The $3,200 ABS Enclosure Order That Taught Me Everything I Didn't Know About Plastics

The Order That Nearly Got Away

It was a Tuesday in February 2024 when I almost threw away a relationship that's now worth tens of thousands of dollars a year. The order seemed small: 50 custom ABS sensor housings for a three-person startup. $3,200 total. I almost said no because of the volume. I'm very glad I didn't.

I've been handling custom molding orders for seven years, mostly SMC composites. SMC is what we know best: sheet molding compound, glass-reinforced polyester, compression molded, tough as nails. We also run polyurethane, nylon, PVC, and a range of thermoplastics. But my expertise sits squarely in SMC. That order tested every bit of that boundary.

The Kickoff Call

The customer sent drawings for a sensor enclosure—180 × 120 × 45 mm, with a recess for a PVC cable gland. The body was specified as ABS, in their signature blue. On the call, the founder, a mechanical engineer, asked: "You work with SMC, right? Can you make this body in ABS?" In my head I heard "plastic" and said "sure." That one word cost us a lot later.

The enclosure was meant to be waterproof. Once their PCB was mounted, they planned to fill the cavity with polyurethane encapsulation. I knew that phrase. My colleague pots electronics with polyurethane compounds every other week. I nodded and moved on.

The First Sign I Should Have Stopped

The numbers supported the substitution. Our SMC grade had higher flexural strength than the ABS they'd originally specified, better heat deflection, and a lower price per part. My gut, though, twitched at the surface finish spec. They wanted a satin texture that matched their existing ABS product. I ignored the twitch. From the outside, ABS and SMC both read as "black plastic parts." The reality is that SMC's glass fiber content and mold release characteristics produce a different surface character—often stronger, but not automatically the same finish as ABS. We quoted a "Class A" SMC finish. What we shipped was closer to Class B.

The color gave us the first visible warning. They specified Pantone 286C, a deep corporate blue. We matched a pigment chip and got their approval, but the first molded sample landed on the wrong side of the Delta E line. Pantone's own color-matching guidance says a Delta E between 2 and 4 is noticeable to trained observers; anything above 4 is visible to most people. The founder caught it in the first thirty seconds. I told him we'd correct the pigment. We did not correct it in time.

Three Failures, One Root Cause

When the customer ran their internal water ingress test, the results came back fast and bad:

• Polyurethane encapsulation lifted. The potting compound separated from the SMC surface around the PCB standoffs, creating a tiny channel for water. A "looks about right" judgment is not a compatibility test, and we paid for learning that.
• SMC and PVC didn't behave as one assembly. The PVC cable gland is thermoplastic; the SMC housing is thermoset. Their coefficients of thermal expansion are different—you can measure them per ASTM D696—and the thread fit that worked at room temperature didn't survive the customer's thermal cycling.
• Production color drifted. The pigment settled differently in the SMC matrix than in our approved chip. The Delta E went from "trained observer" to "everyone noticed."

All three failures trace back to the same decision: I approved a material substitution without validating it against the end-use requirements. I told the customer what I thought was true, not what I'd verified.

I'm not a materials engineer, so I can't speak to the polymer chemistry in detail. What I can tell you from a production perspective is that a favorable data sheet and a gut feeling are not the same as a compatibility test with the customer's actual potting compound.

The Quiet Phone Call

The customer quarantined all 48 remaining parts. I went to the SMC login portal, changed the order status to "On hold," uploaded the failure photos, and picked up the phone.

That call was uncomfortable. I explained what happened without hiding behind excuses. We refunded the order in full and offered to rework the housings with a new surface preparation protocol.

Even after we sent the refund, I kept second-guessing. What if they just took the money and walked? The three days until their reply were the longest I'd had in a while. Then the founder wrote back: they'd had suppliers who blamed the customer when things went wrong. Our willingness to document the mistake and change the process was, in his words, "the reason we didn't leave."

Forty-eight parts quarantined. Rework cost: $2,900. Credibility: damaged but repairable.

What I Check Now

We've run the corrected process on 47 jobs since, with no repeat failures. The checklist isn't complicated:

• Material substitution is a yes/no question. If the drawing says ABS and your core capability is SMC, say "let me verify it's suitable," not "sure." SMC can replace ABS in many housings, but not when polyurethane encapsulation adhesion is critical without surface prep.
• Ask for the exact polyurethane compound name. Not "potting." A specific grade. Request their processing and adhesion data, then run a small coupon test before production.
• Thermal expansion matters when you join SMC with PVC. Different classes of plastics expand at different rates. The standard test method is ASTM D696; ask your material supplier for those numbers before you promise a thread fit.
• HDPE is not "just another plastic" either. The customer's other product used HDPE for an outdoor housing, and at one point they asked me, "how to clean HDPE plastic?" I gave a generic answer—mild soap and warm water. That's safe for most plastics, including SMC. But HDPE specifically tolerates harsher cleaners, including isopropyl alcohol. My generic answer showed that I hadn't checked. Don't answer from intuition; answer from the material data.

Why Small Orders Deserve Big Attention

Here's the part that stuck with me. This was a 50-unit order from three engineers and a dog. The same week, we'd received a 5,000-unit inquiry, and honestly, the smaller one drifted to the bottom of my priority list. That was a mistake.

Small doesn't mean unimportant. It means potential. The startup now orders 300 units per batch, and referrals from their founders brought us two other small hardware companies. If we'd dismissed them because of MOQ, or if I'd covered up the mistake instead of documenting it, we wouldn't have either account.

I've personally made seven significant mistakes in my career, totaling roughly $12,000 in wasted budget. Most came from assuming one plastic will behave like another. The best thing I did was write them down. Our team's checklist now sits next to the material supplier contact list, and every new engineer gets walked through the stories behind each line.

Respect the small order. Verify your material assumptions. And when you're wrong, say so plainly—then fix the process so nobody repeats it.