Three sets, three different ways to fail
A doctor kit, a tool set and a play kitchen look like the same product category on a purchase order. They are not.
A doctor kit fails at the case latch — it pops open in a carton drop and forty small instruments end up loose in the master carton. A tool set fails at the screw thread — the pretend bolt strips on the third assembly and the set is returned as "broken". A play kitchen fails at the housing boss — the seventh screw cracks through 2.2 mm of ABS and the whole run comes back.
None of these show up in a photo sample. All of them are decided in the tool, the material spec and the test plan.
Where each set actually breaks
| Set type | First failure point | Typical cause | Fix at design stage |
|---|---|---|---|
| Doctor kit | Case latch / hinge | Undercut too shallow, PP too stiff | 1.2–1.8 mm latch engagement, PP copolymer |
| Tool set | Threaded bolt and nut | Pitch too coarse for the wall | M8×1.25 profile, 2.4 mm boss wall |
| Play kitchen | Screw boss cracking | Boss wall under 1.5 mm, no radius | 2.0–2.5 mm boss, 0.5 mm fillet at root |
| Any set | Small parts escape | Gate vestige or flash on a 3+ part | Vestige under 0.2 mm, no sharp edge |
The case: hinge, latch and the drop test
The doctor kit case is a one-piece or two-piece moulding, and the hinge decision comes first.
A living hinge needs polypropylene homopolymer with a melt flow rate around 20–35 g/10 min, a web of 0.25–0.40 mm, and a radius on both sides. It survives a million flex cycles in the lab but embrittles with age and UV — acceptable for a playroom toy.
A pinned hinge — PP knuckles plus a stainless or acetal pin — costs one more assembly step and lasts longer. It also lets you use copolymer, which is what you want if the case has to survive a 1.2 m drop onto concrete at −10 °C.
The latch is the part that actually decides the return rate. Two numbers matter:
- Engagement depth 1.2–1.8 mm. Below 1.0 mm the latch releases under its own moulding flash; above 2.0 mm a four-year-old cannot open it and the parent complains.
- Opening force 15–30 N measured at the tab. This is a design target, not a standard, but it is the range buyers ask for once they have had one bad shipment.
Then run the drop. ISTA 1A, ten drops from 760 mm, one on each face and edge, case closed and empty. If the latch opens and instruments spill, the case fails regardless of what the lab report says about the plastic.
Instrument moulding: small parts at high cavitation
A doctor kit is a high-cavity, small-part job: a 32-cavity tool on a 12–18 second cycle turns out a stethoscope head, syringe barrel, plunger, thermometer body, tweezers and tray. Three process details decide whether it ships:
Wall thickness. Keep instruments between 1.2 mm and 2.5 mm. A 0.9 mm syringe barrel short-shots at the far end of a 32-cavity tool on a cold morning.
Gate vestige. Under 0.2 mm, and never on a surface a child grips. A 0.4 mm vestige on a tweezer tip is a sharp point under EN 71-1 and a fail.
Snap-fit strain. The plunger cap and the otoscope head both snap on. Keep ejection strain under 4 % for PP and under 3 % for ABS, or the parts craze at the snap during the first week.
Two-shot moulding is worth it in exactly one place: the stethoscope head, where a TPE ring bonded onto an ABS body gives the soft rim that makes the toy read as "real". Everywhere else, a single shot plus assembly is cheaper and easier to validate.
Play kitchen: the housing is a structural part
A play kitchen is not a moulding with graphics on it. It is furniture assembled by the consumer with eight to twelve screws, often without a second adult present.
The ABS housing should be 2.2–2.8 mm, with screw bosses at 2.0–2.5 mm wall and a 0.5 mm fillet at the root. Pilot hole diameter is roughly 0.8× the screw's major diameter for a self-tapping screw in ABS; too small and the boss splits on the second assembly, too large and the joint rattles.
Watch three things in the sample:
- Boss cracks. Assemble and disassemble the sample three times. Any crack line at the boss root means the wall is too thin or the pilot is too small.
- Oven door hinge. If it is a pinned hinge, check the knuckle wall; if it is a living hinge in PP, check that the door does not droop after 500 cycles.
- Knob torque. A knob that spins freely at 0.1 N·m feels broken. Target 0.15–0.3 N·m of detent torque so it clicks.
Standards you cannot skip
| Standard | Scope | What it catches in these sets |
|---|---|---|
| EN 71-1 | Mechanical and physical properties | Sharp edges, small parts, latch and hinge entrapment |
| EN 71-2 | Flammability | Fabric accessories, doctor coat, chef hat |
| EN 71-3 | Migration of 19 elements | Painted instrument surfaces, printed graphics |
| ASTM F963-23 | US toy safety | Drop, torque, tension, impact tests |
| CPSIA | US chemical limits | 100 ppm lead in substrate, 90 ppm in paint, 0.1 % phthalates |
| ISO 8124-1/-2/-3 | International equivalent | Used for markets outside EU and US |
| EN 62115 / IEC 62115 | Electric toys | Any otoscope or X-ray panel with a lamp or battery |
Two details get missed on the factory floor. EN 71-3 applies to the substrate, not just the paint — a coloured PP instrument is a coated article to most labs. And if the set has a button or coin cell, the compartment must be screw-secured or need a tool; in the US this is enforced under the coin battery rules, and a snap-fit cover fails regardless of age grading.
Age grading drives the rest. A 3+ set may contain small parts, but then the warning must appear on the primary pack and, for the EU, the toy itself where practicable. An under-36-month set may not contain any part that fits the small parts cylinder — 31.7 mm diameter by 57.1 mm depth — and that includes the pretend pills in a doctor kit.
Decoration: rub resistance is the real spec
Three decoration routes, in ascending order of cost and durability:
- Sticker or label. Cheapest, fails first; on a curved instrument it lifts within a month.
- Pad printing. Fine for flat and gently curved surfaces. Specify 500 g at 100 cycles with no substrate showing, and cross-hatch adhesion of 4B.
- In-mould labelling. For flat panels on a play kitchen or the top of a case, IML survives everything and removes a whole class of complaint. It also needs 0.8–1.0 mm of extra wall thickness at the label pocket.
Whichever route, test with artificial saliva as well as water: that is what the migration test simulates.
Packing and the carton drop
Sets are heavy and have awkward shapes, which makes the carton the weakest link. A 3 kg doctor kit in a 40 × 30 × 12 cm carton, ten units per master, will see roughly 1.2 m of drop in a normal parcel network.
Ask for ISTA 1A on the single pack and ISTA 3A on the master, then open the master and count loose parts. Polybags over 200 mm across need a suffocation warning. Specify 3–5 % spare instruments per carton — cheaper than air-freighting one missing syringe.
RFQ checklist
- Age grading and the market (EU, US, or both) — this fixes the standard set.
- Part count and BOM with material and colour per part.
- Case or housing: material grade, wall thickness, hinge type, latch engagement.
- Mould ownership, cavitation, and the tool steel grade.
- Test report scope: which clauses of EN 71-1, ASTM F963-23, and CPSIA.
- Battery type and compartment closure method, if any.
- Decoration method with rub and adhesion targets.
- Carton drop standard, master pack count, spare parts ratio.
- Sample approval stages: pre-production, golden sample, and the retention period.
A pretend play set is a hundred small decisions that all surface as one number: the return rate. The sets that come back are rarely the ones with the wrong colour.
Related product categories
The materials and part types discussed above are grouped in our sourcing directory:
