Last March, a Network Tester Taught Me a Crimp Lesson I Thought I Knew
Last March, I spent a full morning staring at a network tester that kept showing a red failure on a sample of 50 cable assemblies. We had just received a batch of pre-crimped leads for a driver-assistance module. The print called for Molex automotive connectors—MX150 terminals, 22 AWG wire, sealed. Three of the first 50 samples had intermittent opens.
In a quality lab, a network tester is not supposed to be dramatic. It either passes or it doesn't. But this one was acting like a detective. It said a wire was making contact at rest, then losing it when we flexed the harness slightly. On a connector that had never caused us problems before.
I'm a quality compliance manager at a contract electronics manufacturer. I review every incoming connector- and cable-related deliverable before it goes to production—roughly 200 part numbers a year. In Q1 2024, I rejected 5% of first articles because of crimp height or terminal position issues. I've learned to trust test equipment, but not blindly.
And I'll be honest: my first reaction was to blame the connector. The part was a Molex automotive connector, and the certificate traced it to the Molex Lincoln NE location. Molex, right? Should be dependable. But the tester said otherwise.
The Two-Week Decision Before the Failure
Here's the part I'm not proud of. When we first designed this module, I went back and forth between Molex and a cheaper non-automotive connector for two weeks. The cheaper part had a similar pitch and pin count, and it was 30% less. Molex had a real automotive-grade datasheet, global support, and traceable supply. Eventually I chose Molex because the customer required OEM-level reliability. At the time, it felt like over-engineering.
After the network tester issue, I thought maybe I had been wrong. Maybe the cheaper part would have performed just as well. But as the investigation went on, I realized the problem was not the connector.
The visual inspection was fine. The terminal dimensions were within datasheet values. But when the lab cut open a failed sample, the cross-section told a different story. The terminal was crimped too far forward. The conductor grip had only two of the four crimp teeth fully closed. The bellmouth was missing. From the outside, the crimp looked acceptable. Under vibration, it barely held the wire.
According to IPC/WHMA-A-620, the bellmouth and wire brush are acceptance criteria for this type of crimp. If the bellmouth is missing, the wire is more likely to fracture under flex. The network tester had caught something that no visual inspection would have caught. I've seen this pattern many times. But when I say 'many,' I do not mean just a few—I mean consistently across dozens of suppliers and years of audits.
A Time-Pressure Decision and an Expensive Rework
Then the customer's quality engineer emailed. They wanted a corrective action report in three days. Normally, I'd run a full 8D process, re-qualify the tooling, and review every specification line. But there was no time. I had to make a call with incomplete information.
My gut said the terminal was fine, and the process was the problem. So I called the assembler's plant manager. It turned out they had changed the applicator die two weeks earlier because the old one was worn. The new die was intended for a different terminal, and it shifted the crimp position slightly. Nobody updated the work instruction. Basically, the assembler was using a perfectly good die on a connector it wasn't set up for.
We had 1,200 completed assemblies from the bad run. We rejected all of them. The rework cost about $18,000 and delayed our internal launch by two weeks. The replacement parts came from the same Molex Lincoln NE stock, with full lot traceability. The assembler changed the die, re-set the crimp height, and ran 300 samples. The network tester passed every one.
There's something satisfying about a lot that goes from 94% to 100%. After all the stress and coordination, the process finally behaved. But I won't pretend it was a good two weeks. During that same week, we also had to explain to our own procurement team why 'good enough' isn't good enough for automotive-grade terminals.
Why Are Phones So Durable? (And Why It Matters for Automotive Connectors)
During the investigation, a process engineer asked a random question: 'Why are phones so durable? I drop mine all the time.' It sounded out of place, but it was the right question.
A phone survives a drop because the internal connectors don't lose contact. The board-to-board connectors compress, the wire-to-board terminals have friction locks, and the whole assembly is designed to flex without separating. That's what you're paying for with Molex automotive connectors—not just a shell and contacts, but a complete interconnect system engineered for vibration, temperature, and mechanical stress.
Molex technologies like the MX150 line are built around that idea. They are not magic. If you crimp one with the wrong die, it will fail. But the design discipline matters. The same reliability that makes a phone survive a drop is the reliability you need in a vehicle that gets driven through potholes and heat cycles for a decade.
Here's the lesson I keep coming back to: a connector doesn't fail alone. But I do not mean that every connector is equally good—far from it. I mean that a high-quality connector still needs a controlled process. And a low-quality connector cannot be saved by a perfect process.
If you want automotive-grade reliability, you need both: a supplier who understands the boundary between connector design and manufacturing, and an assembler who respects that boundary. The vendor who says 'this isn't our strength' is more credible than the one who says 'we can handle everything.'
If you've ever argued with a supplier over a red test flag, you know how easy it is to blame the part. But the part is rarely the whole story. Crimp height, pull force, wire strip length, terminal position—these are the things that make or break a connector assembly.
We are good at specifying and verifying the final assembly. We are not a custom terminal designer, and we don't pretend to be. So we rely on specialists: Molex for the connector system, our assembler for the crimping. That boundary, when respected, is what makes an interconnect reliable.
I'm not 100% sure why the first bad batch passed our incoming inspection. I suspect we didn't run it through enough thermal cycling. But I learned to look at the crimp cross-section on every new lot. It's kinda embarrassing that it took a failed network tester to teach me that.
So, why are phones so durable? Because every internal connection is designed to stay connected. That's the same logic I want in every Molex automotive connector we use. Pay for the quality, then control the process around it. That's what 'automotive grade' actually means.
