What quality control is for
Quality control is the set of activities that check whether products meet their requirements and keep them doing so. Those requirements might be dimensions, strength, appearance, purity or how well parts fit together. The point is not perfection for its own sake. It is to make sure what leaves the plant is what the customer expects, at a cost that makes sense.
A useful distinction is between detecting defects and preventing them. Inspection at the end of a line detects problems after the cost of making them has already been spent. Control of the process itself aims to stop them being made at all.
Inspection
Inspection means examining a product or material against a specification. It can happen at several points.
- Incoming inspection checks raw material and purchased parts before they enter production.
- In-process inspection checks work between steps, often by the operator.
- Final inspection checks finished products before shipment.
Inspection can be done by eye, with hand gauges and measuring tools, with fixed gauges that simply accept or reject a part, or with machines. Automated systems using cameras and inspection sensors can check large numbers of parts consistently, though they need to be set up and verified carefully. Measuring tools themselves need regular calibration, because an instrument that has drifted gives confident but wrong answers.
Inspection has limits. People get tired, and checking every item by hand is rarely perfect. For that reason, quality programs usually avoid relying on inspection alone.
Sampling
Checking every unit is sometimes impractical, especially when testing damages the product or takes a long time. In those cases plants check a sample and draw conclusions about the whole lot.
The basic idea is that a properly chosen sample can say something about the batch it came from, though never with complete certainty. There is always a chance of accepting a poor lot or rejecting a good one. Sampling plans, often defined in published standards, balance that risk against the cost of inspecting more units. Choosing the sample size and acceptance rules is a statistical question, and the details depend on the product and the consequences of a failure. For products where a defect could injure someone, much stricter checking is normal.
Controlling the process
Process control shifts attention from the product to the way it is made. Every process varies a little. Some variation is natural and small, coming from many minor causes. Other variation signals that something has changed, such as a worn tool, a different batch of material or a drifting setting.
A common tool is the control chart, which plots a measurement over time against limits calculated from the process’s own normal behavior. A point outside the limits, or a pattern such as a run of readings trending in one direction, suggests a cause worth investigating before the process produces out-of-tolerance parts.
Monitoring key process variables such as temperature, pressure, speed or force is the same idea in a different form. If the settings that determine quality are held steady, the output tends to be steadier too. Good product consistency often comes more from stable settings than from sorting good parts from bad.
Preventing defects at the source
Several practices focus on preventing defects rather than finding them.
- Standard work. When everyone follows the same documented method, there is less variation between people and shifts.
- Mistake-proofing. Often called by the Japanese term poka-yoke, this means designing a fixture, sensor or procedure so an error is hard or impossible to make, for example a part that only fits one way.
- Stopping to fix problems. Letting an operator halt the line when something looks wrong prevents a run of bad parts, though it requires a workplace where raising problems is welcomed.
- Finding causes. When a defect occurs, root cause analysis means asking why repeatedly, until the underlying cause is found, which helps prevent repeats instead of just replacing the bad parts.
- Thinking ahead. Using failure mode analysis to work through how a product or process could fail, before it does, helps decide where controls are most needed.
Equipment and quality
Machine condition is a quality issue. Worn bearings, loose fixtures and degraded tooling produce drift in dimensions long before a machine fails outright. Preventive upkeep and checks of critical settings protect quality as well as uptime. Measuring tools like coordinate measuring machines or 3D scanners can capture detailed dimensional data, which is useful for checking complex parts, but they work best as part of a wider system.
How quality connects to lean
In lean thinking, defects are one of the standard wastes. A defective part wastes the material, machine time and labor already spent on it, and rework or scrap adds more. Defects also hide other problems. If a plant keeps big stocks to cover for bad parts, the stock hides the cause.
Lean and quality therefore reinforce each other:
- Smaller batches mean a problem is found after a few parts, not after hundreds.
- Pull systems carry little stock, so bad parts cause an immediate stoppage that forces attention on the cause.
- Standard work reduces variation, which is the root of many defects.
- Quality at the source, meaning each step checks its own work before passing it on, avoids passing defects downstream.
Practical balance
A sensible quality system rarely uses one method alone. Most plants combine incoming checks, process monitoring, operator self-inspection and sample-based final checks, adjusting the mix to the risk of the product. Records matter as well, since data is what allows trends to be seen and causes to be traced. Gathering the data is only useful if someone reviews it and acts.
This article offers general background. Specific products, especially in regulated fields, are subject to standards and requirements that a plant’s quality professionals must apply in detail.