CN116259559A - Chip trimming prediction method - Google Patents

Chip trimming prediction method Download PDF

Info

Publication number
CN116259559A
CN116259559A CN202310328348.2A CN202310328348A CN116259559A CN 116259559 A CN116259559 A CN 116259559A CN 202310328348 A CN202310328348 A CN 202310328348A CN 116259559 A CN116259559 A CN 116259559A
Authority
CN
China
Prior art keywords
chip
trimming
repaired
value
trimmed
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.)
Pending
Application number
CN202310328348.2A
Other languages
Chinese (zh)
Inventor
阳靖
钱向东
卢旭坤
蒋卓怡
李沛东
旷法佳
杨柳
谢凯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Leadyo Ic Testing Co ltd
Original Assignee
Guangdong Leadyo Ic Testing Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong Leadyo Ic Testing Co ltd filed Critical Guangdong Leadyo Ic Testing Co ltd
Priority to CN202310328348.2A priority Critical patent/CN116259559A/en
Publication of CN116259559A publication Critical patent/CN116259559A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • 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
    • 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
    • H01L22/30—Structural arrangements specially adapted for testing or measuring during manufacture or treatment, or specially adapted for reliability measurements
    • H01L22/34—Circuits for electrically characterising or monitoring manufacturing processes, e. g. whole test die, wafers filled with test structures, on-board-devices incorporated on each die, process control monitors or pad structures thereof, devices in scribe line
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Manufacturing & Machinery (AREA)
  • Computer Hardware Design (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Power Engineering (AREA)
  • Automation & Control Theory (AREA)
  • Testing Or Measuring Of Semiconductors Or The Like (AREA)

Abstract

The disclosure discloses a chip trimming prediction method, comprising the following steps: s100: setting the upper limit of trimming times of the first chip to be trimmed; s200: measuring the signal frequency of a first chip to be repaired and obtaining a first measured value; s300: filing the first chip to be repaired based on the first measured value, repairing the first chip to be repaired according to the repairing prediction value of the gear of the first chip to be repaired, and recording repairing times; s400: measuring the signal frequency of the chip to be trimmed after trimming to obtain a second measured value; s500: and comparing the second measured value with the measurement standard value to detect the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as good, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on repair times of the first chip to be repaired, and terminating the repair program.

Description

Chip trimming prediction method
Technical Field
The disclosure belongs to the field of semiconductor testing, and particularly relates to a chip trimming prediction method.
Background
With the rapid development of the chip design and manufacturing industry, the scale of semiconductor manufacturing is larger and larger, and the process size is smaller and smaller, and even if the existing semiconductor production technology and equipment are more and more advanced, defects always occur in the manufacturing process. Thus, most chips with reference voltages, reference currents, and frequency outputs are designed with embedded trimming circuits that are altered during wafer probe testing and final package testing to correct for any process variations in the wafer manufacturing process that affect device parameters.
Two common trimming methods exist in chip testing, fig. 1 is a trimming method one, the method measures an output signal, compares the measured actual value with a target value, determines whether the actual value is larger or smaller, increases or decreases the trimming value, and then configures the trimming value into a trimming register, so as to realize the adjustment of the output signal. Fig. 2 is a second trimming method, which measures an output signal, and according to the measured actual value being larger and smaller than the target value, the tester sends a corresponding adjustment signal (handshake signal) to the chip, so as to implement chip signal adjustment.
The trimming method is that the trimming value is continuously changed in the trimming use process, namely, the trimming value is generally increased by one or decreased by one. In the case of multi-bit trimming values, such as 4-bit trimming value adjustment, the trimming value is adjusted once after each actual value is measured, and the trimming value is configured into the register, the above operations may be looped 16 times at most, that is, 16 measurements and 16 register configuration trimming values may be performed, resulting in a large increase in testing time. At present, a common time optimization method for the trimming method is a dichotomy method and a method for presetting an initial trimming value based on a statistical lookup table, wherein the dichotomy method needs to trim the trimming value from a high position to a low position bit by bit, the trimming is needed for 4-5 times, and the test time is fixed; the method using the lookup table can accurately preset the trimming value, but the measured data is required to be counted manually, so that the operation complexity of production is increased, and the method cannot cope with the trimming performance change caused by the process and physical change of chips in different batches, namely the initial trimming value preset by the lookup table based on statistics in the previous batch may not be matched with the chips in the new batch, so that the preset trimming is inaccurate.
And secondly, measuring an actual value once, comparing the magnitude relation between the actual value and the target value, executing a corresponding trimming handshake protocol, and repeating the measuring and handshake operations until the measured value is trimmed to a required range, or judging the chip as a fail defective product when the trimming frequency reaches an upper limit and the measured value does not meet the pass requirement. However, as shown in fig. 2, the handshake protocol is measured and executed every time trimming is performed, if the trimming frequency reaches 10 times, the chip trimming needs to perform 11 signal measurements and 10 times trimming handshake protocols, so that the test time becomes longer, the test efficiency is reduced, and the test cost is increased.
Disclosure of Invention
Aiming at the defects in the prior art, the invention aims to provide a chip trimming prediction method, by which the predicted value of each batch of products can be automatically adjusted, so that the predicted value of a chip is converged to an optimal range.
In order to achieve the above object, the present disclosure provides the following technical solutions:
a chip trimming prediction method comprises the following steps:
s100: setting the upper limit of trimming times of the first chip to be trimmed;
s200: measuring the signal frequency of a first chip to be repaired and obtaining a first measured value;
s300: filing the first chip to be repaired based on the first measured value, repairing the first chip to be repaired according to the repairing prediction value of the gear of the first chip to be repaired, and recording repairing times;
s400: measuring the signal frequency of the chip to be trimmed after trimming to obtain a second measured value;
s500: and comparing the second measured value with the measurement standard value to detect the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as good, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on repair times of the first chip to be repaired, and terminating the repair program.
Preferably, in step S500, the trimming prediction value of the second chip to be trimmed is obtained by an exponential smoothing based prediction method.
Preferably, the method further comprises the steps of:
s600: in step S500, if the detection is failed, the first chip to be repaired is detected again according to whether the repair number of the first chip to be repaired reaches the upper limit of the repair number set in step S100.
Preferably, the method further comprises the steps of:
s700: in step S600, if the trimming frequency of the first chip to be trimmed reaches the upper limit of the trimming frequency, determining that the first chip to be trimmed is a defective product, and terminating the trimming procedure; and if the trimming frequency of the first chip to be trimmed does not reach the upper limit of the trimming frequency, comparing the second measured value of the first chip to be trimmed with the measured standard value again, and executing a corresponding handshake protocol according to the comparison result.
Preferably, the method further comprises the steps of:
s800: updating the trimming times of the first chip to be trimmed according to the corresponding handshake protocol to obtain the actual trimming times of the first chip to be trimmed.
Preferably, the method further comprises the steps of:
s900: and measuring the signal frequency of the first chip to be trimmed after the trimming times are updated, and obtaining a third measured value.
Preferably, the method further comprises the steps of:
s1000: and comparing the third measured value with the measurement standard value to perform third detection on the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as a good product, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on the actual repair times of the first chip to be repaired, and terminating the repair program.
Preferably, the method further comprises the steps of:
s1100: in step S1000, if the third detection fails, steps S600 to S1000 are repeatedly performed.
The present disclosure also provides a computer storage medium comprising:
a memory for storing a plurality of computer instructions;
a processor for executing computer instructions to implement a method as claimed in any preceding claim.
The present disclosure also provides an electronic device, including:
a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein,
the processor, when executing the program, implements a method as described in any of the preceding.
Compared with the prior art, the beneficial effects that this disclosure brought are:
1. the method can improve trimming efficiency and reduce trimming time;
2. the method can reduce the manual statistical data operation steps;
3. the method disclosed by the disclosure can automatically adjust the predicted value of each batch of products, so that the predicted value is converged to the optimal range.
Drawings
FIG. 1 is a flow chart of a method for trimming by configuring registers;
FIG. 2 is a flow chart of a method of trimming via a handshake protocol;
FIG. 3 is a diagram showing the relationship between the initial measurement frequency of the B chip and the trimming frequency;
FIG. 4 is a flowchart of a chip trimming method based on an exponential smoothing prediction algorithm according to the present disclosure;
FIG. 5 is a schematic diagram showing the predictive effect of exponential smoothing on trim values.
Detailed Description
Specific embodiments of the present disclosure will be described in detail below with reference to fig. 1 to 5. While specific embodiments of the disclosure are shown in the drawings, it should be understood that the disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
It should be noted that certain terms are used throughout the description and claims to refer to particular components. Those of skill in the art will understand that a person may refer to the same component by different names. The specification and claims do not identify differences in terms of components, but rather differences in terms of the functionality of the components. As used throughout the specification and claims, the terms "include" and "comprise" are used in an open-ended fashion, and thus should be interpreted to mean "include, but not limited to. The description hereinafter sets forth the preferred embodiments for carrying out the present disclosure, but is not intended to limit the scope of the disclosure in general, as the description proceeds. The scope of the present disclosure is defined by the appended claims.
For the purposes of promoting an understanding of the embodiments of the disclosure, reference will now be made to the embodiments illustrated in the drawings and specific examples, without the intention of being limiting the embodiments of the disclosure.
In one embodiment, as shown in fig. 4, the disclosure provides a chip trimming prediction method, which includes the following steps:
s100: setting the upper limit of trimming times of the first chip to be trimmed;
in the step, the chip to be repaired is set as a B-type chip, and the upper limit of the repairing times is set as 30 times. The distribution of trimming times of the type B chip is shown in fig. 4, the trimming times are mainly concentrated at 20 times or less, 30 times meet most of chip test requirements, if the upper limit of the trimming times is set to be too large, this means that for judging some defective products (fail, for example, the chip trimming module fails, there is no influence on chip measurement parameters after the handshake protocol is executed), more times of handshake signals need to be measured, and the fail can be judged until the upper limit of the trimming times is reached, which wastes more time.
S200: measuring the signal frequency of a first chip to be repaired and obtaining a first measured value;
s300: filing the first chip to be repaired based on the first measured value, repairing the first chip to be repaired according to the repairing prediction value of the gear of the first chip to be repaired, and recording repairing times;
in the step, the signal frequency of the B-type chip is 135.812KHz, which is measured by a testing machine, and the frequency is taken as a first measured value, namely 135.812. Table 1 divides the gear of the type B chip, and table 1 is as follows:
TABLE 1
Gear position Initial measurement frequency Sign of predicted value Predictive value
L 1 ≤122KHz L 1 ：S n -12.40
L 2 122～123KHz L 2 ：S n -9.39
L 3 123～124KHz L 3 ：S n -7.50
L 4 124～125KHz L 4 ：S n -5.84
L 5 125～126KHz L 5 ：S n -4.16
L 6 126～127KHz L 6 ：S n -2.37
L pass 127～128KHz L pass ；S pass 0.00
L 8 128～129KHz L 8 ：S n 2.45
L 9 129～130KHz L 9 ：S n 4.18
L 10 130～131KHz L 10 ：S n 5.76
L 11 131～132KHz L n ：S n 7.51
L 12 132～133KHz L 12 ：S n 9.28
L 13 133～134KHz L 13 ：S n 10.83
L 14 134～135KHz L 14 ：S n 12.81
L 15 135～136KHz L 15 ：S n 14.58
L 16 136～137KHz L 16 ：S n 16.33
L 17 137～138KHz L 17 ：S n 17.28
L 18 ≥138KHz L 18 ：S n 19.28
According to Table 1, the chip to be repaired can be classified into the 15 th chip, if the chip to be repaired is the 2014 chip divided into the 15 th chip, the repair prediction value symbol L of the 2014 chip in Table 1 is passed 15 :S 2014 The corresponding trimming predicted value 14.58 trims the chip to be trimmed, and specifically comprises the following steps: the test machine rounds the chip to obtain 14 when trimming the chip (rounding the chip refers to the integer part, and rounding the 14.58 decimal part to 15, and rounding the 14.58 decimal part is more accurate, but one more handshake protocol is executed, the handshake protocol is executed for a long time, so that the direct integer part is considered herein), so the test machine and the chip continuously perform 14-time handshake protocol B, namely the trimming number of the chip to be trimmed is recorded as 14.
It should be noted that, the setting basis of the trimming prediction value is as follows: the closer to the gear, the better the average trimming frequency is, but the data is usually obtained from the analysis of the measured data, as shown in fig. 3 and 4, but after the first batch of test and the replacement of the batch, the fluctuation exists in the data, and the analysis result is not necessarily accurate, so that the trimming predicted value is updated by using the prediction algorithm, the purpose of adapting to the products in each batch is achieved by using the learning and adjusting capability of the algorithm, and the predicted value of each gear can be preset to 0 under the condition that the statistical data does not exist on the products, and then the algorithm is automatically adjusted.
It should be further noted that, under the statistical condition of no tested chip (the model chip is not tested before), the predicted value of each program initial stage can be set to 0, and the algorithm is used to realize the adjustment from 0 to a more accurate predicted value, and the specific process is shown in fig. 5.
S400: measuring the signal frequency of the chip to be trimmed after trimming to obtain a second measured value;
after trimming, the signal frequency of the chip to be trimmed is reduced compared with the first measured value 135.812KHz, at this time, the signal frequency of the chip is measured again by the tester, and the measured frequency is 127.6KHz (this data is only used for illustration, and other frequencies are possible), and the value is taken as the second measured value.
S500: and detecting the first chip to be repaired based on the comparison of the second measured value and the measurement standard value, if the detection is passed, judging the first chip to be repaired as a good product, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on the repair times of the first chip to be repaired, and terminating the repair program.
In this step, since the second measurement value is located at the measurement standard value L pass 127-128KHz, so the chip to be repaired is detected as good product, and the detection is passed. At the same time, correct handshake trimming times y of the chip to be trimmed 2014 (i.e., the number of modifications 14 recorded in step S300) the predicted trimming number of the 2015 chip also located in the 15 th stage can be obtained by the prediction method based on exponential smoothing. The prediction method based on exponential smoothing is a time series analysis prediction method developed on the basis of moving average method, and is characterized by that the past observations are given different weights, i.e. more recent observationsThe weight is greater than the weight of the long-term observations. In the same batch of chips with the same model, the actual measurement value and the trimming value are only affected by the physique of the chip, and the test data are irrelevant to the test time and the sequence, so that the time sequence of trimming data of the chips with the same gear does not have obvious trend change, and the chip with the same gear is predicted by using a primary index smoothing algorithm in an index smoothing prediction algorithm.
The prediction method based on the exponential smoothing is specifically expressed as follows:
S n+1 ＝ay n +(1-a)S n
wherein S is n The predicted value of the n period is obtained by predicting the n-1 period through an exponential smoothing algorithm; s is S n+1 The predicted value of the n+1 period is predicted by an exponential smoothing algorithm in the n period; y is n Is the actual observed value during period n; a is a smoothing constant with a value range of [0,1 ]]。
The prediction method is implemented by calculating an exponential smoothing value S n+1 (predicted value of the next period) and predicting the future by matching with a certain time sequence prediction model, the principle is that the predicted value S of the next period n+1 Are all actual observations y in this period n And the current prediction value S n Is a weighted sum of (c). In the formula, the smoothing constant a affects the predicted convergence rate of the algorithm, and if the smoothing constant a is large, the recent weight is large, the predicted value in the next period is greatly affected, and the convergence rate is increased. If the smoothing constant a is too large, the predicted value fluctuates greatly, and the prediction accuracy decreases. In the actual test of chips, the number of chips is larger and the data distribution is more regular, so that the selection of a smaller smoothing constant a is more beneficial to the test prediction adjustment value.
The calculation process for obtaining the trimming number prediction value of the 15 th-grade 2015 th chip according to the above method comprises the following steps:
S 2014+1 ＝ay 2014 +(1-a)S 2014 )，
S 2015 ＝0.1*y 2014 +(1-0.1)S 2014 (smoothing constant a here is taken as an example of 0.1)
S 2015 ＝0.1*14+0.9*14.58
＝14.522
At this time, L is calculated 15 The predicted value of trimming times of the 2015 th chip is S 2015 ＝14.522。
Because the number of trimming times of the 2014 chip is 14, the corresponding trimming prediction value S 2014 =14.58, calculated to obtain S 2015 ＝14.522，S 2015 The term 14.522 is a predicted value of trimming the 2015 th chip of the gear, so it can be seen that, after trimming by one chip, the predicted value of the next chip updated by the prediction algorithm is adjusted according to the latest chip, and the number of trimming times (14) of the new chip is approximated, that is, the prediction method based on the exponential smoothing has the capability of automatically adjusting and adapting to the new batch. As can be seen from FIG. 5, the number of trimming operations of the gear chip is mainly 14 and 15, and 16 trimming operations are required for a very small number of chips. The calculated average trimming number was 14.62. After the initial trimming value (default predicted value S 0 ) With 0, the predicted value can approach the actual value through about 30 chips. After the batch data prediction converges, the error between the predicted value and the actual value is 0.58, and the error source is mainly that the gear chip comprises 14 trimming and 15 trimming, and the predicted average value is 14.58. When the correction frequency is 14 by rounding the predicted value, the data measurement time of the gear chip 13 can be reduced by using an exponential smoothing-based prediction algorithm in actual measurement, the time spent by the chip in correction terms can be obviously reduced, and the overall test efficiency is improved.
In another embodiment, the method further comprises the steps of:
s600: in step S500, if the detection is failed, the first chip to be repaired is detected again according to whether the repair number of the first chip to be repaired reaches the upper limit of the repair number set in step S100.
In this embodiment, if the second measurement value of the first chip to be repaired is 128.2KHz, which exceeds 127-128KHz, the test is failed. At this time, since the trimming number of the chip is 14 times, which is smaller than the trimming number upper limit value (30 times) set in step 1, trimming can be continued.
In another embodiment, the method further comprises the steps of:
s700: in step S600, if the trimming frequency of the first chip to be trimmed reaches the upper limit of the trimming frequency, determining that the first chip to be trimmed is a defective product, and terminating the trimming procedure; and if the trimming frequency of the first chip to be trimmed does not reach the upper limit of the trimming frequency, comparing the second measured value of the first chip to be trimmed with the measured standard value again, and executing a corresponding handshake protocol according to the comparison result.
In this embodiment, since the measured value 128.2KHz is greater than the measured standard value 127-128KHz, it is known that the actual signal frequency of the chip is higher than the target value, so the signal frequency output by the chip needs to be adjusted to a lower frequency, and at this time, a handshake protocol B (including three handshake protocols, namely, protocol a, protocol B and protocol C, when trimming the chip, is needed, wherein protocol a represents the frequency-up adjustment, that is, if the measured signal frequency is lower than the target value, the protocol a is executed to adjust the signal frequency output by the chip to a higher frequency, and protocol B represents the frequency-down adjustment, that is, if the measured signal frequency is higher than the target value, the protocol B is executed to adjust the signal frequency output by the chip to a lower frequency, and protocol C represents the frequency-unadjusted, that is, if the measured signal frequency is within the target value, the protocol C is executed to prevent the signal frequency output by the chip from being continuously adjusted.
In another embodiment, the method further comprises the steps of:
s800: updating the trimming times of the first chip to be trimmed according to the corresponding handshake protocol to obtain the actual trimming times of the first chip to be trimmed.
In this embodiment, since the handshake protocol B is executed in the above embodiment, the number of trimming is added by 1 from the original 14, and the number of trimming after the chip update, that is, the actual number of trimming is 15.
In another embodiment, the method further comprises the steps of:
s900: and measuring the signal frequency of the first chip to be trimmed after the trimming times are updated, and obtaining a third measured value.
In another embodiment, the method further comprises the steps of:
s1000: and comparing the third measured value with the measurement standard value to perform third detection on the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as a good product, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on the actual repair times of the first chip to be repaired, and terminating the repair program.
In another embodiment, the method further comprises the steps of:
s1100: in step S1000, if the third detection fails, steps S600 to S1000 are repeatedly performed.
So far, all the embodiments described above constitute a complete technical solution of the present disclosure, and compared with the existing method, the method of the present disclosure has the following technical effects:
1. the trimming efficiency can be improved, and the trimming time is shortened;
the chip trimming method shown in fig. 2 requires a measurement comparison after each handshake adjustment is performed; while the present disclosure predicts the number of chip trimming by an algorithm, as described in the above embodiments, performs continuous handshake adjustment according to the predicted number of trimming, reduces the number of measurements performed after performing handshake signals, and in the example for L 15 A certain chip of the file is subjected to frequency measurement for the first time and then is subjected to frequency grading to obtain 14.58 of predicted trimming times of the chip, 14 handshaking signals B are continuously executed according to the predicted trimming times, and in the process, only the handshaking signals are subjected to frequency measurement before and after execution, and 2 times of frequency measurement are required; in the prior art, frequency measurement and comparison are required to be performed after each handshake signal is performed, and 15 frequency measurements are required.
In this example, the method can reduce 13 measurement operations of the gear chip, record trimming time of the chip of the model, perform a frequency measurement for 12ms, and perform a handshake protocol for 18ms. Using the prior art to L 15 The chip performs 14 trimming, and the required time is 15 frequency measurement times (12×15=180 ms) plus 14 handshake protocol times (18×14=180 ms)252 ms) and a total time of 434ms. The time required for trimming by using the method is 2 times of measurement time (12×2=24 ms) plus 14 times of execution handshake protocol time (18×14=252 ms), and the total time is 276ms. So that the trimming time reduction ratio of the chip is as follows
Figure BDA0004154021350000121
The gear trimming efficiency can be remarkably improved.
The method disclosed by the disclosure is applied to chip trimming of the configuration register, can take twice measurement time and once configuration register time through a prediction method, achieves the same effect of 14 times trimming, and can reduce 13 times measurement and 13 times register configuration time. By modifying the value S of the predictive chip n The configuration enters a register to carry out rough adjustment, so that the actual measured value of the chip approaches to the target value, and then the fine adjustment times of the chip are 0.58 times on average, thus the time reduction ratio of the trimming of the chip with the gear can be calculated and obtained
Figure BDA0004154021350000131
The absolute value error between the trimming value predicted by the prediction algorithm and the actual trimming value of the rest gear positions of the chip is not more than 1 through analogy calculation, so that the chip for trimming is realized by configuring a register after the prediction algorithm, the trimming frequency in the test process is not more than two times on average, and the chip trimming efficiency and the production cost are obviously improved on the premise of ensuring the trimming effect.
For trimming type chips, the trimming time optimizing effect is influenced by the trimming times required by the chip, and the more the trimming times are required, the more the time is shortened by using the method; the reduction ratio is affected by the measurement frequency, the measurement time for measuring the frequency of the high-frequency signal is shorter, and the lower the measurement frequency is, the more the trimming time is reduced.
2. The manual statistics data operation steps can be reduced;
the existing method for establishing a pre-trimming lookup table according to the measured chip data by adopting statistics needs to adopt the program structure shown in fig. 2 for trimming when detecting the first wafer or the first batch of chips by adopting a conventional method until finishing testing data after testing a batch, establishing a lookup table of the relation between the initial measurement frequency and the trimming times, and modifying the program to call the lookup table to continue the subsequent batch testing. Therefore, the existing method has complex operation and complicated steps, and needs to pause between the first batch test and the subsequent batch to sort data, which can affect the test efficiency.
The method uses an exponential smoothing prediction algorithm to realize automatic adjustment in the test process, the prediction effect of trimming times for the B-model chip is shown in figure 5, the data are obtained by actual measurement of the chip, and one batch of initial measurement frequency is selected to be L 15 Actual trimming times of a gear (135-136 KHz) chip, setting a smoothing coefficient a to 0.1, and initially defaulting to a predicted value S 0 Set to 0 and verified using an exponential smoothing prediction algorithm. As can be seen from the data analysis and FIG. 5, the number of trimming times of the gear chip is mainly 14 and 15, 16 trimming operations are required for a small number of chips, and the average trimming number is 14.62. After the batch data prediction converges, the prediction average value is 14.58, and the error between the predicted value and the actual value is 0.58, wherein the error source mainly comprises 14 trimming and 15 trimming of the gear chip, and the predicted value can approach the actual value after about 30 chips are tested.
The predicted value is updated in real time through the trimming data recorded in the testing process, the step of manually operating and sorting the lookup table is reduced, the program is not required to be paused and modified in the testing process, the use is more convenient, and the operation is simple.
3. The predicted value can be converged to the optimal range by automatically adjusting the products of each batch.
Because the chip has physique difference due to the problems of process and the like in the processing process, the trimming times of products in each batch can be different under the same condition of initial measured values, so that the use of single batch test data to establish a lookup table can not completely adapt to the physique of the chips in each batch, and the lookup table can have deviation, so that the prediction is inaccurate. In the method, the predicted value is updated according to the tested data in the test, and the predicted value can be adjusted and converged along with the chips of the new batch after the chips are replaced, so that the predicted value is more accurate, and the predicting effect of coping with different batches is improved.
In another embodiment, the present disclosure also provides a computer storage medium comprising:
a memory for storing a plurality of computer instructions;
a processor for executing computer instructions to implement a method as claimed in any preceding claim.
In another embodiment, the present disclosure further provides an electronic device, including:
a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein,
the processor, when executing the program, implements a method as described in any of the preceding.
The foregoing description of specific embodiments has been presented only to aid in the understanding of the present disclosure and is not intended to limit the present disclosure. Any local modification or substitution by one of ordinary skill in the art within the scope of the present disclosure is intended to be encompassed within the scope of the present disclosure.

Claims (10)

1. A chip trimming prediction method comprises the following steps:
s100: setting the upper limit of trimming times of the first chip to be trimmed;
s200: measuring the signal frequency of a first chip to be repaired and obtaining a first measured value;
s300: filing the first chip to be repaired based on the first measured value, repairing the first chip to be repaired according to the repairing prediction value of the gear of the first chip to be repaired, and recording repairing times;
s400: measuring the signal frequency of the chip to be trimmed after trimming to obtain a second measured value;
s500: and comparing the second measured value with the measurement standard value to detect the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as good, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on repair times of the first chip to be repaired, and terminating the repair program.
2. The method according to claim 1, wherein in step S500, the trimming prediction value of the second chip to be trimmed is preferably obtained by an exponential smoothing based prediction method.
3. The method of claim 1, wherein the method further comprises the steps of:
s600: in step S500, if the detection is failed, the first chip to be repaired is detected again according to whether the repair number of the first chip to be repaired reaches the upper limit of the repair number set in step S100.
4. The method of claim 1, wherein the method further comprises the steps of:
s700: in step S600, if the trimming frequency of the first chip to be trimmed reaches the upper limit of the trimming frequency, determining that the first chip to be trimmed is a defective product, and terminating the trimming procedure; and if the trimming frequency of the first chip to be trimmed does not reach the upper limit of the trimming frequency, comparing the second measured value of the first chip to be trimmed with the measured standard value again, and executing a corresponding handshake protocol according to the comparison result.
5. The method of claim 4, wherein the method further comprises the steps of:
s800: updating the trimming times of the first chip to be trimmed according to the corresponding handshake protocol to obtain the actual trimming times of the first chip to be trimmed.
6. The method of claim 5, wherein the method further comprises the steps of:
s900: and measuring the signal frequency of the first chip to be trimmed after the trimming times are updated, and obtaining a third measured value.
7. The method of claim 1, wherein the method further comprises the steps of:
s1000: and comparing the third measured value with the measurement standard value to perform third detection on the first chip to be repaired, if the detection is passed, judging the first chip to be repaired as a good product, obtaining a repair prediction value of the second chip to be repaired with the same gear as the first chip to be repaired based on the actual repair times of the first chip to be repaired, and terminating the repair program.
8. The method of claim 7, wherein the method further comprises the steps of:
s1100: in step S1000, if the third detection fails, steps S600 to S1000 are repeatedly performed.
9. A computer storage medium, comprising:
a memory for storing a plurality of computer instructions;
a processor for executing computer instructions to implement the method of any one of claims 1 to 8.
10. An electronic device, comprising:
a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein,
the processor, when executing the program, implements the method of any one of claims 1 to 8.
CN202310328348.2A 2023-03-30 2023-03-30 Chip trimming prediction method Pending CN116259559A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310328348.2A CN116259559A (en) 2023-03-30 2023-03-30 Chip trimming prediction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310328348.2A CN116259559A (en) 2023-03-30 2023-03-30 Chip trimming prediction method

Publications (1)

Publication Number Publication Date
CN116259559A true CN116259559A (en) 2023-06-13

Family

ID=86686320

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310328348.2A Pending CN116259559A (en) 2023-03-30 2023-03-30 Chip trimming prediction method

Country Status (1)

Country Link
CN (1) CN116259559A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116908651A (en) * 2023-07-21 2023-10-20 江阴市华拓芯片测试有限公司 Dynamic Trimming method and device for chip

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116908651A (en) * 2023-07-21 2023-10-20 江阴市华拓芯片测试有限公司 Dynamic Trimming method and device for chip

Similar Documents

Publication Publication Date Title
KR101930420B1 (en) Metrology sampling method with sampling rate decision scheme and computer program product thereof
US5327437A (en) Method for testing electronic assemblies in the presence of noise
US6374084B1 (en) Method and system for calibrating electronic devices using polynomial fit calibration scheme
US6245581B1 (en) Method and apparatus for control of critical dimension using feedback etch control
CN116259559A (en) Chip trimming prediction method
CN108829878B (en) Method and device for detecting abnormal points of industrial experimental data
CN113075527A (en) Integrated circuit chip testing method, system and medium based on Shmoo test
US20080004829A1 (en) Method and apparatus for automatic test equipment
CN112528230B (en) Parameter consistency control method and device based on precision and distribution conversion correction
CN113933672A (en) Method and system for judging correlation of wafer test parameters
US6615157B1 (en) System method and computer program product for automatically assessing experiment results
US5750908A (en) Testing system with real time/off line functionality allocation
US6442499B1 (en) Methods and apparatus for statistical process control of test
JP2021063742A (en) Abnormality discriminating device and abnormality discriminating method
CN111805301B (en) Measuring device and measuring method
JP2002323924A (en) Method and device for detecting defective device, program, and method for manufacturing product
CN110572875B (en) Wireless positioning method based on improved machine learning algorithm
CN114839518A (en) Effective key area parameter vector set reordering method and system
US20040236531A1 (en) Method for adaptively testing integrated circuits based on parametric fabrication data
US20040205052A1 (en) Modified binary search for optimizing efficiency of data collection time
CN115276841A (en) Power correction method for wireless module
CN115774185B (en) Vehicle-mounted chip DPAT detection method and device
CN114839514B (en) Dynamic optimization method and system for chip test engineering
CN110287610B (en) Offline mass production product technological parameter adjusting method and adjusting system thereof
CN117309299B (en) Servo driver vibration test method, device, equipment and medium

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