Statistical Process Control

SPC – The Language of Process Stability We often hear: “The machine is running fine.” But how do we prove it? That’s where SPC (Statistical Process Control) steps in. SPC uses data + statistics to check whether a process is: Stable (predictable, under control) Capable (meeting customer requirements) Step 1: Define the Specification Suppose we are making a motorcycle frame tube joint. Required specification = Diameter = 50.00 mm ± 0.20 mm That means: LSL (Lower Spec Limit) = 49.80 mm USL (Upper Spec Limit) = 50.20 mm --- Step 2: Collect Data from Production We take 5 samples every shift. Example readings: 50.05, 50.02, 49.98, 50.07, 50.01 --- Step 3: Calculate the Average (X̄) and Range (R) Average (X̄) = (50.05 + 50.02 + 49.98 + 50.07 + 50.01) ÷ 5 = 250.13 ÷ 5 = 50.026 mm Range (R) = Highest – Lowest = 50.07 – 49.98 = 0.09 mm This tells us the process is not fluctuating wildly. --- Step 4: Estimate Variation (σ) For simplicity, assume σ = R ÷ d2 (where d2 is a statistical factor). For sample size 5, d2 = 2.326. σ = 0.09 ÷ 2.326 ≈ 0.039 mm --- Step 5: Check Process Capability (Cp & Cpk) 1. Cp (Potential Capability): Formula = (USL – LSL) ÷ (6σ) = (50.20 – 49.80) ÷ (6 × 0.039) = 0.40 ÷ 0.234 = 1.71 Means the process has the potential to meet specs. --- 2. Cpk (Actual Capability): Formula = min[(X̄ – LSL) ÷ (3σ), (USL – X̄) ÷ (3σ)] = min[(50.026 – 49.80) ÷ (0.117), (50.20 – 50.026) ÷ (0.117)] = min[(0.226 ÷ 0.117), (0.174 ÷ 0.117)] = min[1.93, 1.49] = 1.49 Means the process is well within limits, slightly shifted but safe. --- Step 6: Interpret with a Table Cp / Cpk Value Meaning < 1.00 Not capable – high risk of defects 1.00 – 1.33 Marginal – needs improvement 1.33 – 1.67 Capable – industry acceptable > 1.67 World class, highly capable --- Final Takeaway Cp = What the process could achieve Cpk = What the process is actually delivering For our motorcycle frame: Cp = 1.71 → Machine/process is excellent. Cpk = 1.49 → Process is stable, safe, and customer won’t see defects. SPC is not just math – it’s an early warning system to avoid costly rework or recalls. In short: SPC = Early warning system before customers complain. #Quality #SPC #Manufacturing #LeanSixSigma #VIPtalks

ISO 9001, PPAP, APQP, SPC. Most suppliers list these on their capabilities page and never actually use them. They're not compliance checkboxes. They're process control tools that prevent failures if you know how to apply them. What these actually do: ↬ PPAP proves repeatability. Not that you made one good part, but that your process can consistently hit spec across production runs. Most suppliers submit PPAP once and never validate their process again. ↬ SPC catches problems before you scrap parts. Control charts show when variation trends toward limits. You adjust the process, not sort rejects after the fact. ↬ MSA tells you if your measurements mean anything. If your GR&R study shows 30% gage variation on a tight tolerance, you're not measuring the part. You're measuring your equipment's inconsistency. ↬ FMEA maps where processes fail. It prioritizes which risks to control based on severity and likelihood. It's not paperwork. It's risk mitigation before you start production. At Atlas Fibre, I've personally run every process we control. Tool wear drifts dimensions. Temperature cycles affect cure profiles. Fiber direction determines load capacity. If you're not tracking these variables with actual data, you're guessing whether the next batch will pass. Quality systems work when you use them to control processes, not pass audits. Execution is the only valid opinion. → See how we apply quality systems: https://lnkd.in/eB6_6DWY #QualityManagement #Manufacturing #ISO9001 #SPC #ProcessControl

How good is your process? Making a part once and making a part many times are very different tasks. We often create prototypes that lead to larger-volume production. One of the biggest challenges is being confident in your production process so you can guarantee that every part falls within your tolerance band. This is where statistical process control (SPC) comes in. At its simplest, measuring data, analyzing trends, and comparing them against upper and lower limits helps you determine whether you can trust your process. The reliability of your process can be summed up in a single number—Cp, or process capability. At NH Micro, we manufacture a lot of screws, most of them for our in-house wristwatch production. Recently, we started analyzing the dimensional data we collect to better understand our screw manufacturing process. You can see the results in the graphs below! As a quick explainer: Across 750+ parts, one dimension—a 1.90mm diameter—varied within a 6µm band. Against a ±10µm tolerance, we confirmed that our process has a Cp of 1.87 and a Cpk of 1.54. That’s really good! What’s really interesting is the trend over time. - Between 0 and 200 parts, we saw our machine warming up. - Between 200 and 675 parts, it had stabilized, but our cutting tools were slowly wearing. - At the 700th part… Well, I’ll let you guess in the comments what happened! SPC is a superpower—an incredibly useful tool for controlling your manufacturing process and pushing it to its full potential. Josh

1. What is SPC? SPC is a quality control method that uses statistical tools to: Monitor a process Control process variation Improve process stability Prevent defects before they happen 📌 Simple definition: SPC helps us check whether a process is stable or out of control using data. 2. Why SPC is Required? In manufacturing: No process is 100% perfect Variations always exist SPC helps to: ✔ Reduce rejection ✔ Improve consistency ✔ Detect problems early ✔ Reduce inspection cost ✔ Improve customer satisfaction 3. Types of Variations (VERY IMPORTANT) 1. Common Cause Variation Natural variation Always present Machine wear, temperature change ➡ Process is stable 2. Special Cause Variation Due to abnormal reasons Tool breakage Wrong setting Operator mistake ➡ Process is unstable 📌 SPC’s main job is to identify special causes. 4. Key Elements of SPC Data collection Control charts Control limits Process capability Corrective action 5. Control Charts (Heart of SPC) What is a Control Chart? A graph showing: Process data over time With control limits Parts of Control Chart CL (Center Line) → Average UCL → Upper Control Limit LCL → Lower Control Limit 📌 Control limits ≠ specification limits 6. Types of Control Charts A. Variable Control Charts (Used when data is measurable) Chart Used For X̄ – R Small sample size (n ≤ 10) X̄ – S Large sample size I – MR Single observation X – X Individual values B. Attribute Control Charts (Used when data is countable) Chart Used For p-chart % defective np-chart No. of defectives c-chart No. of defects u-chart Defects per unit 7. Selection of Control Chart Situation Chart Dimension measurement X̄-R One part at a time I-MR Visual defect p-chart Casting porosity count c-chart 8. Control Limits Formula (Basic Idea) For X̄-R Chart UCL = X̄ + A₂ × R̄ CL = X̄ LCL = X̄ − A₂ × R̄ (A₂ is constant) 9. Process Stability A process is in control if: ✔ All points within UCL & LCL ✔ No abnormal pattern Out of control if: ❌ Point outside limit ❌ Trend / shift / zig-zag pattern 10. SPC Rules (Western Electric Rules) Some common rules: One point beyond control limit 7 points on one side of CL Continuous upward or downward trend Sudden jump or dip ➡ Indicates special cause 11. Process Capability (Cp & Cpk) Cp (Process Potential) ✔ Shows spread only Cpk (Process Performance) ✔ Shows centering + spread Acceptable Values Value Meaning < 1.0 Not capable 1.33 Minimum acceptable ≥ 1.67 Good ≥ 2.0 Excellent 12. SPC Implementation Steps Select critical characteristic Decide measurement method Collect data Choose control chart Calculate limits Plot data Analyze Take corrective action 13. SPC in Die-Casting & Machining (Your Work) Examples: Bore diameter Flatness Surface roughness Porosity count Weight variation Special causes: Die temperature variation Tool wear Improper lubrication Machine vibration 14. Difference Between SPC & Inspection SPC Inspection Preventive Detective Online control After production Data-based Judgment-based

Why Every Quality Manager Must Track Cp & Cpk Ensuring a manufacturing process consistently meets quality standards is critical in any industry. This is where Process Capability (Cp) and Process Capability Index (Cpk) come into play. These statistical tools help quality managers assess how well a process can produce products within specified limits. What Are Cp & Cpk? Cp (Process Capability Ratio): - Measures the potential capability of a process. - Compares the width of the process variation to the width of the specification limits. - Does NOT consider how centered the process is. - Formula: Cp = (USL - LSL) / (6σ) - If Cp > 1, the process has the potential to meet specifications. Cpk (Process Capability Index): - Measures the actual capability, considering both variation and centering. - Determines how close the process is to the specification limits. - Formula: Cpk = min [(USL - μ) / (3σ), (μ - LSL) / (3σ)] - If Cpk > 1.33, the process is considered capable. If Cpk > 2.0, it meets Six Sigma standards. Why Should a Quality Manager Track Cp & Cpk? - To Ensure the Process Is Capable: If Cp or Cpk is low, the process has excessive variation or is off-target. A capable process reduces defects, rework, and customer complaints. - To Monitor Process Stability: A stable process should have consistent Cp & Cpk values over time. Drastic changes in Cpk indicate process drift or special cause variations. - To Make Data-Driven Decisions: Helps in root cause analysis, predictive maintenance, and continuous improvement. Ensures that critical parameters stay within specifications. - To Reduce Waste and Improve Efficiency: Higher Cpk means fewer defective products, leading to cost savings and higher customer satisfaction. How to Track and Use Cp & Cpk? - Use Statistical Tools: Software like Minitab, Excel, or SPC tools can track and visualize Cp & Cpk trends. - Regular Data Collection: Record process data and calculate Cp & Cpk for critical parameters at defined intervals. - Set Process Capability Targets: Cp ≥ 1.33: Minimum acceptable level. Cpk ≥ 1.33: Process is capable and centered. Cpk ≥ 2.0: Six Sigma-level quality. - Identify & Eliminate Causes of Variability: If Cp is high but Cpk is low, the process is capable but not centered—adjustment is needed. If both Cp and Cpk are low, the process has excessive variation—reduce variability through better control methods. Key Takeaway for Quality Managers - A process is only truly capable if Cpk is high and stable over time. - Tracking Cp & Cpk helps predict potential failures before they impact production. - Process capability studies should be part of a continuous improvement strategy.

🚀 Master Process Capability for Quality Excellence! 🌟 💡 What is Process Capability? It’s a powerful statistical tool used to assess how well your process meets specification limits, ensuring consistent quality output. ✅ 🎯 Purpose: • Align with customer requirements. 👥 • Control and reduce process variation. 📉 • Set realistic tolerances for product performance. 🔧 • Continuously improve operational excellence. 🌟 🔑 Key Metrics: 1️⃣ Cp (Process Capability): Measures the spread of a process relative to specification limits. • Formula: Cp = \frac{\text{USL} - \text{LSL}}{6 \times \sigma} (USL: Upper Spec Limit, LSL: Lower Spec Limit, σ: Process Std. Deviation) • Interpretation: • Cp > 1: 🟢 Excellent process (fits well within spec limits). • Cp = 1: ⚠️ Marginal process (barely fits). • Cp < 1: 🔴 Poor process (doesn’t fit). 2️⃣ Cpk (Process Capability Index): Considers both process spread and how centered the process is within spec limits. • Formula: Cpk = \min \left( \frac{\text{USL} - \bar{X}}{3 \times \sigma}, \frac{\bar{X} - \text{LSL}}{3 \times \sigma} \right) (\bar{X}: Process Mean) • Interpretation: • Cpk > 1: Process is capable and centered. 🎯 • Cpk < 1: Process is off-center or exceeds spec limits. ❌ 🛠️ Tools for Process Capability Analysis: • Histogram: Visualize data distribution. 📊 • Control Chart: Identify special vs. common cause variation. 📉 • Run Chart: Monitor trends over time. 📈 • ANOVA: Compare variations between groups. 🔬 🚗 Car & Garage Analogy for Cp: Think of spec limits as the garage and the process as the car: • Cp > 1: 🟢 The car fits well (excellent process). • Cp = 1: ⚠️ The car barely fits (marginal process). • Cp < 1: 🔴 The car doesn’t fit (redesign needed). 📢 Takeaway: A capable process ensures consistent quality, satisfied customers, and operational efficiency. Aim for Cp & Cpk ≥ 1.33 for optimal performance! 📬 Let’s discuss: How does your team measure process capability? Share your experiences below! 👇

Process Capability - A Practical Guide What is Process Capability Process capability is a statistical tool used to evaluate the ability of a process to produce output within specified limits, reflecting the "potential capability" of the process. It helps organizations determine if their processes can consistently meet customer requirements and specifications. Purpose of Process Capability 🎯Meet customer requirements and specifications. 🎯 Measure and control the spread (variation) of the process. 🎯 Provide realistic tolerances for product dimensions. 🎯 Improve overall process performance capability. Influencing Factors ➡️ Condition of machines and equipment. ➡️ Type and quality of raw materials. ➡️ Operator and inspector skill levels. ➡️ Measurement methods and instrument conditions. ➡️ Operational and environmental conditions. Process Capability Overview 🔈 Voice of the Process: The actual output distribution of the process 🔈 Voice of the Customer: The specification limits set by the customer (USL and LSL) 🔈 Process Width vs. Design Width: The process should fit well within the design (specification) limits to be considered capable Tools Used in Process Capability Analysis 1️⃣ Histogram: Shows the distribution of numerical data 2️⃣ Control Chart: Differentiates between common cause and special cause variation 3️⃣ Run Chart: Graphically represents how a process is performed or how data values change over time 4️⃣ ANOVA: Evaluates the difference between the means of more than two groups Key Indices ↗️ Cp – Process Capability Definition: Measures the ability of a process to produce output within specification limits, considering only the spread (variation) of the process Interpretation: Cp does not account for how well the process is centered. A Cp ≥ 1.33 is generally considered acceptable ↗️ Cpk – Process Capability Index Definition: Measures how close a process is running to its specification limits, accounting for process mean (centering) Interpretation: Cpk considers both spread and centering. A Cpk ≥ 1.33 is desirable Process capability analysis is essential for quality management, helping organizations ensure their processes are capable of consistently meeting customer requirements. By understanding and applying Cp and Cpk, businesses can identify improvement opportunities, reduce defects, and enhance customer satisfaction Image : Learn Fast

Your process can look “in control” and still be unstable. Most teams react only when a point crosses the control limits. But real process problems rarely announce themselves that loudly. That’s why the 8 Nelson Rules matter. They help detect non-random patterns like: • hidden trends • shifts masked by averages • oscillations caused by over-adjustment • false stability created by noise A chart can be green and still be lying. SPC isn’t about reacting to defects it’s about listening to the behavior of the process before defects appear. If you’re only watching Rule 1 (±3σ),you’re missing early warning signals. Which Nelson Rule do you see violated most often in your process and why? Share your experience in the comments. Follow Naveen K for more insights on Quality&CI #SPC #StatisticalProcessControl #NelsonRules #ControlCharts #ProcessStability

I’m a huge believer in using SPC to scale AI in manufacturing. Here’s why. As a data scientist or engineer, your dashboard feels obvious. The signal looks clear. The insight makes sense. And yet… nothing happens. Why? Because insights that require interpretation don’t scale. Let me say that again. 𝐈𝐧𝐬𝐢𝐠𝐡𝐭𝐬 𝐭𝐡𝐚𝐭 𝐫𝐞𝐪𝐮𝐢𝐫𝐞 𝐢𝐧𝐭𝐞𝐫𝐩𝐫𝐞𝐭𝐚𝐭𝐢𝐨𝐧 𝐝𝐨𝐧’𝐭 𝐬𝐜𝐚𝐥𝐞. That’s why I keep coming back to SPC (Statistical Process Control). A well-designed SPC chart is brutally simple: • 𝐁𝐥𝐮𝐞 = 𝐧𝐨𝐫𝐦𝐚𝐥 • 𝐑𝐞𝐝 = 𝐚𝐜𝐭 That’s it. It’s either: → Blue or red → Black or white → Do something, or don’t No debate. No explanation required. SPC scales because it works across every role and education level: → Engineers and data scientists → Executives in boardrooms → Operators on the shop floor → Multiple languages, cultures, and countries Everyone sees the same signal. Now the obvious question: SPC is great for univariate data, but what about multivariate complexity? That’s where MSPC comes in. Multivariate Statistical Process Control looks simple on the surface, but beneath each point may be hundreds of signals. All summarized into one clear action signal. Examples: → Hotelling’s T² → DModX Because simplicity comes after complexity. Not before it. And in manufacturing, simplicity is what turns data into ROI.