A radio frequency coaxial probe assembly and a pulse characteristic measurement method thereof

Abstract

The present invention relates to the field of radio frequency detection technology, and in particular to a radio frequency coaxial probe component and a pulse characteristic measurement method thereof, wherein the component comprises: an SMA female connector, a heat shrink tubing, a radio frequency coaxial cable, a probe mounting flange and a probe component; the method comprises: obtaining a plurality of abnormal historical detection data of the radio frequency coaxial probe on the same inspected workpiece, performing cluster analysis based on the abnormal data to obtain cluster analysis results, generating a plurality of data sets based on the cluster analysis results, and dividing each data set into a training set and a test set, training a model with the training set, verifying the accuracy of the model with the test set, performing data transformation on the abnormal data based on a training model whose accuracy does not meet the standards to improve the accuracy of the training model, and correcting errors caused by the detection process based on the training model whose accuracy meets the standards, thereby improving the detection accuracy of the radio frequency coaxial probe.

Description

Technical Field

[0001] The present invention relates to the field of radio frequency detection technology, and in particular to a radio frequency coaxial probe assembly and a pulse characteristic measurement method thereof. Background Art

[0002] With the rapid development of communication technology, the frequency parameters of electronic equipment are constantly improving, and the requirements for the accuracy and stability of test equipment are becoming increasingly stringent. As an important part of electronic test equipment, RF coaxial probe components can accurately measure RF signals in electronic circuits such as silicon wafers and integrated circuit chips, while the pulse characteristic measurement method further improves its performance and application range in the field of high-frequency signal testing.

[0003] Chinese patent publication number: CN110261651A discloses a test device and test method for a radio frequency probe, which relates to the field of radio frequency detection technology. The technical key points include a radio frequency coaxial housing for the radio frequency probe to be inserted and tested, a test adapter installed at the axial end of the radio frequency coaxial housing for obtaining the output signal of the radio frequency probe, and a host control end electrically connected to the test adapter. The radio frequency coaxial housing includes an inner shaft cylinder with a mounting hole for the radio frequency probe to be inserted, an outer bottom cylinder sleeved and fixed on one side of the inner shaft cylinder, an outer top cylinder slidably sleeved on one side of the inner shaft cylinder, and an elastic member arranged between the outer bottom cylinder and the outer top cylinder with the elasticity to drive the outer top cylinder away from the outer bottom cylinder; a limiting pair is arranged between the outer top cylinder and the inner shaft cylinder to limit the outer top cylinder from sliding out of the inner shaft cylinder; when the outer top cylinder is limited by the limiting pair, the axial end face of the outer top cylinder away from the radio frequency adapter protrudes from the axial end face of the inner shaft cylinder away from the radio frequency adapter. The present invention has the advantage of comprehensively measuring the performance of the radio frequency probe.

[0004] It can be seen that the testing device and testing method of the RF probe have the following problems: while testing the RF probe, the invention changes the connection status of the RF probe in the mounting hole, thereby achieving a more comprehensive detection of the RF probe performance, but the invention lacks monitoring during the testing of the RF probe and cannot guarantee the detection accuracy of the RF probe. Summary of the invention

[0005] To this end, the present invention provides a radio frequency coaxial probe assembly and a pulse characteristic measurement method thereof, so as to overcome the problem in the prior art that the insertion process of the radio frequency coaxial probe is lack of monitoring, resulting in the test results being affected by the insertion process, thereby resulting in low test accuracy.

[0006] To achieve the above object, on the one hand, the present invention provides a method for measuring pulse characteristics of a radio frequency coaxial probe, comprising:

[0007] Acquire several historical test data of the same test workpiece by the radio frequency coaxial probe, and determine abnormal data based on the return loss and the second test result;

[0008] Acquire several process data of the probe insertion process corresponding to the abnormal data and record them as abnormal process data, perform cluster analysis on the abnormal data to obtain cluster analysis results, wherein the historical detection data includes parameters of the inspected workpiece, process data of the radio frequency coaxial probe insertion process and the first detection result, and the process data includes voltage, current, insertion depth and insertion angle of the radio frequency coaxial probe during the insertion process;

[0009] Based on the cluster analysis result, the abnormal data, the abnormal process data and the corresponding first detection result are divided into a plurality of data sets, and the data in each data set is divided into a training set and a test set in a preset ratio;

[0010] Input the training set into the model to obtain the training model, and input the test set into the training model to determine whether the accuracy of the model meets the standard based on the overlap of data features;

[0011] Based on the training model whose accuracy does not meet the standard, determining the data transformation mode in the data set according to the continuity of the pulse characteristic curve of the probe corresponding to the data mutation point;

[0012] Obtain process data of the insertion process of the RF coaxial probe when inspecting the inspected workpiece, and modify the inspection parameters of the probe based on the analysis results of the training model whose accuracy meets the standard;

[0013] The first detection result is the result of using a radio frequency coaxial probe to detect the inspected workpiece, and whether the inspected workpiece is qualified. The second detection result is the result of verifying whether the first detection result is correct.

[0014] Further, the process of determining abnormal data based on the return loss and the second detection result includes:

[0015] Determining, based on the historical detection data, detection data corresponding to the incorrect first detection result;

[0016] comparing the return loss with a preset return loss based on the detection data;

[0017] Based on the comparison result that the return loss is greater than the preset return loss, it is determined that all data of the detection process corresponding to the return loss is abnormal data.

[0018] Furthermore, the process of generating a data set based on the abnormal data includes:

[0019] Obtaining the reason why the abnormal data causes the inspected workpiece to be unqualified;

[0020] Classify abnormal data corresponding to the same reason into the same category;

[0021] Abnormal process data of all probe insertion processes corresponding to the same type of data are acquired, and the abnormal process data and the same type of data are set as a data set.

[0022] Furthermore, the process of determining whether the accuracy of the training model meets the standard based on the data feature overlap includes:

[0023] Comparing the data feature coincidence degree with a preset data feature coincidence degree;

[0024] If the data feature overlap is less than the preset data feature overlap, then it is determined that the accuracy of the model does not meet the standard;

[0025] If the data feature overlap is greater than or equal to the preset data feature overlap, it is determined that the accuracy of the model meets the standard.

[0026] Furthermore, the process of determining the data mutation point includes:

[0027] comparing the pulse characteristic curve with a corresponding standard pulse curve;

[0028] Calculate the difference in slopes between corresponding points of two curves;

[0029] If the slope difference is greater than a preset slope difference, the point corresponding to the slope difference is determined to be a data mutation point.

[0030] Further, the process of determining the continuity of the pulse characteristic curve based on the data mutation point includes:

[0031] Based on the data mutation point, a straight line is drawn on the pulse characteristic curve parallel to the horizontal axis, a span between the straight line and the intersection point of the wave where the data mutation point is located is obtained, and the span is compared with a preset span;

[0032] If the span is greater than the preset span, determining that the continuity of the pulse characteristic curve is a first continuity;

[0033] If the span is less than or equal to the preset span, the continuity of the pulse characteristic curve is determined to be the second continuity.

[0034] Furthermore, based on the first continuity, normalization is used to transform the abnormal data into data of the same dimension.

[0035] Further, data flipping is adopted based on the second continuity to enhance the abnormal data.

[0036] Furthermore, based on the training model whose accuracy meets the standard, process data of the coaxial probe during the detection process is obtained, and the detection parameters are corrected based on the process data to obtain accurate detection results.

[0037] On the other hand, the present invention also provides a radio frequency coaxial probe assembly using the radio frequency coaxial probe pulse characteristic measurement method, comprising:

[0038] SMA female connector, used to connect test instruments;

[0039] A heat shrink tubing connected to the SMA female connector to enhance the connection strength;

[0040] A radio frequency coaxial cable, connected to the heat shrink tubing, for transmitting signals;

[0041] A probe mounting flange, which is connected to the radio frequency coaxial cable and is used to connect the probe assembly and the radio frequency coaxial cable;

[0042] The probe assembly is connected to the probe mounting flange by a detachable thread, and includes a probe compression spring, a probe outer conductor and a probe inner conductor for transmitting and receiving signals. The probe inner conductor includes a center needle, a spring, an inner guide center needle and a supporting insulator.

[0043] Compared with the prior art, the beneficial effect of the present invention lies in that the present invention uses historical anomaly detection data to train the model, establishes a training model for inaccurate detection results caused by improper insertion operation of the RF coaxial probe during the detection process, and corrects the detection results to improve the detection accuracy of the RF coaxial probe.

[0044] Furthermore, the present invention determines abnormal data through return loss. Return loss can accurately reflect the performance state of the RF coaxial probe when transmitting RF signals. When the probe has an abnormality, such as poor contact, it will lead to increased signal reflection, which in turn causes the return loss value to change, resulting in deviations in the detection results. The data with abnormal return loss and normal detection results are recorded as abnormal data, and all data about the detection process corresponding to the abnormal data are obtained. Cluster analysis is performed on the data. Cluster analysis can discover hidden structures and patterns in the data. By calculating the similarity between data objects, similar data points are classified into the same data set to reveal the inherent correlation of the data, thereby improving the generation benchmark of the data set, and further improving the accuracy of the model trained based on the data set, thereby improving the detection accuracy of the RF coaxial probe.

[0045] Furthermore, the present invention verifies the training model through a test set to determine the accuracy of the training model, and uses data transformation to improve the prediction accuracy, stability and generalization ability of the training model whose accuracy does not meet the standards, thereby further improving the accuracy of the model trained based on the data set and improving the detection accuracy of the RF coaxial probe.

[0046] Furthermore, the present invention determines the data transformation method according to the span corresponding to the data mutation point. The larger the span, the greater the error. The regularization method is used to transform the data according to the large span, and overfitting is prevented by limiting the complexity of the model. The data enhancement method is used to transform the data according to the small span, which can improve the model accuracy without increasing the calculation cost, thereby further improving the detection accuracy of the RF coaxial probe.

[0047] Furthermore, the probe assembly in the present invention is a vulnerable assembly, and by arranging the probe assembly to be connected to the probe mounting flange through a detachable thread, the probe assembly can be replaced according to the test usage, thereby further improving the detection accuracy of the RF coaxial probe. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a schematic diagram of the structure of a radio frequency coaxial probe assembly according to an embodiment of the present invention;

[0049] Figure 2 This is a flow chart of a method for measuring pulse characteristics of a radio frequency coaxial probe according to an embodiment of the present invention;

[0050] Figure 3 A flowchart for determining abnormal data according to an embodiment of the present invention;

[0051] Figure 4 A flow chart for determining the accuracy of a training model for an embodiment of the present invention;

[0052] In the figure: 1. SMA female connector; 2. Heat shrink tubing; 3. RF coaxial cable; 4. Probe mounting flange; 5. Probe compression spring; 6. Probe outer conductor; 7. Center pin; 8. Spring; 9. Inner guide center pin; 10. Support insulator. DETAILED DESCRIPTION

[0053] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0054] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0055] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0056] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0057] See also Figure 1-Figure 4 As shown, Figure 1 This is a schematic diagram of the structure of a radio frequency coaxial probe assembly according to an embodiment of the present invention; Figure 2 This is a flow chart of a method for measuring pulse characteristics of a radio frequency coaxial probe according to an embodiment of the present invention; Figure 3 A flowchart for determining abnormal data according to an embodiment of the present invention; Figure 4 A flow chart for determining the accuracy of a training model according to an embodiment of the present invention.

[0058] The embodiment of the present invention provides a method for measuring pulse characteristics of a radio frequency coaxial probe, which is characterized by comprising:

[0059] Step S1, obtaining a number of historical detection data of the same type of workpiece to be detected by the radio frequency coaxial probe, and determining abnormal data based on the return loss and the second detection result;

[0060] Step S2, obtaining a number of process data of the probe insertion process corresponding to the abnormal data and recording them as abnormal process data, performing cluster analysis on the abnormal data to obtain a cluster analysis result, wherein the historical detection data includes parameters of the inspected workpiece, process data of the RF coaxial probe insertion process and a first detection result, and the process data includes voltage, current, insertion depth and insertion angle of the RF coaxial probe during the insertion process;

[0061] Step S3, dividing the abnormal data, abnormal process data and the corresponding first detection result into a plurality of data sets based on the cluster analysis result, and dividing the data in each data set into a training set and a test set in a preset ratio;

[0062] Step S4, inputting the training set into the model to obtain the training model, and inputting the test set into the training model to determine whether the accuracy of the model meets the standard based on the overlap of data features;

[0063] Step S5, based on the training model whose accuracy does not meet the standard, determining the data transformation mode in the data set according to the continuity of the pulse characteristic curve of the probe corresponding to the data mutation point;

[0064] Step S6, obtaining process data of the insertion process of the radio frequency coaxial probe when detecting the workpiece to be detected, and modifying the detection parameters of the probe based on the analysis results of the training model whose accuracy meets the standard;

[0065] The first detection result is the result of using a radio frequency coaxial probe to detect the inspected workpiece, and whether the inspected workpiece is qualified. The second detection result is the result of verifying whether the first detection result is correct.

[0066] Specifically, the workpiece to be inspected includes but is not limited to silicon wafers, tube dies and open microchips.

[0067] Specifically, the preset ratio is: training set / test set = 3 / 7.

[0068] Specifically, the process of determining abnormal data based on the return loss and the second detection result includes:

[0069] Determining, based on the historical detection data, detection data corresponding to the incorrect first detection result;

[0070] comparing the return loss with a preset return loss based on the detection data;

[0071] Determining that all data of the detection process corresponding to the return loss is abnormal data based on the comparison result that the return loss is greater than the preset return loss;

[0072] Based on the comparison result that the return loss is less than or equal to the preset return loss, it is determined that all data of the detection process corresponding to the return loss is normal data.

[0073] Specifically, the preset return loss value range is set to [-70dB, -20dB], and -30dB is preferred in the embodiment of the present invention.

[0074] Specifically, the process of generating a data set based on the abnormal data includes:

[0075] Obtaining the reason why the abnormal data causes the inspected workpiece to be unqualified;

[0076] Classify abnormal data corresponding to the same reason into the same category;

[0077] Abnormal process data of all probe insertion processes corresponding to the same type of data are acquired, and the abnormal process data and the same type of data are set as a data set.

[0078] Specifically, the consistency of the detection results means that the problems of the equipment detected by the RF coaxial probe are consistent, for example, all data corresponding to the short circuit situations of all the inspected workpieces are classified as the same type of data.

[0079] Specifically, the process of determining whether the accuracy of the training model meets the standard based on the data feature overlap includes:

[0080] Comparing the data feature coincidence degree with a preset data feature coincidence degree;

[0081] If the data feature overlap is less than the preset data feature overlap, then it is determined that the accuracy of the model does not meet the standard;

[0082] If the data feature overlap is greater than or equal to the preset data feature overlap, it is determined that the accuracy of the model meets the standard.

[0083] Specifically, the value range of the preset data feature overlap is set to [80%, 100%], and 95% is preferred in the embodiment of the present invention.

[0084] Specifically, the process of determining the data mutation point includes:

[0085] comparing the pulse characteristic curve with a corresponding standard pulse curve;

[0086] Calculate the difference in slopes between corresponding points of two curves;

[0087] If the slope difference is greater than a preset slope difference, the point corresponding to the slope difference is determined to be a data mutation point.

[0088] Specifically, the standard pulse curve is a pulse curve corresponding to normal test results of equipment with the same performance.

[0089] Specifically, the preset slope difference value range is set to [0.1, 0.5], and 0.2 is preferred in the embodiment of the present invention.

[0090] Specifically, the process of determining the continuity of the pulse characteristic curve based on the data mutation point includes:

[0091] Based on the data mutation point, a straight line is drawn on the pulse characteristic curve parallel to the horizontal axis, a span between the straight line and the intersection point of the wave where the data mutation point is located is obtained, and the span is compared with a preset span;

[0092] If the span is greater than the preset span, determining that the continuity of the pulse characteristic curve is a first continuity;

[0093] If the span is less than or equal to the preset span, the continuity of the pulse characteristic curve is determined to be the second continuity.

[0094] Specifically, the value range of the preset span is set to [0.1 μs, 9 μs], and 7 μs is preferred in the embodiment of the present invention.

[0095] Specifically, based on the first continuity, normalization is used to transform the abnormal data into data of the same dimension.

[0096] Specifically, the process of data transformation using normalization includes:

[0097] The original abnormal data is scaled to the same scale range [0, 1] to obtain the transformed abnormal data;

[0098] Transforming the transformation abnormal data normalization formula;

[0099] Specifically, the scaling ratio is determined according to the data value, and is not specifically limited.

[0100] Specifically, the normalization transformation can reduce the scale difference of the data and avoid the deviation caused by the model being biased towards data with larger values ​​and ignoring other important data, thereby improving the overall performance of the model.

[0101] Specifically, a data enhancement method is adopted based on the second continuity to transform the abnormal data.

[0102] Specifically, the method of data enhancement is to take the waves in the pulse characteristic curve as the horizontal axis, mirror the waves below the horizontal axis to the top of the horizontal axis, obtain new feature data based on the new coordinate curve, and enhance the data containing data mutation points, thereby specifically increasing the diversity and quantity of training data, thereby effectively improving the accuracy of the training model.

[0103] Specifically, based on a training model that meets the accuracy standard, process data of the coaxial probe during the detection process is obtained, and based on the process data, the detection parameters are corrected to obtain accurate detection results.

[0104] Specifically, the detection parameters are corrected, for example, the detection voltage of the RF coaxial probe is increased or decreased according to the analysis results of the training model on the process data. The specific details are not limited and are determined according to the analysis results of the training model.

[0105] The embodiment of the present invention further provides a radio frequency coaxial probe assembly using a radio frequency coaxial probe pulse characteristic measurement method, comprising:

[0106] SMA female connector 1, used to connect the test instrument;

[0107] A heat shrink tubing 2 connected to the SMA female connector 1 to enhance the connection strength;

[0108] A radio frequency coaxial cable 3, connected to the heat shrink tubing 2, for transmitting signals;

[0109] A probe mounting flange 4, which is connected to the RF coaxial cable 3 and is used to connect the probe assembly and the RF coaxial cable 3;

[0110] The probe assembly is connected to the probe mounting flange 4 by a detachable thread, and includes a probe compression spring 8, a probe outer conductor 6 and a probe inner conductor for transmitting and receiving signals. The probe inner conductor includes a center needle 7, a spring 8, an inner center needle 9 and a supporting insulator 10.

[0111] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A method for measuring pulse characteristics of a radio frequency coaxial probe, characterized in that: include: Acquire several historical test data of the same test workpiece by the radio frequency coaxial probe, and determine abnormal data based on the return loss and the second test result; Acquire several process data of the probe insertion process corresponding to the abnormal data and record them as abnormal process data, perform cluster analysis on the abnormal data to obtain cluster analysis results, wherein the historical detection data includes parameters of the inspected workpiece, process data of the radio frequency coaxial probe insertion process and the first detection result, and the process data includes voltage, current, insertion depth and insertion angle of the radio frequency coaxial probe during the insertion process; Based on the cluster analysis result, the abnormal data, the abnormal process data and the corresponding first detection result are divided into a plurality of data sets, and the data in each data set is divided into a training set and a test set in a preset ratio; Input the training set into the model to obtain the training model, and input the test set into the training model to determine whether the accuracy of the model meets the standard based on the overlap of data features; Based on the training model whose accuracy does not meet the standard, determining the data transformation mode in the data set according to the continuity of the pulse characteristic curve of the probe corresponding to the data mutation point; Obtain process data of the insertion process of the RF coaxial probe when inspecting the inspected workpiece, and modify the inspection parameters of the probe based on the analysis results of the training model whose accuracy meets the standard; The first detection result is the result of using a radio frequency coaxial probe to detect the inspected workpiece to determine whether the inspected workpiece is qualified, and the second detection result is the result of verifying whether the first detection result is correct; The process of generating a data set based on the abnormal data includes: Obtaining the reason why the abnormal data causes the inspected workpiece to be unqualified; Classify abnormal data corresponding to the same reason into the same category; Acquire abnormal process data of all probe insertion processes corresponding to the same type of data, and set the abnormal process data and the same type of data as a data set; The process of determining the data mutation point includes: comparing the pulse characteristic curve with a corresponding standard pulse curve; Calculate the difference in slopes between corresponding points of two curves; If the slope difference is greater than a preset slope difference, determining the point corresponding to the slope difference as a data mutation point; The process of determining the continuity of the pulse characteristic curve based on the data mutation point includes: Based on the data mutation point, a straight line is drawn on the pulse characteristic curve parallel to the horizontal axis, a span between the straight line and the intersection point of the wave where the data mutation point is located is obtained, and the span is compared with a preset span; If the span is greater than the preset span, determining that the continuity of the pulse characteristic curve is a first continuity; If the span is less than or equal to the preset span, determining that the continuity of the pulse characteristic curve is a second continuity; Based on the first continuity, the abnormal data is transformed into data of the same dimension by normalization; Data flipping is employed based on the second continuity to enhance the abnormal data.

2. The method for measuring pulse characteristics of a radio frequency coaxial probe according to claim 1, characterized in that: The process of determining abnormal data based on the return loss and the second detection result includes: Determining, based on the historical detection data, detection data corresponding to the incorrect first detection result; comparing the return loss with a preset return loss based on the detection data; Based on the comparison result that the return loss is greater than the preset return loss, it is determined that all data of the detection process corresponding to the return loss is abnormal data.

3. The method for measuring pulse characteristics of a radio frequency coaxial probe according to claim 2, characterized in that: The process of determining whether the accuracy of the training model meets the standard based on the data feature overlap includes: Comparing the data feature coincidence degree with a preset data feature coincidence degree; If the data feature overlap is less than the preset data feature overlap, then it is determined that the accuracy of the model does not meet the standard; If the data feature overlap is greater than or equal to the preset data feature overlap, it is determined that the accuracy of the model meets the standard.

4. The method for measuring pulse characteristics of a radio frequency coaxial probe according to claim 3, characterized in that: Based on the training model that meets the accuracy standard, process data of the coaxial probe during the detection process is obtained, and the detection parameters are corrected based on the process data to obtain accurate detection results.

5. A radio frequency coaxial probe assembly using the radio frequency coaxial probe pulse characteristic measurement method according to any one of claims 1 to 4, characterized in that: include: SMA female connector, used to connect test instruments; A heat shrink tubing connected to the SMA female connector to enhance the connection strength; A radio frequency coaxial cable, connected to the heat shrink tubing, for transmitting signals; A probe mounting flange, which is connected to the radio frequency coaxial cable and is used to connect the probe assembly and the radio frequency coaxial cable; The probe assembly is connected to the probe mounting flange by a detachable thread, and comprises a probe compression spring, a probe outer conductor and a probe inner conductor for transmitting and receiving signals. The probe inner conductor comprises a center needle, a spring, an inner guide center needle and a supporting insulator.

Images