Probe Repeatability Testing: A Simple Shop Procedure (That Actually Catches Problems)

If you’re using a touch-trigger probe (on a CMM, shop-floor CMM, or probing on a CNC), “repeatability” is the canary in the coal mine. When repeatability drifts, your results can still look believable—until a critical tolerance goes sideways and nobody can explain why.

This post gives you a fast, shop-friendly repeatability test you can run in minutes, plus a deeper, more diagnostic routine you can run weekly/monthly. The goal is simple: detect change early, identify whether it’s probe/stylus/qualification/environment, and keep your measurement system trustworthy.

---

What “probe repeatability” really means (in plain terms)

Repeatability is the probe’s ability to produce closely similar results when you measure the same thing the same way under the same conditions. Standards commonly express this by repeating measurements of a precision sphere and looking at the range (max–min) of the resulting coordinates—often across X, Y, and Z—with the largest range reported as repeatability. https://cnc-probe.com/cnc-touch-probes/

Why a sphere?
A calibrated reference sphere is stable, symmetrical, and easy to probe consistently. It’s also the standard artifact used for probe performance checks and qualification. https://cnc-probe.com/presition-cnc-calibration-sphere-ceramic-standard-ball-probe-gauge/

---

Two levels of testing you should run

Level 1 (Daily/Shift Start): “10-cycle sphere center repeatability”

• Time: ~3–7 minutes
• Purpose: Detect drift fast (loose stylus, contamination, thermal swings, vibration, probe re-seat issues)
• Output: Repeatability value = largest coordinate range (X/Y/Z)

This mirrors common sphere-repeatability concepts used in practice and in standard-style definitions (repeat center coordinates multiple times and evaluate ranges). https://cnc-probe.com/presition-cnc-calibration-sphere-ceramic-standard-ball-probe-gauge/

Level 2 (Weekly/Monthly): “25-point sphere form / probing performance”

• Time: ~10–20 minutes
• Purpose: More sensitive to true probing performance and qualification quality
• Output: Sphere form / total form deviation (often used as a probing-performance indicator in ISO-style probing tests using ~25 points on a sphere).

Hexagon’s practical guidance aligns with this: qualify normally, then measure the same sphere with at least 25 points, repeat multiple times, and compare to your baseline.

---

Before you test: the “boring” prep that makes the test meaningful

These steps are not optional if you want a repeatability number you can trust:

1. Verify the sphere diameter in software matches the sphere’s certified diameter
   This prevents you from chasing fake errors created by the setup.
2. Clean the reference sphere properly
   Fingerprints and film matter at micron levels. Wipe with a lint-free cloth; denatured alcohol is commonly recommended for cleaning the qualification sphere.
3. Check the stylus assembly
   • Is the stylus tight?Any crash marks?Any long/heavy stylus stack that’s “barely stable”?
   Probe systems are extremely repeatable when configured correctly, but mass/length and setup quality strongly influence real results.
4. Stabilize the environment as much as your shop allows
   Temperature swings and airflow can show up as “repeatability drift” even when the probe is fine. (This is exactly why many performance methods involve repeated measurements and comparing ranges.)

---

Level 1: The Simple Shop Procedure (10-cycle sphere repeatability)

What you need

• A calibrated reference sphere, mounted rigidly (magnetic base or fixture, ideally near the center of the working volume)
• Your normal probe setup (the one you use for real parts)https://cnc-probe.com/cnc-touch-probes/
• A short program or routine that probes the sphere the same way every time

Step-by-step procedure

Step 1 — Qualify the probe (normal qualification)https://cnc-probe.com/cnc-touch-probes/

• Do your normal probe qualification on the qualification sphere.

Step 2 — Measure the sphere using a consistent pattern
A simple, stable pattern is:

• 1 point near the top (Z+ direction)
• 4 points around the equator (roughly 90° apart)

This “5 touches” pattern is commonly used in practical repeatability checks.

Step 3 — Repeat the exact same measurement 10 times

• Same speeds, same approach distances, same stylus, same trigger force setting.

Step 4 — Record the sphere center (X, Y, Z) each cycle
Most metrology software will output the fitted sphere center. If yours doesn’t, export feature results.

Step 5 — Calculate repeatability
For each axis:

• RangeX = max(X) – min(X)
• RangeY = max(Y) – min(Y)
• RangeZ = max(Z) – min(Z)

Repeatability = max(RangeX, RangeY, RangeZ)
This “largest coordinate range” reporting is a standard-style way repeatability is communicated in common practice literature.

Pass/Fail: how to set limits without guessing

Use a baseline-based control limit, because “good” depends on your machine, stylus, and environment:

• Baseline: When the system is known-good (after service or after a verified good run), record 10-cycle repeatability for 3 days and compute the average.
• Alert limit: 2× baseline
• Action limit: 3× baseline (or any sudden step change)

This is often more effective than chasing a spec sheet number, because it catches change—the thing that breaks inspection processes.

---

Level 2: The Deeper Test (25-point sphere form / probing performance)

When you need a more sensitive check (weekly/monthly, after a probe crash, after stylus change, or when Level 1 starts creeping up), run a 25-point sphere measurement, repeat it 3+ times, and compare results to your baseline.

Why 25 points?

Because a denser, well-distributed point set makes it harder for a “marginal” probe qualification or direction bias to hide. Many probing performance approaches evaluate the form deviation of the measured sphere using all points (don’t “throw out outliers” just to make the report pretty).

What to track

Track at least:

• Sphere form (or total form deviation)
• Sphere diameter result
• Sphere center shift vs baseline

If form worsens but center repeatability stays okay, you may be looking at directionality, qualification strategy, or stylus dynamics—not just “noise.”

---

A simple data sheet template (copy/paste)

Probe Repeatability Log — Level 1 (10-cycle sphere center)

• Date/Time:
• Operator:
• Machine:
• Probe / stylus ID:
• Trigger force setting (if applicable):
• Sphere ID / location:
• Qualification program used:
• Measurement program used:
• Ambient temp (if available):

Cycle	X center	Y center	Z center
1
2
…
10

Calculated:

• RangeX =
• RangeY =
• RangeZ =
• Repeatability = max(RangeX, RangeY, RangeZ) =

Disposition:

• ✅ Pass / ⚠️ Alert / ❌ Action

---

If it fails: a no-drama troubleshooting flow that works

1) Clean + re-run (fastest win)

• Clean sphere + stylus tip; re-run Level 1.
• Confirm the software sphere diameter matches the certified diameter.

2) Check the stylus stack

Common failure mode: stylus loosens slightly and repeatability explodes.

• Re-seat and torque properly (per your supplier guidance)
• Shorten stylus / reduce mass if possible (heavy configurations can cause re-arm or repeat issues)

3) Re-qualify with a better strategy

If you qualify with too few points and test with many points, you can create self-inflicted “performance failures.” A higher-density qualification is often recommended before deeper sphere tests.

4) Trigger force / re-seat behavior (touch-trigger probes)

If the probe fails to re-arm reliably, or if trigger force is set too low, repeatability can degrade.
Renishaw’s troubleshooting guidance includes adjusting trigger force to the minimum necessary for reliable triggering and reviewing stylus mass/length.

5) Environment + machine condition

If repeatability drifts with time-of-day, suspect temperature gradients, airflow, vibration, or axis motion issues. Standards-style repeat testing exists for a reason: the system is a combination of probe + machine + environment.

---

How often should you run it?

A practical schedule that doesn’t annoy production:

• Every shift start (or daily): Level 1 (10-cycle)
• After any stylus change: Level 1 immediately + Level 2 within 24 hours
• After a probe crash / tip strike: Level 1 immediately; Level 2 before releasing high-risk parts
• Weekly or monthly (depending on risk): Level 2 (25-point sphere)
• Whenever results “feel weird”: Don’t argue—run Level 1. It’s faster than a meeting.

---

The real competitive advantage: treat repeatability like SPC, not a one-off “calibration ritual”

Most shops run a check, glance at a number, and move on. The shops that win do one extra thing:

They trend it.
A simple run chart of your Level 1 repeatability value will show:

• slow thermal drift
• a stylus getting worse over weeks
• a mounting issue
• a probe re-seat problem starting to appear

That’s how you catch measurement risk before it becomes scrap, RMAs, or customer returns.