In 2024, our incoming inspection rejected 11% of newly delivered test and measurement instruments. None of them arrived visibly broken. Most arrived with a certificate saying they were within tolerance. They still didn't pass.
For measurement equipment, quality is not what the datasheet promises. It's what you can prove before the instrument enters service.
I'm quality/compliance manager at Hexagon, responsible for a portfolio that runs from Hexagon Metrology CMMs to differential pressure transmitters, lab centrifuges, and electrical testers. I review roughly 200 units a month—maybe 210 in heavy months. In 2024, my team rejected 11% of first deliveries. This article is about what I keep seeing when that happens.
Where the 11% comes from
Since January 2022, every instrument in our scope has to be checked against its published specification before it ships. Not 'does it turn on?' — 'does it measure correctly, and can we prove it?' In our Q1 2024 audit, we put 300 instruments through that process across eight product categories. 27 measured outside the manufacturer's tolerance under our test conditions. Another six had certificates we couldn't accept because the reference standard IDs or test points weren't documented. 33 out of 300 — 11%. (We call that 'trust but verify,' which in quality work means we don't trust.)
Those 300 units came from multiple manufacturers. I won't name them, because the point isn't which logo is on the label. Any instrument — including units carrying our own brand — can fail its first real check. Over four years of running this process, I've seen the same gaps in inexpensive and high-end products alike. And the process isn't just a cost center: internal customer-satisfaction scores related to instrument performance improved by 34% between 2022 and 2024. Not a controlled experiment, but it has kept the skeptics quiet.
The failures cluster into four patterns.
Four failure patterns I keep rejecting instruments for
1. Differential pressure transmitters: the wrong question gets asked first
For differential pressure transmitters, the first question buyers ask is 'What's the accuracy?' The question that protects them is 'What is that accuracy traceable to?' A transmitter can have a great sensor and still ship with test data that doesn't cover the points where your process actually runs.
Last year, a batch of DP transmitters arrived with certificates claiming accuracy near ±0.075% of span. On our reference — a process calibrator with NIST-traceable calibration — several units showed 0.2% of span error at the 50% point. That's more than double the claim, and it took a five-point test to see it. The manufacturer's reply was that the deviation was 'within industry standard.' Our contract said otherwise. The batch went back for rework at their cost, and every purchase order since then has required five-point calibration data with reference-standard IDs.
That small error changed a level calculation enough to matter. On a DP transmitter, 50% of span is often where the process runs. Nobody sees the problem unless the full range is verified.
2. CMMs: 'calibrated' is not the same as 'verified on your floor'
When a customer asks whether a Hexagon Metrology CMM is a better choice than the alternative, I always say the machine is only half the answer. For coordinate measuring machines, acceptance and reverification are defined by ISO 10360-2:2009. That standard assumes controlled conditions: temperature, temperature gradients, vibration, and measurement strategy. A 'calibrated' CMM is calibrated under those conditions. The machine you install in a shop that swings several degrees between morning and afternoon is not under those conditions.
In 2023, we installed a CMM at a customer site where the temperature varied by roughly 5 °C across the day. Within a month, the quality team was convinced the machine had lost accuracy. It hadn't. The frame and the workpiece expanded unevenly, and the reported measurements drifted outside the acceptance data. The fix was not a new probe or a software update. It was a week of temperature logging and a room that could hold a stable range.
The counterintuitive lesson: when a CMM starts failing reverification, most teams assume wear. Just as often, the facility has changed. Before adjusting the mechanics, measure the environment.
3. 'Insulation tester vs. Megger': the name is not the spec
One of the most common questions I see is 'insulation tester vs. megger — which should I get?' The premise is wrong. An insulation tester and a megger are the same category of instrument: it applies a DC test voltage — often 250 V, 500 V, 1,000 V, or higher — and reads the resulting insulation resistance. 'Megger' began as a manufacturer's name and became the generic word for the product. Comparing 'insulation tester vs. megger' is closer to comparing a generic name with a brand name, not two different technologies.
The specifications that matter are test voltage, resistance range, and compliance with IEC 61557-2, the standard for insulation resistance measurement in low-voltage systems. If your spec calls for a 1,000 V test and someone sells you a 500 V insulation tester, the brand on the front won't rescue you. A weak insulation layer can look perfectly fine at 500 V because it was never stressed enough to reveal the weakness.
4. Centrifuges: the rotor is half the quality record
Laboratory buyers searching for a centrifuge 5804 usually compare maximum speed, rotor selection, and sample volume. Those are selection criteria, not quality criteria. Once the machine arrives, quality review is about mechanical safety and service records.
We get asked about the centrifuge 5804 almost weekly because it is a proven, popular unit — it earned that reputation. But in pre-shipment verification, the gap isn't the centrifuge; it's the rotor documentation. Rotors have a finite life based on manufacturing date, run cycles, and stress. A rotor with no date stamp or run log is an unknown safety component that no electrical test can detect. Type testing per IEC 61010-2-020 proves the design; it does not certify that the rotor in front of you is safe to run for another five years.
What my rejection rate does not tell you
An 11% rejection rate sounds alarming, but it should be read carefully. Not one of those 33 units was scrap. Some were recalibrated, some were reworked, and some simply needed complete paperwork. That's the entire point of catching problems on the bench instead of after installation.
Passing our verification also does not guarantee an instrument is right for your application. It guarantees that the instrument does what it claims within the tolerance of the points we tested. A clean report on a 500 V insulation tester won't make it suitable for a 1,000 V motor test. A DP transmitter that passes on the bench can still be installed with impulse lines that defeat the measurement. Incoming inspection is one layer, not the whole system.
So don't turn the 11% stat into a religion. Quality requirements should scale with risk. A handheld meter used for routine checks doesn't need the same reverification schedule as a CMM releasing safety-critical parts. Define your tolerances, require traceability where it matters, and document what you measured. If you do that, you will rarely need to rely on a name badge as your answer.
