When the Spare Part No Longer Exists: Reverse-Engineering a Medical Machine Back to Life

🛠️ A service for medical institutions. When an otherwise-healthy machine is sidelined by one small part that is no longer manufactured, I can reverse-engineer that part, generate parametric CAD code and an STL, and 3D-print a working replacement. This post walks through the whole process using one real component.

A yellow 3D-printed filter cap beside the original worn black part — Left: the 3D-printed replacement filter cap. Right: the original, worn and cracked at the hub — the part that had taken a machine out of service.

The problem: an obsolete part, not an obsolete machine

A great deal of medical equipment outlives the supply chain that supports it. The instrument on the bench is mechanically sound, calibrated, and trusted — and then a single moulded plastic component fatigues, cracks, or wears through, and the whole machine stops. When the manufacturer has discontinued the model, or simply no longer stocks that part, the institution is left with an expensive, functional machine and no legitimate way to bring it back.

That is the gap I work in. The part shown above is a small filter cap — a moulded retainer that seats a filter into its housing with a bayonet twist. The original (right, in black) had worn and split at the hub. It is trivial in isolation, and impossible to buy. Without it, the machine it belonged to was finished.

It did not need to be.

The service, in one sentence

Give me a similar part — worn, cracked, or intact — and I will measure it, reverse-engineer it into parametric CAD, and produce a 3D-printable replacement that fits. Every step below is reproducible, the geometry is captured as code rather than a one-off sketch, and the output is a physical part in your hand. Here is exactly how that happened for this filter cap.

Step 1 — Measure the original and capture its geometry

Reverse engineering begins with a careful physical survey. Using digital calipers, I take every dimension that defines how the part fits and functions: the outer disk, the depth and bore of the central hub, the size and spacing of the retaining tabs, and the height of the two locating ribs on top. Nothing is guessed — the disk thickness here, for example, was measured across several points and settled between 1.53 and 1.56 mm.

For this cap the defining measurements were:

Feature	Dimension
Outer disk	Ø 42.60 mm × 1.55 mm thick
Central hub	Ø 18.23 mm outer, 12.37 mm bore, 3.72 mm skirt depth
Bayonet retainer tabs	3 tabs at 120°, 7.22 mm wide, 25.43 mm across
Locating ribs ("wings")	2 opposed ribs, 14 mm long × 4.5 mm tall

Step 2 — Turn the measurements into parametric code

Rather than draw the part once, I express it as parametric CAD code — a small script in which every measured value is a named variable. This is the reusable asset: change a number and the model rebuilds, so a slightly different size, a thicker wall, or a fourth tab is a one-line edit rather than a fresh design. The complete code for this part lives alongside this post (filter_cap.py), written with the open-source build123d library.

Parametric CAD code defining the filter cap dimensions — The parametric definition: each caliper measurement becomes a named parameter, and the solid is assembled from a disk, a hub, three bayonet tabs, and two locating ribs.

Step 3 — Build it in FreeCAD and export an STL

Running the code produces a clean, watertight solid model (a STEP file). I import that into FreeCAD — the free, open-source CAD package — to inspect the geometry from every angle and confirm it matches the original before anything is committed to plastic. The top view shows the two locating ribs on the disk; the underside shows the hub, the central bore, and the three equally-spaced bayonet lugs that lock the cap into its housing.

FreeCAD top view of the filter cap model showing the two locating ribs — Top view in FreeCAD — the disk with its two opposed locating ribs.

FreeCAD underside view showing the hub, central bore, and three bayonet retainer tabs — Underside — the hub with its central bore and three bayonet retainer tabs at 120°, the features that make the cap lock and hold.

Once the model is verified, I export it as an STL — the triangulated mesh format every 3D printer understands.

Step 4 — Slice and print

The STL goes into a slicer (here, UltiMaker Cura), which converts the model into the layer-by-layer instructions the printer follows. For this cap I printed in PETG — a tough, chemically-resistant material well suited to a mechanical retainer — through a 0.8 mm nozzle at a 0.32 mm layer height. Material, wall thickness, and infill are chosen to match how the part is loaded in service.

The filter cap STL loaded in UltiMaker Cura, ready to slice for 3D printing — The model staged in the slicer — PETG, 0.8 mm nozzle, 20% infill — moments before printing on a Creality CR-10 Max.

A short while later, the replacement is in hand (the yellow part at the top of this post): a faithful, functional copy of a component that could no longer be bought — and a machine returned to service.

Why this matters for medical institutions

The economics are stark. A machine worth tens of thousands can be retired by a component worth pennies to mould. Reverse engineering turns that dead end back into a repair:

Downtime becomes days, not never. A discontinued part stops being a reason to decommission working equipment.
The design is preserved as code. Once a part is captured parametrically, it can be reprinted on demand, adjusted, or scaled — indefinitely.
It is sustainable. Keeping a sound machine alive avoids the cost and waste of replacing the whole unit for one failed part.

⚕️ A note on scope and responsibility. This service is for mechanical and structural components — housings, caps, clips, retainers, brackets, and the like. Material selection and any validation are matched to the part's specific role, and suitability for a given machine is always confirmed with the institution's biomedical engineering team. Parts in patient-contacting, sterile, or safety-critical roles carry additional regulatory obligations that are assessed case by case.

If a single part is holding your equipment hostage

If your institution has a machine sitting idle because one small component is no longer available, that is exactly the problem I solve. Send me the worn or broken part — or even a photograph and a few measurements to start — and I will tell you whether it can be reverse-engineered and reprinted.

A discontinued part should not be the end of a good machine.

— Neal