TWI542146B - A Method for Eliminating Zero Crossing Point Noise in Digital Power Supply - Google Patents

Description

Zero crossover point noise cancellation method for digital power supply

The present invention relates to a noise cancellation method for a digital power supply, and more particularly to a zero crossover point noise cancellation method for a digital power supply.

Energy development and application have been accompanied by the growth of human civilization. However, due to the increasing demand for energy from population growth, how to generate new energy and improve energy application efficiency has become an important issue in modern energy. Most of the current transmission and distribution networks use the 50 Hz or 60 Hz AC mode, and most of the devices for regenerative energy conversion are converted by power electronic devices. In order to obtain high conversion efficiency, power electronic devices mostly use switching mode. The technology performs power conversion, and the form converter is susceptible to voltage and current surges due to parasitic inductances and capacitances of the power switching elements and circuits, thereby interfering with the circuit. In order to connect with the transmission and distribution system, the switching mode power converter usually uses high-voltage mode operation. During the operation of the switching element, because the line parasitic inductance and capacitance generate many unnecessary oscillations, it also produces considerable The amount of noise is scattered on the line. DSP (digital signal processor) or MCU (microcontroller unit; microcontroller) for high-speed operation efficiency, operating voltage is reduced by 5V, 3.3V, 1.8V all the way, although higher operating speed However, due to the drop in operating voltage, the resistance to noise is also reduced. Digital power system, all control outputs are detected by the DSP or MCU for external signals (Sensing) and then internal Generated by calculation. If the detected signal is wrong, the resulting control output will be wrong. How to separate the correct signal from the error signal in a series of motion signals, so that the system can operate correctly and stably is very important.

In order to solve the aforementioned problems, we need a novel zero-crossing point noise cancellation method.

In the digital power system, all control outputs are generated by the DSP or MCU's external signal detection (Sensing) and then internal calculations. If the detected signal is wrong, the resulting control command will be wrong. How to separate the correct signal from the error signal in a series of detected signals, so that the system can operate correctly and stably is an urgent problem to be solved. The object of the present invention is to solve this problem.

To achieve the foregoing objective, a zero-crossing point noise cancellation method for a digital power supply is proposed, which uses a corresponding time value of a plurality of zero-point signals as a data source, wherein the plurality of zero-point signals include a plurality of real numbers. a zero point signal, and the time difference between the two adjacent real zero signals is between a minimum period and a maximum period, the method comprising the following steps: the first step: selecting the first of the plurality of zero signals The time point of the zero signal is a starting reference point; the second step: determining whether there is a time difference between the starting zero point and the starting reference point in the plurality of zero signals is between the minimum period and the Determining at least one possible qualified signal between the maximum periods. If the determination result is no, the signal at the initial reference point is defined as noise and the starting reference point is moved to the one closest to the reference point. The time point of the zero signal, and re-execute the second step; if the judgment result is yes, the starting reference point is determined Determining a qualified signal point and setting a time point of the at least one possible qualified signal as at least one subsequent reference point; and a third step: sequentially taking one of the at least one subsequent reference point as the starting reference point Performing the second step; the fourth step: repeatedly performing the third step and generating at least one set of time series data, wherein each of the sets of the time series data is composed of a plurality of the qualified signal points; a fifth step: performing a one-cycle variation amount accumulation calculation on each group of data of the at least one set of timing data to generate at least one period variation amount cumulative value, and using the smallest one of the at least one period variation amount cumulative value A corresponding set of the timing data is used as timing information of the plurality of real zero signals.

The detailed description of the drawings and the preferred embodiments are set forth in the accompanying drawings.

FIG. 1 is a schematic diagram showing the basic principle of the method for debugging the mains zero signal of the present invention.

Figures 2a-2d illustrate the binary arrangement of two zero signals.

Figures 3a-3d illustrate a binary permutation combination of three adjacent zero signals with a starting point of zero.

4a-4d illustrate a binary permutation combination of three adjacent zero signals with a starting point of one.

Figure 5 is a schematic diagram of all the zero signals that are true.

Figure 6 is a schematic diagram of all zero signals being noise.

FIG. 7 is a schematic diagram of real zero signal and noise interaction arrangement.

FIG. 8 is a schematic diagram of signal determination in which the first signal is a real signal.

FIG. 9 is a schematic diagram of the first signal being a real signal and the noise being distributed in the condition determination area.

10-11 are schematic diagrams showing the function of adding a signal re-inspection in the judgment area according to the present invention.

12-14 are schematic diagrams showing the occurrence of more than two noise problems in the determination zone by the method of the present invention.

15a-15d are schematic diagrams showing the occurrence of more than two signals in a plurality of judgment intervals by the method of the present invention.

16-19 are schematic diagrams showing the processing of the first zero signal as a sequence signal of noise by the method of the present invention.

20-21 illustrate the processing of the first zero point noise in the conditional decision area and the second zero point noise also falls within the conditional decision area.

FIG. 22 is a flow chart of an embodiment of the present invention.

Fig. 23 is a graph showing the error percentage of the unprocessed frequency value and the frequency value obtained by the method of the present invention, calculated relative to the simulated mains frequency.

Figure 24 is a graph comparing the error percentages of the frequency values processed by the conventional period comparison method and the frequency values obtained by the method of the present invention with respect to the simulated mains frequency.

The invention adopts the characteristic that the frequency of the commercial power supply system is not easily changed as the judgment basis for separating the true zero signal of the commercial power from the noise, thereby eliminating the frequency misjudgment caused by the noise.

Example Description: FIG. 1 illustrates the basic principle of the method for debugging the commercial zero point signal of the present invention.

The circle n represents the nth zero point signal captured by the hardware circuit, and the corresponding A count value, where circle 1 represents the first zero signal, circle 2 represents the second zero signal, circle 3 represents the third zero signal, circle 4 represents the fourth zero signal, and so on. In addition, if the circle is left blank (except for the number), it means that the zero signal is the correct zero signal; if the circle is filled with a gray, the zero signal is the wrong zero signal.

The arrow curve in the figure indicates that a correct zero signal is used as a reference point, and subsequently One of the zero signals generates a difference in count value, or a time difference; step n F represents a difference in the difference of the nth count value, that is, the difference of the count value is greater than the maximum period allowed or less than The minimum period allowed, and the step n Y represents that the difference of the nth count value is a qualified difference, that is, the difference value of the count value is less than the maximum period allowed and greater than the minimum period allowed.

The range of dashed lines in the figure indicates the allowed time difference, which is in the range Max. Cycle-Min Cycle (maximum cycle - minimum cycle). The invention first uses the first zero point signal as a basic reference point. Since the difference between the count value of the second zero point signal and the first zero point signal is less than the minimum period, the second zero point signal is an unqualified zero point signal, and the judgment result is represented by step 1 F, and still is the first one. The zero signal is the basic reference point. Then, since the difference between the count value of the third zero point signal and the first zero point signal is less than the maximum period and greater than the minimum period, the third zero point signal is a qualified zero point signal, and the judgment result is represented by step 2 Y. Then, the third zero point signal is used as the basic reference point. Since the difference between the count value of the fourth zero point signal and the third zero point signal is less than the maximum period and greater than the minimum period, the fourth zero point signal is a qualified zero point signal, and the judgment result is represented by step 3 Y. That is, when the difference between the count values is less than the maximum period and greater than the minimum period, the basic reference point moves to the relative change point at the time of the judgment, and the next difference judgment is performed. The difference judgment is repeated until the end of the data or the interval is continuously established to reach the set value, and finally the error judgment is performed to determine the optimal signal point and eliminate the error signal. the goal of.

Zero-point signal morphological analysis: Whether it is a real zero-point signal or a signal generated by noise, it is a sampling circuit that sequentially enters the DSP or MCU in a sequence on the time axis. However, the difference in the arrangement of the signals will affect the way of judging, so it is necessary to sort them one by one. The zero signal can be divided into two types, one is the real zero signal, the other is the signal generated by the noise, and is arranged in the time axis. This arrangement can be expressed in binary form, for example, the correct signal is defined as 0, and the wrong signal is defined as 1. 2a-2d illustrate a binary arrangement of two zero signals, wherein (a) the signal is 00, the (b) signal is 01, the (c) signal is 10, and the (d) signal is 11. 3a-3d illustrate a binary arrangement of three adjacent zero signals with a starting point of 0, wherein (a) the signal is 000, the (b) signal is 001, the (c) signal is 010, and the signal is (d). Is 011.

Figures 3a-3d show the arrangement of three adjacent zero signals with a starting point of 0. If the starting point is 1, according to the binary arrangement rule, there will be four more arrangements. 4a-4d illustrate a binary arrangement of three adjacent zero signals with a starting point of 1, wherein (a) the signal is 100, (b) the signal is 101, (c) the signal is 110, and the signal is (d). Is 111.

As can be seen from Figures 4a-4d, the correct zero signal will be correctly in the period of the period judgment, but the wrong zero signal will not necessarily be in which section, so in the subsequent signal processing discussion, It will be discussed separately whether the error zero signal is in the period of judgment. Figures 4a-4d show the starting zero signal as the arrangement of the error zero signal. As can be seen from the figure, the arrangement of the error zero signal and the correct zero signal is still in binary arrangement. If the four-point signal arrangement is performed in binary mode, there will be 16 ways of arrangement, but in the form of these arrangements, it can be classified into three categories without being classified into sixteen types, such as This can effectively analyze and judge the signal.

Figure 5 is a schematic diagram of all the zero signals that are true. In this situation There is no need to perform signal debugging. It is only necessary to continuously judge the zero signal several times to obtain a better average value for the system to use the zero signal.

Figure 6 is a schematic diagram of signals generated by all zero signals being noise. due to The true zero signal is a periodic signal, so a true zero signal will appear in a certain period of time. In Figure 6, it can be seen that if the zero signal generated by the noise is concentrated in two real zero signal periods, the system has extremely high noise, even if the wrong zero signal can be correctly eliminated, in the system other The circuit must also withstand such high noise, and the operation of the system will be extremely unstable. This situation is not caused by a single factor. If the zero signal is processed in the manner discussed in the present invention, considerable computation time and memory will be caused. The consumption of physical resources, which is a multi-factor problem, is not discussed further here.

Figure 5 and Figure 6 are discussed in binary mode, in the order of the arrangement, respectively The front end and the last end of the arrangement are divided into more extreme examples. Figure 5 shows all the real zero signals. This is the situation that the circuit design wants to achieve. However, it can be seen that the binary distribution is not easy to occur. The correct zero signal and the abnormal zero signal will appear. A situation that often occurs.

FIG. 7 is a schematic diagram of real zero signal and noise interaction arrangement. If true The periodic signal is the basis of the demarcation. It can be seen that the abnormal noise is distributed between the real zero signals, and one or several abnormal noises may be distributed in a single interval. However, if there are too many abnormal noises, a phenomenon like that shown in Fig. 6 will occur, and it will not be discussed in the present invention.

True zero signal judgment method: The way the zero signal is judged is still classified in binary mode. This hair First, use the zero-point signal as the reference point, and then use this point to judge the subsequent zero-point signal. However, when the first reference datum point is taken, the real zero signal is not necessarily obtained, so the classification or discussion is performed when the first reference datum point is taken as real or abnormal noise. FIG. 8 is a schematic diagram of signal determination in which the first signal is a real signal. It can be seen from the judgment mark in FIG. 8 that since the real zero signal 1 is used as the reference point, the noise can be deleted and the subsequent true zero signal can be correctly found. In this way, the judgment is continued until the data is judged or the continuous conditional judgment can be obtained X times (X is a positive integer made by the program user, and the higher the number of times, the higher the accuracy rate). If the condition is judged X times, it is indicated in the signal data. Find a set of available data, and then calculate the cycle average based on this data.

The noise is distributed in the conditional judgment area. All the abnormal zero signals in Figure 8 do not fall within the conditional judgment area. The actual zero signal condition is not so ideal. In fact, the abnormal zero signal has considerable size. The opportunity signal will fall within the conditional judgment area. FIG. 9 is a schematic diagram of the first signal being a real signal and the noise being distributed in the condition determination area. It can be seen in Fig. 9 that the zero signal 3 falls within the conditional judgment area, and if the judgment method is as described above, an erroneous judgment will occur. As shown in step 3 in Figure 9, the zero signal 4 in the figure is a real signal. However, due to the misjudgment of the zero signal 3, the zero signal 4 will be considered as an error signal. Therefore, it is necessary to increase the function of re-inspection in the judgment area. Handle this problem.

10-11 are schematic diagrams showing the function of adding a signal re-inspection in the judgment area according to the present invention. In FIG. 10, when the flow proceeds to step 2, although the judgment formula is qualified, the reference datum point is not immediately moved from point 1 to point 3, but the test of step 3 is added again, if the test is performed. If the condition is met, the test result at the time of step 3 is stored in a temporary data matrix of the test result. Then, the reference datum point is changed from point 1 to point 3, and the subsequent judgment is continued until the end of the judgment data or the continuous conditional judgment can be obtained X times, and X is a positive integer.

When the process proceeds to step a of FIG. 10, the test result is temporarily saved. The data in the material matrix is judged. If the number of data is not zero, it indicates that two or more zero signals (including two zero signals) are in the same judgment area. At this time, the data in the test condition temporary storage data matrix is read out, and is returned to the relevant signal judgment index according to the order of storage, and the judgment is performed again. As shown in step a+1 of FIG. 11, it can be seen that the flow of the present invention does not judge the point 2 and the point 3 at this time, but directly judges the point 4, so that the noise distribution can be eliminated in the condition judgment. Problems in the area.

Figure 12-14 shows the presence of more than two noises in the judgment area by the method of the present invention. Schematic diagram of the problem. In FIG. 12, when the data re-inspection function is executed, as shown in step 3 of FIG. 12, the processing index of the current signal and related parameters are stored. After the sequence data is judged, the stored signal processing index and related parameters are restored to the relative working register, which is the same as the state of step a+1 of FIG. 13 but with FIG. Step a+1 is different. Step a+2 of FIG. 13 is in compliance with the signal re-inspection condition in the judgment area, so the program executes the data re-inspection function, just like step a+2 of FIG. 13, the current signal is again The processing indicators and related parameters are stored.

When the determination of the sequence data is completed, as shown in step b of FIG. 13, the flow of the present invention Cheng will once again judge the data in the test condition temporary data matrix. Since the number of data is not zero, the previously stored signal processing indicators and related parameters will be restored to the relative working register. Figure 14 shows the execution steps of step b+1. According to this determination method, no matter how many signals appear in the judgment area, this method can be used to judge one by one without losing any signal. Judgment.

Although the flow of the present invention has a Line judgment, but which sequence is the correct sequence must be discussed. As can be seen from the schematic diagrams of FIGS. 9 and 12, point 3 of FIG. 9 and point 3 and point 5 of FIG. 12 are not in the central area of the determination area, which indicates step 2 of FIG. 9 and step 2 and FIG. 14 of FIG. Although the judgment value of step b+1 meets the judgment condition, it must be much smaller or farther than the signal period value. Therefore, using this feature, a minimum variation determination procedure can be performed for each data sequence that meets the conditional judgment: first, the difference value between each signal point and the point is calculated; and the amount of change of each difference value is calculated; The absolute value of the change is cumulatively calculated; the data sequence group with the smallest change is the most likely real signal sequence. According to this judgment procedure, even if there is noise mixing, the impact on the overall judgment of the system is limited, because the amount of change must be limited, if two or more signals appear in the continuous two sets of signal judgment intervals. Similarly, using the signal re-inspection function, the signal can still be processed correctly, but only one set of processing indicators and related parameter storage arrays must be used to memorize the number of re-inspections and the location of related data. 15a-15d are schematic diagrams showing the occurrence of two or more signals in a plurality of judgment intervals by the method of the present invention, wherein (a) the first sequence data judgment order, (b) the second sequence data judgment order, (c) The third sequence data judgment order, and (d) the fourth sequence data judgment order.

In the judgment interval of points 3 and 5 of Figures 15a-15d, two signals appear to match The condition for judging, whether the signal judged first is abnormal or the signal after judging is abnormal, the order of signal judgment is the same, and finally the judgment is still made by the method of minimum period variation to take out the most stable signal sequence. Steps 3 and 6 of Figure 15a are both performing the signal re-inspection function in the judgment area. The judgment results of the two steps are all in compliance with the limit, so they are stored in the processing index and the related parameter storage array. The next storage array stores two pieces of data, the same After the sequence is judged, the storage array is checked to see if any data is stored inside. If there is, the data of the storage array is read in a last-in, first-out manner, and then returned to the processing index and related parameters. Then, the sequence data judgment as shown in Fig. 15b is performed again, so that the sequence data judgment is sequentially performed until the end.

When the data judgment sequence of Fig. 15b is completed, a storage array has been read previously. The data of the column, so there is still an array of data left inside, so the data of the storage array is read again, and then the processing index and related parameters are restored, and the sequence data is judged again as shown in Fig. 15c. At this time, point 1 is the reference datum point. 4 is a relative change signal point. When the reference reference point moves to point 4, since step b+2 also meets the judgment condition, the signal re-inspection function in the judgment area is performed again, as shown in step b+3 of Fig. 15c. Since the judgment of step b+3 is in accordance with the condition, the condition of the re-inspection is stored, and after the end of the sequence judgment, the processing index and related parameters are restored, and the sequence judgment is performed again, as shown in Fig. 15d. . It can be seen from Fig. 15a to 15d that if there are more repeated signals in the judgment area, the more times the data sequence is repeatedly executed, the more the number of signal abnormalities will be judged one by one, but only executed. More times. Therefore, as long as the minimum period variation amount mentioned above is used, the sequence data of the minimum variation can be effectively found to be used for the frequency calculation or the zero point judgment program, thereby achieving the function of signal debugging and stabilization.

The first zero signal is the sequence signal processing of the noise: FIG. 16-19 is a schematic diagram of the sequence signal of the first zero signal being the noise processed by the method of the present invention. As shown in FIG. 16, the present invention first uses the first zero point signal as a reference point, and then determines the subsequent zero point signal from this point.

Similarly, the sequence signal arrangement is still performed in binary mode, and The different arrangements of the sequence signals are discussed. In Figure 16, the zero signal 1 is a noise, but the subsequent zero signal is a true zero signal. Step 1 is unqualified according to the present invention. The reason for the failure is that the distance between point 1 and point 2 is less than the minimum spacing limit, so the judgment is continued to the next point. Step 2 is unqualified according to the judgment of the present invention. The reason for the non-conformity is that the distance between point 1 and point 3 is greater than the maximum limit period. Since the zero signal is a periodic signal, there must be two true zero signals in the maximum limit period. It can be seen that the first zero signal point must be noise. After knowing that the first zero signal point is noise, the present invention will give up this point, move the first reference signal point to point 2, and then perform subsequent signal judgment, and point 2 is the first reference signal point. The way of judging is just like the signal judgment method described above, as shown in FIG.

The first zero signal is noise and is located at the sequence signal in the conditional decision area. Reason: As can be seen from Figure 18, the zero signal 1 is a noise but falls within the conditional judgment area. Since the judgment result of the step 1 is that the condition is met, the signal reference point is moved to the signal point 2, and then the subsequent sequence signal judgment is performed. If the sequence signal judgment is continued under this condition, a sequence of signals corresponding to the judgment condition will be generated. However, when the minimum period variation of the group signal is judged, the total value of the variation will be greater than the variation of the correct signal. The total value, so the sequence signal can be regarded as the wrong sequence signal and removed.

In Figure 19, the first zero signal is discussed first in the conditional judgment area, but the second The zero signal does not fall within the conditional judgment area. The signal judging method of Fig. 19 is like the above-described judging method, and the signal sequence generated by it is removed in the minimum period variation amount judging program.

The first zero signal is noise and the continuous noise falls within the conditional judgment area.

20-21 illustrate the processing of the first zero point noise in the conditional decision area and the second zero point noise also falls within the conditional decision area. As shown in the figure, when the condition of the zero signal 1 and the zero signal 2 is judged, the result is qualified. The present invention moves the judgment point from the zero signal 1 to the zero signal 2, but the judgment area is first performed before this. Internal signal retest function. After performing the signal re-inspection function in the judgment area, it can be seen that the zero signal 1 and the zero signal 3 are in compliance with the condition, so the condition condition is memorized, and after the judgment of the sequence data is finished, the sequence is again performed according to the condition of the memory. Judgment of the data.

It can be seen from Fig. 20 and Fig. 21 that the zero signal is processed in the same manner as long as it is a true zero signal, as long as the zero signal falls within the conditional judgment area. After that, the final zero point signal can be taken out by a minimum period variation amount judgment program. The reason is that the noise does not have a good periodicity, so even if it meets the period judgment condition, the cumulative value of the change amount must be greater than the cumulative value of the change amount of the periodic signal value, and this is one of the foundations of the operation of the present invention.

In Fig. 21, two consecutive noises are distributed in the two adjacent conditional decision areas, and successive true zero signals appear in the subsequent conditional decision area. Since the real zero signal is unlikely to appear outside the conditional decision area, it can be seen from the calculation formula (3) that Signal _n+2 must fall outside the judgment interval.

Minimum Cycle<Signal _n+1 -Signal _n <Maximum Cycle (1)

Maximum Cycle<Signal _n+2 -Signal _n (2)

Maximum Cycle-Minimum Cycle<Signal _n+2 -Signal _n+1 (3)

In the discussion of the first zero signal as noise, it can be seen that when the sequence data is judged, as long as the sequence data judgment program moves the reference judgment point, one of them is true. The zero point signal, the present invention will use this point as a starting point for subsequent sequence data judgment, and thus is similar to the judgment method that the first signal is a real signal. Here, we can conclude that if the complete sequence signal judgment cannot be completed and the signal re-inspection record in any judgment area is not memorized, the first zero-point signal is known as noise, but it can be ignored and changed to The 2020 signal is the initial judgment signal. If the second zero signal is still noise, and so on, it will be able to find a real zero signal as the starting signal for judgment.

According to the foregoing principle, the present invention proposes a zero crossover point for a digital power supply. The signal cancellation method is characterized in that the corresponding time value of the plurality of zero signals is a data source, wherein the plurality of zero signals comprise a plurality of real zero signals, and the time difference between the two adjacent real zero signals is introduced. For a method between a minimum period and a maximum period, refer to the flowchart of an embodiment of the present invention shown in FIG. 22, including the following steps: First step: selecting the first zero point of the plurality of zero signals The time point of the signal is a starting reference point; the second step: determining whether there is a time difference between the initial reference point and the starting reference point in the plurality of zero signals between the minimum period and the maximum At least one possible qualified signal between the cycles, if the determination result is no, the signal at the initial reference point is defined as noise and the starting reference point is moved to the zero point closest to the reference point And the second step is re-executed; if the determination result is yes, the initial reference point is defined as a qualified signal point and the time point of the at least one possible qualified signal is set to Less a subsequent reference point; a third step of: sequentially taking said at least one point in a subsequent reference point as the starting point and the reference data to perform the second step; a fourth step: repeatedly performing the third step and generating at least one set of time series data, wherein each of the sets of the time series data is composed of a plurality of the qualified signal points; and a fifth step: the at least one Each group of data of the group timing data respectively performs a one-cycle variation amount accumulation calculation to generate at least one period variation amount cumulative value, and uses a set of the time series data corresponding to the smallest one of the at least one period variation amount cumulative value as Timing data of the plurality of real zero signals.

Experimental results: In order to confirm the correct rate of the separation noise and the commercial analog signal of the present invention, the present invention uniformly adjusts the duty cycle of the high frequency noise simulation generating circuit to 25%, and the responsibility cycle of the low frequency noise analog generating circuit is 10 % is gradually adjusted to 90%, and is adjusted every 10%. Each step is to sample 2000 points of analog signals, and perform cumulative and average calculations. Then test and record each step. The results are shown in Table 1. Show. Where Fre is the frequency of the signal output after the error processing by the method of the present invention. Fre1 is a periodical calculation of the input analog signal directly, converted into a frequency value, and accumulated by 2000 signal points and then averaged output value. Fre2 is a point-to-point calculation of analog signals in the traditional way, and then the maximum and minimum possible period comparisons are judged. If the conditions are not met, they are discarded. If the conditions are met, the output frequency values obtained by accumulating and finally averaging are obtained. Err%, Fre1 Err% and Fre2 Err% are the error percentages of the calculated frequency of Fre, Fre1 and Fre2 and the mains simulation frequency.

Figure 23 and Figure 24 are produced according to Table 1. Figure 23 shows the untreated frequency value and the The method of the invention processes the obtained frequency value and calculates a percentage error percentage comparison with respect to the simulated mains frequency, wherein the percentage of the duty cycle of the low frequency noise analog signal is the horizontal axis, and the ordinate is the Fre Err% and Fre1 Err% error percentage. value. It can be seen from Fig. 23 that the error value of Fre1 Err% is much larger than the error value outputted by the method of the present invention, and it is understood that the proposed method of the present invention is effective. Figure 24 is a graph comparing the error percentages of the frequency values processed by the conventional period comparison method and the frequency values obtained by the method of the present invention with respect to the simulated mains frequency. It can be seen from Fig. 24 that the error value of Fre2 Err% is smaller than Fre1 Err% of Fig. 23, but is still much larger than the error value Fre Err% outputted by the method of the present invention, and it can be seen that the proposed method is longer than the conventional cycle. The alignment method can reduce the error more effectively. In addition, as can be seen from the figure, as the noise adjustment ratio increases, the unprocessed signal output value rapidly produces an error, and the signal output value processed by the conventional period limit comparison gradually deviates from the standard value, and The output signal values processed by the method of the invention still maintain a very small error. The number of frequencies processed by the present invention in the measured results The maximum error value of the value is 0.023%, which is compared with the minimum error value of 0.685% obtained by the traditional period interval judgment processing method, and the difference between the two can reach 30 times.

The present invention is the preferred embodiment, and the source is changed or modified locally. Those who are acquainted with the technical idea of the case and who are familiar with the skill can not deduct the patent right of the case.

In summary, this case shows its difference in terms of purpose, means and function. In the technical characteristics of Xizhi, and its first invention is practical and practical, it is also in compliance with the patent requirements of the invention. Please ask your review committee to inspect it and pray for an early patent.

Claims (1)

A zero-crossing point noise cancellation method for a digital power supply, wherein the corresponding time value of the plurality of zero signals is a data source, wherein the plurality of zero signals comprise a plurality of real zero signals, and any two adjacent The time difference between the real zero signals is between a minimum period and a maximum period, and the method includes the following steps: the first step: selecting a first zero point signal of the plurality of zero signals as a start point a reference point; a second step: determining whether at least one of the plurality of zero signals is after the start reference point and the time difference from the start reference point is between the minimum period and the maximum period Possible test signal. If the judgment result is no, the signal at the start reference point is defined as noise and the start reference point is moved to the time point closest to the zero point signal of the reference point, and is re-executed. The second step; if the determination result is yes, the initial reference point is defined as a qualified signal point and the time point of the at least one possible qualified signal is set to at least one subsequent reference a third step: sequentially taking one of the at least one subsequent reference point as the starting reference point and performing the second step; and fourth step: repeatedly performing the third step and performing the third step Generating at least one set of timing data, wherein each of the sets of timing data is composed of a plurality of the qualified signal points; and a fifth step: performing a one-cycle variation accumulation on each of the at least one set of time series data Calculating to generate at least one period of the variation amount cumulative value, and using the set of the time series data corresponding to the smallest one of the at least one period variation amount cumulative value as the time series data of the plurality of real zero point signals.