CN116500426B - Method for calibrating high-temperature test of semiconductor detection equipment - Google Patents
Method for calibrating high-temperature test of semiconductor detection equipment Download PDFInfo
- Publication number
- CN116500426B CN116500426B CN202310769417.3A CN202310769417A CN116500426B CN 116500426 B CN116500426 B CN 116500426B CN 202310769417 A CN202310769417 A CN 202310769417A CN 116500426 B CN116500426 B CN 116500426B
- Authority
- CN
- China
- Prior art keywords
- prediction
- time
- value
- preset
- maintenance
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 69
- 238000000034 method Methods 0.000 title claims abstract description 62
- 238000012360 testing method Methods 0.000 title claims abstract description 50
- 239000004065 semiconductor Substances 0.000 title claims abstract description 48
- 239000000523 sample Substances 0.000 claims abstract description 51
- 238000003780 insertion Methods 0.000 claims abstract description 6
- 230000037431 insertion Effects 0.000 claims abstract description 6
- 238000011835 investigation Methods 0.000 claims abstract description 5
- 230000000875 corresponding effect Effects 0.000 claims description 109
- 238000012423 maintenance Methods 0.000 claims description 83
- 230000035484 reaction time Effects 0.000 claims description 18
- 238000004364 calculation method Methods 0.000 claims description 15
- 238000012549 training Methods 0.000 claims description 12
- 238000007689 inspection Methods 0.000 claims description 8
- 230000009290 primary effect Effects 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000012797 qualification Methods 0.000 claims description 5
- 238000005259 measurement Methods 0.000 claims description 3
- 230000009291 secondary effect Effects 0.000 claims description 2
- 238000011156 evaluation Methods 0.000 abstract description 2
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000005457 optimization Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 2
- 230000008602 contraction Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000013467 fragmentation Methods 0.000 description 1
- 238000006062 fragmentation reaction Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/28—Testing of electronic circuits, e.g. by signal tracer
- G01R31/2851—Testing of integrated circuits [IC]
- G01R31/2886—Features relating to contacting the IC under test, e.g. probe heads; chucks
- G01R31/2887—Features relating to contacting the IC under test, e.g. probe heads; chucks involving moving the probe head or the IC under test; docking stations
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/10—Measuring as part of the manufacturing process
- H01L22/14—Measuring as part of the manufacturing process for electrical parameters, e.g. resistance, deep-levels, CV, diffusions by electrical means
-
- H—ELECTRICITY
- H01—ELECTRIC ELEMENTS
- H01L—SEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
- H01L22/00—Testing or measuring during manufacture or treatment; Reliability measurements, i.e. testing of parts without further processing to modify the parts as such; Structural arrangements therefor
- H01L22/20—Sequence of activities consisting of a plurality of measurements, corrections, marking or sorting steps
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention belongs to the technical field of semiconductor detection, in particular to a method for high-temperature test calibration of semiconductor detection equipment, which comprises the steps of performing prediction operation through an AI model to predict the position of a probe tip and the overall position of a wafer chip, setting compensation precision, generating corresponding prediction compensation operation instructions and sending the corresponding prediction compensation operation instructions to a movement control part, wherein the movement control part drives a relatively moving part of a probe based on the needle position, the position of the wafer chip and the corresponding prediction compensation operation instructions, so that the needle insertion is more accurate, the precise positioning and the offset of a probe test position caused by the change of temperature are avoided, the accuracy of a test result is obviously improved, the damage to the wafer chip in the test process is effectively reduced, and the efficiency evaluation and the condition analysis of the operation process are performed through the prediction analysis process, so that corresponding management personnel can timely conduct reason investigation and judgment and make corresponding countermeasures in a targeted manner, and the high-efficiency stable operation of the semiconductor detection equipment is ensured.
Description
Technical Field
The invention relates to the technical field of semiconductor detection, in particular to a method for calibrating high-temperature test of semiconductor detection equipment.
Background
In a probe testing apparatus using a probe in contact with a semiconductor chip, there is a need for a high and low temperature testing environment, and as the temperature changes, each chip of the probe tip and the wafer generates thermal expansion and contraction to different extents due to different thermal expansion coefficients;
because the precision required by the probe test is in the um level, the temperature change can cause the precision positioning and the deflection of the probe test position, cause the damage such as fragmentation, and the like, which is unfavorable for improving the accuracy of the test result, ensuring the stable and efficient performance of the test process, and is difficult to effectively reduce the damage to the wafer chip in the test process, and needs to be improved;
in view of the above technical drawbacks, a solution is now proposed.
Disclosure of Invention
The invention aims to provide a high-temperature test calibration method of semiconductor detection equipment, which solves the problems that the prior art is easy to cause accurate positioning and offset of a probe test position, is unfavorable for improving the accuracy of a test result, is unfavorable for ensuring the stable and efficient performance of a test process, and is easy to cause the damage of a wafer chip.
In order to achieve the above purpose, the present invention provides the following technical solutions:
a method for calibrating high-temperature test of semiconductor detection equipment comprises the following steps:
step one, aligning a camera at the needle position, wherein the camera collects the probe position corresponding to the semiconductor detection equipment and sends the probe position to a needle position acquisition part, and the needle position acquisition part sends the probe position to a movement control part; the wafer alignment camera is used for acquiring the positions of the wafer chips and sending the positions of the wafer chips to the movement control part;
step two, the temperature sensor collects temperature data corresponding to the test environment and sends real-time temperature data to the input data acquisition part, the positioning data measurement part sends positioning data of the probe and the wafer chip to the input data acquisition part, and the input data acquisition part sends the temperature data and the positioning data to the prediction part;
a third step of the prediction unit performing a prediction operation based on the AI model used as the prediction model, thereby performing prediction and compensation accuracy setting for the probe tip position and the overall position of the wafer chip, generating a corresponding prediction compensation operation instruction, and transmitting the corresponding prediction compensation operation instruction to the movement control unit;
step four, after receiving the corresponding prediction compensation operation instruction, the movement control part drives the relative movement part of the probe based on the needle position, the wafer chip position and the corresponding prediction compensation operation instruction, so that the probe is in contact with the semiconductor chip, and the needle insertion is more accurate; the relative movement part comprises an X movement part, a Y movement part and a Z movement part, and the X movement part, the Y movement part and the Z movement part respectively move based on an X sliding table, a Y sliding table and a Z sliding table.
Further, the prediction model in the prediction part is generated by the prediction model generation part and is trained by the AI model to continuously optimize model parameters, the data is divided into a training set and a verification set in the training process, the training set is used for updating the model parameters, the verification set is used for evaluating the performance of the model, and the unknown data is accurately predicted or decided by the prediction model through continuous training.
Further, in the third step, when the prediction part predicts, the starting analysis time is marked as the prediction starting time, the time for generating the corresponding prediction compensation operation instruction is marked as the prediction ending time, the time difference between the prediction ending time and the prediction starting time is calculated to obtain the prediction time length, the prediction time length is compared with a preset prediction time length threshold value, if the prediction time length exceeds the preset prediction time length threshold value, a prediction symbol YF-1 is generated, and if the prediction time length does not exceed the preset prediction time length threshold value, a prediction symbol YF-2 is generated;
collecting the number of the prediction operations corresponding to the prediction symbol YF-1 in unit time and marking the number as the effective difference prediction number, collecting the number of the prediction operations corresponding to the prediction symbol YF-2 in unit time and marking the number as the effective difference prediction number, carrying out ratio calculation on the effective difference prediction number and the effective difference prediction number to obtain a primary effect value, carrying out numerical calculation on the primary effect value and the effective difference prediction number to obtain a re-effect value, carrying out numerical comparison on the re-effect value and a preset re-effect threshold, generating a model work efficiency unqualified signal if the re-effect value exceeds the preset re-effect threshold, and generating a model work efficiency qualified signal if the re-effect value does not exceed the preset re-effect threshold; and sending out corresponding early warning when generating a model work efficiency unqualified signal so as to facilitate the management personnel to perform model optimization in time.
Further, in the fourth step, after the prediction compensation operation instruction is completed, the time when the movement control part receives the corresponding prediction compensation operation instruction is marked as an instruction receiving time, the time when the corresponding action starts based on the prediction compensation operation instruction is marked as an action starting time, the operation completion time is marked as an action ending time, the action starting time and the instruction receiving time are subjected to time difference calculation to obtain an operation reaction time, and the action ending time and the action starting time are subjected to time difference calculation to obtain an action running time;
marking the actual operation path distance of the current operation as an action length, calculating the ratio of the action length to the action running time to obtain an action efficiency value, respectively comparing the operation reaction time with a preset operation reaction time threshold value and comparing the action efficiency value with a preset action efficiency range, if the operation reaction time does not exceed the preset operation reaction time threshold value and the action efficiency value is in the preset action efficiency range, generating an action symbol DZ-1, and otherwise, generating an action symbol DZ-2.
Further, when the prediction compensation operation instruction is completed to stop the adjustment, acquiring an actual position of the probe and a position required to be reached by the corresponding prediction compensation operation instruction, acquiring an X-position deviation value, a Y-position deviation value and a Z-position deviation value of the probe based on the actual position of the probe and the position required to be reached by the corresponding prediction compensation operation instruction, respectively comparing the X-position deviation value, the Y-position deviation value and the Z-position deviation value with corresponding preset thresholds, and generating a deviation symbol PZ-2 if at least one of the X-position deviation value, the Y-position deviation value and the Z-position deviation value exceeds the corresponding preset threshold;
if the X-position deviation value, the Y-position deviation value and the Z-position deviation value do not exceed the corresponding preset thresholds, carrying out normalization calculation on the X-position deviation value, the Y-position deviation value and the Z-position deviation value to obtain deviation coefficients, carrying out numerical comparison on the deviation coefficients and the preset deviation coefficient thresholds, generating a deviation symbol PZ-2 if the deviation coefficients exceed the preset deviation coefficients, and generating a deviation symbol PZ-1 if the deviation coefficients do not exceed the preset deviation coefficient thresholds; and when the deviation symbol PZ-2 is generated, corresponding early warning is sent out, and position adaptability adjustment is timely carried out.
Further, acquiring a deviation symbol PZ-1 or a deviation symbol PZ-2 and an action symbol DZ-1 or an action symbol DZ-2 corresponding to the prediction compensation operation instruction, marking a corresponding adjustment process as an optimal adjustment process if the PZ-1 and the DZ-1 are acquired, marking the corresponding adjustment process as an extremely poor process if the PZ-2 and the DZ-2 are acquired, and marking the corresponding adjustment process as a poor adjustment process under the other conditions;
collecting the sum of the number of the optimal regulation processes, the sum of the number of the difference regulation processes and the sum of the number of the extremely difference processes in unit time, respectively marking the sum as YT1, YT2 and YT3, carrying out numerical calculation on the YT1, YT2 and YT3 to obtain a regulation analysis value, carrying out numerical comparison on the regulation analysis value and a preset regulation analysis threshold, generating a regulation disqualification signal if the regulation analysis value exceeds the preset regulation analysis threshold, and generating a regulation qualification signal if the regulation analysis value does not exceed the preset regulation analysis threshold; and sending out corresponding early warning when the unqualified regulation signal is generated, so that corresponding management personnel can timely conduct cause investigation and judgment and equipment inspection and maintenance.
Furthermore, in the step, the corresponding semiconductor detection equipment is subjected to maintenance early warning analysis, the maintenance early warning analysis is used for generating an overhaul signal or a preparation test signal, the corresponding early warning is sent out when the signal is generated so as to remind a manager to carry out overhaul and maintenance of the corresponding semiconductor detection equipment, and the manager carries out test preparation when the preparation test signal is generated.
Further, the specific analysis process of the maintenance early warning analysis is as follows:
collecting historical maintenance dates of corresponding semiconductor detection equipment, calculating time differences between two adjacent groups of historical maintenance dates to obtain maintenance time differences, summing all maintenance time differences, taking an average value to obtain a maintenance time coefficient, calculating the time differences between the current date and the last historical maintenance date to obtain a current maintenance interval duration, and calculating the difference between the current maintenance interval duration and the maintenance time coefficient to obtain a maintenance time exceeding value; if the dimension exceeding value is greater than or equal to the preset dimension exceeding threshold value, generating an overhaul signal;
if the maintenance exceeding value is smaller than the preset maintenance exceeding threshold, acquiring the detection times of the corresponding semiconductor detection equipment and the detection time of each detection in the current maintenance interval time, summing the detection time of each detection to obtain a detection total time value, carrying out numerical calculation on the detection times and the detection total time value to obtain a maintenance early warning coefficient, carrying out numerical comparison on the maintenance early warning coefficient and a preset maintenance early warning coefficient threshold, generating a maintenance signal if the maintenance early warning coefficient exceeds the preset maintenance early warning coefficient threshold, and generating a preparation test signal if the maintenance early warning coefficient does not exceed the preset maintenance early warning coefficient threshold.
Compared with the prior art, the invention has the beneficial effects that:
1. according to the invention, the AI model is used for performing prediction operation to predict and compensate the position of the probe tip and the overall position of the wafer chip, a corresponding prediction compensation operation instruction is generated and sent to the movement control part, and after the movement control part receives the corresponding prediction compensation operation instruction, the movement control part drives the relative movement part of the probe based on the needle position, the wafer chip position and the corresponding prediction compensation operation instruction, so that the probe is in contact with the semiconductor chip, the needle insertion is more accurate, the precise positioning and the offset of the probe test position caused by the change of temperature are avoided, the accuracy of the test result is obviously improved, the stable and efficient performance of the test process is ensured, and the damage to the wafer chip in the test process can be effectively reduced;
2. according to the invention, the efficiency evaluation efficiency is carried out on the prediction analysis process, so that symbol calibration of each prediction process is realized, a model work efficiency qualified signal or a model work efficiency unqualified signal in unit time is generated, and corresponding early warning is sent out when the model work efficiency unqualified signal is generated, so that a manager can timely carry out model optimization, and the efficient and stable carrying out of the subsequent prediction analysis operation of corresponding semiconductor detection equipment is ensured; after the prediction compensation operation instruction is completed, carrying out condition analysis on the current operation process to realize symbol calibration of the current operation process, generating a regulation qualified signal or a regulation unqualified signal in unit time, and sending out corresponding early warning when the regulation unqualified signal is generated so as to timely carry out cause investigation and judgment and equipment inspection maintenance on corresponding management personnel;
3. according to the invention, the corresponding semiconductor detection equipment is subjected to maintenance early warning analysis to generate the overhaul signal or the preparation test signal, the corresponding early warning is sent out to remind the manager to carry out overhaul and maintenance of the corresponding semiconductor detection equipment when the signal is generated, and the manager is subjected to test preparation when the preparation test signal is generated, so that the timely maintenance and overhaul of the semiconductor detection equipment can be realized, the high efficiency and stability of the operation process of the semiconductor detection equipment are ensured, and the accuracy of the detection result is ensured.
Drawings
For the convenience of those skilled in the art, the present invention will be further described with reference to the accompanying drawings;
FIG. 1 is a diagram of an overall system framework of the present invention;
fig. 2 is a flow chart of the method of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Embodiment one: as shown in fig. 1-2, the method for calibrating the high-temperature test of the semiconductor detection equipment provided by the invention comprises the following steps:
step one, aligning a camera at the needle position, wherein the camera collects the probe position corresponding to the semiconductor detection equipment and sends the probe position to a needle position acquisition part, and the needle position acquisition part sends the probe position to a movement control part; the wafer alignment camera is used for acquiring the positions of the wafer chips and sending the positions of the wafer chips to the movement control part;
step two, the temperature sensor collects temperature data corresponding to the test environment and sends real-time temperature data to the input data acquisition part, the positioning data measurement part sends positioning data of the probe and the wafer chip to the input data acquisition part, and the input data acquisition part sends the temperature data and the positioning data to the prediction part;
a third step of the prediction unit performing a prediction operation based on the AI model used as the prediction model, thereby performing prediction and compensation accuracy setting for the probe tip position and the overall position of the wafer chip, generating a corresponding prediction compensation operation instruction, and transmitting the corresponding prediction compensation operation instruction to the movement control unit;
the prediction model in the prediction part is generated by the prediction model generation part and is trained by the AI model to continuously optimize model parameters, the data is divided into a training set and a verification set in the training process, the training set is used for updating the model parameters, the verification set is used for evaluating the performance of the model, and the unknown data is accurately predicted or decided by the prediction model through continuous training;
step four, after receiving the corresponding prediction compensation operation instruction, the movement control part drives the relative movement part of the probe based on the needle position, the wafer chip position and the corresponding prediction compensation operation instruction, so that the probe is in contact with the semiconductor chip, and the needle insertion is more accurate; the relative movement part comprises an X movement part, a Y movement part and a Z movement part, and the X movement part, the Y movement part and the Z movement part respectively move based on an X sliding table, a Y sliding table and a Z sliding table.
Embodiment two: in the third step, when the prediction unit predicts, the starting analysis time is marked as the prediction start time, the time when the corresponding prediction compensation operation instruction is generated is marked as the prediction end time, and the time difference between the prediction end time and the prediction start time is calculated to obtain the prediction duration, wherein the larger the value of the prediction duration is, the lower the current prediction analysis efficiency of the prediction model is; the predicted time length is compared with a preset predicted time length threshold value in a numerical mode, if the predicted time length exceeds the preset predicted time length threshold value, the predicted symbol YF-1 is generated, and if the predicted time length does not exceed the preset predicted time length threshold value, the predicted symbol YF-2 is generated;
acquiring the number of the prediction operations corresponding to the prediction symbol YF-1 in unit time and marking the number as the effective difference prediction number, acquiring the number of the prediction operations corresponding to the prediction symbol YF-2 in unit time and marking the number as the effective difference prediction number, calculating the ratio of the effective difference prediction number to obtain a primary effect value, and carrying out numerical calculation on the primary effect value FT1 and the effective difference prediction number FT2 through a formula ZX=a1 x FT1+a2 to obtain a secondary effect value ZX; wherein a1 and a2 are preset weight coefficients, and a1 is more than a2 and more than 0; and the larger the numerical value of the re-effective value ZX is, the better the prediction analysis efficiency in unit time is;
comparing the re-effect value ZX with a preset re-effect threshold value, if the re-effect value ZX exceeds the preset re-effect threshold value, generating a model work efficiency disqualification signal, and if the re-effect value ZX does not exceed the preset re-effect threshold value, generating a model work efficiency qualification signal; and when the model work efficiency disqualification signal is generated, a corresponding early warning is sent out so that a manager can perform model optimization in time, and therefore efficient and stable performance of subsequent predictive analysis operation of corresponding semiconductor detection equipment is ensured.
Embodiment III: in the fourth step, after the prediction compensation operation command is completed, the time when the movement control unit receives the corresponding prediction compensation operation command is marked as the command receiving time, the time when the corresponding operation starts based on the prediction compensation operation command is marked as the action starting time, the operation completion time is marked as the action ending time, and the time difference between the action starting time and the command receiving time is calculated to obtain the operation reaction time, and it is required to be explained that the larger the value of the operation reaction time is, the less agile the reaction corresponding to the semiconductor detection device is, that is, the more insensitive the operation reaction is;
calculating a time difference between an action ending time and an action starting time to obtain an action running time, marking an actual operation path distance of the current operation as an action length, calculating a ratio of the action length to the action running time to obtain an action efficiency value, respectively comparing the operation reaction time with a preset operation reaction time threshold value and comparing the action efficiency value with a preset action efficiency range, if the operation reaction time does not exceed the preset operation reaction time threshold value and the action efficiency value is in the preset action efficiency range, generating an action symbol DZ-1, if the current operation condition is good, generating an action symbol DZ-2, if the operation reaction time exceeds the preset operation reaction time threshold value or the action efficiency value is not in the preset action efficiency range, and if the current operation condition is poor, generating the action symbol DZ-2.
When the prediction compensation operation instruction is finished to stop adjustment, acquiring the actual position of the probe and the position required to be reached by the corresponding prediction compensation operation instruction, acquiring an X-position deviation value, a Y-position deviation value and a Z-position deviation value of the probe based on the actual position of the probe and the position required to be reached by the corresponding prediction compensation operation instruction, respectively carrying out numerical comparison on the X-position deviation value, the Y-position deviation value and the Z-position deviation value with corresponding preset thresholds, and if at least one of the X-position deviation value, the Y-position deviation value and the Z-position deviation value exceeds the corresponding preset threshold, indicating that the adjusted probe position deviation is larger, and not beneficial to guaranteeing the accuracy of detection results of corresponding semiconductor detection equipment, generating a deviation symbol PZ-2;
if the X-position deviation value, the Y-position deviation value and the Z-position deviation value do not exceed the corresponding preset thresholds, normalizing the X-position deviation value PCx, the Y-position deviation value PCy and the Z-position deviation value PCz by the formula py=b1× PCx +b2×pcy+b3× PCz to obtain a deviation coefficient PY; wherein b1, b2 and b3 are preset weight coefficients, and the values of b1, b2 and b3 are all larger than 1; and the magnitude of the deviation coefficient PY is in a direct proportion relation with the X-position deviation value, the Y-position deviation value and the Z-position deviation value, and the larger the magnitude of the deviation coefficient PY is, the larger the position deviation of the probe after adjustment is, the more unqualified the position adjustment operation is, and the accuracy of the detection result is not guaranteed;
comparing the deviation coefficient with a preset deviation coefficient threshold value in a numerical mode, if the deviation coefficient exceeds the preset deviation coefficient, indicating that the position deviation of the adjusted probe is larger, and not beneficial to guaranteeing the accuracy of the detection result of the corresponding semiconductor detection equipment, generating a deviation symbol PZ-2, and if the deviation coefficient does not exceed the preset deviation coefficient threshold value, indicating that the position deviation of the adjusted probe is smaller, generating a deviation symbol PZ-1; and when the deviation symbol PZ-2 is generated, corresponding early warning is sent out and position adaptability adjustment is timely carried out, so that the accuracy of the detection result of corresponding semiconductor detection equipment is ensured, the stable running of the detection process is ensured, and the damage to the wafer chip is reduced.
Furthermore, by collecting a deviation symbol PZ-1 or a deviation symbol PZ-2 and an action symbol DZ-1 or an action symbol DZ-2 corresponding to a prediction compensation operation instruction, if the PZ-1 is obtained, marking the corresponding adjustment process as an optimal adjustment process, if the PZ-2 is obtained, marking the corresponding adjustment process as a very poor process, and otherwise marking the corresponding adjustment process as a poor adjustment process, thereby realizing comprehensive analysis and judgment of the current adjustment operation; collecting the sum of the number of the optimal tuning processes, the sum of the number of the difference tuning processes and the sum of the number of the extremely bad processes in unit time and marking the sum as YT1, YT2 and YT3 respectively;
carrying out numerical calculation on YT1, YT2 and YT3 through a formula TF= (kp2 x YT2+kp3 x YT3)/(kp1 x YT1+0.843) to obtain a regulation analysis value TF, wherein kp1, kp2 and kp3 are preset proportionality coefficients, the values of kp1, kp2 and kp3 are all larger than zero, and kp3 is larger than kp2 and kp1; and, the larger the value of the regulatory analysis value TF, the worse the regulatory condition performance in unit time is shown; performing numerical comparison on the regulation analysis value TF and a preset regulation analysis threshold, generating a regulation disqualification signal if the regulation analysis value TF exceeds the preset regulation analysis threshold, and generating a regulation qualification signal if the regulation analysis value TF does not exceed the preset regulation analysis threshold; and sending out corresponding early warning when the unqualified regulation signal is generated, so that corresponding management personnel can timely conduct cause investigation and judgment and equipment inspection and maintenance.
Embodiment four: the difference between the embodiment and embodiments 1, 2 and 3 is that, during the step, the corresponding semiconductor detection device is subjected to maintenance early warning analysis, so as to generate an overhaul signal or a ready test signal through the maintenance early warning analysis, and when the corresponding early warning is generated, the corresponding early warning is sent out to remind a manager to carry out overhaul and maintenance of the corresponding semiconductor detection device, and when the ready test signal is generated, the manager carries out test preparation, thereby being beneficial to realizing timely maintenance and overhaul of the semiconductor detection device, ensuring the high efficiency and stability of the operation process of the semiconductor detection device, and being beneficial to ensuring the accuracy of the detection result; the specific analysis process of the maintenance early warning analysis is as follows:
collecting historical maintenance dates of corresponding semiconductor detection equipment, calculating time differences between two adjacent groups of historical maintenance dates to obtain maintenance time differences, summing all maintenance time differences, taking an average value to obtain a maintenance time coefficient, calculating the time differences between the current date and the last historical maintenance date to obtain a current maintenance interval duration, and calculating the difference between the current maintenance interval duration and the maintenance time coefficient to obtain a maintenance time exceeding value; the larger the value of the dimension exceeding value is, the more the corresponding semiconductor detection equipment needs to be overhauled and comprehensively maintained in time; comparing the dimension exceeding value with a preset dimension exceeding threshold value, and generating an overhaul signal if the dimension exceeding value is larger than or equal to the preset dimension exceeding threshold value;
if the dimension exceeding value is smaller than the preset dimension exceeding threshold value, acquiring the detection times of the corresponding semiconductor detection equipment in the current maintenance interval time and the time of each detection, summing the time of each detection to obtain a detection total value, and carrying out numerical calculation on the detection times JC and the detection total value SZ through a formula WY=ep1+ep2 to obtain a maintenance early warning coefficient WY, wherein ep1 and ep2 are preset weight coefficients, and ep1 is more than ep2 is more than 1; in addition, the value of the maintenance early warning coefficient WY is in a direct proportion relation with the detection times JC and the detection total time SZ, and the larger the value of the maintenance early warning coefficient WY is, the more the corresponding semiconductor detection equipment is required to be overhauled and maintained in time; and carrying out numerical comparison on the maintenance early-warning coefficient WY and a preset maintenance early-warning coefficient threshold value, generating an overhaul signal if the maintenance early-warning coefficient WY exceeds the preset maintenance early-warning coefficient threshold value, and generating a preparation test signal if the maintenance early-warning coefficient WY does not exceed the preset maintenance early-warning coefficient threshold value.
The working principle of the invention is as follows: when the test device is used, the prediction part is used for performing prediction operation based on the AI model serving as the prediction model, the position of the tip of the probe and the overall position of the wafer chip are predicted and the compensation precision is set, a corresponding prediction compensation operation instruction is generated, the corresponding prediction compensation operation instruction is sent to the movement control part, after the movement control part receives the corresponding prediction compensation operation instruction, the relative movement part of the probe is driven based on the needle position, the position of the wafer chip and the corresponding prediction compensation operation instruction, so that the probe is in contact with the semiconductor chip, the needle insertion is more accurate, the precise positioning and the offset of the test position of the probe caused by the change of temperature are avoided, the accuracy of the test result is remarkably improved, the stable and efficient performance of the test process is ensured, and the damage to the wafer chip in the test process can be effectively reduced.
The above formulas are all formulas with dimensions removed and numerical values calculated, the formulas are formulas with a large amount of data collected for software simulation to obtain the latest real situation, and preset parameters in the formulas are set by those skilled in the art according to the actual situation. The preferred embodiments of the invention disclosed above are intended only to assist in the explanation of the invention. The preferred embodiments are not intended to be exhaustive or to limit the invention to the precise form disclosed. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the invention and the practical application, to thereby enable others skilled in the art to best understand and utilize the invention. The invention is limited only by the claims and the full scope and equivalents thereof.
Claims (4)
1. The method for calibrating the high-temperature test of the semiconductor detection equipment is characterized by comprising the following steps of:
step one, aligning a camera at the needle position, wherein the camera collects the probe position corresponding to the semiconductor detection equipment and sends the probe position to a needle position acquisition part, and the needle position acquisition part sends the probe position to a movement control part; the wafer alignment camera is used for acquiring the positions of the wafer chips and sending the positions of the wafer chips to the movement control part;
step two, the temperature sensor collects temperature data corresponding to the test environment and sends real-time temperature data to the input data acquisition part, the positioning data measurement part sends positioning data of the probe and the wafer chip to the input data acquisition part, and the input data acquisition part sends the temperature data and the positioning data to the prediction part;
a third step of the prediction unit performing a prediction operation based on the AI model used as the prediction model, thereby performing prediction and compensation accuracy setting for the probe tip position and the overall position of the wafer chip, generating a corresponding prediction compensation operation instruction, and transmitting the corresponding prediction compensation operation instruction to the movement control unit;
step four, after receiving the corresponding prediction compensation operation instruction, the movement control part drives the relative movement part of the probe based on the needle position, the wafer chip position and the corresponding prediction compensation operation instruction, so that the probe is in contact with the semiconductor chip, and the needle insertion is more accurate; the relative movement part comprises an X movement part, a Y movement part and a Z movement part, and the X movement part, the Y movement part and the Z movement part respectively move based on an X sliding table, a Y sliding table and a Z sliding table;
in the third step, when the prediction part predicts, the starting analysis time is marked as the prediction starting time, the time for generating the corresponding prediction compensation operation instruction is marked as the prediction ending time, the time difference between the prediction ending time and the prediction starting time is calculated to obtain the prediction time length, the prediction time length is compared with a preset prediction time length threshold value in a numerical mode, if the prediction time length exceeds the preset prediction time length threshold value, a prediction symbol YF-1 is generated, and if the prediction time length does not exceed the preset prediction time length threshold value, a prediction symbol YF-2 is generated;
acquiring the number of the prediction operations corresponding to the prediction symbol YF-1 in unit time and marking the number as the effective difference prediction number, acquiring the number of the prediction operations corresponding to the prediction symbol YF-2 in unit time and marking the number as the effective difference prediction number, calculating the ratio of the effective difference prediction number to obtain a primary effect value, and carrying out numerical calculation on the primary effect value FT1 and the effective difference prediction number FT2 through a formula ZX=a1 x FT1+a2 to obtain a secondary effect value ZX; wherein a1 and a2 are preset weight coefficients, and a1 is more than a2 and more than 0; comparing the re-effect value with a preset re-effect threshold value, generating a model work efficiency disqualification signal if the re-effect value exceeds the preset re-effect threshold value, and generating a model work efficiency qualification signal if the re-effect value does not exceed the preset re-effect threshold value; sending out corresponding early warning when generating a model work efficiency unqualified signal so as to facilitate a manager to optimize the model in time;
in the fourth step, after the prediction compensation operation instruction is completed, the time when the movement control part receives the corresponding prediction compensation operation instruction is marked as the instruction receiving time, the time when the corresponding action starts based on the prediction compensation operation instruction is marked as the action starting time, the operation completion time is marked as the action ending time, the action starting time and the instruction receiving time are subjected to time difference calculation to obtain operation reaction time, and the action ending time and the action starting time are subjected to time difference calculation to obtain action running time;
marking the actual operation path distance of the current operation as an action length, calculating the ratio of the action length to the action running time to obtain an action efficiency value, respectively comparing the operation reaction time with a preset operation reaction time threshold value and comparing the action efficiency value with a preset action efficiency range, if the operation reaction time does not exceed the preset operation reaction time threshold value and the action efficiency value is in the preset action efficiency range, generating an action symbol DZ-1, and otherwise, generating an action symbol DZ-2;
when the prediction compensation operation instruction is finished to stop adjusting, acquiring the actual position of the probe and the position required to be reached by the corresponding prediction compensation operation instruction, acquiring an X-position deviation value, a Y-position deviation value and a Z-position deviation value of the probe based on the actual position of the probe and the position required to be reached by the corresponding prediction compensation operation instruction, respectively carrying out numerical comparison on the X-position deviation value, the Y-position deviation value and the Z-position deviation value with corresponding preset thresholds, and generating a deviation symbol PZ-2 if at least one of the X-position deviation value, the Y-position deviation value and the Z-position deviation value exceeds the corresponding preset threshold;
if the X-position deviation value, the Y-position deviation value and the Z-position deviation value do not exceed the corresponding preset thresholds, normalizing the X-position deviation value PCx, the Y-position deviation value PCy and the Z-position deviation value PCz by the formula py=b1× PCx +b2×pcy+b3× PCz to obtain a deviation coefficient PY; wherein b1, b2 and b3 are preset weight coefficients, and the values of b1, b2 and b3 are all larger than 1; comparing the deviation coefficient with a preset deviation coefficient threshold value in a numerical mode, generating a deviation sign PZ-2 if the deviation coefficient exceeds the preset deviation coefficient, and generating a deviation sign PZ-1 if the deviation coefficient does not exceed the preset deviation coefficient threshold value; when the deviation symbol PZ-2 is generated, corresponding early warning is sent out, and position adaptability adjustment is timely carried out;
collecting a deviation symbol PZ-1 or a deviation symbol PZ-2 and an action symbol DZ-1 or an action symbol DZ-2 corresponding to a prediction compensation operation instruction, marking a corresponding adjusting process as an optimal adjusting process if PZ-1 and DZ-1 are obtained, marking the corresponding adjusting process as an extremely poor process if PZ-2 and DZ-2 are obtained, and marking the corresponding adjusting process as a poor adjusting process if the rest conditions are met;
collecting the sum of the number of the optimal adjustment processes, the sum of the number of the difference adjustment processes and the sum of the number of the extremely difference processes in unit time, respectively marking the sum as YT1, YT2 and YT3, and carrying out numerical calculation on YT1, YT2 and YT3 through a formula TF= (kp2 YT2+kp3 YT1+0.843), wherein kp1, kp2 and kp3 are preset proportionality coefficients, the values of kp1, kp2 and kp3 are all larger than zero, and kp3 is larger than kp2 and larger than kp1; performing numerical comparison on the regulation analysis value and a preset regulation analysis threshold, generating a regulation disqualification signal if the regulation analysis value exceeds the preset regulation analysis threshold, and generating a regulation qualification signal if the regulation analysis value does not exceed the preset regulation analysis threshold; and sending out corresponding early warning when the unqualified regulation signal is generated, so that corresponding management personnel can timely conduct cause investigation and judgment and equipment inspection and maintenance.
2. The method according to claim 1, wherein the prediction model in the prediction unit is generated by the prediction model generation unit and is trained via the AI model to continuously optimize model parameters, the training process divides data into a training set and a verification set, the training set is used to update model parameters, the verification set is used to evaluate the performance of the model, and the prediction model is trained via the continuous training to accurately predict or decide unknown data.
3. The method for calibrating a high-temperature test of a semiconductor inspection apparatus according to claim 1, wherein, when the step is performed, a maintenance pre-warning analysis is performed on the corresponding semiconductor inspection apparatus, a maintenance signal or a ready-to-test signal is generated by the maintenance pre-warning analysis, and a corresponding pre-warning is issued to remind a manager to perform maintenance on the corresponding semiconductor inspection apparatus when the signal is generated, and the manager performs test preparation when the ready-to-test signal is generated.
4. A method for calibrating a high-temperature test of a semiconductor inspection device according to claim 3, wherein the specific analysis process of the maintenance pre-warning analysis is as follows:
collecting historical maintenance dates of corresponding semiconductor detection equipment, calculating time differences between two adjacent groups of historical maintenance dates to obtain maintenance time differences, summing all maintenance time differences, taking an average value to obtain a maintenance time coefficient, calculating the time differences between the current date and the last historical maintenance date to obtain a current maintenance interval duration, and calculating the difference between the current maintenance interval duration and the maintenance time coefficient to obtain a maintenance time exceeding value; if the dimension exceeding value is greater than or equal to the preset dimension exceeding threshold value, generating an overhaul signal;
if the maintenance exceeding value is smaller than the preset maintenance exceeding threshold, acquiring the detection times of the corresponding semiconductor detection equipment and the detection time of each detection in the current maintenance interval time, summing the detection time of each detection to obtain a detection total time value, carrying out numerical calculation on the detection times and the detection total time value to obtain a maintenance early warning coefficient, carrying out numerical comparison on the maintenance early warning coefficient and a preset maintenance early warning coefficient threshold, generating a maintenance signal if the maintenance early warning coefficient exceeds the preset maintenance early warning coefficient threshold, and generating a preparation test signal if the maintenance early warning coefficient does not exceed the preset maintenance early warning coefficient threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310769417.3A CN116500426B (en) | 2023-06-28 | 2023-06-28 | Method for calibrating high-temperature test of semiconductor detection equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310769417.3A CN116500426B (en) | 2023-06-28 | 2023-06-28 | Method for calibrating high-temperature test of semiconductor detection equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116500426A CN116500426A (en) | 2023-07-28 |
| CN116500426B true CN116500426B (en) | 2023-09-05 |
Family
ID=87316947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310769417.3A Active CN116500426B (en) | 2023-06-28 | 2023-06-28 | Method for calibrating high-temperature test of semiconductor detection equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116500426B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117316262B (en) * | 2023-11-30 | 2024-04-09 | 深圳市领德创科技有限公司 | Automatic FLASH chip detection machine |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050047432A (en) * | 2003-11-17 | 2005-05-20 | 삼성전자주식회사 | Temperature controller for probe equipment |
| JP2013140495A (en) * | 2012-01-05 | 2013-07-18 | Hitachi Ltd | Plant state monitoring device and plant state monitoring method |
| TW201350860A (en) * | 2012-06-13 | 2013-12-16 | Taiwan Semiconductor Mfg | Method for aligning probe contacts with wafer contacts |
| WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技（深圳）有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
| CN111198808A (en) * | 2019-12-25 | 2020-05-26 | 东软集团股份有限公司 | Method, device, storage medium and electronic equipment for predicting performance index |
| CN115130702A (en) * | 2022-09-02 | 2022-09-30 | 山东汇泓纺织科技有限公司 | Textile machine fault prediction system based on big data analysis |
| WO2022255030A1 (en) * | 2021-06-04 | 2022-12-08 | 株式会社東京精密 | Prober control device, prober control method, and prober |
| CN115932530A (en) * | 2023-01-09 | 2023-04-07 | 东莞市兆恒机械有限公司 | Method for calibrating semiconductor detection equipment |
| CN116026389A (en) * | 2023-03-07 | 2023-04-28 | 湖南科技学院 | Intelligent sensor operation detection system based on data analysis |
| CN116256544A (en) * | 2023-01-05 | 2023-06-13 | 苏州斯尔特微电子有限公司 | Wafer test probe station with offset correction function |
| CN116300652A (en) * | 2023-04-06 | 2023-06-23 | 合肥元贞电力科技股份有限公司 | Power control cabinet on-line monitoring system based on data analysis |
| CN116301091A (en) * | 2023-05-19 | 2023-06-23 | 浙江农林大学 | Temperature control intelligent management system suitable for radiation cold and warm window |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005072143A (en) * | 2003-08-21 | 2005-03-17 | Tokyo Seimitsu Co Ltd | Probe unit |
| US8311758B2 (en) * | 2006-01-18 | 2012-11-13 | Formfactor, Inc. | Methods and apparatuses for dynamic probe adjustment |
| JP4744382B2 (en) * | 2006-07-20 | 2011-08-10 | 株式会社東京精密 | Prober and probe contact method |
-
2023
- 2023-06-28 CN CN202310769417.3A patent/CN116500426B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20050047432A (en) * | 2003-11-17 | 2005-05-20 | 삼성전자주식회사 | Temperature controller for probe equipment |
| JP2013140495A (en) * | 2012-01-05 | 2013-07-18 | Hitachi Ltd | Plant state monitoring device and plant state monitoring method |
| TW201350860A (en) * | 2012-06-13 | 2013-12-16 | Taiwan Semiconductor Mfg | Method for aligning probe contacts with wafer contacts |
| WO2019056499A1 (en) * | 2017-09-20 | 2019-03-28 | 平安科技（深圳）有限公司 | Prediction model training method, data monitoring method, apparatuses, device and medium |
| CN111198808A (en) * | 2019-12-25 | 2020-05-26 | 东软集团股份有限公司 | Method, device, storage medium and electronic equipment for predicting performance index |
| WO2022255030A1 (en) * | 2021-06-04 | 2022-12-08 | 株式会社東京精密 | Prober control device, prober control method, and prober |
| CN115130702A (en) * | 2022-09-02 | 2022-09-30 | 山东汇泓纺织科技有限公司 | Textile machine fault prediction system based on big data analysis |
| CN116256544A (en) * | 2023-01-05 | 2023-06-13 | 苏州斯尔特微电子有限公司 | Wafer test probe station with offset correction function |
| CN115932530A (en) * | 2023-01-09 | 2023-04-07 | 东莞市兆恒机械有限公司 | Method for calibrating semiconductor detection equipment |
| CN116026389A (en) * | 2023-03-07 | 2023-04-28 | 湖南科技学院 | Intelligent sensor operation detection system based on data analysis |
| CN116300652A (en) * | 2023-04-06 | 2023-06-23 | 合肥元贞电力科技股份有限公司 | Power control cabinet on-line monitoring system based on data analysis |
| CN116301091A (en) * | 2023-05-19 | 2023-06-23 | 浙江农林大学 | Temperature control intelligent management system suitable for radiation cold and warm window |
Non-Patent Citations (1)
| Title |
|---|
| 基于真实开采数据的采运装备虚拟规划方法研究;崔涛;中国优秀硕士学位论文全文数据库工程科技Ⅰ辑(第02期);第62页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116500426A (en) | 2023-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN116500426B (en) | Method for calibrating high-temperature test of semiconductor detection equipment | |
| US20230015731A1 (en) | Real-time abnormity-diagnosing and interpolation method for water regime-monitoring data | |
| CN101246369B (en) | Vehicle element size quality control system and method | |
| CN109918817B (en) | Production line energy consumption analysis method based on time-varying data | |
| CN104700200A (en) | Multivariate product quality monitoring method oriented to digital workshop | |
| CN112904266B (en) | Method and device for predicting service life of electric energy meter | |
| CN108957385B (en) | Method and device for confirming abnormal epitope of automatic verification line of electric energy metering equipment | |
| CN107067100A (en) | Wind power anomalous data identification method and device for identifying | |
| CN116638166B (en) | Intelligent temperature control method and system for high-precision welding processing | |
| CN104647141A (en) | In-place measurement method for thickness of large-sized wall plate | |
| CN113032728A (en) | Voltage transformer state evaluation method and system based on data driving error evaluation result | |
| CN113050018A (en) | Voltage transformer state evaluation method and system based on data drive evaluation result change trend | |
| CN113406558A (en) | Electric meter misalignment detection method and device based on linear regression and electronic equipment | |
| CN103177392B (en) | A kind of correlating method of coil of strip length process | |
| CN111354496B (en) | Nuclear power plant accident online diagnosis and state tracking prediction method | |
| CN103279030A (en) | Bayesian framework-based dynamic soft measurement modeling method and device | |
| CN104732340A (en) | Auxiliary power grid planning decision-making method based on annual time scales | |
| CN101320258B (en) | Two-stage virtual measurement method | |
| CN102478648B (en) | A device measured value precision automatic calibration method and system | |
| CN101709957A (en) | Straightness automatic measurement method based on machine vision | |
| CN109933031A (en) | A kind of system and method automatically correcting soft measuring instrument according to analysis data | |
| CN114925569A (en) | Steel plate quenching temperature prediction method combining finite element and neural network | |
| CN103759606A (en) | System and method for detecting position degree of multiple holes of large-sized workpiece | |
| CN203881239U (en) | System for detecting position degree of multiple holes of large-sized workpiece | |
| CN111443687A (en) | Method for testing response time of power plant decentralized control system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |