FR2951837A1 - Device for transmitting energy supply fall signal over network, has high voltage charger charging capacitor, and voltage reducer permitting auxiliary energy supply to processor outputting energy supply fall signal over network - Google Patents

Abstract

The device has a high voltage charger (2) charging a capacitor (C) during normal mode, where supply energy is greater than threshold in the normal mode. A voltage reducer (3) reduces the voltage produced from the capacitor when the capacitor is discharged during deficient mode in which supply energy is lower than threshold. The reducer permits an auxiliary energy supply to a processor (4) outputting an energy supply fall signal (FDG) over a network (1000). An independent claim is also included for a method for transmitting an energy supply fall signal over a network.

Description

GENERAL TECHNICAL FIELD The present invention relates to a device for transmitting on a network an energy supply drop signal below an energy supply threshold, comprising an energy supply drop detector below the threshold. energy supply, and a capacity forming a reserve of energy. The invention also relates to a method implemented on a device mentioned above. STATE OF THE ART The function "breath of death" or "Dying Gasp" according to the Anglo-Saxon terminology generally used by those skilled in the art, is part of the standards of the digital subscriber line asymmetric flow, or 15 "asymmetric digital subscriber line" (ADSL) and the very high-speed digital subscriber line, or "very-high-bit-rate digital subscriber line" (VDSL). As shown in FIG. 1, the Dying Gasp function, implemented on a known gateway, indicates to network 1000, via a FDG information field 20 of xDSL frames, that the telecommunication gateway is in the process of lose your energy supply. The Dying Gasp function is often required by the operators and requires means within the gateway for its implementation. The gateway comprises for this purpose a comparator 1 which makes it possible to detect the loss of input voltage (for example when the voltage is below a threshold of 10V, for a normal mode input voltage equal to 12V). The detection of the voltage drop is used to warn a main processor 4, which will position the Dying Gasp DG information in the FDG frame 30 leaving on the network 1000. In order to be able to transmit the FDG frame (the standard can also request the emission of several consecutive frames containing the information Dying Gasp), the gateway must have stored enough energy in reserve. For the ADSL standard, it is necessary to be able to maintain sufficient energy for 20 to 40 ms, and for the VDSL standard, it is necessary to be able to maintain a sufficient energy for about 200 ms. The energy reserve is generally contained in C tank capacities, which maintain for the aforementioned necessary time an auxiliary voltage greater than the minimum operating voltage of the gateway. In our example, the detection threshold voltage is 10V for a normal input voltage of 12V, the minimum operating voltage of the gateway being 6V. The energy stored in the capacitor C is therefore equal to: Ec = jCV2 = -C (10ù6) 2 = 8C (EQ1) It is understood that the stored energy is proportional to the value of the C tank capacities used.

Typically, the known gateways have two 6800μF capacities to ensure the transmission of the FDG frame. The disadvantages of these gateways are as follows. The size of the C capacities used is significant because they are usually chemical capacities. Typically, two 6800μF capacities take approximately 40x40x20mm of space congestion, which is important for telecommunication gateways. The price of the capacities is also significant, and is therefore relative to the cost of the gateway. The generally used chemical capabilities are unreliable and reduce the mean time between failures (MTBF). The way in which the capacities are loaded forces the application of a relatively large input voltage (of the order of 12V) having several negative effects: 30 - the total energy efficiency of the gateway is not optimal, because the rail 11 using the input voltage in the gateway is usually not the most energy-intensive; the efficiency of the rails 12 of low voltages (for example of the order of 1 V) of the cores of processors (not shown) is restricted, because of the large conversion ratio of the converter 13 direct current / direct current (DC- DC); - The choice of converters for the converter 13 DC-DC is also restricted, because of the requirement for the converter 13 to support the input voltage. PRESENTATION OF THE INVENTION The invention proposes to overcome at least one of these disadvantages.

For this purpose, there is provided according to the invention a device for transmitting on a network a power supply drop signal below an energy supply threshold, comprising - a power supply drop detector below the power supply threshold, - a capacity forming a power reserve, the device being characterized in that it further comprises - a voltage-boosting charger, for charging the capacity during a normal mode during which power supply is greater than the threshold; and - a voltage step-down, for, during a deficient mode during which the power supply is below the threshold, lowering a voltage from the capacitor when the capacitance discharges, and thus to allow, during a time of discharge of the capacity, an auxiliary power supply of a processor emitting on the network the signal of energy supply drop. The invention is advantageously completed by the following characteristics, taken alone or in any of their technically possible combination: the detector comprises a shunt MOSFET metal oxide semiconductor field effect transistor forming a current detector; ; the device comprises a command enabling, during the deficient mode, deactivating the non-vital elements for the emission on the network of the signal of energy supply drop by the processor; the control is adapted to deactivate all the non-vital elements simultaneously; the charger is adapted so that the charging time of the capacitor is of the order of magnitude of the establishment time of a communication on the network; and the device comprises protections against high voltages. The invention also relates to a method implemented on a device mentioned above. The invention has many advantages. The present invention makes it possible to offer the Dying Gasp function within residential telecommunication gateways, by optimizing their energy consumption, by improving their MTBF, and by reducing the space congestion of the gateways. The invention allows a greater choice of lower voltage input 15 (it is then the charger that increases the voltage for the load capacity), which allows a wider choice of components, especially for converters. Thus, a gateway according to the invention at a comparable cost, or even lower, to known gateways. The invention is not limited to the field of telecommunication gateways, but applies to any device to detect a drop in power supply and compensate for a certain time. PRESENTATION OF THE FIGURES Other characteristics, objects and advantages of the invention will emerge from the description which follows, which is purely illustrative and nonlimiting, and which should be read with reference to the appended drawings in which: FIG. discussed, schematically represents a known gateway; FIG. 2 diagrammatically represents an embodiment of a device according to the invention; - Figure 3 schematically shows a charger according to the invention; FIGS. 4 schematically represent a downcomer according to the invention, FIG. 4A being the preferred embodiment, FIG. 4B schematically showing a low cost version of the down converter; FIG. 5A schematically represents a preferred detector according to the invention; - Figure 5B shows schematically the complement of the detector 5A incorporating a function on / off (On / Off) of the device; FIG. 6 schematically represents a machine of the states implemented on the device; - Figures 7 and 8 schematically show two possible embodiments of the invention; and FIG. 9 shows the ratio between the two required capacitance values, as a function of the voltage used in the invention, at constant energy, between a conventional state of the art approach and the invention.

In all the figures, similar elements bear identical reference numerals. DETAILED DESCRIPTION FIG. 2 schematically shows a possible embodiment of a device according to the invention.

The device allows the emission on a network 1000 of a FDG signal of energy supply drop below an energy supply threshold, also called "breath of death", or "Dying Gasp" (DG) according to the Anglo-Saxon terminology generally used by those skilled in the art.

The signal FDG power supply drop is issued conventionally by a processor 4 on the network 1000, to announce to a correspondent on the network that the communication (on a line ADSL or VDSL for example) will be cut, because of the loss of energy supply.

The device mainly comprises - a detector 1 for lowering the power supply below the energy supply threshold, - a capacitor C forming a power reserve, - a voltage booster charger 2, for charging the capacitor C during a normal mode during which the power supply is above the threshold; and a voltage step-down, for, during a deficient mode during which the power supply is below the threshold, lowering a voltage coming from the capacitor C when the capacitor discharges, and thus to allow, during a discharge time of the capacitor C, an auxiliary power supply at least to the processor 4 for the transmission of the FDG signal.

The solution is based on the high voltage energy storage, also called "High Voltage Energy Storage" (HVES) by the skilled person. As shown by the following formula, Ec =! CV2, the energy Ec stored in a capacitor C being proportional to the square of the voltage V, it is more advantageous to raise the voltage across a capacitor rather than its capacitive value. , to increase the stored energy. At constant energy, the capacitive value of the reservoir capacitors used can be reduced compared to the state of the art.

For comparison, with a voltage of 100V across the capacitor C tank, and with a minimum operating voltage of 5V for the bridge, the formula gives: Ec = j ùCV 2 = 2 C (100 to 5) 2 = 4512, 5C, (EQ2) is a capacity almost 600 times smaller to store the same energy (in the prior art and (EQ1), 8C). Therefore, the gain in place for smaller capacities is important. The reduction of the capacitive value required advantageously makes it possible to use, if desired, a ceramic technology for capacity, instead of a chemical technology as in the state of the art, and thus to improve the MTBF.

The following developments separately describe the elements of the device, taken together in FIGS. 7 and 8. The detector 1 for lowering the power supply below the power supply threshold is advantageously, as shown in FIG. current loss detector entering the device rather than loss of voltage detection as in the prior art. For this purpose, the detector 1 comprises a shunt MOSFET metal oxide semiconductor field effect transistor 10 comprising two terminals A and B, forming a current detector, a voltage divider bridge 102 having four resistors Ra , Rb, Rc and Rc, and connected to the terminals A and B, and a comparator 101 connected to the bridge 102. Advantageously, the transistor 10 has a resistor Rdson. By comparing, through the comparator 101, the voltages VA and VB at the terminals A and B of the transistor 10, the detector 1 can detect the incoming supply fault. In fact, in normal mode, since the power supply supplies the current to the device, the voltage VA at point A is strictly greater than that VB of point B. In fault mode, however, the current flowing from A to B becomes almost zero, ie there is almost equality between VA and VB. The detector 1 can therefore, through the adapted bridge 102 and the comparator 101, detect these changes in modes and deduce the incoming supply fault. As shown in FIG. 5B in combination with FIG. 5A, in order to complete the detector 1, advantageously the case of a voluntary stop, i.e an on / off control (on / off) of the activated transistor 10, can be advantageously provided. The command 14 can be issued from the support of a button or a software command. In this case, the comparison between VA and VB is no longer meaningful because the transistor 10 is open. However, the logical information is available through the on / off signal, thanks to an AND logic gate 142 mounted on a terminal of an OR logic gate 141. With regard to the detector 1, a second low cost approach can be considered, as shown in Figure 8.

The detector 1 relies on the monitoring of the decay of the high voltage across the capacitor C. If the level of the voltage across the capacitor C goes below a certain threshold, the detector 1 can detect a loss of input power because the device uses the energy reserves of the capacity. Since material resources are needed for monitoring the load across the capacity, this approach does not result in material overhead. On the other hand, it has the drawback of being slower than the embodiment of FIG. 5A or FIG.

As will be seen in the rest of this description, the speed of detection of the drop in power supply is advantageous. With regard to the capacitance C, FIG. 9 shows the ratio between the two capacitance values required, as a function of the voltage used in the invention, at constant energy, between a conventional state of the art approach and the invention. Depending on the energy required and the selected high voltage, the value of the capacitance is known. The choice of capacity is a compromise between cost, voltage withstand, preferred technology and size.

Depending on these criteria, it may be necessary to use one or more capacitors connected in parallel to increase the value of the capacitor, and in series to increase the voltage withstand, or even a combination of both. As has already been seen, the voltage-boosting charger 2 makes it possible to ensure the increase of the input voltage towards the high-voltage capacitor C, during a normal mode during which the power supply is greater at the detection threshold. As shown in FIG. 3, the charger 2 thus comprises a switch 21, a resistor 22 of value R, an inductance 23 of value 30 L, and a diode D 24. The switch 21 is advantageously of the n-mosfet type. The diode 24 is mounted in reverse at the input of the capacitor C.

The charger 2 is adapted so that the charging time of the capacitor C is of the order of magnitude of the establishment time of a communication on the network 1000. Indeed, the convergence time of the load of the capacitor for reaching the chosen high voltage is not important. It is not necessary to be able to provide the Dying Gasp function before the xDSL connection for example is established on the network 1000. Thus the charging current of the capacitor C can be relatively small compared to a conventional elevator. The low charging current further generates low noise on the DCin input voltage. In addition, it follows a favorable dimensioning elements 21, 22, 23 and 24 which constitute the charger 2. The charger 2 has a reduced cost. The switch 21 is controlled by a control 100. The control 100 comprises a microprocessor 104, a module 102 forming ADC (Analog-to-Digital Converter) or comparator, for monitoring the high voltage across the capacitor C, and a generator 103 of a pulse-width modulation (pwm) type control signal. The supervised voltage across capacitor C can be continuously through comparators, or discontinuous by cyclic sampling. The charging principle of the capacitor C is as follows. The voltage across the tank capacity C is monitored by the module 102. As long as a high voltage threshold (of the order of 100V as seen in EQ2 or 400V) has not been reached, the generator 103 modulates the switch 21, allowing the charging of the capacitor C and therefore the increase of the voltage across its terminals. When the threshold is reached, the control signal is stopped. Thus in steady state, the high voltage is maintained by regular charging, or "trickkle charging", of the capacitance C, to compensate for the losses of the capacitance (leakage current, conversion or comparison, operation without load of the step-down 3 ).

The width of the closing pulse of the switch 21, together with the values L and R of the inductance 23 and the resistor 22 (possibly zero) respectively, makes it possible to control the intensity of the charge current of the capacitance .

Alternatively, the control of the switch 21 could also be constructed from the processor 4 (dashed lines) with the aid of an external comparator. The step-down switch 3 is adapted for, during a deficient mode during which the power supply is below the threshold, to lower a voltage originating from the capacitor C when the capacitor discharges, and to thus allow, during a discharge time of the capacitor C, an auxiliary power supply of a processor 4 emitting on the network 1000 the signal FDG energy supply drop. As the storage voltage is said to be high voltage, it must indeed be able to lower it to the nominal voltage input and ensure the tilting of the emergency voltage lowered to the input of the device. As shown in FIG. 4A, the step-down 3 comprises a high-voltage converter 31 for DC-DC conversion, and a diode 32, advantageously of the shottky type, for tilting.

Advantageously, to improve the efficiency of the previous assembly, one can add in parallel to the diode 32, or even substitute in certain conditions, a switch 33 mosfet for eliminating losses in the 32 diode shottky. The switch 33 is advantageously controlled by the control 100 to the detection of the loss of voltage input. As shown in FIG. 4B, a low-cost version of the voltage step-down 3 can be made using a linear regulator circuit known to those skilled in the art. The cost saving results in a loss of efficiency over the embodiment of Figure 4A, which may possibly be compensated by the choice of a larger load voltage or a larger capacity.

The control 100 is connected to both the detector 1, the capacitor C, the charger 2, the step-down 3, the non-vital element deactivation switches 8 and 9 to the device, and the processor 4. Thus, , the rapid detection of the embodiment of FIG. 5A enables control 100, as shown in FIG. 2: on the one hand, to launch a fast DG instruction to processor 4 to set the Dying Gasp bit in the xDSL FDG frame, and - on the other hand to disable physically or logically, thanks to the switches 5, the elements (8 and 9 for example) non-vital 10 for the transmission on the FDG signal network of energy supply drop by the processor 4, to gain energy consumption and to optimize the reserves of the capacitor C. Examples of non-vital elements 8 and 9 may be USB interfaces, the Ethernet switch, the subscriber line interface circuit and 15 P, Wifi interface, etc. Advantageously, the command 100 deactivates all the non-vital elements simultaneously. The deactivation of the non-vital elements plays an important role in the final dimensioning of the capacitor C, since it makes it possible to pass from the support of a pessimistic dimensioning, with maximum consumption (for example 30W in the new products) to a consumption just necessary. (for example 5W). The control 100 performs the following steps, as shown in Fig. 6. S1, or initial state, during which the control 100 commands the loader 2 to charge the high-voltage capacitance until reaching the intended high voltage; S2, or state in which the capacitance has reached the threshold voltage, during which the charging of the capacitor is stopped by stopping the control signal coming from the generator 103 and driving the charger 2; S3, or regular recharge, during which the capacitance C is ironed below the high threshold voltage (because of the leakage currents). The loading is again authorized by the reactivation of the generator 103 (the recharge cycle may possibly be different from that used in S1); S4, where the voltage across the capacitance C has fallen below a critical threshold indicating that the system is powering on the energy reserves, during which there is detection of incoming power loss, with the actions associated with this event by the command 100: - generation of a DG signal sent to the main processor 4, - advantageously deactivation of the non-vital elements 8 and 9 to optimize the last energy resources; stopping the control signal of the loading of the capacity, coming from the generator 103; advantageously, short circuit of the 32 shottky diode; - A timer, or "AmIStillAlive Timer" according to the terminology used by those skilled in the art, is launched at the beginning of the deficient mode. The timer is used to return to the initial state S1 if the incoming power supply reappears. The duration of this timer is calculated to be a little longer than the theoretical end-of-life time when power is lost. Advantageously, the device comprises protections 6 against high voltages.

The protections 6 are typically trisils known to those skilled in the art. Voltages above 60V DC are no longer considered as SELV (Very Low Safety Voltage) voltages. If a voltage greater than this limit is chosen, particular precautions must be taken. By construction, the charger 2 and the step-down 3 ensure the isolation between the high voltage zone (around the capacitor as indicated in FIG. 8), and the SELV zones are ensured by the diode 24 in inverse of the charger 2 and the transistor and the inductance of the step-down 3. The vanishing lines are also mastered. Open circuit faults are not a problem. Only the short-circuit faults require additional protections 6 above, so that no voltage greater than 60V is found on the SELV area.

The device of the invention has a high energy efficiency.

Indeed, storing high voltage energy from the input voltage requires the charger 2 voltage booster. Thus, thanks to the charger 2, the choice of the input voltage of the gateway becomes free. Two main advantages emerge: 1. The choice of the input voltage can be calibrated on the most energy-consuming rail 7, 8 or 9. Thus, a conversion stage is eliminated with respect to the prior art, and optimum performance is obtained for the most consumer rail; 2. The input voltage can be decreased (for example up to a value of 5V), compared to an input voltage of the prior art (12V). This therefore makes it possible to favor the efficiency of the stages producing the low voltages (for example the core of the processors at 1v), since one releases a critical DC-DC conversion ratio (for example from 12V to 1V). Here are examples of energy savings: On a gateway with two USB ports and an eSata port with hard disk support, these ports can potentially consume up to 3 amperes at 5V. Assuming a 85% efficiency on a conventional 12V to 5V converter, the losses are 2.25W. Replacing a 12V / 1V converter with a 5V / 1V conversion makes it possible to go from a 70% to 90% efficiency, ie a gain of 200mW on a processor that would consume 1A at 1V.

Claims (7)

REVENDICATIONS1. Device for transmitting on a network (1000) a signal (FDG) for lowering an energy supply below an energy supply threshold, comprising - a detector (1) for reducing the energy supply by below the power supply threshold, - a capacity (C) forming a power reserve, the device being characterized in that it further comprises 10 - a charger (2) voltage booster, for the load capacity ( C) during a normal mode during which the power supply is greater than the threshold; and - a voltage step-down (3), for, in a deficient mode during which the power supply is below the threshold, lowering a voltage from the capacitor (C) when the capacitance discharges, and for thus, during a discharge time of the capacitor (C), an auxiliary power supply of a processor (4) emitting on the network (1000) the signal (FDG) of energy supply drop. 20 2. Device according to claim 1, wherein the detector (1) comprises a shunt MOSFET semiconductor field effect transistor (10) forming a current detector. 3. Device according to one of claims 1 or 2, comprising a control (100) for, in the deficient mode, to disable the non-vital elements (8, 9) for transmission on the signal network (FDG) power supply drop by the processor (4). The device of claim 3, wherein the control (100) is adapted to disable all non-vital elements (8, 9) simultaneously. 5. Device according to one of claims 1 to 4, wherein the charger (2) is adapted so that the charging time of the capacitor (C) is the order of magnitude of the time of establishment of a communication on the network (1000). 6. Device according to one of claims 1 to 5, comprising 5 protections (6) against high voltages. 7. A method for transmitting on a network (1000) a signal (FDG) of energy supply drop below an energy supply threshold, by means of a device comprising a detector (1) lower power supply below the energy supply threshold, - a capacity (C) forming an energy reserve, the method being characterized in that it comprises the steps according to which: - a charger ( 2) voltage booster charges the capacitance (C) during a normal mode during which the power supply is above the threshold; and - a step-down (3) of voltage, lowers a voltage coming from the capacitance (C) when the capacitance is discharged, during a deficient mode during which the power supply is below the threshold, and thus allows, during a discharge time of the capacitor (C), an auxiliary power supply of a processor (4) emitting on the network (1000) the signal (FDG) of energy supply drop. 10. The method of claim 7 including a step wherein the charger (2) periodically recharges the capacitance (C) to compensate for capacitance losses (C). 11. Method according to one of claims 7 or 8, comprising a step according to which a command (100) allows, in the deficient mode, to deactivate the elements (8, 9) which are not vital for the transmission on the network of the signal (FDG) of energy supply drop by the processor (4). The method of claim 9, including a step in which the control (100) deactivates all the non-vital elements (8, 9) simultaneously.