Title of Invention

ENERGY METER ARRAY AND METHOD FOR CALIBRATION

Abstract

The invention relates to an Energy meter array, having a first input (1) for feeding a signal derived from a voltage (V), to which a first analogous/digital-converter (3) is connected, that has an output; a second input (2) for feeding a signal derived from a current (1), to which a second analogous/digital-converter (4) is connected, that has an output; a multiplier (7) that connects the output of both analogous/digital-converters (3,4) to one another; a phase evaluation block (9) with two outputs that are coupled with the outputs of both the analogous-/digital-converters (3,4) and with an output that is coupled with a control input of a correction block (6); the phase correction block (6) that is coupled to an output of the second analogous-/digital-converter (4), designed for correction of a phase fluctuation (∆ϕ) of the digitized signal derived from a current (1) or a voltage (V); whereby the phase correction block (6) consists of a first digital filter and a second digital filter (5) that is switched between the output of the first analogous-/digital converter (3) and the multiplier (7); and means for sample rate control (19) that is respectively coupled with a control input of the phase correction block (6) and the second digital filter (5).

Full Text

Description Energy Meter Array and Method for Calibration The invention pertains to an energy meter array and method for calibrating the energy meter array. Energy meters serve the purpose of measuring consumed or generated electrical energy. Such energy meters are referred to as electricity meters or kilowatt hour meters. In electronic energy meters, generally voltage and current are measured and digitised and multiplied with one another. After multiplication one gets the momentary electrical power. If one integrates or accumulates this electrical power over time, then one obtains a signal that is a measure for the electrical energy generated or consumed in a certain time interval. In order to obtain proportional signals for electrical voltage and electrical current, voltage dividers, voltage converters, current converters or other agents can be used for signal uncoupling. In many applications it is necessary, at least in one of the two channels, to provide a galvanic separation for measuring voltage and current. Such a galvanic separation of current circuits is provided, for example, by a transformer. The problem with such a transformer is however the phase shift caused by the inductive coupling of the transformer. The phase shift occurs, on the one hand, between the output signal and the input signal of the transformer. On the other hand, the phase shift also occurs between the signal representing the current and the signal representing the voltage. This however results in undesired measuring errors during the multiplication of voltage and current. It should be noted here that voltage and current are mostly present not as direct signals but more as alternate current signals with more or less harmonic signal form. The described problem gets additionally intensified, in that even if a transformatory transmitter is used in the voltage and current measuring channel, an exactly predictable phase shift does not happen between both input channels due to production tolerances, temperature effects, aging effects or other unavoidable effects of a mass production. For correcting the described, undesirable phase shift one can for example use RC-networks that consist of resistances and capacitors. However, these have to be generally designed as additional external components and can generally not be integrated. Moreover, the problem of tolerances related to production and temperature are basically not solved thereby. It is the task of this invention to create an energy counter array that can be integrated with least complication, as well as a method for calibration, in such a way that voltage and/or the current can be measured with galvanic separation without any measuring errors. This task is fulfilled according to the invention with the help of an energy meter array that has a first input for feeding a signal derived from a voltage, to which a first analogous/digital converter is connected that has an output; a second input for feeding a signal derived from a current, to which a second analogous/digital converter is connected that has an output; a multiplier that connects the output of both analogous/digital converters with one another; a phase evaluation block with two inputs that are coupled with the output of both the analogous/digital converters, and with an output that is coupled with a control input with a phase correction block; and - the phase correction block that is coupled on to an output of one of both the analogous/digital converters, designed for correcting a phase fluctuation of the digitised signal derived from a current or voltage. According to the suggested principle, a phase shift between the input of the energy meter array, to which a signal derived from an electrical voltage is fed, and the input of the energy meter array to which a signal derived from an electrical current is fed, is measured and compensated. The signal input for feeding the signal derived from the voltage and the signal derived from the current could also be referred to as input channel, namely voltage channel and current channel. The phase evaluation and correction is carried out completely in digital signal processing. With the suggested measurement and compensation of the phase shift between both channels it is advantageously possible to galvanically isolate the channels from one another and/or at least one input from the energy meter array. By balancing the phase shift, measuring errors in the array are avoided. The advantage here is that one does not need any external compensation networks like resistance-capacity networks for the phase shift. The phase evaluation block and the phase correction block effect a so-called on-chip-phase correction of the energy meter array without external components. For example, in a calibration operation mode, the phase shift between both input channels can be measured, in that an identical input signal is placed on both inputs. Preferably a sine signal is fed to both inputs of the energy meter array in the calibration operation mode. The advantage here is that the zero-transmissions of both digitised signals can be compared with the phase evaluation block. Alternatively, it is also possible to evaluate the respective phase positions of the signal peak values of the digitised signal. Thus it is possible to define the relative time interval of the zero-transmissions from one another. The calculation of the phase difference from the time interval of the zero- transmissions of both signals can effected by means of a logic unit in the phase evaluation block. With the measured phase difference it is subsequently possible without any problem to carry out a correction of exactly this phase difference in one of both the channels. The phase correction value can be stored in the phase evaluation block, so that the correction value is available in normal operation even after the calibration operation mode. As no external components are required for phase correction, implementation of the suggested energy meter array is possible in a very cost-effective manner. Therefore, the suggested, integrate-able energy meter array is particularly suited for mass production. An additional advantage is that the time required for calibrating the energy meter array is very little according to the suggested principle. In principle, the same difference between both input channels can be measured within only one period duration of the input signal, preferably the test signal. The period duration can be calculated in a very simple manner from the inverse proportion of the respective signal frequency. The signal frequency in energy meters is generally 50 Hertz or 60 Hertz, depending on the national standards. According to the suggested principle, a phase fluctuation is determined between the input signals of the energy meter array, already available as digital signals. The correction of the phase error is similarly done in the area of digital signal processing. In order to achieve particularly quick measurement of the phase fluctuation of the input channel of the energy meter array, it is advantageous to connect the pulse input of the phase evaluation block with the pulse input of the analogous-/digital-converters and thus use for the phase evaluation, the pulse signal of the analogous-/digital-converters that is required in any case for operating the energy meter array. The phase evaluation block comprises an agent for permanent storing of a phase correction value. The agent for permanent storage of a phase correction value is preferably designed as non-volatile storage, e.g. as EEPROM. Due to the permanent storage of the phase correction value, the phase correction value measured in a calibration operation mode is also available after the energy meter array is switched off and switched on again. The analogous-/digital-converters are preferably designed as sigma-delta-converters or as sigma-delta-modulators. This allows scanning of the input signals derived from voltage and current with high resolution and good integrate-ability. An integrator is foreseen at the output of the multiplier, which integrates the signals provided by the multiplier. The integrator can be designed as accumulator. The integrator is designed in such a way that it integrates the signal provided by the multiplier, representing the momentary electrical power, with a signal that is a measure for the electrical energy consumed or generated. Furthermore, the first and the second analogous-/digital-converters, the phase correction block and the phase evaluation block, are designed in integrated circuit technology. Even the sigma-delta-converters and further functional blocks and/or components in the signal processing chain of the energy meter array can be designed, as far as possible, preferably in integrated circuit technology. The energy meter array can be implemented in a single integrated semi-conductor circuit. To the first input of the energy meter array and/or to the second input of the energy meter array, the output of a transmitter effecting a galvanic separation can be connected. Such a non-galvanic transmitter could ideally be a transformer. The relatively high phase fluctuation of such coupled members effecting a galvanic separation of the input can be compensated with the help of the suggested principle in a very simple, effective and highly precise manner. An agent for generating a test signal is foreseen, that is coupled with the first and the second input of the energy meter array. If launching members are foreseen, for example transformatory transmitters, then the agent for generating the test signal is preferably designed in such a way that the test signal is fed in at the input of the transmitter or launching member. It could be thereby advantageous if a switch-over possibility of the input between an effective signal operation mode and a calibration operation mode is foreseen, in which the inputs are connected to the agent for generating the test signal. A digital filter is connected after the analogous-/digital-converters. The digital filters have a control input each for controlling the scan rate of the digitised signal. Particularly, between the normal operation mode and the calibration operation mode, an alteration of the scan rate can be carried out. Ideally, the phase correction block can include one of the digital filters. The agent for generating the test signal can be activate-able in a calibration/operation mode, while in the normal operation mode, i.e. in the actual energy measuring operation it can be deactivated. With respect to the method, the task of the invention is fulfilled by means of a method for calibrating an energy meter array with the help of following steps: - Launching of a test signal at two inputs of an energy meter array; - digitisation of the test signal at both inputs; - measuring of a phase fluctuation between both the digitised test signals; generation of a phase correction signal and feeding the phase correction signal to one of both the digitised test signals. Measurement of the phase fluctuation can take place, for example, by comparing the phase positions of the signal-zero-transmissions of both signals. Alternatively, the phase fluctuations can be measured by comparing the phase positions of the signal peak values with one another. Determination of the phase position as well as measurement of the phase fluctuation can thereby be done completely in digital signal processing. Even other methods can be used for measuring a phase fluctuation between two digitised signals that can be implemented in digital signal technology. One can measure the phase fluctuations in a particularly precise manner, in that in the calibration operation mode, the scan rate of the digitised signals is influenced. For example, in the calibration operation mode the sigma-delta-over scan rate can be reduced as there is a low dynamics range. One gets a greater number of scan values in a 50-Hertz- pulse period, so that the precision of measuring the phase fluctuation is increased. After digitisation, a digital filtering is foreseen with adjustable decimation rate. In all, the suggested principle provides the advantage of a significant cost reduction, as no external components are required for a phase correction. Moreover, the calibration can be carried out in a particularly short time, even within a 50-Hertz- or 60-Hertz-pulse period. Further details and advantageous extensions of the suggested principle are subject of the sub- claims. The invention is explained below on the basis of a design example and the accompanying figure. The following are shown: The figure - A block circuit diagram of a desired example of the suggested energy meter array. The figure shows an energy meter array with a first input 1 and a second input 2. The first input 1 is designed for feeding a signal derived from an electrical voltage V. The second input 2 is designed for feeding a signal derived from an electrical current I. The electrical voltage V and the electrical current I refer to the same signal. To the first input 1, the input of a first analogous-/digital-converter 3 is connected. To the second input 2 the input of a second analogous-/digital-converter 4 is connected. The analogous-/digital-converters 3, 4 are respectively designed as sigma-delta-modulators. The output of the first analogous-/digital- converter 3 is connected to the input of a multiplier 7 through a first digital filter 5. The output of the second analogous-/digital-converter 4 is connected to a further input of the multiplier 7 through a second digital filer. The second digital filter consists of a phase correction block 6 that has a control input. To the output of a multiplier 7, an integrator 8 is connected that converts a signal at its input, which is a measure for the momentary electrical power P, into a signal that represents the electrical energy E. Additionally, a phase evaluation block 9 is foreseen. The phase evaluation block 9 has a phase position detector 10 for determining the respective phase position of the signal peak values. The phase position detector 10 has two inputs that are connected to the outputs of a digital filter 5 and of the phase correction block 6. Two outputs of the phase position detector 10 are connected to two inputs of a phase difference detector 11 that serves the purpose of measuring phase fluctuation. The phase difference detector has an output that is several bits wide. The output of the different phase detector 11 is connected to the input of a control block 12 that is connected to the control input of the same for activating the phase correction block 6. The control block 12 consists of a non-volatile storage 18, in which the measured phase fluctuation or the corresponding correction value can be permanently stored. A scan rate control 19 has two outputs that are respectively connected to control inputs of the phase correction block 6 with digital filter and the second digital filter 5. In this way, the scan rate of the second digital filter 5 and of the phase correction block 6 with digital filter can be laid down. The adjustment of the scan rate takes place in relationship to the respective operation mode. The energy meter arrangement described so far is arranged in integrated technology on a single chip. For uncoupling the electrical voltage, a voltage divider 14 is foreseen, whose output is connected to the first input 1 of the energy meter array and whose input forms a voltage input 13 for feeding the electrical voltage. For uncoupling the electrical current, a transformer 16 is foreseen, which is switched between a current input 15 and a second input 2 of the energy meter array. The transformer 16 provides a galvanic de-coupling between the current input 15 and the second input 2. To the voltage input 13 and the current input 15, the output of a test signal generator 17 is connected. The test signal generator 17 provides a harmonic, mainly sine-shaped signal with a rated frequency of 50 or 60 Hertz, depending on the country specification. On account of the voltage divider 14 and the transformer 16, for both input channels of the energy meter one obtains different phase shifts. Particularly important is a relative phase difference (Acp) between both the input channels at the inputs 1, 2 of the energy meter array. This phase fluctuation (Acp) is measuredwith the phase evaluation block 9. Measuring of the phase fluctuation (Acp) takes place in a calibration operation mode, in that the test signal generator 17 is activated and thereafter a phase-in, sine-shaped signal is fed to the voltage input 13 and the current input 15. This signal experiences a different phase shift in the voltage divider 14 and the transformer 16. The relative phase fluctuation (Acp) at the inputs 1,2 is measured in the phase evaluation block 9 in the detectors 10, 11, in that the time span between the zero-transmissions or peak values of both signals at the inputs of the phase evaluation block 9 are measured and converted into a corresponding phase shift. A corresponding correction value is provided by the control block 12 and given out at the output of the phase evaluation block 9. Thus a phase correction block 6 is activated in the digital filter that balances the phase difference (Acp). As measurement of the current in the normal operation mode has to take place precisely over a large dynamics range, the over-scan rate of the sigma-delta-modulators is relatively high. In this way, also a suitable signal-noise ratio can be achieved. In the calibration operation mode, in which a harmonic test signal is fed, the requirement of a large dynamics range is not relevant. Rather, peaks or zero-transmissions of the digitised test signal has to be measured and compared in both channels, in order to be able to measure a phase fluctuation. On account of the fact that in the calibration operation mode a lesser dynamics range is required, the over-scan rate of the modulator can be reduced. In this way, more scan values are available within one pulse period. The precision of identification of the phase fluctuation is determined by the frequency of the scan pulse. For a 50 Hz test signal, an over-scan rate of 16 and a scan frequency of 28 KHz, one obtains 559 scan values in one pulse period of the test signal. A scan value thereby corresponds to 0.64 degrees, namely the quotient of 360 and 559. For a phase shift of 0.64 degrees between voltage channel and current channel the measured, electrical power 0.00054 dB would be below the optimum. For an power factor 1 this corresponds to a relative error of 0.00624%, which is negligible. The precision of the phase correction can be further increased by a suitable correction of the scan values. The power factor, in conformity with general convention, is defined in such a way that for a power factor 1 current and voltage are in-phase, i.e. between current and voltage there is a phase shift of 0 degrees. In case of phase fluctuation between current and voltage of say 60 degrees, the power factor will accordingly be 0.5 and thus correspond to the cosine of the phase difference. Accordingly, the relative error for a power factor of 1 and a phase shift of 0.64 degrees can be calculated according to the equation = 0.00624%. The calibration operation mode is activated only once during the production of the energy meter array. With the suggested principle, an automatic phase correction for integrated energy meter arrays is provided, which is realised completely integrated and completely in digital signal processing. A special feature of the suggested energy meter arrangement is also the low calibration time and the low cost of manufacture. Moreover, a galvanic isolation of at least input of the energy meter array is possible without any measuring errors . The galvanic isolation is also of great importance, if more than one channel is measured, as is generally the case in electrical energy meters. Especially, a phase shift that is unavoidably caused by transformers can be compensated. For this no additional external components like resistance-capacitor aer required for phase correction. Due to the non-volatile storage 18 the phase correction value is still available even if the energy meter is switched off. As measurement of phase fluctuation is basically possible within a period duration, one can achieve a particularly quick calibration with the help of the suggested principle. The phase evaluation block 9 has a pulse input that is connected to the pulse inputs of the sigma-delta-modulators 3, 4. In this way, the pulse flanks between two zero-transmissions can be counted and thus the phase fluctuations can be precisely measured in a simple manner. According to the suggested principle, in the digital signal range, measurement of the relative phase fluctuation on both the outputs of the analogous/digital-converters is carried out. Correction of the phase fluctuation similarly takes place in digital signal processing. List reference signs 1. Input 2. Input 3. Analogous-/digital-converter 4. Analogous-/digital-converter 5. Digital filter 6. Digital filter with phase correction 7. Multiplier 8. Integrator 9. Phase evaluation block 10. Phase position detector 11. Phase difference detector 12. Control block 13. Voltage input 14. Voltage divider 15. Current input 16. Transformer 17. Test signal generator 18. Non-volatile storage 19. Scan rate control I Current E Energy P Power V Voltage ∆φ Phase fluctuation WE CLAIM : 1. Energy meter array, having - a first input (1) for feeding a signal derived from a voltage (V), to which a first analogous/digital-converter (3) is connected, that has an output; - a second input (2) for feeding a signal derived from a current (1), to which a second analogous/digital-converter (4) is connected, that has an output; - a multiplier (7) that connects the output of both analogous/digital-converters (3,4) to one another; - a phase evaluation block (9) with two outputs that are coupled with the outputs of both the analogous-/digital-converters (3,4) and with an output that is coupled with a control input of a correction block (6); - the phase correction block (6) that is coupled to an output of the second analogous-/digital-converter (4), designed for correction of a phase fluctuation (∆ϕ) of the digitized signal derived from a current (1) or a voltage (V); - whereby the phase correction block (6) consists of a first digital filter and a second digital filter (5) that is switched between the output of the first analogous-/digital converter (3) and the multiplier (7); and - means for sample rate control (19) that is respectively coupled with a control input of the phase correction block (6) and the second digital filter (5). 2. Energy meter array as claimed in claim 1, wherein said means for sample rate control (19) is respectively coupled with a control input of the phase correction block (6) and the second digital filter (5) and is designed for reducing a over- scan rate in a calibration operation mode with respect to a normal operation mode, in order to increase the precision of calculation of the phase fluctuation. 3. Energy meter array as claimed in claim 1 or 2, wherein the phase evaluation block (9) consists of a control block (12) for activating the phase correction block (6) in relationship to the phase deviation (∆ϕ). 4. Energy meter array as claimed in claim 3, wherein the control block (12) consists of an agent for permanent storage of a phase correction value (18). 5. Energy meter array as claimed in claimed 3 or 4, wherein the phase evaluation block (9) consists of a phase difference detector (11) with two inputs that are coupled with the outputs of both the analogous/digital-converters (3,4) and with an output that is connected to the control block (12). 6. Energy meter array as claimed in claim 5, wherein the phase evaluation block (9) consists of a phase detector (10) that is coupled between the outputs and inputs of the phase difference detector (11). 7. Energy meter array as claimed in claim 6, wherein the phase position detector (10) is designed for measuring signal peak values. 8. Energy meter array as claimed in claim 6, wherein the phase position detector (10) is designed for measuring signal zero transmissions. 9. Energy meter array as claimed in claims 1 to 8, wherein the first and second analogous-/digital-converters 3,4) are respectively designed as sigma-delta-converters. 10. Energy meter arrangement as claimed in one of claims 1 to 9, wherein an integrator (8) is foreseen, which is connected after the multiplier (7). 11. Energy meter array as claimed in one of claims 1 to 10, wherein the first and the second analogous-/digital- converters (3,4), the phase correction block (6) and the phase evaluation block (9) are designed in integrated circuit technology. 12. Energy meter array as claimed in one of claims 1 to 11, wherein at the first input (1) and/or at the second input (2) a non-galvanic coupling transmitter (16) is connected for launching the signal derived from the voltage (V) and/or from a current (1). 13. Energy meter array as claimed in claim 12, wherein the non- galvanic coupling transmitter (16) is designed as transformer. 14. Energy meter array as claimed in one of claims 1 to 13, wherein means for generating a test signal (17) is foreseen that is coupled with the first and the second inputs (1,2) of the energy meter array for feeding the test signal in a calibration operation mode. 15. Method for calibrating an energy meter array comprising the steps of: - coupling a test signal at two inputs (1,2) of an energy meter array; - digitizing the test signal coupled to the first input to generate a first digitized test signal with a reduced oversampling rate compared to normal operation; - digitizing the test signal coupled to the second input to generate a second digitized test signal with the reduced oversampling rate compared to normal operation; - measuring a phase deviation between the first and second digitized test signals (∆ϕ); - generating a phase correction signal; and - applying said phase correction signal to one or more of the first and second digitized test signals. 16. Method as claimed in claim 15, wherein measuring the phase deviation comprises measuring signal peak values of the first and second digitized test signals to establish the phase deviation. 17. Method as claimed in claim 15, wherein measuring signal zero points of the first and second digitized test signals to establish the phase deviation. 18. Method as claimed in claims 15 to 17, wherein said method further comprises the step of digitally filtering the first and second digitized test signals prior to determining the phase deviation. 19. Method as claimed in claim 18, comprising setting a sampling rate of the filtering of the first and second digitized test signals. 20. Method as claimed in claims 15 to 19, wherein coupling the test signal to the first and second inputs comprises inductively coupling the test signal to at least one of the first and second inputs of the energy meter device. ABSTRACT TITLE "ENERGY METER ARRAY AND METHOD FOR CALIBRATION" The invention relates to an Energy meter array, having a first input (1) for feeding a signal derived from a voltage (V), to which a first analogous/digital-converter (3) is connected, that has an output; a second input (2) for feeding a signal derived from a current (1), to which a second analogous/digital-converter (4) is connected, that has an output; a multiplier (7) that connects the output of both analogous/digital-converters (3,4) to one another; a phase evaluation block (9) with two outputs that are coupled with the outputs of both the analogous-/digital-converters (3,4) and with an output that is coupled with a control input of a correction block (6); the phase correction block (6) that is coupled to an output of the second analogous-/digital-converter (4), designed for correction of a phase fluctuation (∆ϕ) of the digitized signal derived from a current (1) or a voltage (V); whereby the phase correction block (6) consists of a first digital filter and a second digital filter (5) that is switched between the output of the first analogous-/digital converter (3) and the multiplier (7); and means for sample rate control (19) that is respectively coupled with a control input of the phase correction block (6) and the second digital filter (5).

Documents:

02784-kolnp-2006-abstract.pdf

02784-kolnp-2006-claims.pdf

02784-kolnp-2006-correspondence others-1.1.pdf

02784-kolnp-2006-correspondence others.pdf

02784-kolnp-2006-correspondence-1.2.pdf

02784-kolnp-2006-correspondence-1.3.pdf

02784-kolnp-2006-description(complete).pdf

02784-kolnp-2006-drawings.pdf

02784-kolnp-2006-form-1.pdf

02784-kolnp-2006-form-2.pdf

02784-kolnp-2006-form-26.pdf

02784-kolnp-2006-form-3.pdf

02784-kolnp-2006-international publication.pdf

02784-kolnp-2006-international search authority report-1.1.pdf

02784-kolnp-2006-international search authority report.pdf

02784-kolnp-2006-pct form.pdf

02784-kolnp-2006-pct others.pdf

2784-KOLNP-2006-ABSTRACT.1.1.pdf

2784-KOLNP-2006-AMANDED CLAIMS.pdf

2784-KOLNP-2006-CORRESPONDENCE 1.1.pdf

2784-KOLNP-2006-CORRESPONDENCE 1.3.pdf

2784-KOLNP-2006-CORRESPONDENCE.1.2.pdf

2784-KOLNP-2006-EXAMINATION REPORT.pdf

2784-KOLNP-2006-FORM 1.1.1.pdf

2784-KOLNP-2006-FORM 1.1.2.pdf

2784-KOLNP-2006-FORM 18.pdf

2784-KOLNP-2006-FORM 2.1.1.pdf

2784-KOLNP-2006-FORM 26.pdf

2784-KOLNP-2006-FORM 3.pdf

2784-KOLNP-2006-FORM 5.pdf

2784-KOLNP-2006-GRANTED-ABSTRACT.pdf

2784-KOLNP-2006-GRANTED-CLAIMS.pdf

2784-KOLNP-2006-GRANTED-DESCRIPTION (COMPLETE).pdf

2784-KOLNP-2006-GRANTED-DRAWINGS.pdf

2784-KOLNP-2006-GRANTED-FORM 1.pdf

2784-KOLNP-2006-GRANTED-FORM 2.pdf

2784-KOLNP-2006-GRANTED-SPECIFICATION.pdf

2784-KOLNP-2006-OTHERS 1.1.pdf

2784-KOLNP-2006-OTHERS PCT FORM.pdf

2784-KOLNP-2006-OTHERS.pdf

2784-KOLNP-2006-REPLY TO EXAMINATION REPORT 1.1.pdf

2784-KOLNP-2006-REPLY TO EXAMINATION REPORT.pdf

2784-KOLNP-2006-SCHEDUAL-FORM 3.pdf

2784-KOLNP-2006-TRANSLATED COPY OF PRIORITY DOCUMENT 1.1.pdf

2784-KOLNP-2006-TRANSLATED COPY OF PRIORITY DOCUMENT.pdf

abstract-02784-kolnp-2006.jpg