Calibration method and device

Abstract

The invention relates to a calibration method for device printing. The calibration method comprises the following steps: acquiring a measurement data set of a product in a production batch; determining a candidate calibration parameter set based on the measurement data set and product related parameters; and selecting a calibration parameter to be applied to another production batch from the candidate calibration parameter set according to a predetermined condition. According to the scheme, the yield can be improved while the printing requirement is met.

Description

Technical Field

[0001] The invention relates to semiconductor packaging, in particular to a solder paste printing technology on a soldering pad of an electronic device. Background Art

[0002] In production processes such as semiconductor device packaging, a solder paste printer is usually used to print solder paste on multiple pads of electronic devices so that the electronic devices can be adhered to substrates (such as FlipChip substrates) or other circuit boards such as PCB circuit boards. Usually, in order to facilitate batch operations, a tray is used to carry multiple electronic devices, wherein the tray has grooves arranged in a row-column matrix for carrying the electronic devices, so that the solder paste printing of the pads on multiple electronic devices can be completed at one time using a solder paste printer. However, in the actual printing process, due to various possible factors, the printing position of the solder paste may deviate from the target pad. When this deviation exceeds a certain limit, the electronic device will be unavailable, that is, it will fail, thereby reducing the yield rate of the product packaging.

[0003] To solve this problem, after a batch of solder paste printing is completed, the printed pads are usually inspected using a solder paste inspection tool to measure the physical offset of the actual printed solder paste relative to the theoretical solder paste position, and the calibration offset of the tray or the whole for the next batch of printing processing (for example, using the average value of the physical offset of all pads) is calculated, such as the rotation angle, horizontal and vertical offset of the tray, so that the solder paste printing tool controls the tray according to the corrected calibration offset, in order to control the printing offset of the next batch of pads within the allowable range. However, in practice, this calibration method still has problems, including:

[0004] (1) The solder paste printing effect obtained according to the currently adjusted calibration offset is still not ideal, which is manifested in that the offset of individual or part of the solder paste may still exceed the allowable limit, so multiple rounds of calibration and printing are required. In extreme cases, even a satisfactory offset may not be obtained, and manual adjustment has to be made, which is time-consuming and labor-intensive, and also leads to increased costs;

[0005] (2) Although the measurement results obtained using the solder paste inspection tool show that the offset is within the limit range, many failed electronic components are still found in actual production, and the yield rate is still a problem. Summary of the invention

[0006] It can be found that the offset calibration method in the prior art takes the average value of the solder paste offset on all pads as the starting point, and its purpose is to overcome the problem that the average value of the offset does not exceed the limit or is as close to the target as possible. However, due to the simultaneous overall printing of multiple devices, various manufacturing deviations such as fixtures will inevitably cause the solder paste on individual pads of some devices to exceed the standard during the solder paste printing process, thereby causing welding abnormalities when welding electronic devices. The present invention aims to improve the final yield rate of electronic devices, and on this basis further meets the requirement that the offset should not exceed the allowable limit as much as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic diagram showing the prior art of determining a calibration offset is shown;

[0008] Figure 2 A schematic diagram of pad offset distribution for different offset values ​​is shown;

[0009] Figure 3 A schematic diagram of calculating the deviation angle according to an example of the present invention is shown.

[0010] Figure 4 A flow chart of a calibration method according to an embodiment of the present invention is shown;

[0011] Figure 5 A flow chart for generating calibration offset parameters is exemplarily shown;

[0012] Figure 6 A flowchart for calculating multiple sets of candidate calibration offset sets is exemplarily shown;

[0013] Figure 7 A flow chart of selecting optimal parameters according to one embodiment is shown;

[0014] Figure 8 A flow chart of selecting optimal parameters according to another embodiment is shown. DETAILED DESCRIPTION

[0015] In the conventional calibration scheme of the prior art, since only the solder paste offset is considered, while the product yield and the error caused by the fixture, etc. are ignored, the calibration parameters obtained for the next batch of printing are difficult to meet the requirements. For example, Figure 1A schematic diagram of determining calibration parameters by a conventional automatic optical inspection (AOI) device in the prior art is shown. As shown in the figure, in the currently completed batch processing, the tray TR rotates by an angle θ to accept the solder paste PA printed by the solder paste printer using a printing template on the pads of each electronic unit (also referred to as an electronic component in this article) DU carried by the tray TR. As shown in the figure, each electronic unit DU may include multiple pads. Subsequently, the offset (dx, dy) of the printed solder paste relative to the pad is measured using the AOI device, where dx and dy represent the horizontal and vertical offsets of the measured pad, respectively, thereby forming an offset measurement data set {dx, dy}. The existing AOI equipment uses the method of calculating the average value (Mean) of all solder paste offsets of all electronic units to determine the pallet calibration offset parameters (off_dx, off_dy), where off_dx = -Ave (∑dx), represents the horizontal offset of the pallet to be applied, and the negative sign '-' represents that the direction of the applied horizontal offset off_dx is opposite to the offset direction of the calculated average value Mean (=Ave (∑dx)); similarly, off dy = -Ave(∑dy), represents the vertical offset of the pallet to be applied. Here, if the calibration offset (off_dx, off_dy) is less than the limit value, the currently calculated (off_dx, off_dy) is provided to the solder paste printer or the previously set (off_dx, off_dy) is continued to be used for the next batch of printing; on the contrary, if the calculated (off_dx, off_dy) is not less than the limit value, the currently calculated (off_dx, off_dy) is fed back to the solder paste printer to obtain the measurement data for the next batch of printing, so as to determine the best offset calibration parameters (off_dx, off_dy) through multiple tests. In another example, when the calculated (off_dx, off_dy) is not less than the limit value, the currently calculated (off_dx, off_dy) can also be provided to the printing simulation software for the solder paste printer, and the simulated printing effect under the current calibration parameters (off_dx, off_dy) can be determined, and the simulated offset of each simulated solder paste under the simulation can be further measured, so as to continue to calculate and evaluate the average offset until the calculated average offset meets the limit requirement.

[0016] Figure 2The figure shows a schematic distribution diagram of defective products - pads exceeding the standard, which is traditionally presented only based on the calibration offset. In this example, the AOI equipment determines three sets of pallet calibration offset parameters (off_dx, off_dy) by calculating the average value Mean of the solder paste offset of all electronic units as described above. To simplify the description, it is assumed that the lateral and longitudinal offsets calculated in each calibration process have the same value. For example, for the first calibration, the calculated average calibration offset Mean = 7.97917μm, for the second calibration, the calculated average calibration offset Mean = -3.6875μm, and for the third calibration, the calculated average calibration offset Mean = -8.6875μm. Figure 2 As shown, for ease of description, a schematic diagram of the distribution of pad offsets measured for each calibration process for six electronic devices in the tray is shown, where UCL represents the upper limit or upper limit of the solder paste allowed to exceed the pad, and LCL represents the lower limit or lower limit of the solder paste allowed to exceed the pad. The positive and negative values ​​here represent the direction in which the solder paste exceeds the pad, for example, positive represents upward or rightward excess, and negative represents downward or leftward excess. In addition, each group of black dots in the figure represents an electronic device, and the black dot represents a pad of the device. It can be seen that for the first calibration process, when the calibration offset is -7.97917μm, there are a total of four black dots that exceed the upper limit UCL, that is, the number of pads exceeding the limit OCC=4, and these black dots come from two electronic devices, so the yield rate Y=(6-2) / 6=67%. For the second calibration process, when the calibration offset is -3.6875, there are 6 black dots exceeding the lower limit LCL, that is, OCC=6, and these black dots come from one device, so the yield rate Y=(6-1) / 6=83%. For the third calibration process, when the calibration offset is -8.6875, all black dots are located between the upper and lower limits UCL and LCL, that is, OCC=0, so the yield rate Y=100%. It can be seen from the figure that according to the prior art, when considering calibration compensation with the minimum offset as the selection standard, off_dx=off_dy=Mean=-3.6875 is usually used as the selected offset calibration parameter for the next batch of printing, but it is obvious that the optimal yield rate cannot be achieved using the minimum offset calibration parameter, and the present invention hopes to find a calibration offset with the best yield rate (that is, as close to 100% as possible), rather than minimizing the calibration offset.

[0017] Figure 3 A schematic plan view of a tray TR is shown, which illustrates the physical properties of the tray. As shown in the figure, the tray is provided with a plurality of pockets in a matrix of R rows and C columns, so that it can carry R×C electronic units DU, for example, Figure 3 The tray can carry 3 rows and 5 columns of unit DU. Figure 3 As shown, the distance between the first groove and the last groove in a row is defined as the length L of the pallet, and the distance between the first groove and the last groove in a column is defined as the width W of the pallet; in addition, the distance between two adjacent grooves in a column is defined as the vertical groove pitch StepY, and the distance between two adjacent grooves in a row is defined as the horizontal groove pitch StepX.

[0018] In production, the groove matrix on the tray is used to accommodate the electronic unit DU to complete the solder paste printing on the pads on the electronic unit. Figure 3 The schematic diagram shows that the electronic unit DU placed in each groove has only one pad PD, and solder paste PA is printed on it. As shown in the figure, for the solder paste printed in the current completed batch, the entire solder paste is tilted relative to the pad, and different degrees of solder paste offset occur between different pads. For example, for the pad in the first row and the first column, the solder paste PA is vertically offset dy relative to the center of the pad. 1,1 , while for the last pad in row 1, the solder paste PA is vertically offset dy relative to the center of the pad 1,5 It is not difficult to see that the solder paste offset dy on these two pads 1,1 With dy 1,5 In the opposite direction. According to an embodiment of the present invention, a solder paste offset data set for each printing batch can be obtained from an AOI device.

[0019] In one embodiment of the present invention, in order to compensate for the solder paste offset to improve the yield rate, the actual solder paste deflection angle θ occurring in the current batch printing is first calculated using the measured offset data set, such as Figure 3 As shown, since the deflection angle of the entire solder paste is the same, the solder paste offset data dy of any two columns in any row can be selected to calculate θ. For example, the offset dy of the first column and the last C column in any row i can be selected. i,1 With dy i,c Calculate the deflection angle θ of the solder paste in the i-th row i ,Right now

[0020] in

[0021]

[0022] Here C represents the total number of columns of the groove matrix on the tray. To reduce the error, the deflection angles θ of all R rows can be calculated separately and their average value can be used as the deflection angle θ of the solder paste offset measured in the current batch and used as the deflection angle applied to the tray during the subsequent solder paste printing, that is,

[0023]

[0024] Based on the determined deflection angle θ, candidate calibration offset parameters, ie, offsets, are calculated through multiple cycles of operation.

[0025] Of course, the present invention is not limited to the above method when calculating the deflection angle θ, and other methods can also be used. For example, the solder paste offset data dx of any two rows in any column can be selected to calculate θ. For example, the lateral offset dx between the first row and the last Rth row in any column i can be selected. 1,i With dx R,i Calculate the deflection angle θ of the solder paste in the i-th column i ,Right now

[0026] in

[0027]

[0028] Similarly, to reduce the error, the deflection angles θ of all C columns can be calculated separately and their average value can be used as the deflection angle θ of the solder paste offset measured in the current batch and used for the deflection angle θ applied to the tray during subsequent solder paste printing, that is,

[0029]

[0030] Of course, according to another example of the present invention, the reference deflection angle θ applied to the tray may be determined by calculating a mixed average of the deflection angles calculated according to the rows and the deflection angles calculated according to the columns.

[0031] like Figure 4 As shown, in one example, the direction angle of each unit DU on the pallet relative to the physical center CC of the pallet TR is first determined. For any unit DU, such as the unit DU numbered k, the position coordinates of its unit center point relative to the physical center CC of the pallet are calculated as follows:

[0032]

[0033] In the formula, (x 1 ,y 1 ) represents the coordinates of a reference point S on the pallet, the symbol % represents the modulus operator, which is used to calculate the remainder after dividing two numbers; and / / represents floor division, which is used to calculate the integer part of the result of dividing two numbers and discard the decimal part. Here, the reference point S can be the reference position of a specific pad on the pallet. For the convenience of calculation, such as Figure 3 As shown, the center position of the groove located at the lower left corner of the tray, that is, the 3rd row and 1st column in the figure, is defined as the reference point S (x 1 ,y 1 ). In addition, in formulas (3) and (4), k represents the position number of each electronic unit DU in the tray. For example, Figure 3The tray shown can carry 3 rows × 5 columns of electronic units DU. The unit in the first row and first column in the lower left corner of the tray, i.e., DU 1,1 When marked as the first unit and assigned the starting number 1, the positions and numbers of each unit are shown in the following table:

[0034] <![CDATA[DU 3,1 :11]]> <![CDATA[DU 3,2 :12]]> <![CDATA[DU 3,3 :13]]> <![CDATA[DU 3,4 :14]]> <![CDATA[DU 3,5 :15]]> <![CDATA[DU 2,1 :6]]> <![CDATA[DU 2,2 :7]]> <![CDATA[DU 2,3 :8]]> <![CDATA[DU 2,4 :9]]> <![CDATA[DU 2,5 :10]]> <![CDATA[DU 1,1 :1]]> <![CDATA[DU 1,2 :2]]> <![CDATA[DU 1,3 :3]]> <![CDATA[DU 1,4 :4]]> <![CDATA[DU 1,5 :5]]>

[0035] Therefore, in this example, for a pallet with 3 rows × 5 columns, the value range of k is: 15 ≥ k ≥ 1. Using the calculated X and Y, the fixed direction angle α of each electronic unit DU on the pallet relative to the pallet center point CC can be calculated, that is,

[0036]

[0037] Since the deflection angle θ has an impact on the solder paste offset, in order to compensate for the impact of the deflection θ, it is assumed that the unit DU needs to be moved to the corresponding DU′ position. Therefore, the offset adjustment factor (Δdx) can be calculated for each unit DU as follows: k ,Δdy k ):

[0038] Δdx k =R×(cos(α)-cos(α+θ)) (Formula 6)

[0039] Δdy k =R×(sin(α+θ)-sin(α)) (Formula 7)

[0040] In the formula, R 2 =X 2 +Y 2 .

[0041] Thus, the offset adjustment factor set {Δdx k , Δdy k}, where k ranges from 1 to R*C.

[0042] It is not difficult to understand that according to Figure 2 As shown in the offset distribution, ideally, it is hoped that all the offsets of the solder paste are between UCL and LCL. Considering that there may be unreliable data due to accidental factors in the measured offset data, therefore, in one example of the present invention, according to the measured solder paste offset data set {dx, dy}, the reference offset range (dx MIN ,dx MAX ) and (dy MIN ,dy MAX), here we can use the percentile commonly used in statistics to determine the benchmark offset range from the measured offset data set. For example, when the percentile is set to 80 percentile, dx MIN The actual lateral offset at the 20th percentile position where the lateral offset is negative in the offset data set can be selected as dx MIN ; and dx MAX The actual lateral offset at approximately the 80th percentile position in the lateral offset positive value in the offset data set can be selected as dx MAX Similarly, dy MIN The actual longitudinal offset of the longitudinal offset in the offset data set at the 20th percentile position can be selected as dy MIN ; and dy MAX You can select the vertical offset in the offset data set as the actual vertical offset at the 20th percentile of the positive value as dy MAX Of course, the above only exemplifies a calculation method of the reference offset range, but the present invention is not limited thereto. For example, weighted UCL and LCL may also be used as the reference offset range.

[0043] Therefore, according to an example of the present invention, the reference offset range (dx MIN ,dx MAX ) and (dy MIN ,dy MAX ) is divided into m groups of reference offsets. Here, the equal division method can be used for division, where the step size of each group of reference offsets is: Δx = (dx MAX -dx MIN ) / m,Δy=(dy MAX -dy MIN ) / m, where m can be a predetermined value. For example, when m=10, the 10 groups of reference offsets are determined as follows:

[0044] For Group 1: (dx 1 ,dy 1 )=(dx MIN ,dy MIN )

[0045] For Group 2: (dx 2 ,dy 2 )=(dx MIN +Δx,dy MIN +Δy)

[0046] …

[0047] For group i: (dx i ,dy i ) = dx MIN +kΔx,dyMIN +kΔy)

[0048] …

[0049] For group 10: (dx 10 ,dy 10 )=(dx MAX ,dy MAX ).

[0050] It should be pointed out here that the present invention is not limited to the above-mentioned equal division range division method, and other division methods can also be used according to the distribution characteristics of the measurement data.

[0051] Therefore, based on the m groups of reference offsets and the offset adjustment factor set {Δdx k , Δdy k}, determine m sets of candidate calibration offset sets (Dx, Dy), where for each candidate offset (Dx i ,Dy i ),

[0052] Dx i,k = {dx i +Δdx k} (Formula 8)

[0053] Dy i,k = {dy i +Δdy k}, (Formula 9)

[0054] Where 1≤i≤m, 1≤k≤R×C. In this way, m sets of candidate offset sets {Dx i,k ,Dy i,k The above data include SUM candidate calibration offsets Dx in the lateral direction, i.e., the X direction, and Dy in the longitudinal direction, i.e., the Y direction, respectively, where SUM = m × R × C. According to the present invention, the best lateral and longitudinal offsets (Dx, Dy) are selected from the m sets of candidate offsets including SUM pairs according to a predetermined condition as calibration offset data (off_dx, off_dy) applied to the solder paste printer.

[0055] According to one embodiment of the present invention, from m groups of lateral candidate offset sets {Dx i,k} and determine the lateral calibration offset data (off_dx) from the m sets of vertical candidate offset sets {Dy i,k To this end, when determining the lateral calibration offset data (off_dx), the vertical offset can be set to any predetermined value, such as zero, and then the offset data (Dx i,k,0) Provide the simulation software of the solder paste printer to perform the simulation printing of the solder paste and obtain the simulated solder paste printing measurement data, that is, the simulated solder paste offset on each pad in each electronic device. Since the vertical offset is not considered at this time, only the horizontal simulated solder paste offset {dum_dx} of each pad can be measured and obtained.

[0056] Based on the lateral simulation offset data {dum_dx}, it can be determined that the current candidate calibration offset parameter (Dx i,k ,0), the yield rate of solder paste printing, in this embodiment, is expressed by calculating the proportion of electronic devices whose solder paste offset does not exceed the upper limit UCL and the lower limit LCL relative to all devices in the tray to express the yield rate Y. For example, in the case where the electronic unit DU has 8 pads, such as Figure 2 As shown, the unit DU contains 8 offset data (i.e. Figure 2 Therefore, in the present invention, if the horizontal offset in any of the solder paste offset data does not exceed the corresponding upper horizontal limit UCL or lower horizontal limit LCL, the device is confirmed to be a good part, otherwise it is a damaged or failed part. i,k ,0) under the yield rate Y i,k .

[0057] In this way, the SUM candidate calibration offset data {Dx i,k}, and the solder paste offset data obtained after the above method is used to calculate the SUM lateral candidate offset parameters {Dx i,k}Resulting in various yield rates:

[0058] [Y 1,1 ,Y 1,2 ,…Y 1,R×C ,Y 2,1 ,Y 2,2 ,…Y 2,R×C ,…,Y m,1 ,Y m,2 ,…Y m,R×C ].

[0059] According to an embodiment of the present invention, from m sets of candidate offset sets {Dx i,k} selects the lateral calibration offset parameter with the maximum yield as the calibration offset data off_dx applied to the solder paste printer. i,j =max(Y 1,1 ,Y 1,2 ,…Y 1,R×C ,Y 2,1 ,Y 2,2 ,…Y2,R×C ,…,Y m,1 ,Y m,2 ,…Y m,R×C ), select Y i,j The corresponding candidate calibration offset data Dx i,j As the lateral calibration offset off_dx applied to the solder paste printer.

[0060] In this way, when determining the longitudinal calibration offset data (off_dy), the lateral offset can be set to any predetermined value, such as zero, and then the offset data (0, Dy i,k ) is provided to the simulation software of the solder paste printer to perform the simulation printing of the solder paste and obtain the simulated solder paste printing measurement data, for example, only the longitudinal simulated solder paste offset {dum_dy} of each pad can be measured and obtained.

[0061] In a similar manner as before, based on the longitudinal simulation offset data {dum_dy}, the yield rate is expressed by calculating the proportion of electronic components whose solder paste offset does not exceed the upper limit UCL and the lower limit LCL relative to all components in the tray to determine the yield rate when applying the current candidate calibration offset parameter (0, Dy i,k )The yield rate of solder paste printing is Y' i,k Therefore, the SUM candidate calibration offset data {Dy in the m sets determined by simulation are i,k}, calculate the corresponding SUM longitudinal candidate offset parameters {Dy i,k}'s various yield rates:

[0062] [Y′ 1,1 ,Y′ 1,2 ,…Y′ 1,R×C ,Y′ 2,1 ,Y′ 2,2 ,…Y′ 2,R×C ,…,Y′ m,1 ,Y′ m,2 ,…Y′ m,R×C ].

[0063] According to an embodiment of the present invention, from m sets of candidate offset sets {Dy i,k} selects the longitudinal calibration offset parameter with the maximum yield as the calibration offset data off_dy applied to the solder paste printer. p,q =Mmax(Y′ 1,1 ,Y′ 1,2 ,…Y′ 1,R×C ,Y′ 2,1 ,Y′ 2,2 ,…Y′ 2,R×C ,…,Y′ m,1,Y′ m,2 ,…Y′ m,R×C ), then select Y′ p,q The corresponding candidate calibration offset data Dy p,q As the longitudinal calibration offset off_dy applied to the solder paste printer, where 1≤p≤m, 1≤q≤R×C.

[0064] Therefore, according to an embodiment of the present invention, the lateral offset and the longitudinal offset (Dx i,j ,Dy p,q ) as the calibration offset parameters (off_dx, off_dy) applied to the solder paste printer to print the next batch of solder paste.

[0065] According to an embodiment of the present invention, when there are multiple candidate offset data having the maximum yield rate at the same time, for example, assuming that the candidate offset Dx 3,4 ,Dx 5,6 , and Dx 8,9 Both correspond to the maximum yield rate of 95%, and the following minimum threshold principle is used to determine the final candidate offset:

[0066] For Dx 3,4 ,Dx 5,6 , and Dx 8,9 For each of the above, the maximum offset max{dum_dx} and the minimum offset min{dum_dx} in the corresponding simulated solder paste offset data set {dum_dx} are measured. As mentioned above, the offset is a direction value with positive and negative signs, so the maximum offset max{dum_dx} is a positive value, while the minimum offset min{dum_dx} is a negative value. Then, the minimum threshold Margin for each candidate calibration offset parameter is determined, where Margin = Min(UCL-max{dum_dx,},min{dum_dx}-LCL) (Formula 10)

[0067] In this example, using the above formula (10), Dx can be determined 3,4 ,Dx 5,6 , and Dx 8,9 The corresponding minimum threshold Margin 3,4 、Margin 5,6 、Margin 8,9 , so you can choose Margin 3,4 、Margin 5,6 、Margin 8,9 The candidate calibration offset parameter corresponding to the maximum value in is used as the calibration offset off_dx applied to the solder paste printer. 3,4>Margin 5,6 >Margin 8,9 When the candidate offset Dx 3,4 , as the final calibration offset off_dx. Similarly, when there are multiple longitudinal candidate offset data with the maximum yield rate at the same time, the above method can also be used to select the threshold parameter with the maximum value as the final calibration offset off_dy. It can be seen that the calibration offset parameters (off_dx, off_dy) determined by the scheme of the present invention first ensure the yield rate, thereby truly meeting the quality target, and can further meet the offset error requirements. Therefore, according to the scheme of the present invention, not only can the yield rate be significantly improved, but also the calibration process can be improved, avoiding multiple rounds of calibration-failure-recalibration cycles, and due to the improvement of the yield rate, even the number of printing product lines can be saved.

[0068] The above shows a scheme for separately determining the horizontal calibration offset data (off_dx) and the vertical calibration offset data (off_dy), but the present invention is not limited thereto. According to another embodiment of the present invention, the offset data (Dx) in the m sets of candidate offset sets may also be determined. i,k ,Dy i,k ) are synchronously determined in pairs. To this end, each pair of offset data (Dx i,k ,Dy i,k ) is provided to the simulation software of the solder paste printer to perform the simulation printing of the solder paste and obtain the simulated solder paste printing measurement data, that is, the simulated solder paste offset on each pad in each electronic device, including the lateral simulation offset dum_dx and the longitudinal simulation offset dum_dy, which is represented by (dum_dx, dum_dy) here.

[0069] Based on the simulated solder paste offset data (dum_dx, dum_dy), it can be determined that the current candidate calibration offset parameters (Dx j ,Dy j ), the yield rate of solder paste printing is expressed by calculating the proportion of electronic devices whose solder paste offset does not exceed the upper limit UCL and the lower limit LCL relative to all devices in the tray. For example, in the case where the electronic unit DU has 8 pads, such as Figure 2 As shown, the unit DU contains 8 offset data (i.e. Figure 2 Therefore, in the present invention, if the horizontal or vertical offset in any of the solder paste offset data does not exceed the corresponding horizontal or vertical upper limit UCL or horizontal and vertical lower limit LCL, the device is confirmed to be a good part, otherwise it is a damaged or failed part.i,k ,Dy i,k ) i,k .

[0070] In this way, the solder paste offset data obtained after applying SUM of the candidate calibration offset data in the m sets to the solder paste printer is determined by simulation, and the overall yield Y″ for the SUM candidate offset data is calculated using the aforementioned method:

[0071] [Y″ 1,1 ,Y″ 1,2 ,…Y″ 1,R×C ,Y″ 2,1 ,Y″ 2,2 ,…Y″ 2,R×C ,…,Y″ m,1 ,Y″ m,2 ,…Y″ m,R×C ].

[0072] According to the present invention, the candidate calibration offset with the maximum yield is selected from the m sets of candidate offsets as the calibration offset data (off_dx, off_dy) applied to the solder paste printer. For example, when Y″ i,j =max(Y″ 1,1 ,Y″ 1,2 ,…Y″ 1,R×C ,Y″ 2,1 ,Y″ 2,2 ,…Y″ 2,R×C ,…,Y″ m,1 ,Y″ m,2 ,…Y″ m,R×C ), select Y″ i,j The corresponding candidate calibration offset data (Dx i,j ,Dy i,j ) as the calibration offset parameters (off_dx, off_dy) applied to the solder paste printer to implement solder paste printing.

[0073] Similarly, according to an embodiment of the present invention, when there are multiple candidate offset data that have the highest yield rate at the same time, for example, assuming that the candidate offset (Dx 3,3 ,Dy 3,3 ), (Dx 5,5 ,Dy 5,5 ) and (Dx 9,9 ,Dy 9,9 ) all correspond to the maximum yield rate of 98%, then the following minimum threshold principle is used to determine the final candidate offset.

[0074] In this example, for (Dx 3,3 ,Dy 3,3 ), (Dx5,5 ,Dy 5,5 ) and (Dx 9,9 ,Dy 9,9 ), the corresponding simulated solder paste offset data set {dum_dx, dum_dy} includes the maximum offset max{dum_dx, dum_dy} and the minimum offset min{dum_dx, dum_dy} including the horizontal offset and the vertical offset data. As mentioned above, the offset is a direction value with a positive or negative sign, so the maximum offset max{dum_dx, dum_dy} is a positive value, while the minimum offset min{dum_dx, dum_dy} is a negative value. Then, the minimum threshold Margin for each candidate calibration offset parameter is determined, where

[0075] Margin=Min(UCL-max{dum_dx,dum_dy},min{dum_dx,dum_dy}-LCL) (Formula 11)

[0076] It should be noted here that in the above formula (11), it is assumed that the allowed horizontal and vertical maximum offset limit values ​​are the same UCL; however, in other embodiments, when the allowed horizontal and vertical maximum offset limit values ​​UCL have different values, when the maximum value of max{dum_dx, dum_dy} is the horizontal offset dum_dx, then the UCL in formula (11) corresponds to the horizontal maximum UCL, similarly, when the maximum value of max{dum_dx, dum_dy} is the vertical offset dum_dy, then the applied UCL corresponds to the vertical maximum UCL. The situation is similar for the minimum limit LCL. When the allowed horizontal and vertical minimum offset limit values ​​LCL have different values, when the minimum value of min{dum_dx, dum_dy} is the horizontal offset dum_dx, then the applied LCL corresponds to the horizontal minimum LCL, similarly, when the minimum value of min{dum_dx, dum_dy} is the vertical offset dum_dy, then the applied LCL corresponds to the vertical minimum LCL.

[0077] In this example, using the above formula (11), it can be determined that (Dx 3,3 ,Dy 3,3 ), (Dx 5,5 ,Dy 5,5 ) and (Dx 9,9 ,Dy 9,9 ) corresponds to the minimum threshold Margin 3,3 、Margin 5,5 、Margin 9,9 , so you can choose Margin 3,3、Margin 5,5 、Margin 9,9 The candidate calibration offset parameter corresponding to the maximum value in is used as the calibration offset (off_dx, off_dy) applied to the solder paste printer. For example, when Margin 5,5 >Margin 3,3 >Margin 9,9 When the candidate offset (Dx 5,5 ,Dy 5,5 ) is provided to the solder paste printer as the final calibration offset (off_dx, off_dy). It can be seen that the calibration offset determined by the solution of the present invention first ensures the yield rate, thereby truly meeting the quality target, and can further meet the offset error requirement. According to the solution of the present invention, not only can the yield rate be significantly improved, but also the calibration process can be improved, avoiding multiple rounds of calibration-failure-recalibration cycles, and due to the improvement of the yield rate, even the number of printed product lines can be saved.

[0078] The offset calibration method implemented according to the present invention can be conveniently installed in the form of a computer executable program on an existing calibration device, such as an AOI device, so that one or more processors on the calibration device execute the program to generate calibration offset parameters suitable for the next batch of solder paste printing based on the solder paste offset measurement data collected on the device. Figure 5 A flow chart of a method for generating calibration offset parameters according to an embodiment of the present invention is shown.

[0079] like Figure 5 As shown, in step 501, a solder paste offset measurement data set of electronic units DU on a tray under a production batch is obtained. The production batch can be the latest completed production batch or a previously completed batch. As an example, these measurement data can be directly obtained from the AOI equipment. These measurement data represent the offset measurement (dx) of the printed solder paste relative to the pad of all device DUs in the tray. i ,dy i ).

[0080] In step 503, based on the measurement data set (dx i ,dy i ) and product-related parameters to determine multiple groups of candidate calibration parameters (Dx, Dy), where the product-related parameters include the properties of the tray, such as the number of rows and columns of grooves for accommodating electronic components and the number of devices accommodated thereby (Panel), tray length L, width W, horizontal and vertical slot spacing StepX and StepY, and horizontal and vertical maximum allowable limits UCL and LCL.

[0081] In another example, the product-related parameters may also include attribute information of the electronic unit, such as pad size information to be received by the solder paste, etc. In this example, before determining the candidate calibration parameters (Dx, Dy), it may be selected to use the pad size information to receive the offset measurement data set (dx i ,dy i ) for filtering. The larger the pad size, the larger the allowable pad offset. Therefore, in order to reduce the amount of data processing, the pad size can be used to filter out measurement data with very little offset or almost negligible offset relative to the pad size. For example, assuming that the pad is circular and has a radius of 50μm, the measurement data of a solder paste with a lateral offset dx and a longitudinal offset dy that are both less than λ·50 can be filtered out by setting a filtering threshold λ, where λ can be set to a predetermined value such as 10%. It should be pointed out here that this data filtering process is not necessary and can be applied when considering reducing the data processing burden.

[0082] As mentioned above, to determine the candidate calibration parameters (Dx, Dy), we first use the measured offset (dx i ,dy i ) calculates the actual deflection angle θ of the solder paste deviation in the current batch, and based on the determined deflection angle θ and the measurement data set {dx, dy}, calculates multiple sets of candidate calibration offset sets (Dx, Dy) through multiple loop operations. Figure 6 The flowchart of calculating multiple candidate calibration offset sets according to one embodiment of the present invention is shown.

[0083] like Figure 6 As shown, in step 601, the reference offset angle θ is determined based on the position offset (dx, dy) of the pads of at least two different electronic units and the properties of the tray using formula (1). The properties of the tray here include the number of rows R, the number of columns C, and the lateral spacing StepX of the grooves in the groove matrix on the tray. In a more preferred example, using formula (2), the reference offset angle is determined for each row of R rows of electronic units on the tray TR and the average value of the offset angles of all rows is used as the reference offset angle θ.

[0084] In step 603, the azimuth angle α of each electronic unit TU carried by the pallet TR is determined based on the pallet attributes; as an example, the azimuth angle of each unit TU relative to a predetermined reference point on the pallet, such as the physical center point CC of the pallet, is calculated using formulas (3), (4), and (5).

[0085] In step 605, based on the reference offset angle θ and the azimuth angle α, an offset adjustment factor (Δdx, Δdy) of each electronic unit DU on the tray TR is generated; as an example, the offset adjustment factor (Δdx, Δdy) caused by the offset of the angle θ can be calculated using formulas (6) and (7).

[0086] In step 607, m groups of reference offsets are determined based on the measured solder paste offset data set {dx, dy}. As an example, to determine the m groups of reference offsets, the percentile commonly used in statistics can be used to determine the reference offset range (dx, dy) from the measured offset data set {dx, dy}. MIN ,dx MAX ) and (dy MIN ,dy MAX ); then divide the reference offset range into m segments as needed, and select the boundary value of each segment as the reference offset of each group. For example, when the reference offset range is divided by the equal division method, the reference offsets of each group are obtained as follows:

[0087] For Group 1: (dx 1 ,dy 1 )=(dx MIN ,dy MIN )

[0088] For Group 2: (dx 2 ,dy 2 )=(dx MIN +Δx,dy MIN +Δy)

[0089] …

[0090] For group i: (dx i ,dy i )=(dx MIN +kΔx,dy MIN +kΔy)

[0091] ...for the mth group: (dx m ,dy m )=(dx MAX ,dy MAX ).

[0092] In step 609, using formulas (8) and (9), based on the m groups of reference offsets and the offset adjustment factor set {Δdx k , Δdy k}, generate m sets of candidate calibration offset sets {Dx, Dy}, where the set contains SUM = m × R × C pairs of candidate calibration offsets (Dx i,k ,Dyi,k )where m≥i≥1,R×C≥k≥1.

[0093] Back to Figure 5 In step 505, according to a predetermined condition, the best set of parameters (Dx, Dy) is selected from the m sets of candidate offset parameter sets as calibration offset parameters (off_dx, off_dy) applied to the subsequent batch, such as the next batch of solder paste printers. As mentioned above, the present invention can adopt multiple implementations to select the best parameters (Dx, Dy). Figure 7 A flowchart of selecting the best parameter in an independent mode according to an example of the present invention is shown, which is used to select the best parameter from m groups of lateral candidate offset sets {Dx i,k} and determine the lateral calibration offset data (off_dx) from the m sets of vertical candidate offset sets {Dy i,k} to determine the vertical calibration offset data (off_dy). To do this, Figure 7 As shown, in step 701, the offset data (Dx i,k ,0) is provided to the simulation software of the solder paste printer to perform the simulated printing of the solder paste and obtain the lateral simulated solder paste offset {dum_dx} of each pad in each electronic device on the tray.

[0094] In step 703, based on the lateral simulation offset data {dum_dx}, the current candidate calibration offset parameter (Dx i,k ,0)The yield rate of solder paste printing is Y i,k In this way, in step 703, the SUM lateral candidate offset parameters {Dx i,k}Resulting in various yield rates:

[0095] [Y 1,1 ,Y 1,2 ,…Y 1,R×C ,Y 2,1 ,Y 2,2 ,…Y 2,R×C ,…,Y m,1 ,Y m,2 ,…Y m,R×C ].

[0096] Then, in step 705, from the m sets of candidate offset sets {Dx i,k} selects the lateral calibration offset parameter with the maximum yield as the calibration offset data off_dx applied to the solder paste printer. i,j =max(Y 1,1 ,Y 1,2 ,…Y 1,R×C ,Y 2,1 ,Y 2,2 ,…Y2,R×C ,…,Y m,1 ,Y m,2 ,…Y m,R×C ), select Y i,j The corresponding candidate calibration offset data Dx i,j As the lateral calibration offset off_dx applied to the solder paste printer.

[0097] In this way, by setting the lateral offset to any predetermined value such as zero, and then executing steps 701-705, the offset data (0, Dy i,k ) is provided to the simulation software of the solder paste printer to perform the simulation printing of the solder paste, thereby selecting the m sets of candidate offsets {Dy i,k} Select the longitudinal calibration offset with the maximum yield, such as Dy p,q As the longitudinal calibration offset off_dy applied to the solder paste printer.

[0098] Thus, in step 505, the determined lateral offset and longitudinal offset (Dx i,j ,Dy p,q ) as the calibration offset parameters (off_dx, off_dy) applied to the solder paste printer to print the next batch of solder paste.

[0099] like Figure 7 As shown, according to a further embodiment of the present invention, step 707 may also be included to determine whether there are multiple candidate offset data (Dx, Dy) that have the maximum yield at the same time. If not, the (Dx i,j ,Dy p,q ) is provided to the printer as the calibration offset parameter (off_dx, off_dy). Otherwise, if it is determined that there are multiple candidate lateral offset data having the same maximum yield at the same time, for example, assuming that the candidate offset Dx is determined 3,4 ,Dx 5,6 , and Dx 8,9 Both correspond to the maximum yield rate of 95%, and the process proceeds to step 709.

[0100] In step 709, for Dx 3,4 ,Dx 5,6 , and Dx 8,9 For each of the above, the maximum offset max{dum_dx} and the minimum offset min{dum_dx} in the corresponding simulated solder paste offset data set {dum_dx} are measured. Then, in step 711, the minimum threshold Margin for each candidate calibration offset parameter is determined. As an example, Dx can be determined using formula (10): 3,4 ,Dx5,6 , and Dx 8,9 The corresponding minimum threshold Margin 3,4 、Margin 5,6 、Margin 8,9 However, in step 713, select Margin 3,4 、Margin 5,6 、Margin 8,9 The candidate calibration offset parameter corresponding to the maximum value in is used as the calibration offset off_dx applied to the solder paste printer. 3,4 >Margin 5,6 >Margin 8,9 When the candidate offset Dx 3,4 , as the final calibration offset off_dx. Similarly, when it is determined in step 507 that there are multiple longitudinal candidate offset data with the maximum yield at the same time, the parameter of the threshold with the maximum value is selected as the final calibration offset off_dy by executing steps 709-713. Thus, in step 715, the offset (Dx, Dy) with the maximum yield and the maximum threshold is output as (off_dx, off_dy).

[0101] Figure 8 FIG. 1 shows a flowchart of selecting the best parameter using the joint mode according to another embodiment of the present invention, for selecting the best parameter from m sets of candidate offset sets {Dx i,k ,Dy i,k} synchronously determine the calibration offset data (off_dx, off_dy). To this end, in step 801, each pair of parameters (Dx i,k ,Dy i,k ) is provided to the simulation software of the printer to perform simulated printing of solder paste and obtain simulated solder paste printing measurement data (dum_dx, dum_dy).

[0102] In step 803, after obtaining the simulated measured offset data (dum_dx, dum_dy), calculate the offset data ((Dx i,k ,Dy i,k )) Under the following conditions, the yield rate [Y″ caused by printing solder paste 1,1 ,Y″ 1,2 ,…Y″ 1,R×C ,Y″ 2,1 ,Y″ 2,2 ,…Y″ 2,R×C ,…,Y″ m,1 ,Y″ m,2 ,…Y″ m,R×C], thereby selecting the candidate offset data (Dx, Dy) with the highest yield rate as the actual calibration offset parameters (off_dx, off_dy) to be applied to the printing tool in step 805. For example, when Y″ i,j =max(Y″ 1,1 ,Y″ 1,2 ,…Y″ 1,R×C ,Y″ 2,1 ,Y″ 2,2 ,…Y″ 2,R×C ,…,Y″ m,1 ,Y″ m,2 ,…Y″ m,R×C ), select Y″ i,j The corresponding candidate calibration offset data (Dx i,j ,Dy i,j ) is used as a calibration offset parameter applied to a solder paste printer to perform solder paste printing.

[0103] In a further embodiment, step 807 is further included to determine whether there are multiple candidate offset data with the maximum yield at the same time. If not, the offset (Dx i,j ,Dy i,j ) is directly provided to the printer as the calibration offset parameter (off_dx, off_dy). Otherwise, if it is determined that there are multiple candidate offset data that have the highest yield rate at the same time, for example, assuming that the candidate offset (Dx 3,3 ,Dy 3,3 ), (Dx 5,5 ,Dy 5,5 ) and (Dx 9,9 ,Dy 9,9 ) all correspond to the maximum yield rate of 98%, then the process goes to step 807, and the final candidate offset is determined using the following minimum threshold principle.

[0104] In step 807, for (Dx 3,3 ,Dy 3,3 ), (Dx 5,5 ,Dy 5,5 ) and (Dx 9,9 ,Dy 9,9 ), measure the maximum offset max{dum_dx,dum_dy} and the minimum offset min{dum_dx,dum_dy} in the corresponding simulated solder paste offset data set {dum_dx,dum_dy}, including the horizontal offset and vertical offset data. Then determine the minimum threshold Margin for each candidate calibration offset parameter according to formula (10): 3,3 、Margin 5,5 、Margin9,9 .

[0105] Then in step 809, select Margin 3,3 、Margin 5,5 、Margin 9,9 The candidate calibration offset parameter corresponding to the maximum value in is used as the calibration offset (off_dx, off_dy) applied to the solder paste printer. For example, when Margin 3,3 >Margin 5,5 >Margin 9,9 When the candidate offset (Dx 5,5 ,Dy 5,5 ) is provided to the solder paste printer as the final calibration offset (off_dx, off_dy).

[0106] Although the above-mentioned embodiments illustrate the general embodiments of the present invention by taking solder paste printing as an example, the present invention is not limited to solder paste printing, but can also be used in other mass production processes, such as miniLED packaging, in which multiple miniLEDs need to be installed on a substrate. Therefore, the solution of the present invention can be used to detect whether the miniLED is installed in place as required, and generate an adjustment value based on the detection result to control the packaging of the miniLED as required.

[0107] Although the exemplary calibration device and method of the present invention are described above in conjunction with specific examples, it will also be appreciated by those skilled in the art that the various illustrative method steps described in conjunction with the disclosure herein can be implemented as electronic hardware, computer software, or a combination of the two. For example, the calibration device according to the present invention can be implemented as a processor or a main controller and a memory, wherein the memory stores various modules in the form of a computer program, wherein the processor can implement the calibration parameter generation method of the present invention by executing these modules. In addition, a machine-readable medium provided in another embodiment of the present invention stores machine-readable instructions, which, when executed by the processor, cause the processor to execute any of the aforementioned methods disclosed herein. These embodiments are also within the scope of protection of the present invention.

Claims

1. A calibration method, comprising: Obtain a measurement data set for a production batch of products; determining a candidate calibration parameter set based on the measurement data set and product related parameters, Calibration parameters to be applied to another production batch are selected from the candidate calibration parameter set according to predetermined conditions.

2. The calibration method of claim 1, wherein the production comprises printing solder paste on pads of electronic units positioned on a tray using a printer, wherein the measurement data set comprises a positional offset of the solder paste relative to the pad for each electronic unit, and the product-related parameters comprise at least one of: tray attributes and pad attributes.

3. The calibration method of claim 2, wherein determining the candidate calibration parameter set comprises: Determining a reference offset angle based on positional offsets of pads of at least two different electronic units and properties of the tray; determining an azimuth angle of each electronic unit carried by the tray based on the tray attributes; The candidate calibration parameter set is determined based on the reference offset angle and the azimuth angle.

4. The method of claim 3, wherein the electronic units are arranged in a row-column matrix in the tray, wherein the at least two different electronic units are located in the same row or the same column of the tray.

5. The method of claim 3 or 4, further comprising: For the measurement data set, determining a reference offset range; selecting a plurality of sets of reference offsets from the reference offset range; generating an offset adjustment factor for each electronic unit on the tray based on the reference offset angle and the azimuth angle, and The candidate calibration parameter set is generated based on the plurality of sets of reference offsets and offset adjustment factors.

6. The method of claim 5, wherein the position offset of the pad comprises a lateral offset and a longitudinal offset; Determining the reference offset range includes processing the measurement data set using a percentile method to determine a lateral reference offset range and a longitudinal reference offset range, respectively.

7. The method of claim 6, wherein selecting calibration parameters for a next product batch from the candidate calibration parameter set according to a predetermined condition comprises: performing a solder ball simulation printing test using each candidate calibration parameter in the candidate calibration parameter set; Determining the printing yield of the electronic unit in the tray obtained by the simulated printing; A candidate calibration parameter with a maximum yield rate is selected from the candidate calibration parameter set as the calibration parameter to be applied.

8. The method of claim 7, wherein the product-related parameters further include: a predetermined range of excursion limits; in Determining the yield rate includes: Calculating the qualified number of electronic units whose simulated offset of the virtual printed solder balls of each electronic unit obtained by the simulated printing relative to the solder pad does not exceed the predetermined offset limit range; The yield rate is calculated based on the qualified quantity.

9. The method of claim 8, wherein the predetermined offset limit range includes a predetermined longitudinal offset limit range and a lateral offset limit range; The simulated offset includes a simulated longitudinal offset and a simulated lateral offset of the virtual printed solder ball relative to the solder pad.

10. The method of claim 9, wherein when there are multiple candidate calibration parameters having the maximum qualified number, the calibration parameter to be applied is selected according to a minimum threshold principle, wherein the minimum threshold principle comprises: For each candidate calibration parameter among the plurality of candidate calibration parameters, calculating a first difference between a larger value of the simulated longitudinal offset and the simulated lateral offset and an upper limit of the limit values, and a second difference between a smaller value of the simulated longitudinal offset and the simulated lateral offset and a lower limit of the limit values, and selecting the smaller of the first difference and the second difference as a threshold value of the candidate calibration parameter; One of the plurality of candidate calibration parameters having a maximum threshold amount is selected as the calibration parameter to be applied.

11. The calibration method of claim 1 , wherein the pad attribute comprises a pad size, the method further comprising: A filtering process is performed on the measurement data set using the pad size.

12. The method of claim 4, further comprising: The offset angles are calculated for the electronic units in different rows and / or columns in the tray, and an average value of the offset angles in the different rows and / or columns is used as the reference offset angle.

13. A calibration device comprising at least one processor, wherein the processor executes a machine-readable program to implement the calibration method of any one of claims 1-12.

14. A computer program product comprising a machine-readable program, wherein when executed by at least one processor, the machine-readable program implements the calibration method of one of claims 1 to 12.

Images