Method and system for automatically identifying faulty components of an electronic device

Abstract

A method to identify a faulty component of an electronic device under test is provided. The method includes receiving a layout of the electronic device that includes a placement of a plurality of probes on the electronic device under test, the plurality of probes including at least a starting probe and an ending probe defining a signal path wherein a flow of an electrical signal flows along the signal path, applying a test signal to a beginning of the signal path, activating two or more of the plurality of probes including at least the starting probe and the ending probe, receiving signal data from each of the activated probes, and evaluating a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and the ending probe in consecutive tests with the test signal until detection criteria are satisfied.

Description

METHOD AND SYSTEM FOR AUTOMATICALLY IDENTIFYING FAULTYCOMPONENTS OF AN ELECTRONIC DEVICEBACKGROUND

[0001] Testing electronic devices can pose significant challenges. One of the challenges is that faults often appear only after a long period of time after start-up and only in very specific situations. A common strategy to perform hardware tests is to position probes connected to a measurement scope on the device under test and let the device run tests. After the tests have been run for a significant amount of time, which can take hours or days, testing personnel inspects the test results. In case of failures, due to a limited number of probes a measurement scope can support, the probes may need to be repositioned in order to get greater insights into the causes of the failure and the entire process repeated. Furthermore, having many scopes analyzing the device in parallel is unfeasible due to the high cost of measurement scopes. Therefore, an alternative testing procedure that automates the process of root cause analysis of component failures of an electronic device to reduce both the time spent testing and the number of measurement scopes is sought.BRIEF SUMMARY

[0002] Methods for automatically identifying faulty components on an electronic device under test are described. According to certain embodiments, anomaly detection and root cause analysis can be accomplished by inputting a configuration of the electronic device under test (DUT) including a placement of probes on the DUT to a computing device and detecting anomalies based on monitored signals from the probes by selectably moving activated probes along the signal path. From the detected anomalies, a faulty component can be precisely identified.

[0003] According to embodiments of the testing methods, the detection of faulty components can be achieved, and the faulty component(s) replaced, in a timely manner. Additionally, according to another aspect, intermittent faults, e.g., faulty components that produce erroneous outputs only under specific and transitory circumstances which may last only a few seconds, can be detected before the fault disappears by performing the proposed testing method.

[0004] A method to identify a faulty component of an electronic device under test is provided. The method includes the steps of: receiving, by a computing device, a layout ofthe electronic device under test that includes a placement of a plurality of probes on the electronic device under test, the plurality of probes including at least a starting probe and an ending probe defining a signal path wherein a flow of an electrical signal flows along the signal path, causing to apply, by the computing device, a test signal to a beginning of the signal path, causing to activate, by the computing device, two or more of the plurality of probes including at least the starting probe and the ending probe, receiving, by the computing device, signal data from each of the activated probes, and evaluating, by the computing device a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and the ending probe in consecutive tests with the test signal until detection criteria are satisfied. The detection criteria include that an anomaly is detected in the next pair of signal data, that, based on the received layout of the electronic device, there are no probes in the signal path between the selected two activated probes, and that no untested paths in the signal path feed the test signal between the selected two activated probes.

[0005] A non-transitory computer readable storage medium, the computer-readable storage medium including instructions that when executed by a computing device perform a method is provided. The method includes the steps of: receiving, by a computing device, a layout of the electronic device under test that includes a placement of a plurality of probes on the electronic device under test, the plurality of probes including at least a starting probe and an ending probe defining a signal path wherein a flow of an electrical signal flows along the signal path, causing to apply, by the computing device, a test signal to a beginning of the signal path, causing to activate, by the computing device, two or more of the plurality of probes including at least the starting probe and the ending probe, receiving, by the computing device, signal data from each of the activated probes, and evaluating, by the computing device a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and the ending probe in consecutive tests with the test signal until detection criteria are satisfied. The detection criteria include that an anomaly is detected in the next pair of signal data, that, based on the received layout of the electronic device, there are no probes in the signal path between the selected two activated probes, and that no untested paths in the signal path feed the test signal between the selected two activated probes.

[0006] This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is notintended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0007] To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

[0008] FIG. 1 illustrates a schematic diagram of an example implementation of a system for identifying faulty components of an electronic device under test.

[0009] FIG. 2 illustrates a process flow describing the method to identify a faulty component of an electronic device under test.

[0010] FIGs. 3A-3H illustrate an example implementation of the proposed method to identify a faulty component of an electronic device under test.

[0011] FIG. 4 illustrates a schematic diagram illustrating components of a computing device.DETAILED DESCRIPTION

[0012] Methods for automatically identifying faulty components on an electronic device under test are described. According to certain embodiments, anomaly detection and root cause analysis can be accomplished by inputting a configuration of the electronic device under test (DUT) including a placement of probes on the DUT to a computing device and detecting anomalies based on monitored signals from the probes by selectably moving activated probes along the signal path. From the detected anomalies, a faulty component can be precisely identified.

[0013] The provided system analyzes an electrical signal on different points on the DUT and follows its propagation throughout the signal path. The signal path is the path that the signal follows through the inside of the DUT similarly to a path that a car follows to travel from an origin to a destination. For example, when the DUT includes multiple functional blocks, e.g., components, the connections among these blocks are the signal paths. From each path, the signal can propagate towards many other paths, and the signal from many paths can be merged in a signal path by, for example, performing arithmetic (sum, multiplication, etc.) or logic (e.g., AND, OR, XOR) operations. The overall connections among all device components form the overall signal path, with the input to the devicebeing the beginning of the signal path and the output of the device being the end of the signal path. The signal path, however, includes only paths where signals, e.g., electrical waveforms carrying information, are routed. The signal path does not include, for example, the connections used to deliver power to the components, or the connections used to reduce electromagnetic interference emissions.

[0014] FIG. 1 illustrates a schematic diagram of an example implementation of a system for identifying faulty components of an electronic device. System 100 includes a measurement scope 104, a computing device 102, a multiplexer 106, and an electronic device under test (DUT) 108.

[0015] Measurement scope 104 is an instrument that collects electrical parameters, e.g., voltage, current, power) in an electronic circuit over time. For example, the measurement scope 104 can be an oscilloscope, power meter, or a measurement module such as an Eaton Measurement Module. The scope 104 includes one or more input channels 110 that can carry signals into the scope 104. The measurement scope 104 is connected to computing device 102 which controls the scope 104. The computing device 102 can be a computer, an edge gateway, or a smartphone, for example. In some cases, the computing device 102 can be integrated into the measurement scope 104. The input channels 110 of the measurement scope 104 can be connected to another device such as multiplexer 106 via a cable.

[0016] The multiplexer 106 is a tool that can connect different probes to an input channel 110 of the scope 104. The selection of which probes to connect to each input channel 110 is instructed by computing device 102 enabling more probes to be connected than the number of input channels 110 on the scope 104.

[0017] DUT 108 is an electrical device containing components 114. DUT 108 can be, for example, a printed circuit board or multiple printed circuit boards including discrete components such as transformers, relays, and switches. Probes are tools that provide an electrical connection between measurement scope 104 and the DUT 108. The probes can each be positioned at a point in the signal path at which it is desired to measure a signal. For example, in the example of FIG. 1, probes 112a, 112b, 112c, 112d, 112e are placed before and / or after various components 114 on DUT 108. When testing DUT 108, a test signal can be applied at the beginning of signal path 116 on DUT 108. The test signal travels along the signal path on DUT 108 and exits the DUT 108 at the end of signal path 118. In some cases, the test signal can be applied directly by an instrument, such asmeasurement scope 104. In other cases, the test signal can be applied by the instrument as directed by computing device 102. The instrument can also be devices external to system 100, such as sensors, signal generators, and the like.

[0018] Probes 112a, 112b, 112c, 112d, 112e connected to multiplexer 106 can be placed on the DUT 108 by a person carrying out the testing. At least one probe, starting probe 112a, is placed at the beginning of signal path 116 and at least one probe, ending probe 112e, is placed at the end of signal path 118. In other cases, the starting probe 112a, can be placed at any desired point on DUT 108 upstream of a desired point for the ending probe 112e.

[0019] FIG. 2 illustrates a process flow describing the method to identify a faulty component of an electronic device under test. Referring to FIG. 2, method 200 can begin (202) with a system 100 set up as described in FIG. 1. Method 200 can be performed by computing device 102 and implemented by system 100. A test signal is applied (206) at the beginning of signal path 116.

[0020] Method 200 includes receiving (204) a layout of the DUT that includes a placement of a plurality of probes on the DUT. In some cases, the layout and placement of the probes is loaded into memory of computing device 102. The layout can include a circuit schematic that is loaded along with the placement of probes on the circuit into computing device 102. Alternatively, the layout and placement of the probes relative to one another on the DUT can be determined by a learning algorithm.

[0021] The learning algorithm determines the location of the probes relative to one another on the DUT by acquiring a sample of signal data from each probe and comparing signals between pairs of signals from respective probes. A similarity score can be obtained by measuring the correlation of the pair of signals for each of the compared pairs of signals. In some cases, the correlation of the pair of signals is determined by calculating a difference between the signals. Other calculations, however, can also be used for the correlation of the pair of signals. A high similarity score indicates that the signals from the compared pair are close to one another in the signal path. A low similarity score indicates that the signals from the compared pair are distant to one another. For example, a pair of signals that include a low similarity score may be in parallel signal branches.

[0022] In some cases, where the similarity calculation includes calculating a difference between the signals, and the signals include two closely related waveforms that are not a linear or affine transformation of each other, the difference calculation may not bepossible. For example, a component that detects when a signal is over a threshold can output a signal oscillating between two values, which may be significantly different from the input waveform. When complex components, such as this, are present on the DUT 108, a learning algorithm based on a library of component models can be used to determine the similarity score. To accurately use a learning algorithm based on a library of component models, each component 114 on DUT 108 should be in the library. For example, component types may include devices such as operational amplifiers, BJT transistors, MOSFET transistors, and logic gates. Instead of calculating the similarity score directly, e.g., when using complex components, the learning algorithm can process received signal data from a first probe through each model in the library and calculate the similarity between the output of each model and the signal data from a second probe. The similarity score that is assigned to the pair of probes is the minimum among the similarity scores obtained across all the models in the library. When the similarity is over a predefined threshold, the model that provided the highest similarity score is most likely the component present between the pair of probes.

[0023] Once the layout is received (204), method 200 further includes activating (208) two or more of the plurality of probes including at least the starting probe 112a and the ending probe 112e. The probes can be activated by multiplexer 106 as directed by computing device 102.

[0024] Method 200 further includes receiving (210) signal data from each of the activated probes by computing device 102 via scope 104. Multiplexer 106 is used to facilitate selection among the probes connected to the DUT 108 as inputs to the scope 104 via the input channels 110. Thus, the received signal data of each of the activated probes is received from multiplexer 106 on a corresponding input channel 110.

[0025] Method 200 further includes evaluating (212) a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and ending probe in consecutive tests with the test signal until detection criteria are satisfied. The detection criteria includes that the anomaly is detected in the next pair signal data, that there are no probes in the signal path between the selected two activated probes, and that no untested paths in the signal path feed the test signal between the selected two activated probes.

[0026] Signal data is evaluated (212) between two probes until the detection criteria are satisfied. Thus, the evaluating step can be performed in a loop until a fault is detected.When an anomaly is detected from the evaluation, a next pair of probes is selected and activated inwards of the starting probe 112a and ending probe 112e. Selecting probes inwards refers to a direction away from a root, e.g., the starting probe 112a or the ending probe 112e. For example, inwards of the starting probe 112a defines a direction according to the flow of the electrical signal on the signal path. Likewise, in this example, inwards of the ending probe 112e defines a direction against the flow of the electrical signal on the signal path.

[0027] In some cases, selecting a next pair of activated probes in the signal path inwards of the starting probe 112a and ending probe 112e includes, deactivating by multiplexer 106 the currently activated probes on the corresponding input channels, choosing two probes to activate wherein at least one of the two probes is an unexplored probe inward from the starting probe and the ending probe, and activating the two probes on corresponding input channels. The term ‘explored’ probe refers to a probe that is currently active or has been activated in the past. Thus, an unexplored probe refers to one that has not been activated previously. The previously activated and tested probes are marked as previously activated so that it can be determined when a probe is unexplored.

[0028] The selection process involves choosing two probes for testing along the signal path inward of the starting and ending probe. A check can be performed to determine if any unexplored probes exist from the currently deactivated probe following the current signal path without crossing a probe previously activated. In this case, the chosen probe can be an adjacent unexplored probe inward of the deactivated probe. When no unexplored probes exist following the current signal path without crossing a probe that has been previously activated, a further check can be performed to determine if there are any other untested paths along the signal path between the starting probe and the ending probe. An untested path includes a path of the signal path starting at starting probe 112a in which none of the probes in the path have been explored. For example, in FIG. 3A, there are three paths that make up the signal path. None of the paths have explored probes (shown as unshaded and unfilled circles). Thus, each of the three paths is untested. If there are untested paths, the probe can be chosen along the untested path.

[0029] The evaluation process of next pair of signal data in the signal path between two probes continues until the detection criteria is satisfied. The detection criteria includes that an anomaly is detected in the next pair signal data. For example, the signal data for each probe will be tested. Detecting an anomaly in the next pair signal will be indicated by thesignal data of a first probe of the pair detecting an anomaly and a second probe of the pair detecting no anomaly. The detection criteria also includes that there are no probes between the selected two activated probes. The detection criteria further includes that that no untested paths in the signal path feed the test signal between the selected two activated probes. When the detection criteria are satisfied, a fault is detected (214) and one or more faulty components can be identified as those between the selected two activated probes. A report with the selected two activated probes, the time of the identification, and the identified one or more components between the selected two activated probes can be generated.

[0030] Detecting an anomaly of the received signal data between two probes includes performing an anomaly detection algorithm on the received signal data for each of the two probes. In some cases, the anomaly detection algorithm can include comparing the signal data of each probe to a reference signal. The reference signal can be a previously captured signal at that point on the DUT when the DUT is behaving normally. In some cases, the reference signal is generated using a machine learning model trained on normal signal data, e.g., signal data acquired when the DUT is performing within typical parameters. In other cases, the reference signal is generated using the library of models of components on the electronic device. The comparison of the anomaly detection algorithm can include calculating a difference between the signal data of a respective probe and the reference signal. The difference can then be compared to a threshold. In response to the difference being greater than the threshold, an anomaly is detected. In some cases, if an anomaly is not detected, the process can restart with new data being received and tested.

[0031] FIGs. 3A-3H illustrate an example implementation of the proposed method to identify a faulty component of an electronic device under test. Referring to FIG. 3A, DUT 108 includes a plurality of components 114 (squares) and a plurality of probes 112 (circles) placed at various points along the signal path of the DUT. In order to illustrate method 200, the faulty component is shown as the filled square. Starting probe 112a is the point at which a signal, e.g., a test signal, enters DUT 108 and ending probe 112e is the point at which the signal exits DUT 108. Arrows indicate the direction of a flow of an electrical signal along the signal path. Probes, e.g., starting probe 112a and ending probe 112e, are activated so that the signal data from selected probes A and B are received by the computing device 102 via scope 104 on corresponding input channels 110. The signal data from probes A and B are evaluated using an anomaly detection algorithm as described.Computing device 102 does not detect an anomaly at probe A, e.g., starting probe 112a (shown by shaded circle), but detects an anomaly at probe B, e.g., ending probe 112e (shown by filled circle). A next pair of probes on DUT 108 are selectively activated inwards of the starting probe 112a and the ending probe 112e in the signal path for the next consecutive test. As there are probes between the selected two activated probes, the detection criteria are not satisfied. Thus, testing continues.

[0032] FIG. 3B illustrates a next pair of probes on DUT 108 that are selectively activated inwards of the starting probe 112a and the ending probe 112e in the signal path for the next (first) consecutive test. Referring to FIG. 3B, Probes A and B are selectively activated inward of the starting probe 112a and ending probe 112e for testing. First, starting probe 112a is deactivated on the corresponding input channel 110. Unexplored Probe A 112 is chosen as one of the two probes. In this case, an adjacent probe to starting probe 112a in an inward direction is chosen. Probe A is unexplored as it has not been chosen previously. Uikewise, ending probe 112e is deactivated on the corresponding input channel 110. Probe B is an adjacent probe to ending probe 112e in an inward direction. Probe B is unexplored as it has not been chosen previously. Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect a fault at probe A but detects an anomaly at probe B as shown in FIG. 3B (shown by the filled circle). As there are probes between the selected two activated probes, the detection criteria are not satisfied. Thus, testing continues.

[0033] FIG. 3C illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (second) consecutive test. Referring to FIG. 3C, Probes A and B are selectively activated inward of the starting probe 112a and ending probe 112e for testing. The previously activated probes are each deactivated on the corresponding input channel 110. In this case, an adjacent probe to previously activated probe A (which is marked as previously activated by the shading) in an inward direction is chosen as Probe A. Probe A is unexplored as it has not been chosen previously. An adjacent probe to previously activated Probe B in an inward direction is chosen as Probe B. Probe B is unexplored as it has not been chosen previously. Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect an anomaly at ProbeA nor does it detect an anomaly at Probe B. As an anomaly is not detected in the next pair signal data, the detection criteria are not satisfied. Thus, testing continues.

[0034] FIG. 3D illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (third) consecutive test. Referring to FIG. 3D, Probes A and B are selectively activated inward of the starting probe 112a and ending probe 112e for testing. The previously activated probes are each deactivated on the corresponding input channel 110. In this case, no unexplored probes exist from the previously activated probes following the current signal path without crossing a probe that has already been activated (see FIG. 3C). Thus, a further check can be performed to determine if there are any other untested paths along the signal path between the starting probe 112a and the ending probe 112e. An adjacent unexplored probe to starting probe 112a in an untested path in an inward direction is chosen as Probe A. Probe A is unexplored as it has not been chosen previously. An adjacent probe to previously activated Probe B in an inward direction is chosen as Probe B. Probe B is unexplored as it has not been chosen previously. Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect an anomaly at Probe A 112 but it does detect an anomaly at Probe B. As there are probes between the selected two activated probes, the detection criteria are not satisfied. Thus, testing continues.

[0035] FIG. 3E illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (fourth) consecutive test. Referring to FIG. 3E, Probes A and B are selectively activated inward of the starting probe 112a and ending probe 112e for testing. The previously activated probes are each deactivated on the corresponding input channel 110. An adjacent unexplored probe to starting probe 112a in an inward direction is chosen as Probe A. Probe A is unexplored as it has not been chosen previously. When choosing a new Probe B, an unexplored probe exists inward of ending probe 112e without crossing a probe that has already been activated. Probe B is unexplored as it has not been chosen previously. Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect an anomaly at Probe A, but it does detect an anomaly at Probe B. While detection criteria that there are no probesbetween the selected two activated probes and an anomaly is detected in the next pair signal data are satisfied, there are untested paths in the signal path that feed the test signal between the selected two activated probes. Thus, testing continues.

[0036] FIG. 3F illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (fifth) consecutive test. Referring to FIG. 3F, Probes A and B are selectively activated inward of the starting probe 112a and ending probe 112e for testing. The previously selected probes are each deactivated on the corresponding input channel 110. An adjacent unexplored probe to starting probe 112a in an inward direction is chosen as Probe A. Probe A is unexplored as it has not been chosen previously. In this case, no unexplored probes exist from previously activated Probe A (see FIG. 3E) following the current signal path without crossing a probe that has already been activated. Thus, a further check can be performed to determine if there are any other untested paths along the signal path between the starting probe 112a and the ending probe 112e. An adjacent probe to starting probe 112a in an inward direction in an untested path is chosen as Probe A. Probe A is unexplored as it has not been chosen previously. Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect an anomaly at Probe A nor does it detect an anomaly at Probe B. As an anomaly is not detected in the next pair signal data, the detection criteria are not satisfied. Thus, testing continues.

[0037] FIG. 3G illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (sixth) consecutive test. Referring to FIG. 3G, Probes A and B are selected inward of the starting probe 112a and ending probe 112e for testing. The previously selected probes are each deactivated on the corresponding input channel 110. An adjacent probe to previously activated probe in an inward direction is chosen as Probe A. An adjacent probe to previously activated ending probe 112e in an inward direction is chosen as Probe B. Probe B is unexplored as it has not been chosen previously. In this case, no unexplored probes exist following the current signal path without crossing a probe that has already been activated (for Probe A). Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detectan anomaly at Probe A but detects an anomaly at Probe B. As there are probes between the selected two activated probes, the detection criteria are not satisfied. Thus, testing continues.

[0038] FIG. 3H illustrates a next pair of probes on DUT 108 that are selectively activated inward of the starting probe 112a and the ending probe 112e in the signal path for the next (seventh) consecutive test. Referring to FIG. 3G, Probe B is selected inward of the starting probe 112a and ending probe 112e for testing. Probe A remains selected. Previously selected probes are each deactivated on the corresponding input channel 110. An adjacent probe to previously activated probe in an inward direction from ending probe 112e is chosen as Probe B. Probe B is unexplored as it has not been chosen previously. In this case, no unexplored probes exist following the current signal path without crossing a probe that has already been activated (for Probe A). Both probes A and B in the selected pair are activated on corresponding input channels 110. Using the anomaly detection algorithm on the signal data from probes A and B as previously described, computing device 102 does not detect an anomaly at Probe A but detects an anomaly at Probe B. All the detection criteria are satisfied: that there are no unexplored probes between the selected two activated probes and an anomaly is detected in the next pair signal data are satisfied, there are no untested paths in the signal path feed the test signal between the selected two activated probes. Thus, testing ends, and the faulty component can be identified.

[0039] FIG. 4 illustrates a schematic diagram illustrating components of a computing device that may be used in certain implementations described herein. It should be understood that aspects of the computing device described herein are applicable to both mobile and traditional desktop computers, as well as server computers and other computer systems. Referring to FIG. 4, system 400 may represent a computing device such as but not limited to personal computer, a tablet computer, a reader, a mobile device, a personal digital assistant, a wearable computer, a smart phone, a laptop computer (notebook or netbook), a gaming device or console, an entertainment device, a hybrid computer, a desktop computer, a smart television, or an electronic whiteboard or large form-factor touchscreen as some examples. Accordingly, more or fewer elements described with respect to system 400 may be incorporated to implement a particular computing device.

[0040] Referring to FIG. 4, system 400 can include at least one processor 410, a memory 420, software 430 that includes operating system 440 and application 450, network interface 460, and user interface system 470. Processor 410 processes data according toinstructions of the software 430. The instructions of application 450, e.g., method 200, may be loaded into computing device 102 and run on or in association with the operating system 440. System 400 can further include a user interface system 470, which may include input / output (I / O) devices and components that enable communication between a user and the system 400. System 400 may also include a network interface 460 that allows the system to communicate with other computing devices, including server computing devices and other client devices, over a network.

[0041] Although the subject matter has been described in language specific to structure features and / or acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed examples of implementing claims and other equivalent features and acts are intended to be within the scope of the claims.

Claims

CLAIMSWhat is claimed is:

1. A method to identify a faulty component of an electronic device under test, the method comprising: receiving, by a computing device, a layout of the electronic device under test that includes a placement of a plurality of probes on the electronic device under test, the plurality of probes including at least a starting probe and an ending probe defining a signal path wherein a flow of an electrical signal flows along the signal path; causing to apply, by the computing device, a test signal to a beginning of the signal path; causing to activate, by the computing device, two or more of the plurality of probes including at least the starting probe and the ending probe; receiving, by the computing device, signal data from each of the activated probes; and evaluating, by the computing device, a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and the ending probe in consecutive tests with the test signal until detection criteria are satisfied, wherein the detection criteria include: that an anomaly is detected in the next pair of signal data, that, based on the received layout of the electronic device, there are no probes in the signal path between the selected two activated probes, and that no untested paths in the signal path feed the test signal between the selected two activated probes.

2. The method of claim 1, wherein the anomaly is detected in the next pair of signal data for each probe of the two probes by: calculating a difference between the signal data of the respective probe and a reference signal, comparing the difference to a threshold, and determining an anomaly of the signal data of the respective probe when the difference is greater than the threshold.

3. The method of claim 2, wherein the reference signal is generated using a machine learning model trained on normal signal data from the electronic device under test.

4. The method of claim 2, wherein the reference signal is generated using a library of models of components on the electronic device.

5. The method of claim 1, wherein received signal data from each activated probe is received from a multiplexer on a corresponding input channel of a measurement scope.

6. The method of claim 5, wherein the two probes are selectively activated inward between the starting probe and the ending probe by: causing to deactivate currently activated probes on the corresponding input channels; choosing two probes to activate wherein at least one of the two probes is an unexplored probe inward from the starting probe and the ending probe; and causing to activate the two probes on corresponding input channels.

7. The method of claim 1, wherein the placement of the plurality of probes on the electronic device is determined by a learning algorithm based on a similarity score determined from signal data received from a pair of activated probes.

8. The method of claim 7, wherein the similarity score is determined by measuring a correlation between the signal data from the pair of activated probes.

9. The method of claim 7, wherein the similarity score is determined using a library of models of components on the electronic device.

10. The method of claim 1, wherein when the detection criteria are satisfied, identifying the faulty component as the one or more components between the selected two activated probes.

11. The method of claim 10, further comprising generating a report with the selected two activated probes, a current time, and the identified faulty component.

12. A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a computing device perform a method, the method comprising: receiving, by the computing device, a layout of the electronic device under test that includes a placement of a plurality of probes on the electronic device under test, the plurality of probes including at least a starting probe and an ending probe defining a signal path wherein a flow of an electrical signal flows along the signal path; causing to apply, by the computing device, a test signal to a beginning of the signal path; causing to activate, by the computing device, two or more of the plurality of probes including at least the starting probe and the ending probe; receiving, by the computing device, signal data from each of the activated probes; and evaluating, by the computing device, a next pair of signal data in the signal path between two probes that are selectively activated inward between the starting probe and the ending probe in consecutive tests with the test signal until detection criteria are satisfied, wherein the detection criteria include: that an anomaly is detected in the next pair of signal data, that, based on the received layout of the electronic device, there are no unexplored probes in the signal path between the selected two activated probes, and that no untested paths in the signal path feed the test signal between the selected two activated probes.

13. The non-transitory computer readable storage medium of claim 12, wherein the anomaly is detected in the next pair of signal data for each probe of the two probes by: calculating a difference between the signal data of the respective probe and a reference signal, comparing the difference to a threshold, and determining an anomaly of the signal data of the respective probe when the difference is greater than the threshold.

14. The non-transitory computer readable storage medium of claim 13, wherein the reference signal is generated using a machine learning model trained on normal signal data from the electronic device under test.

15. The non-transitory computer readable storage medium of claim 13, wherein the reference signal is generated using a library of models of components on the electronic device.

16. The non-transitory computer readable storage medium of claim 12, wherein received signal data from each activated probe is received from a multiplexer on a corresponding input channel of a measurement scope.

17. The non-transitory computer readable storage medium of claim 16, wherein the two probes are selectively activated inward between the starting probe and the ending probe by: causing to deactivate currently activated probes on the corresponding input channels; choosing two probes to activate wherein at least one of the two probes is an unexplored probe inward from the starting probe and the ending probe; and causing to activate the two probes on corresponding input channels.

18. The non-transitory computer readable storage medium of claim 12, wherein the placement of the plurality of probes on the electronic device is determined by a learning algorithm based on a similarity score determined from signal data received from a pair of activated probes.

19. The non-transitory computer readable storage medium of claim 18, wherein the similarity score is determined by measuring a correlation between the signal data from the pair of activated probes.

20. The non-transitory computer readable storage medium of claim 18, wherein the similarity score is determined using a library of models of components on the electronic device.

Images