Quality Maintenance — Hinshitsu Hozen in the original JIPM framework — is the TPM pillar that treats every defect as the fingerprint of an equipment condition that quietly drifted out of specification. In a facility setting, chilled water supply temperature, compressed air dew point, room pressure cascade and incoming power quality are not background services; they are process inputs that stamp themselves onto finished product. Most engineering groups still manage those utilities as uptime assets rather than as quality assets, which is why scrap and rework stay stubborn even when availability charts look healthy. A zero-defect facility closes that gap by defining, measuring and holding the exact utility conditions under which the defect physically cannot occur. To see that program running inside a working CMMS, start with OxMaint free or book a live demo.
TPM quality maintenance software for a zero-defect facility
Availability and speed are only two thirds of OEE. The third term — quality rate — is where facility utilities do their damage, and it is the term most CMMS deployments never instrument.
Availability
Breakdowns, changeovers, utility outages
×
Performance
Minor stops, reduced speed, throttling
×
Quality rate
Scrap, rework, start-up loss, out-of-spec batches
=
OEE
The number your plant reports
Quality maintenance moves inspection upstream — from catching bad output to controlling the equipment conditions that produce it.
The utility-to-defect chain most facilities never map
Quality maintenance begins with a physical claim: a defect mode always traces back to a measurable condition on a specific asset. The chain below shows how facility parameters become product defects in real plants.
Chilled water supply drifts 2°C above setpoint
leads to
AHU coil cannot pull latent load; room RH climbs above 55%
produces
Hygroscopic powder caking, tablet sticking, coating defects
Compressed air dryer regeneration cycle fails
leads to
Pressure dew point rises; moisture and oil carryover downstream
produces
Paint fisheyes, pneumatic valve hunting, contaminated fills
Room pressure cascade collapses during filter loading
leads to
Airflow reverses across a door; particle count breaches class limit
produces
Cross-contamination, batch rejection, deviation investigation
Voltage sag from upstream switching event
leads to
Drives ride through unevenly; line speed desynchronises for seconds
produces
Seal defects, registration drift, start-up scrap on restart
Steam trap fails open on a jacketed vessel line
leads to
Jacket temperature undershoots; heat-up ramp extends unpredictably
produces
Viscosity variance, incomplete reaction, off-spec first cut
Why the chain stays invisible
Quality teams own the defect. Facilities teams own the utility. Nobody owns the arrow between them. The defect is logged in a quality system, the utility excursion is logged in a BMS alarm list, and the two records never meet in the same review. Quality maintenance software exists to put both ends of that chain in one asset record.
Five root causes behind a stalled quality maintenance pillar
Most facilities that launch TPM reach autonomous maintenance and planned maintenance, then lose momentum at the quality pillar. These are the failure patterns that show up repeatedly in audits.
1
Defects are recorded by symptom, not by condition
A reject log that says "surface blemish" carries no maintenance information. Until the defect taxonomy names the suspected 4M condition — machine, material, method or man — the maintenance team has nothing actionable to schedule against.
2
Standards exist for tasks, not for values
PM procedures say "inspect dryer" rather than "verify pressure dew point at or below minus 40°C at the point of use." Quality maintenance requires a numeric standard, a tolerance band and a recorded reading on every execution.
3
Condition data has no retention path
Readings taken on paper or in a spreadsheet cannot be trended. Degradation rate — the slope, not the single value — is what predicts the next quality excursion, and it is only visible when readings live in a structured asset history.
4
Restoration work is treated as closed once the line runs
A corrective repair that returns the asset to running condition is not the same as returning it to the condition that guarantees defect-free output. Without a verification step tied to the quality standard, the defect returns on the next campaign.
5
The matrix is built once and never revised
A quality maintenance matrix is a living document. Every new defect mode, every equipment modification and every changed material specification should push a revision. Static matrices decay into shelf documents within two production quarters.
Building the Q-M matrix that drives your PM schedule
The quality maintenance matrix links each defect mode to the equipment conditions that cause it, then converts every condition into a maintainable standard with an owner and a frequency. This is the artefact that turns TPM theory into work orders.
| Defect mode |
Causal 4M condition |
Facility asset |
Standard value |
Verification |
Frequency |
| Coating surface pitting |
Machine — moisture carryover in supply air |
Refrigerated dryer, point-of-use filter |
Dew point at or below −40°C; filter dP under 0.35 bar |
Handheld hygrometer plus inline gauge photo |
Weekly |
| Batch reject on particle count |
Machine — pressure cascade loss |
AHU, terminal HEPA, door interlocks |
Minimum 10 Pa differential, corridor to suite |
Magnehelic reading logged per shift |
Per shift |
| Weld porosity |
Machine — shielding gas contamination |
Gas manifold, regulator, hose assembly |
Flow within ±5% of setpoint; no leak at joints |
Flowmeter check plus leak-detect spray |
Monthly |
| Viscosity out of band |
Machine — jacket heat transfer loss |
Steam traps, condensate return, jacket valve |
Trap discharge within design cycle; no cold leg |
Ultrasonic trap survey with thermal image |
Quarterly |
| Dimensional drift on machined parts |
Machine — thermal growth from ambient swing |
Zone AHU, chilled water valve, controls |
Ambient held at 21°C ±1.5°C over the shift |
Continuous sensor trend with excursion flag |
Continuous |
| Start-up scrap after power event |
Machine — ride-through inconsistency |
Switchgear, UPS, drive cabinets |
No sag below 85% nominal for over 20 ms |
Power quality meter event log review |
Monthly |
Turning matrix rows into scheduled work
Every row above is a preventive maintenance definition waiting to be written. In OxMaint, the standard value becomes a required numeric field on the inspection checklist, the frequency becomes the PM trigger, and an out-of-band reading raises a corrective work order automatically against the same asset record — with the defect mode already attached as context.
Quality maintenance in OxMaint
Give every quality-critical condition a standard, an owner and a due date
Digital inspection checklists with numeric tolerance fields, automatic corrective work orders on out-of-band readings, full asset condition history, and dashboards that show which utilities are trending toward the next defect. Configure your first Q-M matrix in an afternoon.
The quality maintenance ladder, step by step
JIPM sets out quality maintenance as a staged progression. Each rung depends on the one below it, and skipping a rung is the most common reason programs produce documentation instead of results.
Step one
Confirm the current defect picture
Pull twelve months of reject, rework and deviation data. Sort by defect mode and by monetary loss, not by count. The top three modes by cost usually account for most of the recoverable value.
Step two
Investigate the process conditions
For each priority mode, walk the process and list every point where the product touches a utility, a surface or a controlled environment. This produces the candidate condition list for the matrix.
Step three
Restore assets to baseline condition
Fix the accumulated deterioration first — leaking joints, fouled coils, worn seals, drifting sensors. Setting a standard on a degraded asset only guarantees that the standard is breached on day one.
Step four
Run PM analysis on the remaining modes
Where the cause is unclear, break the phenomenon down physically: what must be true for this defect to form? Each physical precondition becomes a testable equipment condition rather than an opinion.
Step five
Set numeric standards and tolerance bands
Each condition receives a target, an acceptable band and an action limit. Standards without a band are unenforceable; standards without an action limit produce readings nobody responds to.
Step six
Build the inspection and PM schedule
Assign frequency by degradation rate, not by convenience. A condition that drifts over weeks needs a weekly check; one that can shift in minutes needs a sensor and an alarm, not a clipboard.
Step seven
Move from result control to condition control
Once conditions hold, inspection of finished output can be reduced and eventually sampled. This is the point where quality maintenance pays for itself twice — in avoided scrap and in released inspection labour.
Result control against condition control
The entire pillar rests on one shift in where control is applied. The comparison below is the fastest way to explain that shift to a leadership team.
Controlling the result
Defects are found after the product is made
Inspection headcount scales with output volume
Maintenance is called after quality raises an issue
Root cause is reconstructed from memory and paperwork
Rework and scrap are budgeted as a fixed cost of doing business
Equipment condition is unknown between breakdowns
Controlling the condition
Conditions are verified before the product is made
Inspection moves to sampling as capability is proven
Maintenance is triggered by an out-of-band reading
Root cause is read from a timestamped condition trend
Scrap becomes a tracked variance with an owner
Equipment condition is a live, reportable number
What this means for the maintenance function
Under condition control, maintenance stops being a cost centre that restores availability and becomes a quality function that protects margin. That reframing is usually what unlocks budget for the sensors, the instrumentation and the CMMS configuration the pillar needs.
The measures that prove a zero-defect program is working
Quality maintenance needs its own scoreboard. Availability metrics will not show progress, and finished-goods reject rate moves too slowly and too noisily to steer by.
Condition compliance rate
Share of scheduled condition checks completed on time with a reading inside the tolerance band. This is the leading indicator; it moves weeks before reject rate does.
Quality rate within OEE
Good units divided by total units started, reported per line and per shift. Isolating this term from availability and performance is what makes the pillar visible.
Defects traced to a mapped condition
Proportion of quality events that resolve to a condition already on the Q-M matrix. A rising figure means the matrix is complete; a falling figure means new modes are emerging.
Excursion-to-restoration time
Elapsed time from an out-of-band reading to verified restoration of the standard. Every hour in this window is production made under a known defect risk.
Repeat defect ratio
Share of defect modes recurring within ninety days of a closed corrective action. High values point to restoration without verification rather than to analysis failure.
Cost of poor quality attributable to assets
Scrap, rework, investigation labour and expedited freight assigned to the responsible asset. This is the figure that justifies the next tranche of instrumentation.
How OxMaint carries the quality pillar day to day
A quality maintenance program lives or dies on whether the data collection survives contact with a real shift. These are the capabilities that do the work.
Digital inspections
Checklists carry numeric fields with target, band and action limit. Technicians enter the reading on a phone or tablet, attach a photo of the gauge, and the entry is timestamped against the asset.
Preventive scheduling
Frequencies run on calendar, runtime or meter triggers, so a condition tied to operating hours is not checked on a fixed date that ignores how hard the asset actually worked.
Corrective work orders
An out-of-band reading raises a corrective order automatically, pre-populated with the defect mode, the standard that was breached and the recent trend for that condition.
Asset management
Every utility asset carries its full condition history, linked documents, spare parts and the defect modes it is responsible for — so handover between shifts and contractors keeps the context.
Condition-based workflows
Sensor and IoT feeds compare live readings against the same standards used on manual checks, so continuous conditions and periodic conditions share one definition of normal.
Inventory
Quality-critical spares — filter elements, desiccant charges, seals, sensors — are held against minimum levels so restoration is not delayed waiting on a purchase order.
Reporting and dashboards
Condition compliance, excursion counts and cost of poor quality by asset are available without exporting to a spreadsheet, which is what keeps the pillar alive past the pilot phase.
Compliance records
Signed, dated inspection records with attached evidence satisfy internal reliability audits and external quality system reviews from the same dataset the engineers use daily.
Questions teams ask before starting the quality pillar
Do we need the other TPM pillars in place before starting quality maintenance?
Autonomous and planned maintenance should be running, because quality maintenance assumes assets are already restored to baseline. You can start the matrix in parallel, but do not set standards on equipment with unaddressed deterioration. Book a demo to review your readiness against a live configuration.
How many defect modes should the first matrix cover?
Three to five, selected by monetary loss rather than frequency. A narrow first matrix proves the method, produces a measurable result inside one quarter, and gives you a template the rest of the site can copy without a consultant.
Can quality maintenance work without sensors on every utility?
Yes. Many conditions are best held by a disciplined manual reading on a defined frequency. Instrument continuously only where the condition can shift faster than your inspection interval — you can set up manual checklists free and add sensor feeds later.
Who owns the Q-M matrix, maintenance or quality?
Quality owns the defect definitions, maintenance owns the conditions and standards, and operations owns daily verification. The matrix should be reviewed jointly each month; single-function ownership is the most reliable way to make it stale.
How long before the program shows in scrap numbers?
Condition compliance improves within weeks of deployment. Scrap and rework typically respond over one to two production quarters, once restoration work from the baseline phase is complete and standards are being held consistently across shifts.
Start your zero-defect program
Your next defect is already forming in a condition nobody is measuring
Map the modes, set the standards, schedule the checks, and let out-of-band readings raise their own work orders. OxMaint gives the quality maintenance pillar a place to live — with the inspection records, asset history and dashboards the program needs to survive past month three.
