Dead on Arrival: The G310 5G Prototype and a Bad First Few Days
Back in February 2023, I was staring at a dead G310 5G prototype with a power rail that was fine, a microcontroller that was fine, and an analog front end that was delivering complete garbage. The problem, I was sure, was the precision resistor network.
If you've ever built a mixed-signal circuit where 0.01% tolerance actually matters, you know the sinking feeling when your measurements don't line up with the datasheet. I had a batch of Vishay precision resistors that looked wrong. Not a little wrong—whole ohms off. And I was about to spend about $3,200 learning that the resistors were fine and I was the one who was wrong.
For context: I'm a product engineer. I've been handling component sourcing for prototype builds for about seven years. In that time, I've personally made (and documented) eleven significant mistakes totaling roughly $28,000 in wasted budget. This one was the most instructional, because it involved the structure of Vishay itself, the limits of my test equipment, and the dangerously ambiguous word "switches."
Our project was the G310 5G, a next-gen industrial vibration sensor platform. The entire concept depends on detecting very small resistance changes—micro-ohms matter. The reference resistor network is the foundation of every measurement the device makes. If those resistors drift even 0.05%, the product's accuracy is compromised.
I've known for years that Vishay is one of the go-to names for precision resistor technology. What I didn't fully appreciate is that Vishay is not one monolithic company with one catalog. It's more like a family of specialized groups. The precision foil resistor side is a different world from the standard chip resistor catalog—different systems, different lead times, different websites.
I spent about a week going back and forth with a distributor, trying to get matched to the right part. I was in a hurry and I clicked approve on the first viable match. (Should mention: the confirmation did say "Vishay Intertech" on it. I thought that was just legal boilerplate. It wasn't.)
The parts arrived in three days, shipped straight from Vishay's Yankton SD facility. I remember thinking, "These are coming right from the source." In hindsight, that should have made me look more carefully at what I'd actually ordered.
Bad Readings and Worse Word Choices
Here's where the real lesson starts.
I grabbed my trusty bench multimeter—a well-known brand, nothing fancy, but it had served me well for years. I set up a quick test jig and started measuring the reference resistors.
The 10kΩ parts read 9.79kΩ on one, 10.24kΩ on another. Out of a batch of twenty, about half looked marginal and a few looked flat-out bad. I sorted them into "good" and "suspect" piles. I even made a color-coded spreadsheet, because that's the kind of nerd I get when I'm angry about something I don't fully understand yet. My anger grew with each measurement. I assumed Vishay shipped a bad lot.
I emailed the distributor for a return authorization and scheduled a call with an engineer from Vishay Intertech—the precision foil group. I was prepared to be extremely not-nice on that call.
Meanwhile, our technician walked into the lab and asked a simple question: "Should I use the same switches for the G310 as last time?"
"Sure," I said, without looking up. I was deep in resentment mode, staring at my suspect pile.
That one word cost us $1,800.
Here's the thing about procurement language: words need to work in multiple contexts. When our technician said "switches," he meant the little tactile push-button switches that go on a PCB. The kind that cost about 40 cents each. The kind that appear in a BOM as "SW1, SW2, SW3."
When the purchasing system went looking for "switches," though, it found what its catalog considered the best match: Cisco switches. Enterprise network switches. Forty-eight ports of blinking, rack-mounting networking hardware. The box that arrived had a distinctly IT-department energy.
Switches vs. Cisco switches—obviously different, right? But without a manufacturer part number, the system had no way to tell. The word "switches" was the problem, and the system was just doing its job with an ambiguous input. (Our IT guy was briefly excited, thinking we were finally upgrading the office network. He was very disappointed.)
"What's the Accuracy of Your Meter?"
On my call with the Vishay Intertech engineer, I walked her through the resistor readings with the full confidence of someone who has definitely, 100%, been handed defective components.
She listened. Then she asked a question that rewired my entire testing philosophy:
"What's the rated DC accuracy of your multimeter?"
I pulled up the spec sheet. My trusty bench meter: ±0.5% basic DC accuracy.
Let me put that in perspective. 0.5% of 10,000Ω is 50Ω. The resistors I was testing were spec'd at ±0.01%—meaning ±1Ω. My meter's margin of error was about 50 times larger than the tolerance I was trying to verify. I was using a yardstick to measure the thickness of paper.
"Oh," I said. "Oh, no."
(around that exact moment, my theory of my own competence quietly collapsed)
The Vishay engineer was polite. "Happens all the time," she said. I could hear the smile across the line. She suggested I send the two worst-looking resistors to Vishay's Yankton SD facility for an independent verification.
The report came back a week later: both resistors well within spec. One measured 10,000.4Ω. The other, 9,999.7Ω. When I compared those numbers to what my multimeter had shown, I finally understood why test equipment precision matters more than part number confidence. The parts were perfect. My process was the defect.
She explained that Vishay's foil resistor testing at Yankton, SD uses statistical sampling and multi-point verification—standard practice for a facility that's been making precision products for decades. The spec isn't a suggestion; it's a baseline. Technology leaders like Keithley, in their low-level measurements handbook, recommend that your meter be at least ten times more accurate than the parameter you're verifying. I now repeat that rule so often my team groans when I start the sentence.
The Bill: $3,300, Give or Take
Here's the final damage report:
- $1,800 for a stack of Cisco switches sitting in a lab closet (we later used one on the test network—so not a total loss, but the $1,700 premium over the tactile switches we actually needed still stings)
- $1,100 in expedite fees to overnight the actual tactile switches from an electronics supplier
- $400 to rent a proper micro-ohmmeter with a 4-wire Kelvin setup, because I could no longer trust my bench multimeter for precision verification
- One week of schedule slip on the G310 5G prototype phase, pushing our launch target from May to mid-June
Total: about $3,300, plus one credibility hit that took a while to overcome. But the real cost was the pattern it revealed.
What I Now Do Differently
It took me about two years and half a dozen wrong orders to understand that component sourcing is not just about choosing the right part number. It's about the ecosystem around it: which division makes it, what tools you need to verify it, and how you communicate it to the people who need to buy it.
Here are the changes that actually stuck:
1. Respect the structure inside Vishay
Vishay Intertechnology (shorthand: "Vishay Intertech") runs multiple specialized product groups. The precision foil resistor group, with manufacturing in Yankton, SD, is a different world from the commodity resistor catalog. According to Vishay's official site (vishay.com), foil resistors target applications where high precision and stability are non-negotiable. The ordering process, lead times, and technical contacts for that group are specific to it. Don't assume one part of Vishay knows what the other is doing—or that the distributor's first match is the right one. If you order precision foil resistors and expect commodity resistor tolerances, you'll be paying too much. If you order commodity resistors expecting foil-grade stability, you'll be disappointed. The part number is the contract, but the division's technology is the context.
2. Make your meter ten times better than your spec
If the spec calls for 0.01%, your measurement tool must be rated for at least 0.001% (or better). For most precision resistor work, that means 4-wire Kelvin measurement, not a standard 2-wire multimeter lead setup. My old bench meter isn't bad—it just can't see down to the level this project needs. (I still use it daily for all the non-precision stuff. It's a good tool. It was just the wrong tool.)
3. Ban ambiguous component descriptions
When a technician asks "which switches?" the correct answer is "SW1 on the BOM, part number this." Manufacturer part numbers remove all doubt. The Cisco switch mixup happened because "switches" in the purchasing system had a different default meaning than "switches" on a PCB design. Now I teach every engineer I work with: if it's not in the BOM with a part number, it's not a component.
4. Keep a pre-order checklist and actually use it
After the G310 5G mess, I built a pre-order checklist for our team. It covers four things: manufacturer part number, verified measurement capability, the specific Vishay division (or any supplier division) that owns the part, and a sanity check on component descriptions before they go to purchasing. That checklist has caught 47 potential errors in the last 18 months. Not bad for a lesson that cost $3,300 to learn.
I still kick myself about that switch order. If I'd spent thirty seconds actually answering the technician's question, we'd have saved $1,800 and a very awkward conversation with the CFO. At least now, "switches vs. Cisco switches" is a running joke around the lab whenever someone is being vague about a component request.
The G310 5G shipped mid-June. It's been rock solid in field testing for over a year now. And the micro-ohmmeter I rented? I bought it (note to self: account for test equipment upgrades in future project budgets). Some lessons cost money. But the right measuring tool is a cost you only pay once—and if you're ordering precision Vishay parts, make sure you've got the precision to check them properly.
Take it from someone who made every mistake in the book in one project: check your tools, clarify your language, and know the steps between you and the sign-off. The vendor is usually not the villain. The measurement equipment just might be.
Prices and lead times referenced here are from the author's experience in early 2023. Verify current specifications, availability, and test requirements with Vishay or an authorized distributor.