Method and apparatus for detecting and suppressing a parallel ARC fault

Information

  • Patent Application
  • 20030063420
  • Publication Number
    20030063420
  • Date Filed
    September 28, 2001
    23 years ago
  • Date Published
    April 03, 2003
    21 years ago
Abstract
Electrical equipment (14) is safeguarded from damage due to parallel arc faults by a circuit that provides several levels of protection. A semiconductor switch (18) and a current sensor (24) are placed in series with the electrical equipment (14). When the current to the equipment exceeds a first threshold for a predefined period of time, the semiconductor switch (18) is rendered non-conductive until the circuit is specifically reset. When the current to the equipment exceeds a greater second threshold, a pulsed signal alternately places the semiconductor switch (18) in conductive and non-conductive states so that the average current applied to the equipment (14) is within an acceptable level. The pulses are measured to determine whether a parallel arc fault has occurred. When the measured pulses (74) are within a predetermined range, a parallel arc fault is declared and the semiconductor switch (18) is rendered non-conductive. Should the current exceed an even greater third threshold, the semiconductor switch (18) is immediately rendered non-conductive until the circuit is manually reset.
Description


BACKGROUND OF THE INVENTION

[0001] The present invention relates to mechanisms that protect equipment from damage due to electrical faults and short circuits; and particularly to such devices which electronically monitor performance of the electric equipment and take protective action in the event of a parallel arc fault.


[0002] It is important that electrical apparatus be protected from damage when electrical failures occur. For example, conventional fuses and electro-mechanical circuit breakers are commonly employed to disconnect equipment from an electrical supply upon detection of excessive current when a short circuit occurs. Nevertheless, these conventional protection devices are relatively slow in disconnecting the current flow to the apparatus being protected. As a consequence, enough excessive electrical current can flow into the equipment to cause damage during a fault.


[0003] Furthermore, various electrical apparatus require different response characteristics for the protection device. For example, electronic equipment may draw a substantially constant current level from initial start-up through a normal shut-down and be very intolerant of excessive current levels of even short duration. The protection device for such equipment has to respond very quickly to even relatively small over current conditions. Other types of electrical equipment draw large instantaneous current levels at certain times, such as upon start-up, in comparison to the current level drawn during remainder of their operation. Thus, a circuit protection device that responds too rapidly to an high current condition may inadvertently shut-off current to the equipment during normally occurring events. As a consequence, the protection device for this type of equipment must respond in a manner that tolerates brief high currents. The manner in which a protection device responds to over currents is referred to as the trip response characteristic or trip curve, and has to be matched to the particular type of electrical apparatus being protected.


[0004] This usually means that a manufacturer of protection devices must design, manufacture and stock in inventory, a large variety of protection devices that have different trip response characteristics in terms of current level and duration.


[0005] In the case of parallel arcs, it is difficult to detect a parallel arc because the arcs are often intermittent, unlike a bolted fault, or short circuit, where the surge of current is continuous.


[0006] Thus it is desirable to provide a protection device that can identify a parallel arc fault. It is also desirable to provide a parallel arc protection device that can be customized easily with different trip response characteristics.



SUMMARY OF THE INVENTION

[0007] An apparatus for protecting an electrical load from a parallel arc fault employs a semiconductor switch to connect the electrical load to a source of current. A current sensor is coupled in series with the semiconductor switch and produces a sensor signal that indicates the magnitude of current flowing to the electrical load.


[0008] A control circuit is connected to the current sensor and the semiconductor switch. The control circuit responds to the sensor signal by producing a control signal that is applied to a control input of the semiconductor switch. In a first mode of operation when the magnitude of current is less than a first threshold, the control circuit maintains the semiconductor switch in a continuous conductive state. When the magnitude of current is greater than the first threshold and less than a second threshold, the control circuit in a second mode of operation renders the semiconductor switch non-conductive after a predefined period of time. In a third mode of operation when the magnitude of current is greater than the second threshold, the semiconductor switch is alternately pulsed conductive and non-conductive by the control circuit to apply an average current through the load that is within an acceptable level wherein damage does not occur. When the control circuit pulses the semiconductor switch, the pulses are measured to determine whether a parallel arc fault has occurred. When the measured pulses are within a predetermined range, a parallel arc fault is declared and the control circuit renders the semiconductor switch non-conductive.







BRIEF DESCRIPTION OF THE DRAWINGS

[0009]
FIG. 1 is a block diagram of a solid state circuit protector according to the present invention;


[0010]
FIG. 2 illustrates the details of an instant trip circuit in the protector;


[0011]
FIG. 3 is a graph of an exemplary trip response characteristic of the solid state circuit protector; and


[0012]
FIG. 4 is a diagram of exemplary pulses generated by the control circuit measured within time windows according to the present invention.







DETAILED DESCRIPTION OF THE INVENTION

[0013] With initial reference to FIG. 1, a solid state circuit protector 10 controls application of a direct current to an electrical load 14, depicted as capacitance in parallel with a resistance. The solid state circuit protector 10 has a positive voltage terminal 12 which is connected to the electrical source for powering a load 14. Current flows from the positive voltage terminal to the load through a back-up fuse 16, a semiconductor switch 18 and an inductor 20 to a load terminal 22. The load is connected between the load terminal 22 and the negative side of the voltage supply, represented as ground.


[0014] The back-up fuse 16 is a conventional device with a conductor which heats-up and ultimately breaks when excessive current flows for a given period of time. Standard devices, such as glass tube encased fuses or an appropriate trace on printed circuit board, can be employed for the back-up fuse 16. The back-up fuse 16 provides redundant protection in case the semiconductor switch 18 fails in the conductive state or upon failure of electronic circuits controlling the semiconductor switch. As will be understood, the trip response time of the back-up fuse is considerably slower than the trip response characteristic of the electronic circuit protection.


[0015] The semiconductor switch 18 must be able to interrupt the load current and handle transient currents, over currents and in-rush at a specified operational voltage range as dictated by the particular load 14 to be controlled. An n-channel field effect transistor (FET), such as model IRF1404 from International Rectifier of El Segundo, Calif. 90245 USA may be used as the semiconductor switch 18. The channel resistance in the conductive state has to be relatively low to minimize the voltage drop across the FET and the heat dissipation. Although the preferred embodiment employs the semiconductor switch 18 between the positive voltage terminal 12 and the load 14, alternatively the switch could be placed on the ground side of the load. However, this alternative approach has the disadvantage that a fault from load to ground would be unprotected.


[0016] A voltage sensor 28 produces an analog signal which indicates the voltage level at the load terminal. That analog signal is applied to an analog input of a microcontroller 26. As will be described, the microcontroller 26 responds to an indication from the sensor 28 that the voltage across the load 14 is too low by turning off the semiconductor switch 18.


[0017] A current sensor 24 is provided to detect the level of current flowing between the positive voltage terminal 12 and the load 14. This sensor must have a dynamic range which is large enough to cover the current extremes for the desired trip response characteristic of the protection device and have a transient response that is sufficiently fast to implement the desired trip response characteristic. The current sensor 24 may be a Hall effect sensor that produces an output voltage indicative of the DC current magnitude and which output voltage can be applied via line 31 directly to an analog input of a microcontroller 26. Other types of conventional current sensors, such as a shunt resistor, may be used to provide a current magnitude indication to the microcontroller 26.


[0018] The microcontroller 26 is microprocessor based and includes an internal analog-to-digital converter with a multiplexed input for signals from the current and voltage sensors. Digital input/output circuits of the microcontroller handle signals for other components of the solid state circuit protector 10. For example, a user control panel 25 has a keypad 27 and light emitters 29, such as LED's. The keypad 27 has separate momentary contact switches that supply input signals to the microcontroller 26 to manually turn the solid state circuit protector 10 on and off, as well as reset a trip condition. The light emitters 29 are powered by signals from the microcontroller to indicate the operational states of the circuit protector. One of those light emitters 29 indicates when the circuit protector 10 is tripped. The microcontroller 26 also has an internal non-volatile memory which stores a software program defining the protection function and which stores data, such as the trip response characteristic, for use by that software program. The microcontroller 26 and the control panel 25 optionally can control additional poles of a circuit protector as indicated by a second pole 11 drawn in phantom lines.


[0019] The microcontroller 26 operates the semiconductor switch 18 through a trip circuit 36 that generates a drive voltage which is adequate to control the FET 19 in the preferred embodiment of the semiconductor switch 18. Because the voltage driving the gate of an N-channel FET 19 has to be approximately ten volts greater than the voltage at the source electrode of the FET, the trip circuit 36 includes a charge pump or similar circuit to generate voltage greater than that found on the positive input terminal 12.


[0020]
FIG. 2 illustrates the details of the trip circuit 36 wherein the output signal ISENSE on line 31 from the current sensor 24 is applied to a first voltage comparator 40. The sensed current level ISENSE is compared to a second threshold ITH2 which is produced on an analog output line 37 of the microcontroller 26. A fixed value for the second threshold ITH2is programmed into the microcontroller 26 depending upon the over current tolerance of the specific load 14. The result of that comparison at the output of the first comparator 40 is applied to the RESET input of a flip-flop 42. The reset input also is connected to a positive supply voltage V+ by a pull-up resistor 44.


[0021] The SET input of the flip-flop 42 is connected to the output of a dual input NAND gate 46, having both inputs tied together to function as an inverter. The inputs of first NAND gate 46 are connected to a digital output line 33 from the microcontroller 26 which carries a pulsed signal at a fixed frequency in excess of 15 kHz., specifically in the range of 20-80 kHz. and preferably at 50 kHz. The pulsed signal has a fixed duty cycle thereby forming a train of constant-width pulses. As will be described, the pulse train periodically sets the flip-flop output which is tied to one input of a second NAND gate 48 having three inputs. Another input of the second NAND gate 48 receives an ON signal on another digital output line 33 from the microcontroller 26. Whether the ON signal is active or inactive is determined by manual operation of switches on keypad 27 of the control panel 25.


[0022] The third input of the second NAND gate 48 receives an output signal from an instant trip mechanism formed by a second voltage comparator 50 and a second flip-flop 51. Specifically, the second comparator 50 compares the current sensor output signal ISENSE to a third threshold ITH3. The third current threshold ITH3 is generated on another analog output line 38 by the microcontroller 26 and is defined by a fixed value programmed into the solid state protection circuit 10. The third current threshold ITH3 is greater than the second current threshold ITH2. The precise relationship between those two current thresholds will become apparent from a subsequent description of the operation of the solid state current protection circuit. The second and third current thresholds ITH2 and ITH3, instead of being programmable, may be set by conventional voltage dividers at the inputs to the respective comparators 40 and 50. The output of the second comparator 50 is latched by the second flip-flop 51 with an output connected to another input of the second NAND gate 48. The set input of the second comparator 50 is connected to RESET output line 35 from the microcontroller 26.


[0023] The components of the trip circuit 36 described thus far, provide input signals to the second NAND gate 48. The output of that gate is fed through a third NAND gate 52 which is connected as an inverter. The signal emanating from the third NAND gate 52 is coupled by a resistor 56 to an isolation circuit 54, such as a standard opto-isolator. The isolation circuit 54 produces an output on line 58 that is applied to an input of a conventional FET gate driver circuit 60. A charge pump 62 provides a voltage level that the FET gate driver 60 uses to bias the gate of the FET 19 via line 39.


[0024] The operation of the solid state circuit protector 10, in FIG. 1, commences with the operator pressing an appropriate switch on keypad 27. The microcontroller 26 responds to this switch activation by applying a high level, or active ON signal, via line 34 to the second NAND gate 48. At this time, the microcontroller 26 also begins producing a pulse train on digital output line 33 connected to the first NAND gate 46. A high logic level of that pulse train causes the output of the first flip-flop 42 to go high, applying another high level to another input of the second NAND gate 48.


[0025] During normal operation of the load 14, the output signal ISENSE from the current sensor 24 is less than the third threshold ITH3. As a consequence, the second voltage comparator 50 produces a high logic level at the third input of the second NAND gate 48. Thus the second NAND gate 48 produces a low level output signal that, upon inversion by the third NAND gate 52 and conduction through isolator 54, activates the FET gate driver 60. This causes the gate driver 60 to bias the gate of the FET 19 into a conductive state, thereby applying current from the positive voltage terminal 12 through the inductor 20 to the load 14.


[0026] The level of current through the semiconductor switch 18 rises rapidly and soon exceeds the second threshold ITH2. At that time, the output of the first comparator 40 goes low resetting the flip-flop 42 and causing the second NAND gate 48 to change output states. This results in the FET gate driver 60 rendering the semiconductor switch 18 non-conductive. The energy stored in the inductor 20 produces a decaying current that flows through the load 14 and the fly back diode 21.


[0027] When the next positive pulse occurs in the pulse train on line 33 to the first NAND gate 46, the flip-flop 42 will be SET to produce another high logic output level which once again turns on the FET gate driver 60 and the semiconductor switch 18. This on-off cycling of the semiconductor switch continues chopping the current at the rate of the signal on line 33 until the capacitance in the load 14 adequately charges, at which time the load current becomes substantially constant at a level less than the second threshold ITH2. Thus the load current during start-up is limited to being less than the second threshold ITH2 while still applying current to initialize the load operation. Once the excursions of current through the semiconductor switch 18 fall below this threshold, the flip-flop 42 no longer is reset and the FET gate driver 60 maintains the semiconductor switch 18 in a conductive state. That conductive state continues as long as the load 14 functions normally.


[0028] If there is a fault with the load during start-up, the load current does not drop below the second threshold ITH2. The current chopping could continue indefinitely in this case. To prevent that, the duration of the current chopping is limited by counting the current pulses applied to the load and terminating the chopping upon the occurrence of given number of pulses that normally is sufficient to charge a typical load capacitance. Specifically, the microcontroller 26 monitors the input line 31 from the current sensor 24 which indicates alternate high current and zero current conditions and counts the number of high current pulses.


[0029] That count is compared to a reference number and the chopping mode is terminated when that reference number of current pulses has occurred. At that time, the microcontroller 26 sends a low logic level signal on line 34 to the trip circuit 36, which renders the semiconductor switch 18 non-conductive until a person presses the RESET switch on the control panel 25 and resets the microcontroller 26.


[0030] A parallel arc fault is more difficult to detect than a short circuit, or “bolted fault”. A parallel arc fault may result from worn insulation on a wire, as an example, resulting in an intermittent short. The microcontroller 26 of the circuit protector 10 detects a parallel arc fault by measuring the pulses applied to the load 14. The microcontroller 26 may measure the pulses by monitoring the input line 31 from the current sensor 24 or by monitoring the pulses sent to the FET gate driver 60.


[0031] A number of algorithms may be applied to declare a parallel arc fault. For example, a number of pulses within an infinite time period, a number of pulses within a predetermined time period, a number of periods with a declared parallel arc fault, when the number of pulses are within a predetermined range, or any combination of the above. It should also be noted that “measured” when referring to the pulses may mean integrated, counted or any other mathematical operation. It should be understood that pulses sent to the FET gate driver 60 are a representation of the percentage of time that ISENSE is greater than ITH1. When the measured pulses meet a predetermined criteria, a parallel arc fault is declared and the semiconductor switch 18 is rendered non-conductive.


[0032] Alternatively the voltage sensor 28 can be employed to safeguard against operating in the current chopping mode for too long a time period. During a short circuit condition when the load 14 is drawing excessive current, the voltage across the load will be significantly lower than during normal operation. The voltage across the load 14 is detected by the voltage sensor 28 which applies an analog voltage level indication to the microcontroller 26. If that sensed load voltage remains below a given threshold for greater than a predefined time interval during the current chopping mode, the microcontroller 26 turns off the trip circuit 36 by applying a low logic level, (an inactive ON signal) to the ON/OFF line 34.


[0033] The operation of the solid state protection circuit 10 during an over current condition after a normal start-up may best be understood with respect to an exemplary trip response characteristic, such as the one depicted in FIG. 3. A load current which is below a first threshold ITH1 can be tolerated indefinitely by the load 14 and thus will be conducted continuously by the semiconductor switch 18. The first threshold ITH1 is set between 100% and 125% of the current rating for the load 14 being protected. Load currents between levels ITH1 and I2 can be tolerated by the load for an amount of time which is inversely proportional to the current magnitude. In other words, small deviations above the first threshold ITH1 can be tolerated for a longer period of time than over currents which approach level I2. This produces a linear trip response characteristic in portion 70 of the response curve. This portion of the trip response characteristic is programed into the microcontroller 26 and stored in its memory either as a linear equation or as a data table. That data table has pairs of values with one value being a current magnitude and the other value defining a time interval during which that current magnitude can be tolerated before the solid state circuit protector 10 must trip.


[0034] Current between level I2 and the third threshold ITH3 can be tolerated by the load for a period designated T1. Current above that higher level ITH3 cannot be tolerated by the load 14, even momentarily, and thus the current protection device will trip immediately. It should be noted, that load current within the cross-hatched region 72 between second threshold ITH2 and a third threshold ITH3 while tolerated by the load 14, can damage the FET 19. Thus, when operation within this region is determined to occur, the solid state circuit protector 10 enters a current chopping mode of operation. In this mode the semiconductor switch is pulsed on and off at a rate which produces an average current that is less than the second threshold ITH2. Thus the load remains powered so that the load capacitance remains charged but the current applied to the load is limited to that second threshold level.


[0035] When the sensed current ISENSE is between the first threshold ITH1 and a second threshold ITH2, the trip circuit 36 initially maintains semiconductor switch 18 in a conductive state because that current is below the two comparator thresholds ITH2 and ITH3. However, the microcontroller 26 receives the output signal ISENSE from the current sensor 24 on line 31 utilizes the programed trip response characteristic for section 70 to determine whether to turn-off the semiconductor switch 18. Specifically, the microcontroller 26 determines whether the over current magnitude has occurred for the time period defined by the trip response characteristic. Once that has occurred, the microcontroller 26 turns off the trip circuit 36 by applying a low logic level, inactive ON signal, to the digital line 34. This constant low logic level toggles the output level from the second NAND gate 48 which turns-off the FET gate driver 60 and thus the semiconductor switch 18. The microcontroller 26 also illuminates the light emitter 29 on the control panel 25 which indicates the tripped condition. The OFF signal continues to be applied by the microcontroller 26 to the trip circuit 36 until a manual reset switch on the control panel 25 pressed.


[0036] When the sensed load current signal ISENSE is between the current thresholds ITH2 and ITH3, the microcontroller 26 does not utilize the trip response characteristic data to determine whether to turn-off the trip circuit 36. Instead, the solid state protection circuit 10 enters a current chopping mode in which the FET 19 is pulsed on and off at the rate of the pulsed signal on line 33.


[0037] Specifically with reference to FIG. 2, when the current sensor 24 produces an output signal ISENSE on line 31 which is greater than the second threshold ITH2 on line 37, the output of first comparator 40 goes low. That low output resets the flip-flop 42, thereby applying a low logic level to an input of the second NAND gate 48. This produces a high logic level at the output of the second NAND gate 52 which is inverted by the third NAND gate 52 thus applying a low logic level to the opto-isolator 54. This in turn deactivates the FET gate driver 60 which renders the semiconductor switch 18 non-conductive. At that time current from the inductor 20 flows through the load 14 and a fly back diode 21.


[0038] The semiconductor switch 18 remains off until the next high logic level pulse in the pulse train from the microcontroller 26 that is applied to the trip circuit 36 on digital line 33. That pulse upon inversion by the first NAND gate 46 sets the flip-flop 42 which produces a high output level that is applied to the second NAND gate 48. This high logic level activates the FET gate driver 60, once again rendering the semiconductor switch 18 conductive.


[0039] When the FET 19 turns on again, the inductor 20 limits the rate at which the current rises so that the current level does not immediately exceed the second threshold ITH2. Thus, a small amount of current will be applied to the load 14 and charging its capacitance. However, the current through the semiconductor switch 18 eventually rises above the second current threshold ITH2 which will be detected by the first comparator 40. When this occurs the first comparator 40 changes output states and resets the flip-flop 42, which in turn applies a signal to the second NAND gate 48 that ultimately results in the FET gate driver 60 shutting off the semiconductor switch 18. This cycling of the semiconductor switch 18 off and on continues which results in an average load current that is below the first threshold ITH1.


[0040] Although the semiconductor switch 18 is not exposed to as great a degree of thermal stress as with linear current limiting, damage to the FET 19 or the load still may occur if the current chopping mode continues for too long a time period. As discussed previously with respect the start-up operation of the solid state circuit protector 10, the duration of the current chopping can be limited by the microcontroller 26 counting the number of current pulses applied to the load and sending a low logic level OFF signal on line 34 to the trip circuit 36 when a given number of pulses has occurred. Alternatively the voltage sensor 28 can be employed to detect a short circuit and inform the microcontroller 26 to turn-off the trip circuit 36.


[0041] The parallel arc fault detection logic may be best understood with reference to an exemplary pulse signals detected by the microcontroller 26 depicted in FIG. 4. A series of pulses 74 is shown indicative of the number of current pulses applied to the load 14. A parallel arc fault may be declared when a number of pulses 74 are detected by the microcontroller 26. Additionally, the microcontroller 26 may count the number of pulses 74 over a period of time. Alternatively, a range for declaring a parallel arc fault may be programmed into the microcontroller 26 or be hard coded into the logic.


[0042] In the preferred embodiment, the first pulse 74 initializes the microcontroller. The pulse train has a frequency of 50 kHz, rendering a maximum of 50 pulses per millisecond. It should be noted that the frequency range of the pulse train is application specific and should not be deemed to be limiting upon the present invention. Each pulse 74 causes an interrupt which increments a counter. The counter decrements once every 100 milliseconds to reduce any nuisance trips. The count is analyzed every millisecond. The microcontroller 26 is set to declare a bolt fault if 36 or more pulses are counted within a 1 millisecond window. If the microcontroller 26 counts between 1 pulse and 36 pulses in a 1 millisecond window a parallel arc fault is declared, sending a low logic level OFF signal on line 34 to the trip circuit 36.


[0043] The microcontroller 26 may evaluate a series of windows. The results of a previous number of windows will remain in the memory of the microcontroller 26. For example, when four of six previous 1 millisecond windows declare a parallel arc fault, the microcontroller 26 will declare a parallel arc fault and render the semiconductor switch 18 non-conductive. The memory of the microcontroller 26 is reset when no pulses are detected for 6 milliseconds. Those skilled in the art will recognized that the aforementioned time segments and numbers of pulses are for exemplary purposes and should not be deemed to be limiting upon the scope of the present invention.


[0044] The current chopping should ensure that the load current never exceeds the third threshold level ITH3. However, in the event that a malfunction occurs, the second comparator 50 detects a load current above that third threshold level ITH3 and produces an output that renders the semiconductor switch 18 continuously non-conductive. Specifically, the output of the second comparator 50 goes low which resets the second flip-flop 51 thereby applying a low logic level to the second NAND gate 48. This results in the FET turning off.


Claims
  • 1. An apparatus (10) for protecting an electrical load (14) from a parallel arc fault, comprising: a semiconductor switch (18) for connecting the electrical load to a source of current (12) and having a control input; a current sensor (24) coupled to the semiconductor switch (18) and producing a sensor signal indicative of a magnitude of current flowing to the electrical load (14); and a control circuit (26) connected to the current sensor (24) and the semiconductor switch (18) and responding to the sensor signal by producing a control signal that is applied to the control input, the control circuit having a first mode of operation when the magnitude of current is less than a first threshold wherein the semiconductor switch (18) is maintained in a continuous conductive state, a second mode of operation when the magnitude of current is greater than the first threshold and less than a second threshold wherein the semiconductor switch (18) is rendered non-conductive after a predefined period of time, and a third mode of operation when the magnitude of current is greater than the second threshold wherein the semiconductor switch (18) is alternately pulsed conductive and non-conductive to limit the current through the load (14) to less than the second threshold, the control circuit (26) measures the pulses (74) to determine whether a parallel arc fault has occurred, when the measured pulses (74) are within a predetermined range a parallel arc fault is declared wherein the semiconductor switch (18) is rendered non-conductive.
  • 2. The apparatus (10) of claim 1, wherein the number of pulses is counted.
  • 3. The apparatus (10) of claim 2, wherein the number of pulses is counted over a predetermined period of time.
  • 4. The apparatus (10) of claim 2, wherein the number of pulses is counted over successive predetermined periods of time.
  • 5. The apparatus (10) of claim 1, wherein the range has a maximum value.
  • 6. The apparatus (10) of claim 3, wherein the predetermined period of time is about 1 millisecond.
  • 7. The apparatus (10) of claim 4, wherein the pulses are counted over six successive predetermined periods of time.
  • 8. The apparatus (10) of claim 5, wherein the range of counted pulses is between 1 and 36 pulses.
  • 9. The apparatus (10) of claim 1, wherein the pulse frequency is between about 15 kHz and 75 kHz.
  • 10. The apparatus (10) of claim 9, wherein the pulse frequency is about 50 kHz.
  • 11. A method for protecting an electrical load (14) from a parallel arc fault, comprising: placing a semiconductor switch (18) into a conductive state to apply current from a source (12) to the electrical load (14); sensing the magnitude of the current; comparing the magnitude of the current to a first threshold; comparing the magnitude of the current to a second threshold that is greater than the first threshold; determining when the magnitude of the current exceeds the first threshold for a predefined period of time and when that occurs issuing a termination signal; in response to the termination signal, placing the semiconductor switch (18) in a non-conductive state to terminate application of current to the electrical load (14); and when the magnitude of the current exceeds the second threshold, alternately placing the semiconductor switch (18) conductive and non-conductive states to apply current pulses to the electrical load (14) and limit the average current through the load (14) to less than the second threshold and measuring the pulses (74) to determine whether a parallel arc fault has occurred, when the measured pulses (74) are within a predetermined range a parallel arc fault is declared wherein the semiconductor switch (18) is placed in a non-conductive state to terminate application of current to the electrical load (14).
  • 12. The method of claim 11, wherein the number of pulses is counted.
  • 13. The method of claim 12, wherein the number of pulses is counted over a predetermined period of time.
  • 14. The method of claim 12, wherein the number of pulses is counted over successive predetermined periods of time.
  • 15. The method of claim 11, wherein after a first pulse is detected, the pulse energy is summed.
  • 16. The method of claim 13, wherein the predetermined period of time is about 1 millisecond.
  • 17. The method of claim 14, wherein the pulses are counted over six successive predetermined periods of time.
  • 18. The method of claim 16, wherein the range of counted pulses is between 1 and 36 pulses.