- sometimes referred to as outer vessel or sleeve - which contains the discharge vessel and the coupler.

The outer bulb which is typically made of quartz glass should enclose applicator and lamp in a gas-tight way and therefore has to have a feedthrough to provide RF power to the applicator. The feedthrough (Fee) - as shown in Figure 2 - has to be considered as an additional element in the RF power chain because it can cause high power loss in the system if its electrical characteristic is not designed appropriately.

Object

It is an object of the invention to enhance handling of

EHID-Systems and to increase the efficiency of those systems .

The invention contributes to such a system making it easier serviceable and adding more features. The invention con- tributes to a more efficient EHID system by having high efficacy due to the use of an outer bulb (containing coupler and discharge vessel) without having a degraded (electrical) efficiency in providing the RF power to the coupler . The invention proposes an EHID-System, where the Coupler and discharge vessel are integrated in an un-dismountable lamp-coupler-unit. The lamp-coupler-unit may be separable from the electronic driver (ED) . The invention constitutes of three main features: 1. Coupler and lamp are tight together to a lamp-coupler- unit (LCU) in an un-dismountable way. The EHID system consists of two main parts: The LCU and an electronic driver (ED) which are separable by the user. 2. The coupler contains an impedance-controlled feedthrough. 3. The LCU incorporates a data memory which contains data about itself. If power is necessary for operating the data memory this power is delivered by the ED.

4. There is a communication between LCU and ED.

Description of the invention

Coupler and discharge vessel are integrated in an un- dismountable lamp-coupler-unit which has a non-volatile data memory keeping information about the lamp-coupler- unit. The data is used by the electronic driver for operating the EHID system. Preferably a (bidirectional) communication path is used between electronic driver and lamp- coupler-unit.

An impedance-controlled feedthrough is proposed to provide RF power into the outer vessel (e.g. to the applicator) with good efficiency (e.g. due to the outer bulb reducing heat loss) . To generate this feedthrough an arrangement incorporating the wall of the outer vessel and additional material inside the wall (e.g. metal wires or stripes) is used. In addition lumped elements (e.g. capacitors welded to the wires going through the glass) and distributed elements (e.g. transmission lines) may be involved in the arrangement to come to an impedance-controlled feedthrough The invention contributes to an EHID system having the following advantages: more efficient EHID system by having high efficacy due to the use of an outer bulb (containing coupler and dis- charge vessel) without having a degraded (electrical) efficiency in providing the RF power to the coupler, increased lifetime (due to reduced power and thereby wall-loading in the beginning of life) , increased availability of the system for the customer, reduced operating cost due to the fact that lamp- coupler-unit (LCU) and electronic driver (ED) can be replaced by the customer so the whole system is not to be sent to the manufacturer for service or even a new system has to be purchased, simplified warehousing, higher efficacy over life due to a reduced aging (e.g. reduced frosting in case of a discharge vessel made of quartz glass in the beginning of life, which means in the first several hundred to thousand operating hours) , better lumen maintenance over life achievable (in particular reduced drop of luminous flux in the beginning of life) possibility to reliably (regarding production spread and interchangeability of system components) increase the system's performance by employing acoustic resonances (e.g. the luminous intensity may be increased for protection systems by excitation of specific acoustic modes inside the discharge vessel) , - no additional feedthroughs for the communications in the outer bulb (containing coupler and discharge vessel) are necessary.

By using a single cable for communications and power distribution only little additional expenses are necessary to realize the invention.

It is possible to use the same RF connectors within a whole family of EHID systems (regardless of their system power, optics, etc.) . There is no need for mechanical coding of LCU and ED for different versions or models to pre- vent an improper combination. Otherwise one would have to realize mechanical coding e.g. by grooves, salients, different connector shapes etc. In the end by use of the invention a reduction in variants and hence cost is possible. additional features are possible by means of a commu- nication between LCU and ED:

► Temperature control of LCU The temperature measured inside the LCU is communicated to the ED. In case of over-temperature the ED has the opportunity to reduce power and/or shut down. ► If a bidirectional communication is available be- tween LCU and ED the ED can command actions inside the LCU. Such an action might be tuning via filter (Fi) inside the LCU which would lead to an increased efficiency of such a system. protection functionality for the ED in case of operat- ing the system without load (without LCU or with a broken or disconnected cable) . The ED can be destroyed by such a fault condition. Especially for EHID systems having expensive ED compared to EDs for electroded lamps such a feature is of special interest. - The invention allows the construction of a very compact lamp based on EHID technology (without cable between ED and LCU) which might be designed as a retrofit e.g. in a PAR38 housing or with an E27-socket.

Every discontinuity in the impedance (in the line imped- ances) along the RF path the wave is travelling leads to a reflection of a part of the wave. The reflected wave travelling back to the source leads to additional losses. Some of the reflected power returning at the source is transformed into forward power by the source again but the rest of it is lost as heat. For those two reasons it is desirable to have no reflected power at all. No reflected power means that Vmax = Vmin yielding to SWR = 1 and hence an exact impedance matching of the ^λ load' impedance with ^source' impedance is necessary. An impedance-controlled feedthrough as described above is proposed.

A procedure is given to check whether the feedthrough is an ^λ impedance-controlled' feedthrough according to this invention or not: Step 1: The SWR is measured right before (regarding the forward power path) the feedthrough (e.g. between the end of the cable and the feedthrough according to figure 2) .

Step 2: The feedthrough is removed from the power path. In the example shown in figure 2 the cable would be directly connected to the applicator. The rest of the system has to be operated under the same conditions as before (preferably in steady state) .

To operate the system under the same conditions as before is difficult in case of an EHID system. To come to reliable results the steady state operation should be chosen. In addition the discharge vessel has to operate under conditions where it electrically behaves in the same way as before (with the feedthrough which means with the outer bulb influencing the thermal balance of the discharge vessel) . This might be achieved by operating the whole arrangement in a vacuum chamber. Alternatively coupler and lamp may be replaced by an equivalent network (e.g. a dummy load and some reactive components) to simulate the coupler- lamp-unit.

Step 3: The SWR is measured at the ^λ same place' again. This means the meter is placed between the end of the cable and the applicator in the example above.

Step4 : From the SWR the quotient of reflected power Pref and maximum achievable power Pmax (with ideal tuning corresponding to an SWR equal 1) can be determined using formula (3) or figure 3 or 4. Besides the reflected power figure 3 illustrates the quotient of load power Pl and maximum achievable power Pmax which can be calculated according to formula (2) .

In case of an ^λ impedance-controlled' feedthrough the reflected power measured in step 3 increases by less than 50% compared to the reflected power measured in step 1 :

(4) stepλ

In increase of less than 50% is an acceptable amount of reflected power for an feedthrough in an EHID system as it degrades the overall efficiency not too much and this efficiency drop on the electrical side of the system can be overcompensated by the increase of efficacy of the discharge arrangement due to the outer bulb as experiments show.

Example 1: In step 1 a SWR of 1.6 is measured corresponding to 5.33% of reflected power. In step 3 a SWR of 1.4 is measured corresponding to 2.78% of reflected power. Therefore the reflected power was reduced by (5.33-2.78) /5.33 = 48% what yields to the result that the feedthrough was an ^λ impedance-controlled' feedthrough .

Example 2: If a SWR of 1.4 is measured in step 1 whereas a SWR of 1.6 is measured in step 3 no calculation would be necessary to be able to conclude that the feedthrough is an impedance-controlled feedthrough because the SWR got ^λ worse' by taking out the feedthrough. Such an increase may be observed if the feedthrough is used as impedance match- ing network between e.g. the cable and the applicator have significantly different impedances (see the paragraph after example 3) .

Example 3: In step 1 a SWR of 1.4 is measured corresponding to 2.78% of reflected power. In step 3 a SWR of 1.2 is measured corresponding to 0.83% of reflected power. There- fore the reflected power was reduced by (2.78-0.83) /2.78 = 70% what yields to the result that the considered feedthrough is not an ^λ impedance-controlled' feedthrough.

Preferably the impedances of the feedthrough seen from the outer and from the inner side of the outer vessel have impedances coming close to the impedances of the elements connected on both sides of the feedthrough. The element on the outer side of the outer vessel may be a 50 ohm cable coming from the power amplifier. Therefore the ^λ input im- pedance' of the feedthrough (seen by the cable) should have impedance coming close to 50 ohms. The element on the inner side of the outer vessel may be the coupler having an im ^¬ pedance of 600 ohms. Therefore the ^λ output impedance' of the feedthrough (seen by the coupler) should have impedance coming close to 600 ohms. The ^λ input impedance' and the ^λ output impedance' of the feedthrough may be not the same.

Feedthrough arrangements can be subdivided into two catego ^¬ ries :

A) Galvanically isolating feedthrough arrangements

Galvanically isolating feedthrough arrangements come with the advantage that there is no need for an electric connec ^¬ tion by a wire etc. between the inside and the outside of the outer vessel. Therefore the vessel can be manufactured without conventional feedthrough systems yielding to low cost and long life of the vessel. Drawbacks might be high EMI and reduced overall efficiency of the system but the existence of such drawbacks is highly dependant on the individual embodiment.

B) Galvanically connecting feedthrough arrangements

Galvanically connecting feedthrough arrangements are based on conventional feedthrough systems which establish an electrical connection by use of wires and/or foils between the inside and the outside of the outer vessel. Preferably the feedthrough constitutes of strip line or coplanar line or coaxial arrangement inside the vessel's wall. In addition elements like capacitors, strip lines, coaxial cables, etc. on the inner and/or on the outer side may be involved. Those elements may be partially included into the feedthrough.

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 an example of a known EHID system,

Fig. 2 shows an EHID system using a feedthrough in the outer bulb, Fig. 3 shows the load power Pl and reflected power Pref normalized to the maximum achievable power Pmax as a function of the SWR,

Fig. 4 shows the load power Pl and reflected power Pref normalized to the maximum achievable power Pmax as a function of the SWR in a different scale, Fig. 5 shows a galvanically isolating feedthrough ar- rangement using a magnetic coupler,

Fig. 6 shows an EHID system with an unidirectional communication between ED and LCU,

Fig. 7 shows a section of the EHID system with a bidirectional communication between ED and LCU, Fig. 8 shows a galvanically isolating feedthrough arrangement using feedthrough capacitors,

Fig. 9 shows the top view of two feedthrough capacitors in concentric arrangement, Fig. 10 shows an exploded view of two feedthrough capacitors in concentric arrangement including the outer vessel's wall as dielectric material,

Fig. 11a shows side view of feedthrough capacitors without a holding recess,

Fig. lib shows side view of feedthrough capacitors with a holding recess,

Fig. 12 shows an exploded view of magnetically and ca- pacitive coupled feedthrough including the outer vessel's wall as dielectric material,

Fig. 13 shows a glass feedthrough system using molybdenum foil and wires,

Fig. 14 shows a glass feedthrough system using a feedthrough part Fig. 15 shows a glass feedthrough system using molybdenum foil and wires, where the two outer foils are rotated compared to the arrangement in figure 13,

Fig. 16 shows a top view of quasi-coaxial feedthrough using five Mo-foils with two wires welded to each foil.

Preferred embodiment of the invention

1. The Feedthrough

A. Galvanically isolating feedthrough arrangements A.I Magnetic feedthrough (e.g. a feedthrough coupler)

Coils on the inside and outside of the outer vessel are arranged in such a way that the magnetically flux in the two coils is coupled well. Especially at lower frequencies (16 kHz - 10 MHz) coils with magnetic cores e.g. ferrite cores may be used to achieve high coupling which leads to low EMI and good efficiencies of the system. At higher frequencies air cores are sufficient regarding coupling but an EMI shielding e.g. two hemispheres on both sides of the vessel's wall might be necessary. Figure 5 depicts a magnetic feedthrough that uses a magnetic coupler made up of two coils (nl, n2) which may have different number of turns wound on ferrite E-cores (El, E2) made of 4Fl material from Ferroxcube and having additional capacitors (Cl, ..., C4) used for impedance matching on the inner side (Obi) and on the outer side (Obo) of the outer vessel in an EHID system operating at 8 MHz. The vessel is made of a quartz glass which forms the wall (Wa) separating both parts of the feedthrough arrangement. The (input) impedance of the outer part of the feedthrough is Zo and the (output) impedance of the inner part of the feedthrough is Zi. The feedthrough is designed in such a way that Zo comes close or is identical to the impedance of the element to which it is connected on the outer side to the feedthrough (e.g. the cable) whereas Zi comes close to the applicator impedance or whatever is connected an the inner side. Therefore the impedances Zo and Zi match the adjacent impedances. The impedances do not have to have the same value even though the two impedances should have the same value and be identical to e.g. the cable impedance if an outer vessel will be added afterwards to enhance an already designed EHID system. The impedance matching may also include the conversion from a symmetrical line to an unsym- metrical line and vice versa (e.g. coaxial line - strip line) .

Instead of an E-core other geometries might be used. Half a ring core on each side or two U cores might be beneficial due to the simplicity of their geometries. In addition instead of ferrite material iron powder might be used. The two parts of the feedthrough are glued to the wall using a thin layer of temperature stable adhesive. Alternatively a spring construction may hold each part of the feedthrough in place. The latter makes the replacement of the outer vessel including applicator and bulb easy (this should be necessary only very rarely due to the long life of the EHID system) . A.2 Capacitive feedthrough

Capacitive feedthrough is based on feedthrough capacitors in the wall of the outer vessel. Especially at moderate frequencies and high frequencies (10 MHz to several GHz) feedthrough capacitors using the wall material e.g. quartz glass as dielectric material are getting small and are simple to build yielding to low EMI and cost. The electrodes of the feedthrough capacitors may be made of - metallization on the wall (e.g. on both sides of the wall as shown in figure 8) metal strip pressed into the glass (e.g. 2 strips in the glass facing each other or just one strip in the glass and a metal sheet on the outside as shown in figure 9) - stripes of sheet metal (preferably the sheet metal is formed as a clip or spring keeping itself in place) or a combination of the above mentioned ways.

Additional passive elements like capacitors and inductors on the inner and/or on the outer side of the feedthrough capacitors may be added to make up an impedance-matched feedthrough.

To ignite electroded HID lamps the use of feedthrough capacitors using sleeve material has been described in US 7,271,547B2 and US 7,378,800B2. In both documents the elec- troded HID lamp has a third electrode for starting purposes which is connected via a single feedthrough capacitor to the starting circuit. During the starting process a small amount of energy is coupled into the lamp to initiate starting. In contrast to the current invention the power for completing the starting process as well as for operating the lamp (run-up and steady-state) is delivered to the two main electrodes of the electroded HID lamp without passing any feedthrough capacitor. In the current invention the power for starting, run-up and steady-state operation is running through two feedthrough capacitors.

Figure 8 illustrates an arrangement having two feedthrough capacitors. The first capacitor is made up of the plates PIl, P12 and the wall (Wa) as dielectric material. The second capacitor is formed by plates P21, P22 and the wall. Impedance matching is achieved by the additional capacitors Cl, ..., C4 and the inductors Ll and L2.

Especially for feeding an unsymmetrical line (e.g. coaxial cable) through the outer vessel two feedthrough capacitors in a concentric arrangement (circle + circular ring) as shown in figure 9 and 10 are well suited. The capacitors are made of metallization on both sides of the vessel's wall which are electrically connected by metal springs. The contact areas of the two springs on one side, e.g. the outer side of the vessel are shown as black areas. The outer circular ring on the outer side of the vessel is electrically connected to ground potential. The ground may be used as a shielding for reducing EMI and the shielding may enclose most of the outside of the outer vessel.

Figure 11 (a) depicts two feedthrough capacitors lying behind each other therefore only the upper one can be seen. Each capacitor consists of a metal strip pressed into the glass (Pi) and a second metal strip (Po) on the outer sur- face of the vessel surrounding the inner metal strip. The outer plate may hold itself in one ore more recesses of the wall as shown in figure 11 (b) .

A.3 Magnetically and capacitively coupled feedthrough Figure 12 shows a feedthrough arrangement made of metallization on both sides of the outer vessel. The input and the output of the feedthrough are electrically connected to the adjacent system components as in one of the previous examples by metal springs. Three contact points of the four springs can be seen in the figure. The coil formed on the outer surface is magnetically coupled to the coil on the inner surface. In addition the potential differences along the coil are capacitively coupled to the other side of the vessel. Therefore both ways of coupling - magnetically and capacitively - are employed.

B. Galvanically connecting feedthrough arrangements

Preferably the galvanically connecting feedthrough arrangements constitute of strip line or coplanar line or coaxial arrangement inside the vessel's wall. In addition elements like capacitors (e.g. capacitors in series and in parallel (analogously to figures 5 and 8) on both sides of the wall are welded to the metal mounting structure to achieve impedance matching), strip lines, coaxial cables, etc. on the inner and/or on the outer side may be involved. Those elements may be partially included into the feedthrough.

A glass feedthrough system using molybdenum foil and wires seems to be well suited because it can be manufactured by conventional press seal techniques for quartz glass. A feedthrough using incorporating such a glass feedthrough can be manufactured with low tolerances leading to electrically well defined parameters in a low-cost process. A geometrical symmetric arrangement with a ^λ hot' middle conductor and two outer conductors as shown in figure 13 form a electrical unsymmetrical line which may be connected e.g. to a coaxial cable. Such an arrangement comes close to a coplanar line. Besides the simplicity of manufacturing it has the advantage of low RF radiation leading to low EMI .

By choosing appropriate geometric dimensions and material constants (e.g. foil width and thickness, wire width, distances between the parts, wall thickness and dielectric constant of the glass) the impedance of the line can be adjusted. A fist estimate is possible by using the formulas in Andreas Thiede, Introduction to High Frequency Technol- ogy (in German) on page 6 and 7. The dielectric constant is and is not calculated according to equation (1.16) because the metal is fully confined by glass in contrast to a conventional coplanar line where only one half-space is filly with dielectric material. Impedance variations due to a not infinite geometrical extension of the ground plane are covered in M. Houdart : ^λ Coplanar Lines: Application to Broadband Microwave Integrated Circuits', 6th European Microwave Conference Proceedings, 1976, pp. 46-53. Preferably the impedance is chosen in such a way that the impedance matches the impedance of the adjacent element (s) on the inside and/or on the outside. In case all three impedances are the same no additional measures for matching are necessary (e.g. the same type of cable is used on both sides the feedthrough impedance is designed to match the cable impedance) . If the impedances are different capacitors (preferably through-hole devices) in series and in parallel (analogously to figures 5 and 8) on one or on both sides of the wall are welded to the metal mounting struc- ture to achieve impedance matching.

Figure 13 depicts an embodiment where the cable cannot be disconnected from the feedthrough. The joints are made by soldering, welding or crimping. The coaxial cable Cab consists of a copper-plated core Co, an inner dielectric insu- lator Di, a woven copper shield Sh and an outer plastic sheath (not shown) . The cable is connected to wires Wo at the outside of the outer vessel. Those are connected to the molybdenum foils Mo which are pressed into the quartz glass of the wall Wa. At the inside of the outer vessel the wires Wi are providing electrical and mechanical support to the coupler (not shown) . To avoid any discontinuity in the epsilon between cable and glass feedthrough the connecting area outside the outer vessel is filled with adhesive material Bp (e.g. a silicone which is temperature-resistant). The adhesive comes with the additional advantage of me- chanically supporting the joining areas leading to increased mechanical robustness.

Instead of using organic materials like adhesive materials or potting materials the epsilon discontinuity can be avoided by attaching additional metal foils (e.g. sheet metal welded to the wires) in the area of the air between the outer vessel and the cable (not shown) . Those metal parts increase the capacity in this area. An embodiment may use two identical metal foils welded to the two outer wires (and an optional third metal foil on the inner wire) . The two outer foils are symmetrically attached to their wires with respective to the orientation of the inner wire (or foil) .

The two outer ground connections at the inside of the ves- sel are made of a single wire Wi (U-shaped ground wire) . This has the advantage that both ^λ legs' are electrically connected and the arrangement has an increased mechanical stability and may be used as a mounting support e.g. for the coupler. In contrast to embodiment shown in figure 13 a (cylindrical) connector could be attached to wires coming out of the vessel realizing a dismountable connection between cable and vessel.

As described above the impedance can be adjusted not only by the dimensions of the parts it can also be adjusted by their orientations. Figure 15 illustrates an embodiment where the two outer foils Mo are rotated compared to the arrangement in figure 13. By the rotation angle the impedance can be adjusted (in particular due to the change in the capacity per unit length) .

Figure 15 shows some dashed lines indicating optional ground connections. Those connections elongate the ground plane (and the design rules for coplanar lines match better) and give additional mechanical strength. By attaching the four ground lines around the circumference of the man- tie of the cable a highly symmetrical arrangement is achieved leading to good RF properties.

Figure 14 shows an alternative embodiment with advantages in respect of manufacturing. All connections are estab- lished in a feedthrough part FP, where the feedthrough is made as described above using outer Wires Wo, Metal Foils Mo and inner Wires Wi. The feedthrough part is a ready prepared part and glued in the Wall Wa of the outer Vessel by a sealing adhesive SL like a glass solder or an epoxy. The ready made feedthrough part can be made of a composite material of different glass types or oxides to match the required ε _r . The ready made feedthrough part may also be composed of a partly porous glass part with bubbles included for improved adjustment of the required ε _r . A quasi-coaxial feedthrough using 5 Mo-foils with two wires welded to each foil is shown in figure 16. Such a feedthrough might particularly be suited to feed a coaxial cable through the vessel's wall.

Instead of using quartz glass an outer vessel constructed of polycrystalline alumina or sapphire (for UV generation) or a transparent plastic (which might be part of the fixture) could be used.

2. The System The components of an EHID system according to the first embodiment is shown in figure 6. Coupler (or ^λ applicator' ) and discharge vessel (or ^λ bulb' or ^λ lamp' ) are integrated in a lamp-coupler-unit (LCU) during production by mechanical interconnection (therefore additional mechanical parts like a frame, housing or bulb surrounding both components might be used) . The LCU is operated by the electronic driver (ED) . The dash-dotted lines indicate RF paths whereas solid or dashed line indicate low-frequency or optional low-frequency paths. The first embodiment realizes an unidirectional communication between ED and LCU. The cable connecting both components is used for power delivery (RF in the frequency range of 100 MHz and several GHz) and communication. The dual use of the cable comes with several advantages: no cost for an additional communication path, the power amplifier is protected against 'no load' (which may bow up the amplifier due to thermal stress caused by severe mismatch) because the software inside the ED at fist establishes communication with the LCU and later on starts up the amplifier, making both connections (comm. and power) at the same time it is impossible that only power delivery or communication is working which makes the system more robust and simple (software program running in the ED) , it can be guaranteed that the temperature protection of the LCU as an optional feature is always active (important to the manufacturer regarding guaranteeing a warranty to the customer) . The LCU comprises non-volatile data memory (DM) used as a Specification plate' which can be accessed by the electronic driver (ED) . Preferably this is done by electrical signals directly (e.g. by a transmitter (TX) sending the data to ED) but it could also use mechanical (grooves, fingers, etc. coding the information), electromagnetic

(e.g. RFID technology) or optical (e.g. fiber optics, data matrix, etc.) ways to read out the information stored in DM. In further descriptions only the direct electrical way for communication is considered. The driver electronics (ED) is powered by an external energy source e.g. the mains (M) . The ED may comprise a power supply (PS) , control unit (Ctrl) , oscillator (Osc) , power amplifier (Amp) and power meter (Mtr) to measure forward power and reflected power (or SWR, the standing wave ratio) besides the receiver (RX) for receiving the information coming from DM. The control unit (Ctrl) controls the whole system. It encodes the information coming from

(a) DM as well as

(b) information like dimming levels, etc. from an inter- face (DL) communicating with the 'outside world' or

(c) other commands (shut down, in system programming during manufacturing, etc.).

The lamp-coupler-unit (LCU) comprises impedance tuning circuit or filter (Fi) for matching the applicator-lamp- unit (App-La) to the cable or to the amplifier (in case there is no cable) . All electronics inside the LCU are supplied with electric power by the ED. The LCU has no own power supply. The power may be delivered over the same cable that provides the RF power. In addition the same cable may be used for the (low-speed) communication between ED und LCU. A splitting of RF power and low frequency signals (auxiliary power supply and communication) is realized by high-pass and low-pass filters (HP, LP) . Besides reading data from DM the communication is used to get information about the coupler temperature, luminous flux (photodiode in the LCU) , reflected power (in case of a power meter in the LCU), and operating hours (from an operating hours counter in the LCU which may use DM for data storing) .

The filter (Fi) may not be incorporated in the LCU as shown in figure 6. Instead it may be placed between Mtr and HPl.

In another embodiment a bidirectional communication is used which gives a more effective way of communicating in contrast to the unidirectional communication (e.g. EC can ask for a specific data and doesn't have to filter the data stream coming from the LCU) and enables more features (e.g. commanding tuning measures to Fi, writing aging state (or at least operating hours) to the LCU' s DM unit) .

Depending on the particular embodiment of the EHID system the customer may be able to replace the whole lamp-coupler- unit. If this is possible it is only possible as a single part. This unit is not dismountable which means not further dismountable by the end-user without damaging the unit. Therefore under all circumstances (in assembly of the system as well as for replacement at the customer's site) it is made sure that the data provided by the LCU to the ED is valid for the coupler and the lamp currently being used.

The data memory DM is programmed during production. In case of a bidirectional communication this may be accomplished through the particular ED unit belonging to the same system via ^λ in system programming' using the input DL (communication bus) .

A variety of data is held in the DM:

Product number: All LCUs with the same wattage, optics, etc. do have the same number. The ED reads this num- ber and operates the LCU accordingly. This makes the operation of a 100 W LCU on a 200 W ED possible. Without this feature the bulb might explode when the user hooks up the wrong lamp which might be easily possible due to standardized coaxial power connectors. A recognition of the lamp operated on the electronic driver by characteristic properties as described in EP 1 519 638Al for low-pressure electroded fluorescent lamps (e.g. by the resistance of filaments) is not necessary making the driver simpler and low priced. A communication telling the 'outer world' about the lamp type (and the version of the electronic driver) currently present in the system by the communication bus DL which might be a DALI or DMX interface, etc. is possible. EP 1 566 989Al describes such a communication for low-pressure electroded fluorescent lamps employing DALI.

Serial number of the LCU (makes dedicated callback of products possible as well as measures against product piracy or enables an effective warranty management) , Impedance over time Z (t) of the LCU for tuning and operating (e.g. frequency and impedance variation necessary by Osc and Fi during ignition, run-up and for steady state may be held in DM) , - Resonance frequencies for the particular lamp to take advantage of acoustic resonances in the lamp. These frequencies are measured during production and then stored in DM for each LCU individually. Therefore an elaborate measurement of resonance frequencies during the operation as described in EP 0 708 579Bl for electroded HID lamps is not necessary.

Dimming data: light output and impedance depending on RF power

Run-up characteristic of the lamp over time (e.g. lumens output over time at several power levels - necessary for fast run-up time or dimming during run-up)

Aging status: In the most primitive implementation this is the actual value of operating hours counter

Aging parameters: Parameters describing the aging of the lamp-coupler-unit (e.g. deterioration of luminous flux, change in impedance, weariness - leading to a reduction of life, shift of acoustic resonance frequencies) over life. There might be correction factors included taking dimming level, number of starts, etc. into account. This information is used to operate the lamp at the beginning of life only with e.g. 80% of the nominal power (and still having nominal light output) . This measure elongates life and reduces power consumption. Later on e.g. between 2.00Oh and 5.000h the power is ramped up to nominal power and after 15.000h the power is reduced again to further elongate life (while scarifying luminous flux) .

In addition the aging parameters may be used to calculate the aging status of the LCU. In this case the aging status is not only the amount of operating hours. It might be something like the effective operating hours taking dimming, starts, etc. into account. The aging status may be calculated in the ED and written to DM. Alternatively this may be done by the LCU which is fed with operating parame- ters like the actual power during lamp operation.

By knowing the aging parameters and knowing the actual aging state the system predicts the remaining service life (regarding the user's habits) until a replacement of the LCU is necessary. This information may be provided to the user or service personnel e.g. by the communications bus DL. a further embodiment uses a bidirectional communication between ED and LCU. The overall system design of this embodiment is identical to the design shown in figure 6. The section of this system belonging to the communication and powering the LCU is shown in figure 7 in detail.

The high-pass filters HPl and HP2 are realized by capacitors CIl and C21, respectively. The capacitor CIl may already be part of a state of the art amplifier circuit (used to extract RF power from the final power stage of the amplifier) . C21 may be part of the filter according to the state of the art and hence no additional power parts (for the high-pass filters) might be necessary. The low-pass filters LPl and LP2 are realized by LIl, C12, L12 and L21, C22, L22, respectively.

The power supply (PS) of the electronic driver (ED) provides the (auxiliary) voltage VCC feeding the communication and the control unit of the ED. In addition PS supplies auxiliary power to the LCU. The latter is achieved by the resistor RIl which is connected to VCC and supplies power to the cable Ca. RIl is used as a short-circuit protection. By use of resistor R21 and diode D21 the auxiliary voltage VDD is generated which is filtered by C23. VDD supplies the LCU. R21 should have a low value and is only used as an over-current protection for the diode D21. Transmitting data is accomplished by gating the transistors Ql and Q2 which will pull the 'signal line' to ground or (in case neither Ql nor Q2 is conducting) let its 'low- frequency potential' unchanged. Receiving data is managed by comparing the attenuated voltage (attenuation is accomplished by resistor dividers R12, R13 and R22, R23) on the 'signal line' (the core of the coaxial cable Cab) with the reference voltage Vrefl or Vref2 (the reference voltages have not to be identical) by the use of comparators Col and Co2. The comparators are protected against perturbances (e.g. over-voltage, surges, EDS, etc.) by the components R12, C14, D12, DIl and D21, R22, C24, D22, respectively (DIl, D12 and D22 have been added only for this purpose) .

For the communication an appropriate coding scheme or pro- tocol should be used to prevent the occurrence of turning on Ql or Q2 too long (independently of the data which is transmitted) because this would risk the proper power supply of the LCU (as C23 gets discharged too much) . Alternatively a minimal pause time (e.g. between the frames or packages transmitted) has to be assured (during which C23 can be charged to a sufficiently high voltage) .

By the use of the optional Zener diodes D14 and D24 the transistors Ql and Q2 won't pull the 'signal line' to ground even though the transistors are still able to modu- late a low-frequency signal on the 'signal line' . By adding D14 and D15 the afore mentioned restrictions on the coding scheme or protocol do not apply any more as the voltage VDD is generated no matter what state (low or high) is modulated on the 'signal line'. The embodiment shown in figure 7 is useful for a galvani- cally connecting feedthrough arrangement. For galvanically isolating feedthrough arrangements modified arrangements can be used modulating the 'signal line'. To keep the cost low a communication frequency lower than the operating frequency of the EHID system (the fundamental frequency of the amplifier's output) is used and high-pass and low-pass filters are applied as described e.g. in figure 6. To generate the supply voltage VDD in the LCU a rectification of the signal on the 'signal line' is used.

List of reference designations

Amp power amplifier

App applicator or coupler App-La applicator-lamp-unit

Bp adhesive material used as gap filler

Cab cable

Ctrl control unit

Co core of coaxial cable Di inner dielectric insulator of coaxial cable

DL communication bus

DM non-volatile data memory

El, E2 magnetic cores

ED electric drive Fee feedthrough

Fi impedance tuning circuit or filter

HP, LP high-pass and low-pass filters

La lamp or discharge vessel

LCU lamp coupler unit M mains

Mo molybdenum foil

Mtr power meter nl, n2 coupler coils

OB outer bulb Obi, Obo inner and outer side of the outer vessel

Osc oscillator

PIl, P12 plates of first capacitor

P21, P22 plates of second capacitor

Pi, Po inner and outer metal strip PS power supply

RX receiver

S switch

Sh shield of coaxial cable

SL sealing TX transmitter

Wa wall of the outer vessel

Wi wires inside the outer vessel

Wo wires outside the outer vessel