Torque Testing a 48-Cavity Closure Mould: Why You Can't (and Shouldn't) Test Every Cap

By Amy Zhao, Factory Technical Lead, KHT Instrument

Cavity-based torque sampling is a quality control method where caps are tagged by their mould cavity number and tested in rotating batches, so that every cavity gets covered over a short run cycle instead of testing every single cap produced.

If you run a multi-cavity injection mould for closures, you already know the anxiety this creates: with dozens of cavities producing caps simultaneously, how do you know that every cavity is producing a cap within spec — without testing all of them?

The Real Question Behind "Test All 48"

A closures manufacturer running a 48-cavity injection mould raised exactly this question with us: their mould produces caps 1 through 48 in every batch, and each one needs to pass torque inspection — so how does a torque tester help capture all 48 samples in one batch?

It's a good question, and it comes from the right instinct: a bad cavity — say, one with a worn thread-forming pin — can silently produce out-of-spec caps for that entire cavity position, batch after batch, while the other 47 cavities look fine. The real risk isn't "one bad cap slipping through." It's one bad cavity slipping through undetected for days.

Why You Physically Can't Test All 48 (One-Hand Experience)

Torque testing is a destructive test: the cap is turned to its opening or breakage point to capture peak force, which means the cap can't be resealed and returned to the batch afterward. This isn't a limitation of any particular instrument — it's inherent to how torque measurement works. So "test all 48 caps from every batch" isn't actually an option once you think it through: you'd be destroying 100% of what you just tested, which defeats production yield entirely. This is standard across the packaging QC industry, which is exactly why sampling-based inspection — not full-batch testing — is the accepted method.

The Method That Actually Protects You: Cavity-Tagged Rotating Sampling

The method that addresses the real concern — a single bad cavity going unnoticed — is cavity-based sampling:

1. Tag each pulled sample by its mould cavity number at the point of collection (this is a labeling step in your sampling procedure, not something the torque tester does automatically).
2. Test a rotating subset each shift, cycling through cavity numbers so that across a defined number of shifts, every one of the 48 cavities has been sampled and tested at least once.
3. Track results by cavity number, not just by batch. If one cavity's torque values start drifting or falling outside your window while the others stay stable, you can trace the deviation straight back to that specific mould position — often before it becomes a customer-facing problem.

This turns "we tested some caps and they were fine" into "we know which of our 48 cavities are behaving and which one needs mould maintenance" — a materially stronger quality signal, even though you're testing fewer total samples than a naive "test everything" approach would require.

Common Mistake: Treating Cavity Tracking as an Instrument Feature

A mistake worth naming directly: cavity-to-cavity comparison is only meaningful because of how you sample and tag, not because the torque tester itself tracks which mould cavity a cap came from. A digital torque tester like the NLY-20S records torque values, peak force, and run history with high precision — but the cavity number is information your sampling procedure attaches, typically by labeling or batching samples before they reach the instrument. If your sampling procedure doesn't tag by cavity, cavity-level traceability is lost no matter how good the instrument is.

What a Consistent Instrument Adds to This Method

Cavity-based sampling only works if the instrument's own readings are trustworthy enough that a shift in the number reflects the cap, not test-to-test noise. This is where an automatic tester earns its keep: on the NLY-20S, the mechanical gripper applies and releases each cap identically, with locking deviation under 0.001 N·m between cycles — so when your cavity-tagged sample from cavity #23 reads differently from cavity #7, you can trust that the difference is real, not an artifact of how the test itself was run (an operator-driven manual test would not give you that same confidence).

Sampling Approach at a Glance

Frequently Asked Questions

Q1. Can a torque tester automatically identify which mould cavity a cap came from? No. Cavity identification is a labeling/sampling-procedure step done before or during sample collection, not a built-in torque tester function. The instrument measures torque; your process ties that measurement to a cavity number.

Q2. How many cavities should I sample per shift? There's no single universal number — it depends on your production volume, your risk tolerance, and how frequently your standard operating procedure requires full 48-cavity coverage (e.g., covering all cavities within a day vs. within a week). This is worth defining explicitly in your sampling SOP rather than leaving it ad hoc.

Q3. Why can't I just retest a cap that already passed the destructive test? Because the test itself deforms or breaks the closure at its measured peak force — that's what "destructive" means. Once tested, that specific cap can no longer represent an untested, sealed unit.

Q4. Does this sampling approach apply to manual torque testers too, or only automatic ones? The cavity-tagging and rotation method itself works with either instrument type — it's a sampling procedure, not an instrument feature. An automatic tester simply makes the resulting cavity-to-cavity comparisons more trustworthy, since it removes operator-to-operator variation from each reading.

See full specifications and configurations on the NLY-20 Digital Cap Torque Tester product page.