‹ Modern TPM class

Lesson 4 of 20 · Chapter 4

Quality Maintenance and Defect Prevention

Most plants try to make good product by catching bad product: inspect, sort, rework, scrap. This chapter turns that around. A defect is a sign that some condition on the machine was not controlled. Control the condition, and the defect is never made. This lesson shows you how to find those conditions, watch them in real time, stop early without blame, and learn from every defect that still gets through.

Choose how you want to learn

You can change this any time. Every level does the same hands-on steps. Press Listen on any part to hear it read aloud.

Part 1

Quality is maintained into the machine

When a plant relies on inspection, it pays twice: once to make the bad part, and again to find it, sort it, fix it or scrap it. If the customer finds it, the plant pays a third time, in lost trust.

Quality at the source means the machine makes good parts because it is in the right condition to make them, not because someone catches the bad ones afterwards.

No inspector can sort a worn guide back into line. Only looking after the machine can do that.

Defects are signals that equipment conditions, process settings and people’s responses are not controlled tightly enough to produce right first time. The early signs are usually there: a contaminated sensor causes intermittent misreads before a hard stop; a drifting temperature loop makes marginal seals before it trips a fault.

Treating those signs as a quality-department problem instead of an equipment-condition problem misses the cheapest window for prevention. Inspection stays, but as a safety net, not the strategy.

“When organizations lean on inspection, they pay twice: once to make the defect and again to detect, sort, rework, scrap, or explain it.”

“The quality department cannot inspect a worn guide back into alignment or sort a drifting temperature loop into stability.”

“But the foundation is here: quality is not inspected into product; it is maintained into the equipment.”

Read the whole chapter free in the book reader.

Move the catch point. The earlier the problem is caught, the fewer times the plant pays.

Part 2

Name the conditions that make defects impossible

Ask one question about each product: what must be true at the machine for a defect to be unlikely? Those are its quality-critical conditions.

For a seal, it is things like jaw temperature, pressure, time and a clean sealing surface. For a label, it is a clean sensor, steady web tension and rollers that do not slip.

Then beware the operator trap. A skilled operator who tweaks the machine all day to keep it in spec hides the real problem. When that person is on holiday, the defects come back.

These conditions are different from general “keep it running” conditions. Quality maintenance identifies the quality-critical characteristics of the process and links each one to equipment conditions that can be checked, controlled and restored.

Operator technique makes the process depend on individuals and hides the true state of the machine. Often what looks like craftsmanship is really missing standardisation or missing basic conditions. The chapter asks plants to stop accepting technique as a permanent quality control method.

“In practical terms, ensuring equipment quality at the source begins with a question that many plants skip: what conditions must be true, at the machine, for defects to be physically impossible or at least highly unlikely?”

“Quality at the source also requires an uncomfortable but liberating discipline: stop accepting “operator technique” as a permanent quality control method.”

“It makes the process dependent on individuals, it hides the true state of the machine, and it guarantees that defects will spike during vacations, shift changes, or turnover.”

Four defects. Which family of machine conditions is most likely behind each one?

Part 3

Control points, and the right to stop

A control point is one simple check on one quality-critical condition. A good one passes three tests: it is predictive (it catches drift before bad parts pile up), actionable (everyone knows what to do if it is out), and practical (it can be done at the right frequency without stopping the line).

A lab test that answers tomorrow fails the first test on a fast line. An end-of-shift inspection fails it too: the bad parts have already travelled.

And operators need a protected right to stop. If people get blamed for stopping to protect quality, they will keep running, and defects will hide.

Each control point needs decision rules, not just a measurement: if the temperature is low, does the operator adjust, wait or stop? Without rules, one supervisor pushes to keep running, another stops, and the process becomes unpredictable.

Stopping for quality often gets framed as “we cannot afford downtime.” The real comparison is controlled downtime versus uncontrolled loss. The bench below lets you design one control point and see which choice breaks it.

“For example, a check that requires a lab test that returns results tomorrow is not a quality-at-the-source control point for a fast-moving line.”

“A decision to stop for quality is often framed as “we cannot afford downtime.” But the real comparison is not uptime versus downtime; it is controlled downtime versus uncontrolled loss.”

“If operators are punished for stopping early to protect quality, the plant has built a system that teaches defects to hide.”

The Control Point Bench. Pick a condition, then design its check, response, cadence and stop rule.

Part 4

Root cause: ask what condition made the mistake possible

When a bad part is found, the easy question is “who made the mistake?” TPM asks a better one: “what condition made the mistake possible?”

Work through three questions. What exactly is the defect (not “bad quality” but “label shifted 3 millimetres to the right”)? What condition directly created it? And why did our checks let it travel?

Gather facts at the machine: measurements, trends, the parts themselves. Evidence is what separates root cause from storytelling.

Treat any adjustment as a red flag: if a machine needs nursing to make good parts, the true root cause is the condition that makes it need nursing. Treat any restart as a suspect event until proven otherwise.

The countermeasure follows restore, stabilise, improve: restore the condition now, then change a condition or a control, with an owner and a due date, so the same drift cannot come back. And the investigation must be psychologically safe; if it is used to assign fault, people hide information and quietly adjust to protect themselves.

“The plant stops asking, “Who made the mistake?” and starts asking, “What condition made the mistake possible?” That is a profound cultural shift.”

“Evidence is what separates root cause analysis from storytelling.”

“The true root cause is the condition that makes the process require nursing.”

A real defect report. Answer the three questions in order.

Part 5

Running a quality maintenance system every day

A system does two jobs. Prevention: keep the quality-critical conditions inside their normal window. Containment: when one drifts out, catch it fast, respond the same way every time, and stop the spread.

Match how often you check to how fast things drift. A sensor that gets dirty within hours must be checked within hours. Tool wear that builds up over weeks can be checked less often, as long as it has a clear trigger.

And make first-off a proper gate: confirm the critical conditions before running at full speed after a restart.

Write each condition set down in operating language, assign an owner and a cadence (a “chain of custody,” so nobody assumes someone else is watching), and make it visible at the machine: reference marks, labelled adjustment points, posted ranges, photos of normal and abnormal.

A practical start is a short list of the top defect families on each line, each with its likely condition causes, leading indicators, control points and first response. That shifts behaviour from “tweak until it looks right” to “verify conditions, restore baseline, then run.”

“A condition that can change within hours, like sensor contamination or web tension drift, must be checked within hours, not once per week.”

“If the plant tries to control everything equally, it will control nothing well.”

“A quality maintenance system treats first-off as a stabilization step: confirm critical-to-quality conditions before full-rate production resumes.”

Five conditions. Choose how often each should be checked.

Part 6

Try it where you work

Do these on a real machine: at work if you work with equipment, or at home on a car, washing machine, furnace or lawn mower if you do not. Tick each one when it is done. Safety first: look and clean only what you are allowed to touch, never reach into moving equipment, and follow your site’s lockout rules.

Part 7

Your “My Control Point” card

Design one control point for one defect you care about. Nothing you type here leaves this device.

My Control Point

Global Sovereign University · Modern TPM, lesson 4

Defect I am preventing
Quality-critical condition
How I check it
How often
Normal band
First response / stop when

Three tests for a control point

  1. Predictive: it catches drift before defects pile up.
  2. Actionable: a written first response restores the condition.
  3. Practical: it can be done at the needed frequency.

Ask what condition made the mistake possible, not who made it. Stopping early is controlled downtime, not failure.

Part 8

Check yourself

Six real situations. Pick what you would do. You will see why after each one.

Lesson 4 complete.

You can now find a product’s quality-critical conditions, design a control point that passes all three tests, work a defect back to the condition behind it, and match checks to drift speed. Next: Lesson 5, early equipment management and design for maintainability.

Modern TPM, cover

Modern Total Productive Maintenance (TPM) by Dr. Gene A Constant. This lesson comes from chapter four. Read the whole book free in the reader, or keep a copy.

Buy a copy — it funds a free seat

GSU is a 501(c)(3) nonprofit and book sales are its only income. A purchase is not a tax-deductible donation. This lesson is free and always will be.

Accreditation notice: Global Sovereign University is not an accredited institution and does not grant degrees, academic credit, licenses, or industry credentials. Learners who pass GSU's online exams earn GSU certificates of comprehension.