Verification of the placement machine process capability index CPK

For Qat (Qaulity Acceptance Test), the challenge is to ensure that the parameters to be measured accurately represent the long-term performance of the machine. The measurement must quantify and verify the offset of the X-axis, Y-axis, and q-rotation offset from the ideal placement position. One method used to verify placement accuracy uses a glass core that is stamped with a "perfect" high pin count QFP pad for machine mounting (see pin) Figure). By placing an ideal component, here is a 140-pin, 0.025" pitch QFP, the accuracy of both the camera and the mounting mandrel can be measured consistently. In addition to specific machine performance data, intrinsic usability Measurements of production capacity and reliability should be provided on the basis of cumulative data from multiple machines. After completing the pre-dry cycle and setup steps, including transformation and calibration, the Quality Acceptance Criteria (QAC) step begins. It is.

Eight stages of steps

QAC is the exact performance parameter that the placement machine must meet. The first step in the eight-stage QAC step is the initial 24-hour dry cycle, during which the machine must work continuously and without errors.

The second stage requires components to be accurately placed on two boards, each of which includes 32 140-pin glass core components. There are six global reference points on the main board, which are used as a reference for checking the placement accuracy of components before machine placement and visual measurement systems. The number of mounting plates depends on the specific head of the machine being tested and the configuration of the camera. For example, if the machine has two placement heads and two cameras, then 8 blocks must be placed with a total of 256 components (35,840 pins). board. This includes all possible combinations of patch heads and cameras.

Use all four mounting mandrels in all four directions: 0°, 90°, 180°, 270° mounting components. Following this step, each board is scanned with the measurement system to get a complete list of any offsets. Each 140-pin glass core contains two circular reference points with an accuracy of ± 0.0001" relative to the corresponding corner of the component and is used to calculate the offset of the X, Y, and q rotations. All 32 patches are The system measures and calculates the offset of each patch. This predetermined parameter is ± 0.003" in the X and Y directions, and the q rotation direction is ± 0.2. The machine must be placed for each component placement.

In order to pass the initial “jogging”, all 32 components placed on each side of the board must meet four test specifications: during operation, any placement position must not exceed ± 0.003” or ± 0.2. In addition, X The average deviation from the Y offset cannot exceed ± 0.0015”, their standard offset must be in the range of 0.0006”, the standard offset of q must be less than or equal to 0.047°, and the average offset is less than ± 0.06° , Cpk (Process capability index) is greater than 1.50 in all three quantized regions. This translates to a minimum of 4.5s or a maximum of approximately 3.4 defects per million (dpm, defects per million).

Typically, the performance factor now achieved exceeds the process capability index of 2.0, or approximately 2 defects per billion (6s performance). This measurement step allows the manufacturer to measure how satisfied his production requirements are.

After the accumulation is completed, a single performance data is used to calculate the average and standard offset of all components mounted on the board, and then Cpk is determined. The final QAC summary should be provided by the measurement system, listing the target location, offsetting the target amount, and calculating the pin-to-pad coverage for various pitches, in units of one thousandth of an inch, (Figure 1)

Machine property running

After the machine passed these tests, the machine attributes of the 3000 components began to run, with standard SMDs of all specifications, including: 0603, 0805, 1206, SOT23, SOIC08, PLCC44 and QFP100. It is a production ready operation, with no pass/fail criteria, which allows for the reset of any feeder or other mechanical part. Tracked defects may include upper facing components, side or upright components, offset components and any component with a pad footprint of less than 75%.

The next step in the evaluation process is another 12-hour non-stop dry cycle.

Next is the second glass core component running, requiring a 32-glass panel, camera and patch head, the same process as the second stage. (The number of patches is determined by the total number of combinations of head and camera.) Again, the requirements established for the first glass component must be met to allow the machine to continue QAC.

After completing the scheduled QAC step, the machine must undergo a final run of 3000 component placements, allowing only one error. If there is more than one placement error, the root cause must be found and corrected. At this stage, the intrinsic usability of the machine cannot be less than 98%.

The last 12 hours of dry cycle operation does not allow any failure. The final placement series required the placement of two plates with a 140-pin QFP glass core component, camera and patch head, scanning all placement locations with a measurement system, recording data, and charting.

Process capability index

The design geometry of the board and the required quality level determine the performance (accuracy and repeatability) requirements of the assembly equipment in a given application. For example, consider the following case where the component pins are placed on the pads of the board (Figure 2): Based on their geometry, the placement of the component (only in the X direction) is defined by the pin on the pad. ± {Wp – Wl)/2 + C (C is 0.001"). That is, when the pin placement position on the pad exceeds the specified limit in the direction of +X (prescribed upper limit) or -X (prescribed lower limit) , the board is not accepted.

Figure 2 Figure 2, Lead-to-pad coverage and deviation permitted

Once the board's worst-case specified limits are determined, the statistics provided by the equipment supplier can be placed in the following equation:

Cpk = minimum of {(USL – m ) / 3 s , (m - LSL) / 3 s }

Here, m and s are the mean and standard deviation, respectively, obtained by the device in many program-driven experiments. These data are provided along with a detailed test method description, and data is collected by this test method to ensure that the machine performance meets the specified performance requirements. The resulting value of Cpk is proportional to the standard deviation value, which falls between the distribution average and the most recent specified upper and lower limits. Cpk is important because it conveys the specified defect rate for a given placement accuracy (Figure 3).

Figure 3 Figure 3, The distribution of deviation from normal X, Y and Theta.

Support service

Providing support services to customers is a must for equipment manufacturers, as well as providing a sense of responsibility to keep the communications lines open to get accurate, understandable information about the machines. In order to meet informed needs, a complete machine performance documentation should be provided at the time of shipment, which should include a table containing all the machine specification data collected during the QAC test, as well as a document explaining the terms and types of information collected.

To help customers understand the QAC process and provide a contact address for answering technical questions about the device, the document must also include the name and signature of the person directly involved in the QAC assessment.

In addition to this information, operations, interruptions, failures, and total repair time plus total patch quantity, number of patch errors, intrinsic usability, intrinsic throughput, and placement performance at each PPM level, complete technical terminology and The definition of statistical relationships, etc., is also in the machine performance package.

Finally, to provide good prospects for long-term processing performance of equipment and the ability of equipment manufacturers to maintain control of internal manufacturing processes, historical data on the latest 70 machines should also be provided. This information can be used to calculate the average time between failures, between interruptions, and downtime as a continuation average. The data for 70 machines increases the total number of tiles from 10,000+ to 500,000+, resulting in associated cumulative benefits. In fact, these provide a long-term tracking process for machine performance and equipment manufacturer's internal measurement process.

in conclusion

The requirements associated with PCB assembly are becoming more and more intense. Due to the increased complexity of the process and the wide range of component types, placement machine performance must be more sophisticated in all aspects. Collecting and providing machine performance data in an understandable way is a key part of the partnership process. The measurement itself can only answer a certain aspect of the problem; a proper understanding and application of the results can achieve a complete understanding.