Gauge Calibration and MSA Guide — Gauge R&R, %GRR and ndc

Q: What do bias, linearity and stability mean in MSA?

Bias is the difference between the observed average measurement and a reference (true) value — a consistent offset, the accuracy of the system. Linearity is how bias changes across the gauge's operating range — a gauge might read true at the low end and drift at the high end. Stability is how bias changes over time — the same gauge measuring the same reference part drifting from week to week. Together these three describe the location behaviour of a measurement system, while Gauge R&R describes its spread.

Q: How do you read a Gauge R&R result (%GRR and ndc)?

%GRR expresses the measurement-system variation (repeatability plus reproducibility) as a percentage of the total variation or of the tolerance. As a common guideline, under 10% is acceptable, 10% to 30% is conditionally acceptable depending on the importance and cost of the application, and over 30% is unacceptable — the gauge cannot be trusted for that characteristic. ndc, the number of distinct categories, is how many separate groups the system can reliably distinguish across the part range; it should be 5 or more. A high %GRR or an ndc below 5 means the measurement system, not the process, is producing much of the scatter you see.

Q: What is the difference between repeatability and reproducibility?

Repeatability is the variation you get when the same operator measures the same part with the same gauge several times — it is equipment variation, the gauge's own inability to repeat. Reproducibility is the variation between different operators measuring the same parts with the same gauge — it is appraiser variation, the difference people introduce through technique. Gauge R&R combines the two. If repeatability dominates, look at the gauge or fixture; if reproducibility dominates, look at operator method and training.

Q: What should you do when a gauge fails calibration?

Quarantine the gauge immediately so it cannot be used, then assess the impact: every characteristic and inspection that relied on it since the last good calibration is suspect and should be flagged for review — a recall of suspect measurements, which may mean re-checking product or notifying a customer. Investigate the cause, repair or replace the gauge, re-calibrate and re-run its Gauge R&R before returning it to service. The whole point of tracking gauges in a register with due-dates is to make this recall bounded and fast rather than open-ended.

A quality system runs on numbers, and every one of those numbers came out of a gauge. If the gauge is out of calibration, the number is wrong by an unknown amount. If the gauge cannot resolve the feature it is measuring, the number is mostly noise. Two disciplines keep the numbers honest: calibration, which proves a gauge reads true against a traceable standard, and MSA, which proves the measurement system can actually tell parts apart. Skip either and every downstream decision — accept, reject, capable, not capable — is built on sand.

This guide is for metrology owners, quality engineers and MRs working to IATF 16949 or ISO 9001. It covers calibration due-dates and traceability, MSA studies, how to read a Gauge R&R result, and what to do on a failed gauge. For the full picture see the pillar guide, What is quality management software?

Calibration is accuracy; MSA is capability

A gauge can be perfectly calibrated and still fail MSA — reading true against a standard says nothing about whether it can distinguish a good part from a marginal one in real use. Calibration answers "does it read the right value?"; MSA answers "can it reliably tell my parts apart?" You need both, and in that order.

1. What are calibration and MSA?

Calibration is the periodic comparison of a gauge against a reference standard of known, traceable value, to confirm it reads true within its permitted error — and to adjust or flag it if it does not. Traceability means that reference chains back, through an accredited laboratory, to a national or international standard. This is the requirement behind IATF 16949 and ISO 9001 clause 7.1.5 on monitoring and measuring resources.

MSA — Measurement System Analysis — is the set of studies that qualify a measurement system for a given characteristic: bias, linearity and stability (its location behaviour) and Gauge R&R — repeatability and reproducibility (its spread). MSA uses the real gauge, the real operators and the real method, because a measurement system is more than the instrument: it includes the person, the procedure and the environment.

Both feed the gauge register and calibration discipline, and both are required before a gauge is used on a control-plan characteristic.

2. Calibration, due-dates and clause 7.1.5

Calibration only protects you if no gauge silently drifts out of date between audits. That is what the gauge register and its due-date tracking exist to prevent.

Each gauge is registered once with its identity and its cycle: gauge code, type, least count and range, location or department, last-calibration date and calibration frequency — from which the system derives the next-due date. A gauge-type master classifies types (vernier, micrometer, plug gauge, height gauge, CMM) each with a default frequency, so a new gauge inherits its cycle rather than being set by hand. From there the calibration follow-up drives the cycle:

#	Stage	What happens
1	Register	Gauge added with type, range, last-cal date and frequency; the next-due date is derived.
2	Due & alert	The follow-up lists gauges by next-due date and alerts as one approaches or passes due.
3	Sent	The gauge is sent for calibration — in-house or to an accredited external lab.
4	Calibrated	The result and certificate are recorded; a pass rolls the next-due date forward.
5	Returned	The gauge returns to service with its new sticker; a fail is quarantined (see section 6).

The audit-ready output is the calibration register / MIS: every gauge with its calibration status, history and overdue count. Clause 7.1.5 asks for exactly this evidence — measuring resources verified and kept fit for use, with records retained — and a register produced from daily calibration work is far more defensible than a spreadsheet assembled the week before an audit. See Gauge, MSA & Calibration.

"An out-of-calibration gauge does not announce itself. The only defence is a register that raises its hand before the due date, not after the audit finding." — Fast Technology Team

3. MSA: bias, linearity and stability

Calibration checks a gauge at a point; MSA studies how it behaves across range, over time and in real hands. Three studies describe its location — how far off, and how consistently:

Bias

The difference between the observed average and a reference value
A consistent offset — the accuracy of the system
Found by measuring a known reference part many times

Linearity

How bias changes across the gauge's operating range
Reads true at the low end but drifts at the high end
Checked with reference parts spanning the range

Stability

How bias changes over time
The same gauge and reference drifting week to week
Tracked by periodic checks on a master part

These three answer "is the system on target, and does it stay on target?" The fourth study — Gauge R&R — answers a different question: "how much does the system scatter?" A gauge with tiny bias can still be useless if its scatter swamps the tolerance, which is why R&R gets its own section.

4. Gauge R&R: %GRR and ndc

Gauge Repeatability & Reproducibility measures the spread a measurement system adds. It splits that spread into two sources:

Repeatability — the variation when the same operator measures the same part with the same gauge several times. This is equipment variation (EV): the gauge's own inability to repeat.
Reproducibility — the variation between operators measuring the same parts with the same gauge. This is appraiser variation (AV): the difference technique introduces.

A typical study uses about 10 parts, 3 operators and 3 trials each, spanning the range of the process. The results are read through two figures:

%GRR	Verdict	What it means
Under 10%	Acceptable	Measurement system is fine for the characteristic.
10% – 30%	Conditionally acceptable	May be acceptable depending on the importance and cost of the application.
Over 30%	Unacceptable	The gauge cannot be trusted for this characteristic — improve or replace it.

%GRR expresses the measurement-system variation (repeatability + reproducibility) as a percentage of the total variation or of the tolerance. ndc — the number of distinct categories — is how many separate groups the system can reliably distinguish across the part range; it should be 5 or more. A high %GRR or an ndc below 5 tells you the measurement system, not the process, is producing much of the scatter you are seeing — so tightening the process would be chasing a problem that lives in the gauge. And the split matters: if repeatability dominates, look at the gauge, fixture or datum; if reproducibility dominates, look at operator method and training.

5. A worked example (illustrative)

Here is an illustrative Gauge R&R on a micrometer measuring a shaft diameter. The numbers are illustrative, for teaching only.

Gauge R&R — shaft diameter micrometer, illustrative

Characteristic: shaft ⌀ 25.00 mm, tolerance 0.02 mm; gauge: 0–25 mm digital micrometer.
Study: 10 parts × 3 operators × 3 trials.
Result: %GRR = 18% — conditionally acceptable; ndc = 6 — acceptable (≥ 5).
Split: repeatability 16%, reproducibility 8% — the gauge/fixture dominates, not the operators.
Reading: usable for now given the characteristic's importance, but the repeatability share says the micrometer and setup are the limit — a bench-mounted bore gauge or a fixtured setup would cut EV.

The lesson from the split: because repeatability (equipment) dominates, operator training would barely help — the improvement is a better gauge or a more repeatable fixturing method. Had reproducibility dominated instead, the fix would have been a standard work instruction and operator training. Reading the split, not just the headline %GRR, is what turns an R&R study into an action. This is the shape of a real deployment such as Kakade Laser or Shree Engineering.

6. What to do when a gauge fails

A gauge that fails calibration — or fails Gauge R&R — is not just a gauge problem; it is a data problem, because every reading it produced is now in question.

Quarantine it immediately

Remove the gauge from service so it cannot be used again
Tag it clearly and record the failure against the gauge in the register

Assess and recall suspect measurements

Every characteristic and inspection that used it since the last good calibration is suspect
Flag those for review — re-check product, and notify the customer where shipped product is affected

Fix, re-qualify and return

Investigate the cause, repair or replace the gauge, and re-calibrate
Re-run Gauge R&R before it goes back on a control-plan characteristic

The reason a gauge register with due-dates matters so much is precisely this recall: when a gauge is tracked, the set of suspect measurements is bounded — everything since its last good calibration — and the review is fast. When gauges live on a wall of stickers, a failure means an open-ended, frightening question about how much data is bad.

7. Running the gauge register well

A few habits separate a metrology function that passes audits comfortably from one that scrambles.

One register, derived due-dates

Every gauge in one register, with the next-due date derived from frequency, not typed
Type-level default frequencies so a new gauge inherits its cycle
Alerts before due — email, SMS or WhatsApp — not a monthly manual sweep

Link the gauge to the control plan

Every control-plan characteristic points at a specific gauge from the register
That gauge must be MSA-qualified and in calibration to be used
So an out-of-date gauge is visible against the checks that depend on it

Keep MSA current, not one-off

Re-run Gauge R&R on a schedule and after any repair or major event
Store the study with the gauge as a PPAP element
Read the EV/AV split each time, not just the headline %GRR

8. How Fast Quality Software runs gauges, MSA and calibration

Fast Quality Software is the automotive QMS of the Fast Suite, built in Pune by Improsys under the Fast Technology brand and deployed cloud or on-premise. It runs the full gauge lifecycle so every inspection number is defensible:

Capability	How Fast Quality Software does it
Gauge register	Every gauge registered with code, type, least count/range, location, last-calibration date, frequency and a derived next-due date; a gauge-type master sets default frequencies. See gauge, MSA & calibration.
Calibration follow-up	Lists gauges by next-due date and drives the cycle — due, sent, calibrated, returned — rolling the next-due date forward and alerting before a gauge falls due.
MSA / Gauge R&R	Bias, linearity, stability and Gauge R&R recorded per gauge and attached as a PPAP element; a gauge that fails R&R is flagged before use on a control-plan characteristic.
Calibration register / MIS	An audit-ready register of every gauge with calibration status, history and overdue count — the evidence clause 7.1.5 asks for.
Failed-gauge recall	A gauge that fails is quarantined and the characteristics and inspections that used it are flagged for review — a bounded recall of suspect measurements.
Alerts & AI	Calibration-due alerts by email, SMS and WhatsApp; Dhruv AI summarises calibration-due and overdue gauges in plain English.

Part of the Fast Suite — the automotive QMS

Make every measurement defensible — calibrated, traceable, MSA-qualified.

Fast Quality Software runs the gauge register, calibration follow-up and MSA on one platform, linked to the control plan and inspection. A control-plan characteristic can only be measured by a gauge that is in calibration and has passed Gauge R&R — and when a gauge fails, the recall of suspect measurements is bounded and fast. Clause 7.1.5 evidence is a by-product of daily work.

Gauge register with derived next-due dates and pre-due alerts

MSA / Gauge R&R per gauge, attached as a PPAP element

Cloud or on-premise, for IATF 16949 and ISO 9001 makers worldwide

Get a demo

9. Frequently asked questions

What is the difference between calibration and MSA?

Calibration proves a gauge reads true against a traceable reference standard on a defined frequency — the requirement behind clause 7.1.5. MSA proves the measurement system is capable of telling parts apart in real use, through bias, linearity, stability and Gauge R&R with the actual operators. A gauge can be perfectly calibrated yet still fail MSA if it cannot resolve the characteristic. You need both — calibration for accuracy, MSA for capability.

What do bias, linearity and stability mean in MSA?

Bias is the difference between the observed average and a reference (true) value — a consistent offset, the accuracy of the system. Linearity is how bias changes across the gauge's range — true at one end, drifting at the other. Stability is how bias changes over time — the same gauge and reference part drifting week to week. Together they describe the location of a measurement system, while Gauge R&R describes its spread.

How do you read a Gauge R&R result (%GRR and ndc)?

%GRR expresses the measurement-system variation as a percentage of the total variation or tolerance: under 10% is acceptable, 10–30% conditionally acceptable depending on importance and cost, and over 30% unacceptable. ndc — the number of distinct categories — is how many groups the system can reliably distinguish, and should be 5 or more. A high %GRR or ndc below 5 means the measurement system, not the process, is producing much of the scatter.

What is the difference between repeatability and reproducibility?

Repeatability is the variation when the same operator measures the same part with the same gauge several times — equipment variation, the gauge's own inability to repeat. Reproducibility is the variation between different operators measuring the same parts — appraiser variation, the difference technique introduces. Gauge R&R combines the two. If repeatability dominates, look at the gauge or fixture; if reproducibility dominates, look at operator method and training.

What should you do when a gauge fails calibration?

Quarantine it immediately, then assess the impact: every characteristic and inspection that relied on it since the last good calibration is suspect and should be flagged for review — a recall of suspect measurements that may mean re-checking product or notifying a customer. Investigate the cause, repair or replace, re-calibrate and re-run Gauge R&R before returning it to service. Tracking gauges in a register with due-dates makes this recall bounded and fast.

See a gauge register and Gauge R&R on your own gauges

A 30-minute demo — your gauge register with next-due dates, calibration follow-up, MSA studies and a control plan that only uses qualified gauges. Cloud or on-premise.

Get a demo View pricing Read the pillar guide