In June 2024, our line stalled. Not with a crash — the machines kept running, the screens stayed on, but the test fixture stopped producing usable readings. Three times in two weeks. Each time, the techs swapped cables, rebooted everything, and it worked again. For a day. Maybe two. Then it would die at the worst possible moment, usually during a customer order.
I'm the office administrator for a 40-person manufacturing company. When I took over purchasing in 2020, I managed everything from office supplies to capital equipment — about 60 to 80 orders a year across 8 regular vendors. I report to both operations and finance, which means I feel production delays and budget overruns equally.
So when our test fixtures started acting up, my first instinct was to replace them. I had a vendor ready to sell us a new data acquisition system for about $12,000. I was one signature away from placing the order when Doug, our senior engineer, asked a question nobody had asked yet:
"Did anyone check the cable assemblies?"
The problem I thought we had
For three weeks, everyone was convinced the issue was the test equipment itself. Intermittent signal loss basically screams "data acquisition card failure." The vendor who sold us the system even hinted at that. I had the quote in hand and had already started building the case for finance.
But Doug wasn't convinced. He spent two afternoons with a continuity tester, chasing the signal path from the fixture to the DAQ card. And then he found it: a broken contact inside one of the cable assemblies we'd been buying for the past 18 months.
That contact had a part number stamped on it: 8110. When I looked it up to order a replacement, the search led me to TE Connectivity — a company I knew almost nothing about, but now understand I should have researched much earlier.
The deeper problem: nobody asked who actually made the parts
Here's what I didn't understand in my first few years of industrial purchasing: the quality of a finished product is only as good as its components, and the quality of a component is only as good as the company that designed and manufactured it.
We had been buying replacement cable assemblies from a general-purpose electronics vendor. Price was about 30% lower than the original manufacturer's. Delivery was fast. And the failure rate was terrible — we just didn't see it because the failures were intermittent. Connectors that worked most of the time. Crimps that looked fine under a flashlight but failed under thermal cycling. Shielding that wasn't really shielding.
I remember standing in Doug's office while he showed me the broken contact. "This is a TE part," he said. "The original spec. But what we got... this isn't it."
That sent me down a rabbit hole. I spent an evening reading about TE Connectivity, Inc. and its history. The company came out of Tyco Electronics. They manufacture connectors, sensors, relays, and cable assemblies for industrial, automotive, medical, and telecom applications. Component engineering is the entire focus. Not consumer gadgets, not "everything for everyone." Their whole business model is built around doing connection and sensor technology properly.
What really caught my attention was their medical division. The TE Connectivity medical division makes components that go inside the human body — micro connectors, sensor catheters, implantable-grade wire. When the tolerance for failure is literally a patient's health, the quality standard is in a completely different league. A lot of that capability came from acquisitions. When I looked into TE Connectivity acquisitions, I saw a consistent pattern: they acquire specialist manufacturers to deepen expertise in specific areas, not to fill gaps in a product catalog.
That's the opposite of the vendor we'd been buying from. Our vendor sourced cable assemblies from whichever factory quoted the lowest price that week. No design ownership. No traceability. No consistency. They weren't a component manufacturer at all. They were a middleman with a catalog.
A manufacturer who claims to do everything usually ends up doing none of it particularly well. TE doesn't claim that. They don't make multimeters or test fixtures, and they'll tell you so. But they own the design, the tooling, and the process for the components they do make. And for someone whose job depends on those components not failing, that boundary is exactly what makes them trustworthy.
The cost of ignoring who made the parts
Let me break down what those "cheap" cable assemblies actually cost us between March and July 2024:
- $3,200 in scrap — boards and subassemblies damaged by intermittent connections during testing
- 47 engineer hours spent troubleshooting — roughly $6,400 at fully loaded cost
- $4,100 in emergency shipping and priority rework for two late customer shipments
- $2,300 in credits and discounts to hold those accounts
Total: about $16,000 — to save maybe $600 over 18 months on cheaper cable assemblies.
And the worst part? We weren't saving money at all. We were paying almost the same price for significantly worse products, and there was no one to hold accountable. The vendor didn't manufacture anything, so they couldn't answer questions about materials, tolerances, or quality control. When the parts failed, the only answer was "we'll send replacements." Which also failed.
The conventional wisdom in procurement is that "equivalent" parts are a safe way to cut costs. My experience with 80+ orders suggests otherwise. If the original part was designed and manufactured by a specialist like TE Connectivity, buying a no-name equivalent isn't cost savings. It's deferred risk.
What I do differently now
I'm not going to tell you to switch all your suppliers overnight. But this experience changed three things in how I handle component purchasing:
- I verify who actually makes the part before ordering. If a component was designed by TE Connectivity, I buy genuine from TE or an authorized distributor. Genuine parts cost more upfront, but they bring documentation, traceability, and application engineers who know what the part is actually supposed to do. The industry standard IPC/WHMA-A-620 spells out what a proper crimp looks like — our "generic equivalents" failed that standard visually.
- I look at a manufacturer's history and focus before trusting their catalog. TE's acquisition history showed me a company building depth, not just width. And the existence of their medical division told me a lot about their internal quality culture. If a manufacturer can make a sensor that goes into a human body, their industrial connectors are probably okay.
- I bought a proper multimeter for incoming inspection. This one sounds basic, but we now spot-check every batch of cable assemblies and connectors when they arrive. After two weeks of research and honest conversations with our engineers, the best multimeter for electronics in our budget was a Fluke 87V. It does continuity, capacitance, resistance, frequency — everything we need to catch a bad batch before it reaches production.
One more thing changed, and it's the one I'm most proud of. I now ask every vendor what they don't do.
When I talked to TE's application engineering team about a custom requirement, they told me something I didn't expect: "This is outside our standard product range. For a fully custom hermetically sealed assembly, you'd probably be better served by a specialist."
That honesty built more trust than any price discount or sales pitch ever has. A supplier who tells you where their expertise ends is a supplier who is confident about what they deliver. That's a far cry from the vendor who says "we do everything" and then quietly fails at most of it.
The lesson
I still buy office supplies from whoever is cheapest. But components? Never again.
The next time your test equipment fails intermittently, don't ask "which machine do we replace?" Ask "who made the parts inside?" — and then ask whether that manufacturer has a good reason to be good at it.
That question would have saved us $16,000, 47 engineering hours, and a very uncomfortable meeting with my VP. So glad we caught it before signing off on that new data acquisition system. We almost replaced an entire test rig because of one broken contact from a vendor with no accountability.
(Note to self: verify part origin on every PO, no matter how small.)