Short answer: most people on the line learned the record from a busy colleague, and nobody reads their entries while the unit is still in the fixture. The errors come from how the habit is taught and timed. Carelessness plays a minor role. A structured first lesson on records, a supervisor who checks entries at the station and a sheet designed for gloved hands cure most of it.
Picture the query that starts it. A dealer writes back about a nebuliser whose compressor tripped on its second use. Quality pulls the build sheet for that serial number. The diaphragm torque column is ticked, not filled. The leakage reading is in the same ballpoint as every other entry that day, at the same slant. The initials belong to an assembler who had the afternoon off. Nobody lied. The sheet was jotted from memory at half past five, exactly the way his trainer always did it.
What the errors look like
- Ticks where a value was asked for, such as a diaphragm torque, a leakage figure or a crimp pull result.
- Entries made after the fact, recognisable by one pen, one hand and identical times down a whole shift.
- Values pencilled on a glove or a scrap of card, then copied across later, sometimes into the wrong row.
- Lot numbers carried over from the previous carton of silicone tubing, O-rings or diaphragms because a fresh carton was opened mid-shift.
- Serial numbers transposed between the build sheet, the test printout and the carton sticker.
- Initials of the wrong person, usually the senior who owns the station, even when a trainee did the fitting.
- Rework left off the sheet, so a board resoldered at final test looks identical on paper to one that passed first time.
Each slip is trivial at the bench and serious later. A hospital, a distributor or an export buyer asks who built a unit, with which parts, and whether it was tested. The device record exists to answer that within an hour. Gaps turn the hour into a week of guesswork and phone calls to people who have since left.
Where the habit comes from
New assemblers learn the work standing beside someone. They pick up the paperwork the same way. If the person beside them scrawls the sheet at clocking-off, so will they. Induction in many units covers the gate, the canteen and the fixture. It seldom includes ten minutes with a blank build sheet and a pen.
The sheet itself frequently adds to the trouble. Someone drew it up in the quality office. It asks for a dozen fields at a station where hands are gloved and the reference gauge sits two benches away. It lives in a folder on the supervisor's desk instead of hanging beside the jig. Output targets make it worse, since every pause to write costs a unit. In a tender rush, newcomers arrive faster than anyone can teach them. Errors climb in exactly the weeks when buyers look hardest.
Why blaming the last initials backfires
When a query lands, the reflex is to find the initials and issue a warning. Two things follow. Careful operators decide the record is a trap and stop signing anything they are unsure of, so blanks multiply. The rest learn to back-fill tidily, so the sheet looks spotless and tells you less. Good assemblers with options find a unit where a missed box is not a disciplinary matter. The fault sat in the system, and the system stays untouched.
A fix that works at the bench
- 1
Pull ten recent records and sort the errors
Take ten build or batch sheets from the past month. Mark every blank, tick, carried-over lot and late entry. Group them by column and by station. Two or three columns usually produce most of the trouble.
- 2
Redesign the columns that keep failing
Ask the assemblers why each one fails. A field may need a gauge kept elsewhere, or a value nobody explained. Delete what is never used. Print lot numbers on peel-off stickers from the carton, and hang the sheet on a clip beside the fixture.
- 3
Teach the record on day one
Give every newcomer a practice sheet and a dummy sub-assembly in the first session. They fill it while the trainer watches and corrects. Explain who reads it later, and why that reader matters to them.
- 4
Check entries at the station, at fixed times
The supervisor reads open sheets mid-morning, after lunch and before the shift closes, initials the check and corrects gaps on the spot. Fixed checkpoints beat a hunt at month end.
- 5
Separate who builds from who signs the test
A second named person reads the test value and enters it. Self-checked records are where optimism creeps in.
- 6
Review five sheets a week, aloud
Pick them at random and go through them with the team for ten minutes. Praise clean records by name. Change the form when the same column fails again.
What owners ask next
- Should we move to tablets or barcode scanners? Only once paper entries are reliable. A scanner records a wrong lot as faithfully as a ballpoint does.
- How quickly do errors drop? Fixed-time checks change behaviour within weeks. Long-serving staff take longer, so begin with the stations that feed the most queries.
- Is a dedicated records clerk worth hiring? Usually not. A clerk filling sheets away from the bench recreates the original fault. Whoever does the work makes the entry, and a colleague verifies it.
- What happens when an error is found? Correct it with a dated, initialled note and never overwrite. Then find out what made the slip easy, and change that.
- Does this apply to temporary hands brought in for a peak? Most of all. Give them the same practice sheet and keep them off columns that need judgement until a supervisor signs them off.
Where to look further
The medical devices and equipment hub sets this problem beside the other people patterns in device units. Our piece on traceability as a daily habit covers the wider routine. For the first-day lesson, see induction on a disposables line. The onboarding process service helps build one in your own plant. If your build sheets need an outside look, describe the problem to us.
