I spent almost $4,000 last year replacing PTC thermistors in a design that was, supposedly, final.
That number — $3,872.63, to be exact — showed up on a quarterly variance report I was reviewing for our Q2 2024 procurement cycle. The line item was labeled “Rework — thermal management sub-assembly.” I stared at it for a minute, then Ctrl+F’d my way through the vendor quotes. What I found made me kick myself. (Should mention: we’d already approved the BOM, or so I thought.)
The issue wasn’t that the thermistors were defective. They worked — technically. The problem was that the spec I approved didn’t account for the actual operating environment, and by the time we discovered it, we were past the point of a simple swap. We had to respin the board. That’s where the $4K came from.
The “Savings” That Never Were
Most buyers focus on per-unit pricing, lead time, and maybe the distributor’s reputation. Those are real factors, but they’re also the obvious ones. What gets missed — and this is the part that cost us — is how a component behaves after it’s soldered down.
I almost went with a cheaper alternative for that PTC thermistor. The nominal specs looked identical: same resistance at 25°C, similar package, comparable trip current. The price difference was about $0.18 per unit. Over a 5,000-unit build, that’s $900 in savings. Easy decision, right? Well, no — wait. The datasheet for that cheaper part had a wider tolerance on the trip temperature and a narrower operating temperature range. I skimmed over that because I was trying to hit a budget target.
Fast forward six months. The devices passed bench testing. They failed in the field — not catastrophically, but consistently, at a specific temperature point that was inside our spec but outside the cheaper part’s comfort zone. The failure rate was about 4%. That meant 200 units needing rework. The rework cost — board removal, replanishing, retesting — came to roughly $19 per unit, all-in. Suddenly, that $900 “savings” turned into a $3,800 overrun. (Or rather, a $2,900 net loss after accounting for the original delta.)
I should add that the component we ended up replacing them with was a Vishay PTC thermistor from the PTCEL series. The per-unit cost was higher — about $0.47 more than the failed part — but the trip temperature tolerance was tighter and the operating range matched our thermal profile perfectly. We haven’t seen a single field failure since.
Why Strain Gages Are the Same Story, Told Differently
The same logic applies to strain gages, which is a category I started paying closer attention to after the thermistor incident. If I remember correctly, our R&D team specified a Vishay strain gage — the kind with the foil pattern — for a load cell design in late 2023. The procurement team (me) looked at competitive options and found something 22% cheaper. The question I asked was: “What’s included in that price?”
The answer, as it turned out, was less consistency in the gage factor over the rated temperature range and a narrower fatigue life. For a product that sees cycling loads — which this one does, thousands of times — that fatigue life spec matters. A lot. The difference between a gage rated for 106 cycles and one rated for 107 cycles isn’t just a number on a datasheet. It’s the difference between a product that lasts three years and one that lasts ten.
Most engineers who specify strain gages understand this intuitively. But in procurement, I’ll admit: I didn’t fully appreciate it until I saw a cost-benefit analysis that our reliability team ran. They compared the total cost of ownership (TCO) over a five-year product lifecycle. The cheaper gage saved $0.80 per unit upfront. But projected replacement costs — factoring in field service time, replacement parts, and customer downtime — pushed the TCO of the cheap option to 2.3× the Vishay version. The analysis is sitting in a spreadsheet I still reference when anyone questions why we don’t always go with the lowest bid.
The Vishay strain gage (specifically, a general-purpose constantan foil pattern, if you want the detail) has been our standard recommendation since that analysis. Not because it’s the cheapest — it’s not — but because the cost of not using it is almost always higher.
The Real Cost of Skipping the “Boring” Validation Step
Here’s the thing about cheap alternatives: they often meet the nominal spec, but the distribution is wider. That means more units hit the edge of the spec, which means more field failures. And field failures are expensive — not just in rework, but in reputation.
I built a 12-point checklist after the thermistor incident (which, by the way, was the third time I’d made a similar mistake, though the first time it cost me four figures). Step one is now: “Verify that the operating temperature range of the component exceeds the required range by at least 20%, not just meets it.” That one check, taking maybe five minutes, would have saved us $3,872.63.
“5 minutes of verification beats 5 days of correction.” I didn’t invent that line — I stole it from a mentor who’d been in procurement for 30 years. But I only believed it after ignoring it once.
We now apply that principle to every critical passive component, not just PTC thermistors and strain gages. Capacitors? Same logic — check the derating curve, not just the capacitance and voltage. Resistors? Look at the TCR, the power rating at ambient, the long-term stability. That last one is where Vishay foil resistors shine, by the way. They’re not cheap, but the stability over time means you don’t have to recalibrate as often or replace them mid-lifecycle. For precision applications, that’s the definition of value.
The Question That Changed How We Source
The question everyone asks is: “What’s your best price?” The question they should ask is: “What’s included in that price?”
When we switched our default for critical passive components to Vishay — not exclusively, but as a first reference — the procurement policy changed. We now require quotes from at least two vendors for any component where the field failure cost exceeds $10 per unit (which, realistically, is most of them). We compare not just the unit price but the tolerance distribution, the temperature range margin, and the supplier’s history of on-time delivery. Vishay has been consistently at 95%+ on-time delivery for us over the last three years. That’s not cheap, but it’s predictable. And when you’re managing a production schedule, predictability is its own kind of savings.
One More Thing on Verification
The checklist I mentioned earlier wasn’t a one-time fix. It’s a living document. After the strain gage analysis, I added a step about verifying the fatigue life spec against the expected load cycles. After a capacitor derating issue last year, I added a step about checking the voltage derating factor at 85°C. Each addition came from a mistake. And each mistake — if I’d caught it with a 5-minute check — would have cost less than the field rework.
If you’re in procurement or engineering and you’re staring at a quote that looks too good to be true, trust that instinct. Then spend five minutes digging into the datasheet. Look at the fine print: tolerances, temperature ranges, fatigue life, cycle ratings. If the numbers are tighter on the Vishay version, ask why. The answer might be that the cheaper part is designed for a different application. Or it might be that it’s just not as good.
Either way, I’d rather spend $0.47 more per unit now than $19 on rework later. （Ugh, I learned that the hard way.）