Resources/Glossary/
Source Inspection
Quality at Source

Source Inspection

Check the conditions that cause defects, not the defects themselves.

Updated
·
4
min read
Definition

What is Source Inspection?

Source inspection is the lean practice of checking the conditions that cause defects before the defects can occur, rather than checking the output after the fact. Coined by Shigeo Shingo, source inspection moves the check upstream of the failure point: verifying setup parameters, tool wear, fixture position, or material condition. Done well, source inspection makes most downstream inspection unnecessary because the defects never get a chance to be made.

Source inspection is the inspection mode Shigeo Shingo argued was the only one that could actually produce zero defects. Most inspection looks at output and decides what to keep. Source inspection looks at the input conditions that determine output and decides whether to start running at all. The shift sounds small but it changes the economics. A shop that masters source inspection spends almost no time on downstream sorting because the defects never get made in the first place.

"Inspect the conditions that make defects, not the defects themselves."

How source inspection works

Source inspection rests on a specific cause-and-effect logic. For any operation that produces defects, there is a small number of input conditions that determine whether the output will be good or bad. Tool condition. Fixture position. Material variation. Machine parameters. Setup geometry. Identify the conditions, verify them before running, and the output follows.

The process is straightforward.

  1. Map the defect causes. For each common defect type on the operation, work backward to the input condition that produces it. A dimensional defect usually traces to tool offset, fixture position, or material variation. A surface defect usually traces to tool condition, speed, or feed. A cosmetic defect on a molded part usually traces to temperature, pressure, or cycle time.

  2. Define the verification check. For each cause, define exactly how to verify it before the run. A tool wear check uses a counter and a sample part. A fixture check uses a gauge. A material check uses an incoming hardness or thickness verification.

  3. Build the check into the setup. The verifications go into the setup procedure or shift-start checklist. They happen before the first part is made, not after.

  4. Stop on a failed verification. Any failed check stops the work until the condition is corrected. The operator does not start running and hope for the best.

The discipline that makes source inspection work is the cause map. A shop that has not done the work to identify which conditions produce which defects ends up with a generic checklist that misses the real causes. The first source inspection effort in a new shop usually requires a FMEA or a focused defect study to identify the real causes.

Where source inspection fits on the shop floor

Picture a 20-person plastics injection shop running parts for a small appliance customer. The shop has historically run a heavy downstream inspection program: every fifth part off each press gets a visual check at a bench station, and every twentieth part gets a dimensional measurement. Despite this, sink-mark and short-shot defects keep appearing in customer shipments.

A source inspection rebuild starts by mapping the causes. Sink marks trace to melt temperature, pack pressure, and cooling time. Short shots trace to material moisture, melt temperature, and injection speed. Each of those conditions can be verified at machine startup with the existing instrumentation. The setup procedure gets a six-line check sheet: melt temp in band, material dried within the last 4 hours, mold temp at setpoint, pack pressure in band, cycle time correct, scrap chute clear.

The downstream visual inspection stays for a few months as a backstop. Within two months, the source-inspection checks are catching the issues before they produce defects, and the downstream visual inspection finds almost nothing. The shop eventually cuts the visual inspection to once per shift as a sanity check rather than a 20 percent sample. Defect rate at the customer drops to under 0.3 percent.

Common mistakes with source inspection

  • Generic checklists. A setup checklist that does not target the actual causes of defects produces theater without prevention. The work of mapping causes has to happen first.
  • Skipping the checklist when the schedule tightens. Source inspection only works if the checks happen every time. The first skip teaches the team they are optional.
  • No response when a check fails. A failed verification has to halt the work and trigger a fix. Otherwise the checklist becomes documentation, not prevention.
  • One-time mapping. Causes drift. A source-inspection program built on a 2022 defect study needs to be refreshed every year or two as tools, materials, and operations change.

Source inspection and related Lean tools

Source inspection is one of three core modes of quality at the source, alongside self-inspection by the operator and successive-inspection by the next station. Of the three, source is the earliest and most preventive. The most natural companion technique is poka-yoke, error-proofing fixtures and devices that physically prevent the wrong conditions from being set. Together, source inspection and poka-yoke are the two strongest tools for moving a shop from defect detection toward genuine defect prevention.

Common questions

The questions we hear most about this term.

How does source inspection actually work?
It works by identifying the small number of process conditions that produce most defects and verifying those conditions before each run, each shift, or each part. A worn tool, an out-of-position fixture, an incoming material lot with the wrong hardness, a temperature outside the band. Each is a leading indicator that a defect is about to happen. The source inspection checklist verifies those indicators are in range. When they are, the run proceeds with confidence. When they are not, the issue gets fixed before any part is produced.
How is source inspection different from successive inspection?
Successive inspection happens after the part is made: the next operator checks the previous step's work. Source inspection happens before the part is made: the conditions that would produce a bad part are verified upstream. Successive inspection catches defects that already exist. Source inspection prevents defects from being made in the first place. Both are quality-at-the-source modes; source inspection is the earlier, more powerful one when the cause-to-effect relationship is well understood.
How does source inspection compare to self-inspection?
Self-inspection has the operator check their own output against a standard, usually right after each part or each batch. Source inspection has the operator (or a setup person) verify the conditions that will determine output quality, usually before the work starts. Self-inspection catches errors after the fact at the bench. Source inspection prevents the errors from happening. The two work together: source inspection at setup, self-inspection during the run.
Why does Shigeo Shingo argue source inspection is the most powerful inspection mode?
Because it is the only mode that produces zero defects, not just a low defect rate. Every other inspection mode is a downstream filter, deciding which defects to catch and which to let through. Source inspection acts before the defect can exist. Shingo's argument was that if you can identify the cause of a defect and check that cause directly, you can prevent the defect with near certainty. The downstream inspection apparatus, expensive and imperfect, becomes optional rather than essential.
What does source inspection look like on the shop floor?
In a 25-person stamping shop, it looks like a setup checklist with five items the operator runs through before each run: die alignment within tolerance, scrap chute clear, press tonnage in band, material gauge verified on the incoming coil, lubrication system operational. The checklist takes two minutes. Compared to the alternative, finding a bad part later in the run and tracing back to whichever of those conditions was off, the two minutes is the cheapest insurance in the shop.

Ditch the whiteboards and spreadsheets.

Same-day setup. No distributor lock-in. Zero stockouts. Top teams double revenue in 9 months.