Metrology article header with calibrated measuring tools
Metrology Notes

Industrial Instrument Audits: Why We Reject 12% of Deliveries (Even Endress+Hauser)

Posted on 2026-11-20 by Marcus Feld

Industrial Instrument Audits: Why We Reject 12% of Deliveries (Even Endress+Hauser)

12% of the industrial instrument deliveries we audited in 2024 needed rework or replacement. That includes Endress+Hauser, which performed better than most at roughly 5%. But 5% still meant about 9 units out of 180 that had to be fixed before they could go on the line. A reliable brand doesn't eliminate the need for your own acceptance process—it just lowers the odds slightly.

Bottom line: you don't avoid instrument problems by picking a good brand. You avoid them by building a receiving check that catches the outliers.

I'm the quality and brand compliance manager at our plant. Over four years, I've reviewed every instrument that comes through our receiving dock—Endress+Hauser flow meters and turbidity sensors, safety sensors for our process safety loops, and lab equipment. Our rejection rate was 18% in 2022. It dropped to 12% by 2024. Not because suppliers improved. Because our checklist finally got serious.

We didn't have a formal incoming inspection process until 2022. The third time a mis-calibrated flow meter nearly made it to the floor, I finally created one. Should've done it after the first time.

What we actually check

Endress+Hauser flow meters. These are our bread and butter—electromagnetic, Coriolis, vortex, ultrasonic. They're generally solid. But in a Q1 2023 batch of electromagnetic units rated ±0.2% accuracy, we tested six with our portable calibrator and found two drifting to about ±0.45% under actual conditions. Still within 'industry standard' as the vendor pointed out. Not within our spec for trade-transfer applications. We rejected the batch, they recalibrated and reshipped at their cost. Now every contract requires a factory calibration certificate with matching serial numbers.

Endress+Hauser turbidity sensors. Similar pattern. We bought a unit rated 0.001 NTU resolution for drinking water monitoring. When we tested it against zero-turbidity water, readings bounced between 0.003 and 0.005 NTU. The vendor called it 'normal baseline noise.' Maybe for some applications. But our drinking water compliance reporting can't absorb that noise without a footnote. We replaced it. The next unit tested clean.

I hear people ask whether the cheaper alternative would've worked. We actually ran that comparison once—two turbidity sensor models, one Endress+Hauser, one cheaper with comparable specs on paper. The spreadsheet said go cheap. My gut said stick with E+H. I went with my gut. Six months later, I learned the cheaper model had baseline drift issues that weren't in any datasheet. Sometimes the gut is just pattern recognition doing its job.

Safety sensors. This is where the gloves come off. Process safety sensors need valid SIL certification, and the certificate number has to match the product batch. We had a vendor ship a safety sensor where the product was current-generation but the certificate was from the previous version. That's a red flag you don't see unless you're cross-checking documents at the dock. We caught it. Now every safety sensor gets a document audit before it leaves receiving.

We also started using a FLIR E54 advanced thermal imaging camera for quarterly control cabinet inspections. It's not a replacement for fixed sensors—it's a supplement. It caught loose terminal connections running 8–10°C above normal. The sensor on that loop was reading fine. Its wiring wasn't. Fixed instruments don't tell you when their own installation is degrading.

Lab equipment. Same discipline applies to our QC lab. When someone asks how often to change your columns—the HPLC Agilent ones—the answer isn't a calendar number. Watch the data. Replace when theoretical plates drop more than 20% or system pressure fluctuates past 15%. We once ran a dirty sample matrix that killed a column in 300 injections. Clean samples lasted 800. Fixed intervals are the easiest policy to write and the most expensive to follow.

Under FTC advertising guidelines (ftc.gov), performance claims need substantiation. When a datasheet says ±0.2% accuracy, there should be a traceable test method behind it. If the spec sheet omits test conditions entirely, that's a red flag before the instrument even ships.

One more thing about time pressure. During a Q2 2024 shutdown, we had about four hours to approve six safety sensors before the line needed to restart. Normally I'd run the full 12-point check. No time. I did a 4-point quick check—certificate match, serial number verification, basic function test, physical inspection. One of those sensors failed two weeks later. It was within spec at the time of testing—barely. In hindsight, I should've pushed back on the timeline instead of signing off with incomplete data.

When you can ease up

Not every instrument needs the full two-hour review. If you're buying a non-critical sensor for warehouse ambient monitoring, spot-checking plus supplier qualification is enough. But critical flow meters, turbidity sensors, and safety sensors? Those get the full checklist every time. Five minutes of verification beats five days of rework.

And I'm not saying Endress+Hauser has a quality problem. Their first-pass acceptance rate is among the highest in our supplier base. But 'among the highest' still means there are others, and every shipment needs eyes on it. A rigorous receiving process protects both sides—you know the instrument is right, and they don't carry the blame for a field failure that could've been caught at the dock.

Share this technical note Email Permalink
Marcus Feld

Marcus Feld

Marcus Feld is an electrical test and measurement analyst specializing in multimeters, oscilloscopes, clamp meters, insulation testers, spectrum analyzers, and data loggers. He applies IEC 61010-2-030 and IEC 61010-031 concepts while examining measurement category, bandwidth, true-RMS response, input loading, and stated uncertainty. His work helps maintenance engineers and test teams choose safe instruments with performance suited to the signals and environments they actually measure.

Leave a Reply