Cutting 38% off a remanufacturing line's cycle time

The situation

The line processed remanufactured diesel aftertreatment units — colloquially "one boxes" — through pressure testing, catalyst repair, bake, dynamometer test, scrapping, capping, and shipping. Each station had a quota. The quotas were largely derived by dividing available shift hours by an assumed unit rate, which meant nobody knew how much of the standard was work and how much was slack.

My brief was to establish what the work actually took, then find what could be removed.

Method

I timed each station directly and observed the operators working normally, then compared measured time against quota allocation per box — active shift time divided by current quota. That comparison is where the story lives: it separates stations where the standard is genuinely tight from stations where the standard has quietly absorbed a problem nobody escalated.

Every observed loss was written up as a discrete problem with a proposed solution and an estimated time saving, expressed as a percentage of average station time. Operator input was treated as primary evidence — two operators had independently complained about the same hoist, which turned out to be worth twenty-two seconds a unit and a meaningful ergonomic exposure.

What the study found

Pressure test

Measured 34:01 per box against a 1:00:00 quota allocation. The station was carrying an awkward single-hook hoist that raised the unit at a poor angle, a pallet-jack shuffle between sanding and hoisting, threaded bung fixtures that consumed a large share of station time, a leaking o-ring fixture requiring repeated reseating, and paper documentation at every step.

The fixes: adopt the two-hook hoist design already proven elsewhere in the plant, bring the unit directly onto the work table so all modifications happen in one place, engineer threadless bung fixtures, replace the seal fixture with a larger-diameter o-ring or a single-piece rubber snap seal, and move documentation to a portable electronic system.

Metric	Before	After
Time per box	34:01	27:17
Quota	9	18
Throughput change	—	+66%

Bake oven

The largest single opportunity, and the one nobody had looked at because the oven was "just how long it takes." Four changes: a purpose-designed rack raising oven capacity from six units to ten, enough pallets to unload an oven in one pass rather than in stages, running both ovens on alternating schedules to produce continuous flow instead of batch starvation, and running the exhaust hood during cool-down so both ovens could operate. A non-supervisor was assigned the transfer to pressure test — no time saved, but a materially lower labour cost on the move.

Metric	Before	After
Time per box	2:08:01	30:00
Quota	6	20
Throughput change	—	+108%

Dynamometer

Measured 1:56:36 against a 1:07:50 allocation. Losses came from attaching and detaching the probe box on every unit rather than mounting it to the dyno cart, threaded probes, unnecessary box handling between table, pallet, and cart, and data entry performed entirely after the test rather than during it.

The dominant item was cool-down: the procedure held a constant 20% throttle idle. Changing to a graduated 30% → 15% decrease was estimated to save 32:47 per unit and cut the overhead burn that goes with it. That single change was worth more than every other dyno fix combined — a good reminder that the biggest opportunity is usually inside a procedure nobody has revisited since commissioning.

Metric	Before	After
Time per box	1:56:36	1:12:32
Quota	8	10
Throughput change	—	+22.2%

Flow routing

The most valuable finding wasn't a station fix at all. Units were being fully processed before anyone verified the catalyst was functional — but functionality became apparent within the first ten minutes of scrapping. Reordering the flow to scrap first meant non-functional units could be routed to welding and on to the second plant immediately, rather than consuming a full process cycle before being discovered.

That change removed overhead cost, removed travel cost, and ensured the downstream plant only received units worth processing. It cost nothing to implement.

Result

Measure	Value
Quota allocation per unit	6:05:28
Measured time per unit	6:28:15
Projected time with fixes applied	3:59:26
Difference	2:28:49 removed per unit

Note the first two rows: the operation was running over its own quota allocation before the study. The standard wasn't tight, it was fictional — and the gap was absorbed invisibly by the schedule.

What transfers

Two stations produced no recommendations. Catalyst repair and shipping were examined and found sound, and the report said so. A study that finds an opportunity everywhere it looks should be treated with suspicion; the value of the exercise depends on being willing to report that something is already working.