Real APTE spec used in 2020 to design a unified test bench for an industrial workshop assembling medical equipment. Names and proprietary details anonymized. The bench was built and is still in production 6 years later.
A French distributor of imported medical equipment assembled and tested each unit on-site before shipping to end users. Each test required two separate benches — one for pneumatic + water circuits, one for electrical — physically distant in the workshop. This caused unnecessary movement between stations and missed procedure steps. Total test time per unit : 3 h 50.
The mission given to me (intern, BTS year 2) : propose and build a redesign that reduces test time and procedure errors. Budget cap : ~500 €.
| Question | Answer |
|---|---|
| Who does it serve ? | The workshop technicians and the company (faster + more reliable testing). |
| What does it act on ? | The assembled equipment units, before they ship. |
| In what purpose ? | Validate every unit's pneumatic, water, and electrical function in one continuous procedure, without station switching. |
The bench interacts with :
- Technicians (use it, follow the procedure)
- Compressed air (workshop compressor, ~6 bar)
- Water (autonomous bottle, 2 bottles for purge cycle)
- Mains electricity (230 V workshop)
- Equipment under test (the assembled unit, all configurations)
- Test instrument (electrical leakage tester, Fluke-class)
- Workstation PC (test report archive)
- Workshop floor (~30 m², bench footprint constraint)
- Production schedule (any install must not stop production)
FP1 : Let the technician validate every unit's full operation in a single continuous procedure.
FC1 : Adapt to the 4 product families in scope (2 chair groups × 2 unit types). FC2 : Use the existing energy sources — 230 V mains, workshop compressor, water from bottle. FC3 : Respect the budget : ~500 €. FC4 : Integrate the test procedure for each product variant (no manual cross-referencing). FC5 : Include a second water bottle for circuit purge / cleaning cycle. FC6 : Integrate an assembly-tracking sheet (per-unit BOM with serial numbers).
| Function | Criterion | Level / value | Flex |
|---|---|---|---|
| FP1 — electrical | Functional test, leakage test, insulation test | Pass per Fluke procedure | F0 |
| FP1 — pneumatic | All instruments respond, no air/water leak | Visual + auditory + pressure check | F0 |
| FP1 — duration | Full cycle one unit | ≤ 1 h | F0 |
| FC1 | Unit types supported | 4 families (group 500 : 533/532/542 ; group 300 : 333/332) | F0 |
| FC2 — electrical | Mains input | 230 V | F0 |
| FC2 — pneumatic | Air source | Workshop compressor | F0 |
| FC2 — water | Source | Autonomous bottle | F0 |
| FC3 | Total cost | ~500 € (components + tubing) | F0 |
| FC4 | Procedure support | One A4 fiche per variant | F0 |
| FC5 | Purge cycle | 2nd bottle with cleaning tablet | F0 |
| FC6 | Tracking sheet | Computer file + per-unit fiche | F0 |
- The bench does not test mechanical strength (the chair frame is validated upstream by the OEM).
- The bench does not replace the post-install QA on customer site.
- The bench does not automate report generation (manual print of first page only, 2020 limitation — see lessons below).
The bench completes one full test cycle (pneumatic + electrical + Fluke leakage check + procedure validated step by step) for any of the 4 supported unit families, in under 1 h, without the technician switching between two physical stations.
| Role | Name | Date |
|---|---|---|
| Asked by | Product manager (workshop redesign initiative) | 2020 |
| Built by | Me (intern, BTS year 2) | 2020 |
| Validated by | The two senior technicians who use it daily | 2020 |
- One bench (built from 2 recovered bathroom tables, drilled and integrated).
- Lovato 3-position selector switch at the center (0 = off / 1 = pneumatic + electrical simultaneously / 2 = electrical leakage test).
- Step-down transformer 230 V → 24 V integrated.
- Autonomous water reservoir with 2 bottles (operation + purge).
- Color-coded tubes following European convention (blue = air-spray, green = water, transparent = air-drive).
- Custom enclosure with grommets, brass T-fittings, electrical interface on one face, pneumatic interface on the other.
Total build : ~500 € of components + ~2 weeks of workshop time during a partial production pause.
Outcome : the bench has been in continuous production use for 6 years (as of 2026).
This is where the toolkit becomes honest. Three gaps I see now that I missed in 2020 :
-
No quantified gain measurement. The 2020 spec aimed for "reduce test time" but never recorded the actual before / after. I should have instrumented the bench from day 1 (technician timestamps each step, average over N units, publish the delta). Without numbers, the "success" is anecdotal.
-
Data was inert. The Fluke generated a PDF report per unit, saved to a workshop PC folder ; only the first page was printed and archived. In 2026 terms, that report should have been parsed into a queryable store (serial, technician, leakage current, date) instead of sitting as dead PDFs. The system did the test ; it didn't learn from it. Today I'd pipe the Fluke output into SQLite or a structured store, with a basic dashboard. This is the gap between "a system that works" and "a system that gets better over time".
-
No version-controlled spec. The APTE doc lived in a Word file that probably got lost. The procedure fiche existed on paper. Today the whole spec + procedure + bench schematic + assembly tracker would live in a Git repo, evolve via commits, be reviewable.
This apte-toolkit and the companion banc-test-2020 case study are partly an effort to fix gap #3 retroactively : make the methodology version-controlled, reusable, and visible.
Anonymized example. License : MIT. Method : APTE. Author : Nevil Bofumbo, 2026.