Two production batches can look the same and wear differently.
An average hides that difference. Consistency means that the same test on the same face gives the same number twice.
The Short Answer: Consistency Is a Repeat, Not an Average
Grit retention is a ratio. It compares how much texture and how much spin a face keeps after a set wear run.
A batch is consistent when that ratio repeats. Three things make it repeat.
- A fixed protocol. Same speed, angle, hits, measurement points, cleaning and climate.
- A sampling plan. Same number of paddles, pulled from the same places in the run.
- An acceptance range. A range written down before the first batch, instead of after the third.
If a line misses one of the three, the numbers move for reasons unrelated to grit.
What “Grit Retention” Means as a Number
Two readings per set: one before the wear run, one after.
- Texture retention: roughness after ÷ roughness before, as a share. Use Rz and Rt.
- Spin retention: RPM after ÷ RPM before.
Both are ratios, so both travel between batches. A raw RPM figure cannot do that, because one rig difference moves it. For where the grain sits in the face, see how grit is anchored in the face.
Step 1 — Fix the Test Conditions Before You Compare Anything
A brand should write the protocol once and then leave it alone for the season. Seven values decide the result.
- Ball speed
- Impact angle
- Number of hits
- Measurement points on the face
- Number of shots measured
- Cleaning between runs
- Room temperature and humidity
Two runs with different values cannot be compared, even when both report RPM. The machines and the evidence levels behind each test sit in test protocol and proof levels.
Step 2 — Set the Sampling Plan: Which Paddles, How Many, From Where
Samples should come from set positions in the run rather than from wherever it is convenient.
- How many: at least six measured readings per batch, across at least three paddles.
- From where: one from the start of the run, one from the middle, one from the end.
- Which batches: three in a row, because two batches only show a direction.
Paddles pulled from the open carton are the ones the line checked most, so they pass most often. Pull from sealed cartons instead. A paddle sitting beside the press has already been handled by the crew that knows the surface best.
Weight needs the same discipline on its own spec, because mass drift and texture drift come from different causes. That side of the file is covered in weight consistency in bulk production.
Step 3 — Set Acceptance Limits, and Write Them as Ranges
A single pass or fail line hides a slow drift. A band of two limits shows it.
Measure | What to write |
Texture retention | A minimum share, plus a maximum |
Spin retention | A minimum share of the batch baseline |
Roughness on arrival | Inside the certification limits (Rz ≤ 30 µm, Rt ≤ 40 µm) |
Spread | A ceiling on the gap between the highest and lowest reading |
A maximum earns its place next to the minimum. A face that arrives far coarser than the approved reference wears balls faster and pushes against the roughness cap.
A breached limit turns the row into a hold rather than a reject. A re-test comes first, because roughness readings move with the room. A second run on the same paddles settles most disputes before they reach the buyer.
Step 4 — Keep One Batch Table With Fixed Columns
The table keeps one row per batch, and the columns stay the same every time.
Column | Why it earns its place |
Batch and date | Ties the row to a carton |
Sand lot and resin lot | Explains a step change |
Line and shift | Explains a step change |
Roughness before (Rz, Rt) | The baseline |
Roughness after | Texture retention |
Spin before and after, mean | Spin retention |
Spread of readings | Whether the batch is stable |
Verdict | Accept, hold or re-test |
Our own face keeps this table, and the pickleball paddles with durable grit line is what the table protects. A table with empty columns is a promise rather than a record.
Reading Batch Data: Drift, Step Change and Scatter
Three patterns, three different causes.
- Drift. Numbers slide slowly. Usually a worn tool, a mold losing pressure, or a resin lot change.
- Step change. One batch jumps. Usually a new sand lot, a different line, or a new operator.
- Scatter. Readings inside one batch spread wide. Usually a process that is not controlled, or a test method that is not fixed.
Each pattern needs a different answer. Drift needs a maintenance schedule. A step change needs a lot check. Scatter needs a tighter protocol before it needs a tighter spec.
What Our Own Batch Table Looks Like
We keep those fields per run rather than per month, because a month can hold several sand lots. A sample paddle stays in the file for the length of the warranty. A claim then traces back to a lot number instead of someone’s memory.
That file, and what a buyer can ask to see, is described in how a factory documents grit capability.
Traceability: Sand Lot, Resin Lot, Line and Shift
Retention numbers only help if you can trace them.
- Sand lot: mesh and coating change between lots.
- Resin lot: cure behavior shifts between lots.
- Line: cold press, thermoformed and unibody build different faces.
- Shift: small process changes show up here first.
Traceability is the difference between fixing a batch and guessing about one.
When a Batch Fails: Contain, Rework, Re-Test
A failed row needs three actions, in order.
- Contain. Hold that batch and its cartons, not the whole run.
- Find the variable. Check the lots, the line and the shift in that row.
- Re-test after the change. Same rig, same protocol, fresh readings.
Two clear limits apply here. Six readings per batch work as a sample rather than a full inspection. Three matching batches still leave the fourth batch unproven. They show that the process can repeat while every input stays the same.
FAQ
How often should grit retention be measured?
At least once per production batch. Weekly is not the same thing, because one week can hold several sand lots.
How many readings per batch?
Six or more. Below that, the spread between readings is wider than the change you are testing for.
Do I need to test every paddle?
No. Testing every paddle destroys the order. Sample the run, then keep the samples.
What if two batches match?
Two matching batches are encouraging, and they are still not proof. Keep the third batch in the plan.
Can I compare retention numbers from two factories?
Not unless both protocols match on all seven values. Most do not, so the comparison does not hold.
Does grit retention drop as a mold ages?
It can. Mold pressure and surface wear change how much texture a face keeps. A buyer should watch the drift rather than one reading.
Should appearance defects get an AQL too?
Yes. Set the AQL for visual defects in the quality agreement, and keep it separate from wear data.
What should happen to the sample paddles?
Keep them. A retained sample per batch is the fastest way to settle a warranty claim later.
Final Checklist: Batch Terms for the Quality Agreement
- One written protocol with all seven values fixed
- Sampling plan: three paddles, start, middle and end of the run
- Minimum and maximum limits for texture and spin retention
- A ceiling on the spread inside one batch
- One batch table with fixed columns, delivered with each run
- Sand lot and resin lot recorded per batch
- Line and shift recorded per batch
- Containment and re-test procedure for a failed batch
- Sample paddles retained for the warranty period
A supplier that answers all nine items turns the third batch into a known quantity. Send the list to our engineering team and we will answer every line.