EP2326144B1 - Lighting device with LED and regulated power supply - Google Patents

Description

• Avoir un encombrement extrêmement faible, qui interdit l'usage de composants électroniques passifs tels que des condensateurs électrochimiques de forte capacité ;
• Mettre en oeuvre un très grand nombre de diodes (plusieurs centaines) de façon à assurer le niveau de luminance nécessaire ;
• Assurer un balayage temporel de l'alimentation des diodes synchronisé sur le balayage vertical de l'image vidéo appliqué à l'imageur matriciel de façon que l'allumage des diodes puisse être synchronisée sur le balayage vidéo ;
• Posséder une très grande dynamique de luminosité. Pour assurer cette dynamique, les diodes sont allumées pendant une certaine durée d'un cycle. Au minimum de lumière, on démontre que la durée d'allumage ne doit pas excéder une à deux microsecondes.

The field of the invention is that of light-emitting diode lights used to illuminate liquid crystal matrix imagers. The invention relates more particularly to lighting having to have both a very large dynamic and a very high level of luminance. This type of lighting is used in aeronautics to illuminate micro-imagers helmet visors that must be used day and night. For this type of application, the lighting must have the following characteristics:

• Have an extremely small footprint, which prohibits the use of passive electronic components such as electrochemical capacitors of high capacity;
• Implement a large number of diodes (several hundred) to ensure the necessary level of luminance;
• Providing a temporal scan of the synchronized diode power supply on the vertical scan of the video image applied to the array imager so that the diode firing can be synchronized to the video scan;
• Possess a very great luminosity dynamic. To ensure this dynamic, the diodes are lit for a certain period of a cycle. At minimum light, it is shown that the duration of ignition should not exceed one to two microseconds.

In general, the diodes are organized in a matrix of N ramps each having M diodes, N and M being greater than one. In the remainder of the description, the ramps are referenced R _i , i being an index varying from 1 to N and the diodes are referenced D _j , j being an index varying from 1 to M. To carry out the illumination of the matrix, several feeding solutions have been proposed.

A first solution exposed in figure 1 consists in using a power supply R ramp, the supply controlling M diodes D _j arranged in series. Each power supply comprises a boost circuit also called a "booster" 20. This circuit 20 is powered by a low DC voltage V _IN through an inductor 30 and controlled by a PWM type control, acronym for "Pulse Width Modulator". By way of example, the reference circuit ISL97634 of the company INTERSIL _® can be used as a booster. Refer to the technical data sheet "Data Sheet FN6234.3 March 2008" of this company for all technical information on this circuit. This first solution unfortunately leads to an excessively large footprint in that it requires N risers 20 to drive N ramps R of M diodes D _j .

A second solution is to use a single high voltage boost converter delivering a voltage V _HT , each ramp being controlled by a circuit called lowering buck called "buck". As for the previous solution, N inductances are necessary but these are much weaker because their current is of value close to the current of the diodes and the value of the cyclic ratio of cutting HF quite high, of the order of 90%. This solution is still not satisfactory in terms of size.

A third solution is to mount the ramps 10 in parallel. It is illustrated in figure 2 . The diode ramps R _i are powered by a single voltage V _HT . Since the direct voltages of the diodes are not exactly equal, the voltages of the ramps R _i are not exactly identical. To properly balance the currents, it is necessary to use voltage sources C _i independent of the voltage. The nominal voltage at the terminals of these sources corresponds to a lost power. It must therefore be of just sufficient value in the worst case.

An example of an assembly of this type is shown in figure 3 . In order to ensure the "dimming" of the light box, the current sources C _i must be switched temporally with a cyclic ratio dependent on the required luminance by dimming control means DIM _i driving transistors T _i type "MOS". In the case of a non-time-scanning light box, all the transistors T _i can have the same command. The collector-emitter voltage V _CE of the bipolar transistors T _i must not exceed a few volts, but be sufficient to withstand the disparity of the voltages between the different ramps R _i . To obtain this, the voltage is measured on the collectors of the transistors and the supply voltage V _HT of the ramps is slaved accordingly. But this can only work properly if the voltage V _HT is properly dynamically regulated and its ripple is reasonable. The scheme of the figure 3 all ramps have independent control allows any combination of time scanning, one to N ramps can be lit simultaneously.

• Un circuit logique programmable 100 de type FPGA qui commande les principales fonctions du dispositif d'éclairage ;
• Trois unités d'éclairage 110 comportant chacune six rampes R de diodes D, les six rampes étant disposées en parallèle ;
• Un circuit « DCM » 120 de type convertisseur-élévateur générant la haute tension d'alimentation ;
• Un circuit d'asservissement 130 à vide de cette haute tension d'alimentation ;
• Un premier circuit de « dimming » 140 à six commandes indépendantes permettant de piloter simultanément une rampe de chaque unité d'éclairage, soit au total d'éclairer trois rampes simultanément ;
• Un second circuit de dimming 150 à trois commandes permettant de piloter une et une seule rampe de chaque unité d'éclairage.

The figure 4 represents an example of complete electronic realization of a lighting device of this type. In this diagram, three units each comprising six ramps of diodes D are driven by an electronic circuit FPGA (Field-Programmable Gate Array) type. The complete electronic circuit comprises six main sets framed by a dashed rectangle on the figure 4 and who are:

• A programmable logic circuit 100 of the FPGA type which controls the main functions of the lighting device;
• Three lighting units 110 each comprising six ramps R of diodes D, the six ramps being arranged in parallel;
• A "DCM" circuit 120 of converter-boost type generating the high supply voltage;
• A servocontrol circuit 130 of this high power supply voltage;
• A first "dimming" circuit 140 with six independent controls for simultaneously controlling a ramp of each lighting unit, ie in total to illuminate three ramps simultaneously;
• A second dimming circuit 150 with three controls for controlling one and a single ramp of each lighting unit.

This diagram thus allows two modes of operation: three ramps lit simultaneously or one ramp at a time. This provision makes it possible to halve the number of transistors and control signals to achieve the current sources.

As previously mentioned, the voltage V _HT must be controlled to the nearest volt in order to minimize the voltage across the transistors of the current sources. The voltage V _HT must not be subject to transient variations when switching the charging current during the "dimming" of the diodes. The conventional solution to this problem is to size the capacitor C _HT filtering the voltage V _HT to a value such that the upconverter does not feel the load variations. This is the solution, a priori, natural for applications where the volume is not critical and where it is not forbidden to use capacitors of several hundred microfarads with a service voltage greater than 100 volts. In this case, it is possible to use a conventional "current mode" integrated control circuit for a step-up converter, and the converter control loop is very slow. In the case of an architecture with a converter by diode ramp, the best specialized integrated circuits on the market make it possible to dispense with a high capacitor value as has already been mentioned. However, for some applications, the light source must necessarily have an extremely small footprint, which prohibits both the use of passive electronic components such as electrochemical capacitors of high capacity and the multiplication of risers.

• Obtention de la haute tension par un convertisseur élévateur à conduction discontinue dit DCM ;
• Asservissement individualisé pour chaque mode de charge par circuit d'asservissement dédié ;

The diode lighting device according to the invention does not have these disadvantages. Indeed, the electronic circuit controlling the voltage source is based on the combination of two main characteristics which are:

• Obtaining high voltage by a DCM step-down boost converter;
• Individualized servo control for each load mode by dedicated servocontrol circuit;

Thus, thanks to this individualized servocontrol and to the particular response of the DCM DCL, it is easy to control the stability and the speed of the servocontrol. for each mode, without having to use a specialized integrated circuit with current loop and counter-ramp.

More precisely, the subject of the invention is a light-emitting diode lighting device, said light-emitting diodes being arranged in a first plurality of ramps arranged in parallel, each ramp comprising a second plurality of light-emitting diodes arranged in series, said ramps being fed with a plurality of electroluminescent diodes. by a DC voltage of several tens of volts said high voltage, said voltage being generated by a boost converter circuit from a low DC voltage of a few volts, the value of said voltage being a function of the duty cycle of the boost converter circuit, said voltage being slaved to a constant average value by means of a substantially analog servocontrol device controlling said duty cycle, said servocontrol device comprising several modes of operation, a mode being defined either by a particular electronic addressing of the high voltage, or by a defined number of diode ramps lit, characterized in that the upconverter circuit is discontinuously conductive and that the servocontrol device comprises a plurality of electronic servocontrol circuits connected to an electronic multiplexer, each electronic circuit of servo-control being dedicated to a particular mode of operation, the electronic characteristics of said servocontrol electronic circuits depending on said mode of operation, said servo electronic circuit being operational only when the operating mode is selected.

Advantageously, the servo device comprises means for storing the different cyclic ratios dedicated to each mode of operation.

Advantageously, when the servo-control device is made in analog mode, each electronic servocontrol circuit comprises an operational transconductance amplifier known as OTA, an activation command and an integration circuit arranged in series.

Advantageously, the gain of the operational transconductance amplifier of each electronic servocontrol circuit depends on the mode of operation to which said electronic servocontrol circuit is dedicated and the different integration circuits of the different electronic circuits are all identical.

Advantageously, the first plurality of ramps is structured into a first number N of ramp units, each unit comprising a second number M of ramps, the illumination of the diodes comprising the ramps being controlled in a matrix manner by two control circuits also called "dimming" circuits, the first circuit comprising N first control means, each first control means for controlling simultaneously one and a single ramp of all the ramp units, the second circuit comprising M second control means, each second means command to simultaneously control all ramps of one and only one unit.

The invention also relates to a lighting device with light-emitting diodes, said light-emitting diodes being arranged in a first plurality of ramps arranged in parallel, each ramp comprising a second plurality of light-emitting diodes arranged in series, said ramps being supplied by a DC voltage several tens of so-called high voltage volts, said voltage being generated by a boost converter circuit from a low DC voltage of a few volts, the value of said voltage being controlled by the duty cycle of the boost converter circuit, characterized in that said voltage is slaved to a constant average value by means of an essentially digital servocontrol device controlling said duty cycle, said servocontrol device having a plurality of operating modes, a mode being defined either by a particular electronic addressing the high voltage, or by a defined number of diode ramps lit, the upconverter circuit "being discontinuous conduction".

• La figure 1 représente un premier dispositif d'éclairage selon l'art antérieur comportant une seule rampe de diodes électroluminescentes, lesdites diodes étant disposées en série;
• Les figures 2 et 3 représentent un second dispositif d'éclairage comportant plusieurs rampes de diodes électroluminescentes, lesdites rampes disposées en parallèle ;
• La figure 4 représente un schéma électronique détaillé d'un second dispositif d'éclairage comportant 18 rampes de diodes électroluminescentes ;
• La figure 5 représente, en réalisation analogique, le principe de l'alimentation électrique et de son asservissement d'un dispositif d'éclairage à rampes de diodes électroluminescentes selon l'invention ;
• La figure 6 représente le schéma électronique d'un circuit convertisseur élévateur selon l'invention ;
• La figure 7 représente le schéma électronique d'un dispositif d'asservissement selon l'invention ;
• La figure 8 représente les différents gains de la boucle d'asservissement électronique selon l'invention ;
• La figure 9 représente la valeur ces différents gains en fonction de la fréquence ;
• La figure 10 représente, en réalisation numérique, le principe de l'alimentation électrique et de son asservissement d'un dispositif d'éclairage à rampes de diodes électroluminescentes selon l'invention.

The invention will be better understood and other advantages will become apparent on reading the description which follows given by way of non-limiting example and by virtue of the appended figures among which:

• The figure 1 represents a first lighting device according to the prior art comprising a single ramp of light-emitting diodes, said diodes being arranged in series;
• The figures 2 and 3 represent a second lighting device comprising a plurality of light-emitting diode ramps, said ramps arranged in parallel;
• The figure 4 represents a detailed electronic diagram of a second lighting device comprising 18 light-emitting diode booms;
• The figure 5 represents, in analog embodiment, the principle of the power supply and its servocontrol of a lighting device with ramps of light-emitting diodes according to the invention;
• The figure 6 represents the electronic diagram of a boost converter circuit according to the invention;
• The figure 7 represents the electronic diagram of a servo device according to the invention;
• The figure 8 represents the different gains of the electronic servocontrol loop according to the invention;
• The figure 9 represents the value of these different gains as a function of frequency;
• The figure 10 represents, in digital production, the principle of the power supply and its servocontrol of a lighting device with light emitting diode bars according to the invention.

• Mode M₁ dit « 3 rampes » : trois rampes de diodes sont allumées simultanément avec le courant de diode nominal ;
• Mode M₃ dit « 1 rampe » : une seule rampe est allumée avec un courant de diode atténué ;
• Mode M₀ dit « presque à vide », c'est-à-dire que la tension VHT est contrôlée par un pont de résistance qui charge légèrement le convertisseur et nécessite son fonctionnement. Le mode presque à vide facilite le démarrage du convertisseur et permet la surveillance de la tension de sortie.

The light-emitting diode lighting devices concerned by the invention and as illustrated in FIG. figure 5 comprise a first plurality of ramps R _i , said ramps being arranged in parallel, each ramp comprising a second plurality of light-emitting diodes D _j arranged in series, said ramps being fed by a DC voltage V _HT of several tens of so-called high-voltage volts, said voltage being generated by a DCM boost converter circuit from a low DC voltage of a few volts, the value of said voltage V _HT being controlled by the duty cycle of the boost converter circuit, said voltage being slaved to a constant average value at by means of a servo device controlling said duty cycle provided by a control circuit of the PWM type, acronym for "Pulse Width Modulator", said servocontrol device comprising several operating modes M _k , a mode M _k being defined either by a special electronic addressing of the high voltage, either by a defined number of lit diode booms. A particularity of a light-emitting diode lighting device comprising a plurality of ramps connected in parallel is that the charging current is known. Indeed, there exists a finite number K of operating modes M _k . Thus, a first mode M ₁ can correspond to a ramp in operation R _i , a second mode M ₃ to three ramps in operation, a third mode M ₀ to no diode in conduction at a given moment with a known amplitude value. For example, there are typically three charging modes:

• Mode M ₁ says "3 ramps": three diode ramps are lit simultaneously with the nominal diode current;
• Mode M ₃ says "1 ramp": only one ramp is lit with attenuated diode current;
• Mode M ₀ says "almost empty", that is to say that the voltage VHT is controlled by a resistance bridge which slightly charges the converter and requires its operation. The almost empty mode facilitates the start of the converter and allows the monitoring of the output voltage.

• Mode « zéro courant » avec le pont de résistance de la haute tension d'alimentation VHT déconnecté. Dans ce mode, le circuit convertisseur élévateur est éteint. Ce mode est préférable au mode « presque à vide » pour une application à basse lumière qui fonctionne sur batterie ;
• Modes comprenant d'autres combinaisons éventuelles du nombre de rampes avec des intensités de courant dans les diodes égales ou différentes de l'intensité nominale.

But other modes can be added:

• "Zero current" mode with the VHT high voltage supply voltage bridge disconnected. In this mode, the boost converter circuit is off. This mode is preferable to the "almost empty" mode for a low-light battery operated application;
• Modes comprising other possible combinations of the number of ramps with current intensities in the diodes equal to or different from the nominal intensity.

The light emitting diode device according to the invention uses this feature. Indeed, as represented in figure 5 the servocontrol device comprises a plurality of electronic servocontrol circuits CA _k connected to an electronic multiplexer MUX, each electronic servocontrol circuit CA _k being dedicated to a particular operating mode M _k , the electronic characteristics of said electronic servocontrol circuits dependent on said mode of operation, said servo electronic circuit being operational only when the operating mode is selected. In addition, in the device according to the invention, the boost converter is discontinuously conductive.

To each of these operational modes, there corresponds a servo loop which controls the cyclic ratio of the upconverter, symbolized by the small slots on the figure 5 . When leaving a mode, the value of the corresponding duty cycle is memorized. This allows the immediate establishment of the correct duty cycle when returning to the mode in question.

For example, the figure 7 represents an electronic diagram of an assembly comprising three servocontrol circuits CA ₁ , CA ₃ and CA ₀ . These circuits are dedicated to three modes of operation which may be the modes known as "3 ramps", "1 ramp" and "almost empty" as described above. As indicated on the figure 7 each servocontrol electronic circuit comprises an operational transconductance amplifier known as "OTA", a C _ACT activation control and a Cl _INT integration circuit arranged in series. The activation commands select the servo circuit corresponding to the selected operating mode. Each integration circuit comprises an integration capacitor noted C _INT and a resistance noted R _ZERO arranged in series. It can be shown that the time constants can be the same for the three servo loops, only the gains noted G _M transconductance amplifiers changing in the M mode to optimize the bandwidth and stability of the loop. A simple formula gives the approximate value of the gain of the amplifiers to be corrected by simulation or experimentation for the idle mode where the losses of the converter are totally preponderant. At power up, we can force the mode to empty, the time that the voltage V _{HT is} established. In case of overvoltage, the other modes are forbidden. When switching from one mode to another, the duty cycle is immediately switched to the good value previously delivered by the servo.

It is known that in order to adequately balance the currents flowing in the diodes of the ramps, it is necessary to use voltage sources C _{i that are} independent of the voltage. The analog diagram of the figure 7 includes also a servocontrol circuit CA _c current sources comprising transistors Ti. The collector-emitter voltage V _CE of the transistors must not exceed a few volts, sufficient to withstand the disparity of voltages between the different ramps R _i . The adjustment of the collector voltage is done by retouching the voltage V _{HT set point} with the aid of an additional loop whose speed is of little importance insofar as the phenomena having the most influence on the Direct voltage diodes are temperature variations necessarily slow. Also, only an integrable action is useful for this loop.

It is possible to imagine other variants of regulation schemes, it is possible in particular to adapt the principle of the device according to the invention to digital regulation. In the operating modes in which one or more diode ramps are activated, the converter can be directly controlled by measuring the voltages of the collectors of the transistors without being concerned with the voltage V _HT . In this type of architecture, in almost no-load mode, except for the start-up phase of the power supply, the voltage V _HT must be enslaved to the value retouched in the other modes. The important thing is that the VHT voltage does not change significantly when changing mode to ripple on the C _HT capacitor near (see figure 2 ).

As has been said, the high voltage is obtained by means of a DCM DCL boost converter. The figure 6 represents the electronic diagram of such a boost converter circuit. As indicated in this figure, it essentially comprises an inductance L _DCM arranged in series with a diode D _DCM , a charge capacitor C _DCM , a control transistor M _DCM driven by a PWM type signal in the form of a time slot. It further comprises a network composed of a resistance R _SNUB and a capacity C _SNUB to dissipate the residual energy at the end of the cycle and a diode D _SNUB if necessary.

• L'amplitude crête du courant dans l'inductance est élevée, le double du courant moyen ;
• L'évacuation de l'énergie restante dans l'inductance avant d'amorcer un nouveau cycle nécessite un circuit d'amortissement dit « damping » de type RC qui rajoute des pertes.

One of the essential characteristics of the device of the invention is the use of a discontinuous cutting mode for the DCM up-converter. In this mode of cutting, the energy stored in the L _DCM inductance is completely discharged before starting a new cycle. At each cycle, the current starts from zero. This mode of cutting is very marginally used and is generally decried for the following reasons:

• The peak amplitude of the current in the inductor is high, twice the mean current;
• The evacuation of the energy remaining in the inductor before starting a new cycle requires a damping circuit called "damping" type RC which adds losses.

In the particular case of the supply of light-emitting diode booms, the use of this mode of cutting makes it possible to achieve a quite suitable efficiency, independently of the other advantages provided by this solution.

In most applications of DCM converters, virtually no diode is arranged in series with the control MOS M _DCM . The absence of the diode D _SNUB then allows this transistor to lead back by its so-called "body diode" several times during the dead time of the cycle, which is not favorable for the dissipation of the residual energy. Despite the small additional loss that it causes, the diode D _SNUB allows damping or "damping" more energetic and reproducible. At each beginning of the cycle, the current is zero in the inductor, which happens during one cycle therefore has no impact on the next. The advantage is the cycle-to-cycle independence behavior obtained. Thus in a single cycle, a desired mode of operation can be achieved. This is essential for the device according to the invention because this gives a perfect switching from a charging mode to another mode without the voltage VHT is disturbed.

In addition, a property of the DCM mode has particularly advantageous practical consequences. Its index response for example to a setpoint variation is very close to a first order. This is exactly true in small signals. The internal current control loop of the boost converter control ICs which is usually mandatory with a DCM converter is no longer necessary in the case of DCM. Indeed, in the case of a CCM converter, if there was not the internal loop of current, the open loop transmittance of the power stage would be second order and stability would be extremely difficult to obtain.

Regulation as described on the figures 6 and 7 is digitally feasible within a programmable logic circuit of the FPGA type. This type of component is a priori already available to control the display micro-display illuminated by the light-emitting diodes of the lighting circuit. The "FPGA" has analog resources including an analog-to-digital converter or "ADC". This converter usually works on 12 bits.

$${\mathrm{C}}_{\mathrm{power}}\mathrm{\sim }{\mathrm{VHT\; C}}_{\mathrm{HT}}\mathrm{/}\mathrm{Iout}$$

$${\mathrm{F}}_{\mathrm{power}}\mathrm{\sim }\mathrm{Iout}\mathrm{/}\left(\mathrm{2}\mathrm{*}\mathrm{\pi \; VHT}\mathrm{*}{\mathrm{C}}_{\mathrm{HT}}\right)$$

In small signals, the power stage of the DCM converter has a first-order characteristic with a gain denoted G _power and a time constant denoted C _power to which corresponds a cutoff frequency denoted F _power whose values can be determined by the following approximate formulas, neglecting the losses: $${\mathrm{BOY\; WUT}}_{\mathrm{power}}\mathrm{=}\mathrm{DVHT}\mathrm{/}\mathrm{dTon}\mathrm{~}\mathrm{Wine}\mathrm{*}{\left(\mathrm{2}\mathrm{*}\mathrm{VHT}\mathrm{/}\mathrm{The}\mathrm{1}\mathrm{*}\mathrm{Iout\; To}\right)}^{\mathrm{1}\mathrm{/}\mathrm{2}}$$

$${\mathrm{VS}}_{\mathrm{power}}\mathrm{~}{\mathrm{VHT\; C}}_{\mathrm{HT}}\mathrm{/}\mathrm{lout}$$

$${\mathrm{F}}_{\mathrm{power}}\mathrm{~}\mathrm{lout}\mathrm{/}\left(\mathrm{2}\mathrm{*}\mathrm{\pi \; VHT}\mathrm{*}{\mathrm{VS}}_{\mathrm{HT}}\right)$$

• To : période de découpage ;
• Ton : durée de conduction du transistor M_DCM;
• Vin : tension d'entrée sur L_DCM ;
• Iout : courant consommé sur VHT.

With

• To: cutting period;
• Tone: conduction duration of the transistor M _DCM ;
• Vin: input voltage on the _DCM ;
• Iout: current consumed on VHT.

In large signals, despite the non-linear character of the transfer function, a function of the square of Vin * Ton, the response of the DCM converter remains aperiodic and similar to a first order. The approximate formulas have sufficient precision to parameterize the regulation, but it is possible, by the numerical simulation of the electronic circuit to determine values closer to reality. As an example, the "SPICE" simulation software can be used.

In figure 8 when an operational mode is shown, the entire supply and servo chain of the lighting device with the different gains is shown, G1 being the gain of the bridge of resistance of the servo-control device, G _M being the gain of the transconductance amplifier, G _C1 the gain of the integration circuit, G2 the gain of the PWM signal generating circuit and finally G _power being the gain of the converter. In figure 9 the different gains of this chain are represented as a function of frequency.

si Chf est petit devant Cint $$\mathrm{BP}=\mathrm{Fdécoupage}/\left(\mathrm{marge}1*\mathrm{marge}2\right)$$

$$\mathrm{Fzéro}=\mathrm{Fdécoupage}/\left(\mathrm{marge}1*\mathrm{marge}2*\mathrm{marge}3\right)=1/\left(2*\mathrm{\pi \; Rzéro}*\mathrm{Cint}\right)$$

$$\mathrm{Atténuation\; en\; fréquence\; à\; BP}=\mathrm{produit\; des\; gains\; statiques}$$

$$\mathrm{BP}\mathrm{/}\mathrm{Fpower}\mathrm{=}\mathrm{G}\mathrm{1}\mathrm{*}{\mathrm{G}}_{\mathrm{M}}\mathrm{*}\mathrm{Rzéro}\mathrm{*}\mathrm{G}\mathrm{2}\mathrm{*}{\mathrm{G}}_{\mathrm{power}}$$

Various frequency margins are needed for the stability of the loop. First, it is necessary that its gain is very low at the frequency of cutting of the duty cycle. This is even more essential in the case of digital regulation, firstly because the sampling frequency of the signals does not exceed this value and secondly to minimize the temporal micro-variations of the clipping ratio or "Jitter" and make effective the function of "electronic screening" better known under the Anglo-Saxon term of "dithering" final. The last filtering can be more energetic than a simple first order. Margins 2 and 3 must be sufficient to overcome the variations of gain, they must not be less than an octave. $$\mathrm{HF\; Fole}=\mathrm{Fdécoupage}/\mathrm{<figure-callout\; id="1"\; label="margin"\; filenames="imgf0004.png,imgf0007.png"\; state="\{\{state\}\}">margin</figure-callout>}1~1/\left(2*\mathrm{\pi \; RZero}*\mathrm{Chf}\right)$$

if Chf is small in front of Cint $$\mathrm{BP}=\mathrm{Fdécoupage}/\left(\mathrm{margin}1*\mathrm{margin}2\right)$$

$$\mathrm{fzero}=\mathrm{Fdécoupage}/\left(\mathrm{margin}1*\mathrm{margin}2*\mathrm{margin}3\right)=1/\left(2*\mathrm{\pi \; RZero}*\mathrm{Cint}\right)$$

$$\mathrm{Frequency\; attenuation\; at\; BP}=\mathrm{produces\; static\; gains}$$

$$\mathrm{BP}\mathrm{/}\mathrm{fpower}\mathrm{=}\mathrm{BOY\; WUT}\mathrm{1}\mathrm{*}{\mathrm{BOY\; WUT}}_{\mathrm{M}}\mathrm{*}\mathrm{RZERO}\mathrm{*}\mathrm{BOY\; WUT}\mathrm{2}\mathrm{*}{\mathrm{BOY\; WUT}}_{\mathrm{power}}$$

The gain G _{M is} deduced as a function of the operating point or the charging mode.

The figure 10 represents an exemplary digital embodiment of a device for controlling the control voltages according to the invention. It essentially comprises two digital assemblies 210 and 220. The first set 210 calculates the error signal on the voltage levels V _HT and V _CE . The second set 220 is an integrator that drives the control signal generator of the upconverter circuit not shown in this figure.

The first set 210 comprises a first multiplexer 211, an analog-digital converter 212, a reference channel comprising the initial voltage setpoints 213 and two multiplexers 214, a comparator 215 making it possible to compare the values of the voltage setpoints with the measured values in order to deduce the error signal therefrom. This chain also comprises a security comparator 216.

The integrator assembly 220 comprises a state machine 221 which controls the various charging modes and the transition from one mode to another. This machine is dependent on the brightness settings of the micro-display, the video vertical sync signal and the initialization circuit. Unlike the analog schema, there are not as many integrators as modes but only one with backup and recall of integral values for each mode.

Fsw is the converter switching frequency as well as the sampling frequency of the ADC converter. Given the large ratio between the clock frequency Clk of the digital part and the switching frequency Fsw, certain gain or multiplication functions can be performed sequentially with a single adder. This adds at least Fsw latency for these blocks.

High frequency filtering can be done with a simple recursive filter 222 equivalent to an analog low pass filter. Although the ratio between the clock frequency Clk and the switching frequency Fsw is large enough, to obtain a mean temporal resolution that is finer than the clock period Clk, it is possible to add a temporal dithering device which simply consists of to report the rounding error on the next cycle. The effectiveness of dithering circuit is all the better that the signal has been filtered beforehand.

Since the digital set of the error signal uses only one ADC converter, its electronic scheme is a little different from its analogue counterpart.

As before, the start is performed in so-called "almost empty" mode with an initial VHT setpoint by default. As soon as one is in operational operation, in modes 2 or 3, the servocontrol is done directly by measuring the voltage of the collectors of the transistors of the current sources. This servocontrol will cause the voltage VHT to be different from the default value. It is necessary in mode 1 to maintain the voltage VHT at this same value. The new setpoint is obtained by storing the value of VHT when exiting mode 2 or mode 3.

During night use, the required brightness is very low. Thus, the converter is in mode 1 most of the time. The power consumed in this mode is essentially that of the losses of the converter circuit or "booster". In this condition of use, it may be requested that the micro-display operates on battery in certain circumstances. Thus, when the imager illuminated by the lighting device belongs to a helmet visual worn by a pilot, it is possible that this pilot needs to use his helmet outside his aircraft. To minimize consumption in this low light mode, it is better to replace mode 1 with a current zero mode such that the booster circuit is stopped with the load of the VHT measurement bridge disconnected using a MOS type transistor. Mode 1 still keeps its usefulness at power-up because it allows a naturally progressive start that does not require an auxiliary circuit called "soft-start".

Claims (6)

1. A lighting device with light-emitting diodes, said light-emitting diodes being arranged in a first plurality of ramps that are disposed in parallel (R_i), each ramp comprising a second plurality of light-emitting diodes (D_j) that are disposed in series, said ramps being supplied with a DC voltage (V_HT) of several tens of volts, referred to as high voltage, said voltage being generated by a step-up converter circuit (DCM) from a low DC voltage of several volts, the value of said voltage being controlled by the duty factor of said step-up converter circuit, said voltage being servo-controlled at a constant average value by means of a substantially analogue servo-control device (D.A.) controlling said duty factor, said servo-control device comprising a plurality of operating modes, a mode being defined either by specific electronic addressing of the high voltage or by a defined number of ramps of illuminated diodes, characterised in that said step-up converter circuit has discontinuous conduction and said servo-control device comprises a plurality of electronic servo-control circuits (C.A._k) that are connected to an electronic multiplexer (MUX), each electronic servo-control circuit being dedicated to a specific operating mode, the electronic characteristics of said electronic servo-control circuits depending on said operating mode, said electronic servo-control circuit being operational only when the operating mode is selected.
2. The lighting device with light-emitting diodes according to claim 1, characterised in that said servo-control device comprises means for storing the different duty factors that are dedicated to each operating mode.
3. The lighting device with light-emitting diodes according to claim 1, characterised in that, when said servo-control device is realised in an analogue manner, each electronic servo-control circuit comprises an operational transconductance amplifier, designated OTA, an activation command and an integration circuit that are connected in series.
4. The lighting device with light-emitting diodes according to claim 3, characterised in that the gain of said operational transconductance amplifier of each electronic servo-control circuit depends on the operating mode to which said electronic servo-control circuit is dedicated and in that the different integration circuits of the different electronic circuits are all identical.
5. The lighting device with light-emitting diodes according to any one of the preceding claims, characterised in that the first plurality of ramps is structured in a first number N of units of ramps, each unit comprising a second number M of ramps, the lighting of the diodes composing the ramps being controlled in a matrix manner by two control circuits that are also referred to as "dimming" circuits, the first circuit comprising N first control means, each first control means allowing the simultaneous control of only one ramp from all of the ramp units, the second circuit comprising M second control means, each second control means allowing the simultaneous control of all of the ramps of only one unit.
6. The lighting device with light-emitting diodes, said light-emitting diodes being arranged in a first plurality of ramps (R_i) that are disposed in parallel, each ramp comprising a second plurality of light-emitting diodes (D_j) that are disposed in series, said ramps being supplied with a DC voltage (V_HT) of several tens of volts, referred to as high voltage, said voltage being generated by a step-up converter circuit from a low DC voltage of several volts, the value of said voltage being controlled by the duty factor of said step-up converter circuit, characterised in that said voltage is servo-controlled at a constant average value by means of a substantially digital servo-control device (210, 220) controlling said duty factor, said servo-control device comprising a plurality of operating modes, a mode being defined either by specific electronic addressing of the high voltage or by a defined number of ramps of illuminated diodes, said step-up converter circuit (DCM) "having discontinuous conduction", each of said operating modes corresponding to a servo-control circuit that controls the duty factor of said step-up converter circuit, the value of the duty factor of the on-going operating mode being stored before each change in said on-going mode.

Images