SPC Management Specification: Ensuring Process Capability and Preventing Defects

"SPC Management Specification\n\n1. Purpose\nTo use control charts and process capability calculations in relevant processes to ensure that the processes are capable and to prevent defects from occurring.\n\n2. Scope\nApplicable to the management of control charts specified in control plans.\n\n3. Responsibilities\n3.1 The project team is responsible for selecting statistical techniques, including the use of control charts, in the product quality advanced planning and control plan.\n3.2 The production department is responsible for implementing corrective measures for control chart abnormalities.\n3.3 The quality department is responsible for the use of control charts, calculation of PPK and CPK values, and the rate of non-conforming products. They are also responsible for training personnel on the use of control charts and, if necessary, guiding the production department in developing corrective measures.\n3.4 The personnel department is responsible for planning and organizing the training of personnel involved in control chart usage and process capability calculations.\n\n4. Definitions\n4.1 -R Chart: Control chart for mean and range.\n4.2 P Chart: Control chart for non-conforming product rate.\n4.3 PPK: Short-term process capability index.\n4.4 CPK: Long-term process capability index.\n4.5 5M1E: Man, Machine, Material, Method, Environment, and Measurement.\n\n5. Job Content\n5.1 Control Chart Format:\n5.1.1 Use the standard format specified in the external document "Statistical Process Control (SPC)."\n5.1.2 If the customer has a specified format, use the customer's format.\n5.1.3 To ensure the reliability and accuracy of control, control charts should be automatically generated from external documents. The production staff is prohibited from manually drawing control charts. (Manual recording is allowed on the production line, and the recorded content is entered into electronic files for automatic generation.)\n\n5.2 Steps for Plotting Control Charts (Measurement -R Chart):\n5.2.1 Collect Data:\n5.2.1.1 Select subgroup size, frequency, and data:\na. Subgroup Size: The subgroup size for -R charts in our company is 5, meaning that 5 consecutive products are sampled each time.\nb. Subgroup Frequency: Sample once every 2 hours.\nc. Subgroup Count: Normally 25 subgroups.\n5.2.1.2 Establish control charts and record raw data. The control chart user records relevant dates, times, readings, etc. in the corresponding columns of the -R chart.\n5.2.1.3 Calculate the mean () and range (R) for each subgroup using the following formulas and record the results in the corresponding columns of the -R control chart:\nX = (X1 + X2 + \u2026 + Xn) / n\nR = Xmax - Xmin\n5.2.1.4 Determine the width of the control chart: The range of the coordinates on the X chart is max - min = 2X(max - min), and the maximum scale value on the R chart is 2Rmax.\n5.2.1.5 Plot the mean and range R on the control chart.\n\n5.2 Calculate Control Limits:\n5.2.2.1 Calculate the average range () and process average ():\nR = (R1 + R2 + \u2026 + Rk) / k\nX = (X1 + X2 + \u2026 + Xk) / k (k is the number of subgroups)\n5.2.2.2 Calculate the control limits:\nUCLR = D4 \u00D7 \nUCLX = + A2 \u00D7 \nLCLR = D3 \u00D7 \nLCLX = - A2 \u00D7 \nwhere D4, D3, and A2 are constants.\n5.2.2.3 Plot the control limits: Draw horizontal dashed lines representing the control limits on the mean chart () and range chart (R). In the initial study phase, these limits are called trial control limits.\n5.2.2.4 Use the calculated UCL and LCL from the previous X-R chart as the control limits for the next X-R chart.\n\n5.3 Process Control Interpretation:\nWhen the following situations occur, it indicates an abnormality in the control chart. The quality department should record the process factors (5M1E) and, if necessary, conduct a root cause analysis in conjunction with the production department (refer to the SPC textbook) and take corrective measures.\n5.3.1 Points beyond the control limits.\n5.3.2 Seven consecutive points increasing or decreasing.\n5.3.3 Seven consecutive points above or below the centerline.\n5.3.4 Clear non-random patterns.\n\n5.4 Process Capability Interpretation:\n5.4.1 Explanation of process capability under assumptions:\n- The process is in a state of statistical stability.\n- The measurements of the process follow a normal distribution.\n- Engineering and other specifications accurately represent customer requirements.\n- The target value is located at the center of the specification.\n5.4.2 In general, compare the distribution of process outputs with engineering specifications to determine if they consistently meet these specifications.\n5.4.3 If the distribution is not known to be normal, a normality test should be conducted, such as examining a histogram or plotting points on normal probability paper.\n5.4.4 The process standard deviation is estimated using the average range (R) to evaluate the process capability, as long as both the range and mean of the process are in statistical control.\n5.4.5 Calculate process capability:\n5.4.5.1 Short-term process capability index:\n5.4.5.2 Long-term process capability index:\nwhere:\nUSL - Upper Specification Limit\nLSL - Lower Specification Limit\n- Mean of the measured process\n- Average range of the process\nd2 - Constant, see table above\n\n5.4.6 Improving Process Capability:\nTo improve process capability, it is necessary to focus on reducing common causes and directly target the fundamental factors that cause process variation, such as machine performance, consistency of input materials, basic methods of process operation, training methods, or work environment. Generally, correcting systemic causes of unacceptable process capability may go beyond the capacity of operators or their on-site managers. Instead, management intervention is needed to make fundamental changes, allocate resources, and coordinate the overall performance improvement of the process. Correcting the system with short-term local measures alone will not be successful.\n\n5.4.7 Plotting Control Charts and Analyzing Modified Processes:\nAfter implementing systemic measures, the effects should be reflected on the control chart, which becomes a way to verify the effectiveness of the measures. Control charts should be carefully monitored when making changes to the process.\n\n5.4.8 Process Capability Acceptance Criteria:\n5.4.8.1 When customers have requirements for process capability, follow their requirements.\n5.4.8.2 When there are no customer requirements, ensure that the process is capable with PPK > 1.67 and CPK > 1.33. If these criteria cannot be met, full inspection is required.\n\n6. Relevant Documents:\n6.1 External Document - Statistical Process Control (SPC) Textbook\n\n7. Generated Records:\n7.1 "-R" Control Chart\n7.2 PPK/CPK Calculation Data