TroubleshootingPublished 2026-09-22

Too Many or One Short? 5 Pitfalls of Wrong Card Counting

Customers open a pack and find one too many or one short — lost margin and lost trust. Five causes of counting failures and how to fix them so every pack is exact.

A customer returned a batch: in almost every box a few packs had one too many or one short. He said, "the machine shows 10, doesn't it?" Showing 10 and actually having 10 are two different things, and counting is the gap between them. In card packaging counting looks trivial, but at volume — a smudged sensor, a missed double, a beat that drifts — miscounts creep in. One short and the customer complains; one extra and you lose film and card. Here are the five pitfalls our engineers hit most.

1. Counting Sensor Dirty or Misaligned

Counting rides on the photo-eye, and when film dust or debris coats the window — or the sensor drifts out of position — the number stops matching reality. Fix: wipe the sensor window with alcohol pad every shift; set through-beam to trigger on the card edge, not the whole card, so a thinner card doesn't get missed; run a stack of scrap cards and check the displayed count against the real one.

2. Double or Overlapped Cards Not Detected

Two cards fed overlapped count as one, and the pack ends up with an extra. Fix: add double-sheet detection at the feeder (ultrasonic or thickness-based) to reject overlapped cards; set the separator-wheel gap to card thickness so only one passes at a time; keep the feed stack loose so cards aren't pressed together.

3. Collection Beat Out of Sync with the Line

Line fast, collector slow — a card slips past uncounted or gets counted twice, and the count drifts. Fix: drive the collector off the line encoder, counting pulses not clock time; after tuning, count 50 packs and check the error sits near zero, not just one or two.

4. Film Flutter Trips the Sensor

When the film flutters, the reflection shifts and the sensor trips, adding a phantom card. Fix: steady the wrap tension and add film guides where flutter is worst; keep the sensor off high-glare zones, or switch to a diffuse type with background suppression that sees the card, not the film shine.

5. Logic Doesn't Filter Bad or Stuck Packs

A card tilts or two sit too close and the logic counts them as two — or one. Fix: define a "valid card" by size plus spacing, and treat two pieces closer than the threshold as one; run a batch and measure the misjudge rate, only scaling up once it's below one in a thousand.

Why Counting Is Worth Watching

Counting accuracy hits both cost and reputation — one short is a complaint, one extra is real loss. One line: a correct display isn't enough; every pack actually matching its count is what makes the line stable.

See Accurate Counting Live

These are the exact checks our engineers run on every CHANFER card line before it ships. PRINTING United Expo 2026 opens tomorrow — Booth N7411, Las Vegas, Sept 23–25 — bring a pack that gives you counting trouble and we'll get the count right, seal tight and shrink flat on the spot.

Book Sept 23-25 US Show

FAQ

The machine shows 10 cards — why is the actual count wrong?

The displayed count and the actual count are different things — the counting step is the gap. Common causes: the counting photo-eye coated with film dust or misaligned, overlapped or double cards not detected, collector beat out of sync with the line, film flutter tripping the sensor, or logic not filtering bad or stuck packs. Check them one by one.

A pack has one too many or one short — how do I troubleshoot?

First wipe the sensor window and set through-beam to trigger on the card edge; add double-sheet detection at the feeder to reject overlapped cards; drive the collector off the line encoder; steady the wrap tension to cut flutter; define a valid card by size plus spacing to filter stuck packs. After tuning, count 50 packs and check the error sits near zero.

How do I push counting error to the minimum?

Count pulses off the encoder rather than clock time, so collection stays in sync with the line; measure the misjudge rate over a batch and only scale up once it is below one in a thousand; clean the sensor every shift and keep double-sheet detection on.