The parts were dimensionally correct. They passed inspection. Two production batches went out the door before the problem surfaced.
When it did, it was at the pin mounting points, during assembly, at the customer’s own bench.
That detail matters more than the crack itself. A part that fails a dimensional check failed on the molding floor, and the molder owns it. A part that measures correct and cracks in assembly passed your control plan and still does not work. The fault is in the design decision, and the design decision that most often hides there is the material.
First, the Geometry Fix — and What It Proved
This was a black outer cover for a phone grip — the core outer structural part of the brand’s grip range. A Canadian brand, sold through Amazon, its own site and retail across North America.
The early work looked like a normal tooling project. Mold built, samples out, and the client came back with two specific dimensional changes after their own assembly trial:
- Add 0.2 mm of wall thickness on both pin installation sides, to strengthen the anti-cracking performance
- Raise the overall height by 1 mm and shift the assembly holes up by 1 mm, to match their assembly fixtures
We adjusted the mold accordingly and shipped 15 finished sample sets — pinned, with the green adhesive stickers fitted, because the client’s test needed to be a real assembly test, not a plastic-part test.
That second requirement is worth pausing on. Moving assembly holes by a millimetre is not a tolerance tweak. It is the client telling you their fixture is the fixed reference and the part has to come to it. When a supplier treats a fixture mismatch as “the part is in tolerance, the fixture is wrong,” the project stalls. The part has to meet the fixture.
Then two production batches of PPO ran and shipped. And the cracking started.
Why a Press-Fit Pin Is a Stress Test
A pin pressed into a plastic boss is not a joint. It is a controlled interference, and the boss pays for it in hoop stress.
Three things stack up at that feature:
Interference stress. The pin is larger than the hole. Pressing it in stretches the boss wall permanently, and the wall carries that load for the life of the part.
Cycle load. This is a reusable phone grip — the product is bought precisely because it attaches, detaches and re-attaches. Every cycle flexes the boss. The stress is low, but it is repeated, and repeated stress finds notches.
Notch concentration. The root of a boss, where the wall meets the base, is a geometric stress riser. Sharp radii there multiply the local stress well above the nominal value. This is where the crack starts — not at the pin, at the root below it.
Add those three together and the demand on the material is not “be stiff.” It is “strain without cracking, repeatedly.” That is a toughness requirement, and toughness is exactly where a general-purpose PPO blend and polycarbonate diverge.
PPO blends — modified PPE, usually with polystyrene — are dimensionally excellent. Low moisture uptake, stable across temperature, good flow, and a lower density than PC, which makes the part cheaper per unit. For a cosmetic outer cover, that profile reads as the correct choice on a datasheet. It is stiff, stable and economical.
What it is not is forgiving at a notch under repeated strain. Under impact and press-fit stress, the failure is brittle and sudden: no whitening, no warning, just a crack.
PC trades some of that stiffness and flow for strain capability. It deforms further before it breaks, and it recovers. On a feature that is loaded by a press-fit pin and cycled by the end user, that is the property that matters.
A material that measures perfect and cracks in assembly was selected for stiffness when the joint needed toughness. That is the whole diagnostic.
Two Fixes, One Decision
We put two options in front of the client rather than one recommendation:
- Local rib redesign. Add structure around the pin area to reduce local strain on the boss wall, and keep PPO.
- Switch to high-toughness PC. Change the material, keep the tool geometry largely as-is.
Both are legitimate. They are not equally expensive, and the expense is not in the same place.
| Rib redesign, keep PPO | Switch to PC | |
|---|---|---|
| Tooling | Cavity rework at the boss — real steel work | Minimal geometry change, but re-qualification |
| Material cost | Unchanged, and PPO stays cheaper per part | Higher resin cost per part |
| Schedule | Rework, re-sample, re-approve | Sample and approve on the existing tool |
| Dimensional risk | Low — material unchanged | Real — shrinkage and flow both shift |
| Compliance | Existing reports still valid | New RoHS & REACH reports required |
| Residual risk | Strain reduced, not eliminated at a notch | Toughness solves the failure mode directly |
We prioritised PC sample testing, because it attacks the failure mode instead of reducing it, and because it does not require pulling the tool apart. That was a schedule decision as much as an engineering one — the client was mid-launch, and every week of rework was a week of not selling.
What Actually Changes When You Change the Material
This is where material substitution projects go wrong. The resin swap is the easy part. What changes around it is the project.
Shrinkage moves the hole. A press-fit is a designed interference. Change the material and the same cavity produces a different hole diameter — which means a different interference, which means a different stress in the boss you were trying to protect. The very feature that cracked is the one most sensitive to the change. Re-measure it, do not assume it.
Viscosity changes the pressure. PC needs more injection pressure than a PPO blend. That has consequences upstream: gate size, weld line position, and whether a thin section fills at all. A tool validated on one resin is not automatically validated on the other.
Melt temperature changes the surface. Higher barrel temperatures shift cooling behaviour. Cycle time moves, and on a textured cosmetic surface — which this is, because it is the visible outer cover — the texture can read differently. Cosmetic approval has to be re-run, not inherited.
Toughness is not free. PC buys strain capability with lower stiffness and higher notch sensitivity of its own, which means radii at the boss root still matter. A material change does not license a bad corner.
Adhesive contact deserves a second look. PC is susceptible to environmental stress cracking when it sits under load in contact with certain solvents and adhesives. On a part where a sticker goes on and a pin goes in, that combination is worth testing rather than assuming — stressed PC plus the wrong adhesive is a slow crack that passes every mechanical test on day one.
The Compliance Part Nobody Plans For
There is a second deliverable in a material switch that is not plastic at all.
RoHS and REACH declarations are issued per material, not per part. Switch resin and the old report no longer describes what you are shipping. For a brand importing into North America, that is not paperwork — it is the thing that clears customs and satisfies retail compliance.
We had accredited third-party labs test both materials on this project, PPO and PC, and issued formal RoHS and REACH reports for each. Doing both was deliberate: the client kept a valid declaration for what had already shipped, and gained one for what would ship next, without a gap in coverage during the transition.
Most buyers discover this requirement after the material decision, when the shipment is already booked. It belongs in the comparison table, next to the tooling cost.
The Sequence We Ran
- Initial mold built for the black covers, on PPO as specified.
- Targeted cavity modification, paid as a revision, from the client’s assembly test feedback.
- Samples revised to the two dimensional requirements — 0.2 mm wall added at both pin sides, height +1 mm, assembly holes shifted +1 mm — and 15 sets shipped pinned, with adhesive stickers fitted, for a genuine assembly trial.
- Two PPO production batches molded and delivered.
- Cracking reported at the pin mounting points during the client’s assembly of the first bulk shipment.
- Two solutions proposed — local rib redesign or high-toughness PC — with PC prioritised for testing.
- First PC trial batch produced for toughness and assembly verification.
- Additional PC sample batches produced on request, shipped together with MagSafe and regular grip units so the client could run one comparative test instead of several.
- RoHS and REACH reports issued by accredited third-party labs for both materials.
- Volume production on the corrected specification.
What the Whole Scope Cost
The programme ran to USD 56,388, and it covers the full shape of the problem rather than a single part price: initial tooling, multiple paid mold revisions, the PPO production batches already run, the PC trial batches, the mixed sample shipments and third-party compliance testing on both materials.
That is the honest shape of a material problem found late. The resin was never the expensive line item. The rework, the batches already produced, the re-testing and the compliance duplication were.
Which is the argument for testing the joint before volume, not after: the cheapest place to discover that a boss needs toughness is on fifteen sample sets, not on two production batches.
What Came Out of It
- Cracking at the pin mounting points resolved by the PC switch, verified through assembly testing
- Multiple rounds of dimensional refinement completed against the client’s actual fixture, not a nominal drawing
- Two production batches delivered on PPO while the fix was developed, so the launch was not held
- PC trials run in parallel rather than sequentially, protecting the development schedule
- RoHS and REACH declarations issued for both materials by accredited third-party labs
- Volume production established on the corrected specification, with repeat orders following
What the Client Said
“You adjusted mold dimensions rapidly after receiving our structural feedback, the revised samples fit our assembly fixtures perfectly.”
“After we found cracking issues on PPO mass goods, you immediately launched PC material trial without delaying our product development schedule.”
“Multiple batches of test samples were delivered on schedule, mixed MagSafe and regular grips were packed together to save our testing time.”
“It’s very convenient that you can supply formal environmental test reports for all plastic materials to meet our North America market entry standards.”
Four Questions Before You Change a Material
- What is the joint actually being asked to do? Stiffness and toughness are different purchases. Press-fits, snaps, clips and hinges are buying toughness.
- Which feature is most sensitive to the change? It is almost always the one that already failed, or the tightest interference in the part. Re-measure that one first.
- Does the tool survive the switch? Ask about gate size, pressure and weld lines — not just “will it fit in the same mold.”
- Who is paying for the second compliance report? If the answer is “nobody thought about it,” budget for it now.
Send us the part that failed
The most useful thing you can send is not a drawing. It is the broken part, and where it broke.
What to send
- The failed part, with the crack location marked
- The current material, and the requirement it was chosen for
- The assembly it goes into: what presses into it, how often, under what load
- Target market, so compliance scope is clear from the start
- Annual volume and the schedule you are working to
What comes back
- A read on the failure mode: is this geometry, material, or the joint between them
- The options, priced separately, including tooling and re-qualification
- A material recommendation tied to the loading case, not to a datasheet ranking
- The compliance implication of each option, stated up front