The Measurement Was Wrong. The Instrument Was Fine. (Turck Sensors, Calipers, Pipettes)

Last month, a process engineer sent me a photo of a Turck proximity sensor sitting on his bench. The sensor, he said, was triggering about 4 mm too early. He was about to order a replacement. I asked two questions: What is the target material? And what are you measuring against? The sensor wasn't the problem. The steel stop plate had a brass wear strip brazed onto it. Different material, different sensing distance. No recalibration could fix that. I'm not a metrologist. I'm the guy who has made enough expensive measurement mistakes to write this from experience. Over the past eleven years, I've personally documented fourteen significant errors, totaling roughly $78,000 in wasted budget—parts scrapped, lines stopped, clients frustrated. Most were not caused by broken sensors, calipers, or pipettes.

The surface problem: you just stopped trusting the tool

When a reading doesn't match expectations, the first reaction is always the instrument is wrong. The next reaction is send it for recalibration, and the third is replace it. I understand that. But those reactions skip the most common cause of bad measurements: the setup.

An engineer sees a Turck level sensor report 63% when a sight glass says 58%. A QC tech sees an 8 inch caliper read 8.0002 on a part that a CMM says is 8.0010. A lab analyst sees a pipette dispense 9.4 µL when the dial says 10 µL. All three people blame their device. And all three might be wrong.

The deeper cause: a measurement is not a property of the device

Here is the counterintuitive thing I learned the hard way: a calibrated instrument can still give you a perfectly wrong reading. The reading is a function of the whole system: device, target, operator, environment, and reference values. Change one element and the number changes. The instrument can be perfect and still lie.

Calibration certificates tell you the device was within tolerance in a lab, under controlled conditions, at one point in time. They do not tell you what happens when the cable runs next to a VFD, the caliper drops on the floor, or the pipette sits in a warm hand for five minutes. I do not mean you should ignore certificates. I mean you should treat them as a starting point, not a guarantee.

For Turck proximity sensors

Turck proximity sensors—like most inductive sensors—are specified using a standard target defined in IEC 60947-5-2. That target is typically a mild steel plate of a defined size and thickness. Swap that target for aluminum, brass, or a small bolt head, and the sensing distance shifts. The sensor didn't fail; the target material did. I once watched an engineer replace a good sensor three times before someone asked about the stainless steel plate that came from a different supplier. The plate was slightly magnetic. That changed everything.

For level sensors

When people say Turck level sensor, they usually mean one of several technologies: capacitive, radar, ultrasonic, or conductivity. Each has application limits. A capacitive level sensor is sensitive to the dielectric constant of the liquid. If you calibrate it for water and then run a cleaning solution with different dielectric properties, the reading shifts. Same instrument, different universe. The sensor is telling the truth about its electrical world; it's not telling the truth about the liquid level unless the settings match. And here's the nasty part: it looks normal while it's wrong. It doesn't send an alarm. It just gives you a number you can rely on until you can't.

For calipers

An 8 inch caliper is a workhorse. But it is not the same as a depth caliper. The depth rod on a general-purpose caliper is narrow, and it is easy to tilt. A dedicated depth caliper has a wide base that seats more reliably on the reference surface. I have seen operators measure the same step at 0.028 inch and 0.015 inch using the same 8 inch caliper, simply because they held the rod at a slightly different angle. Actually, it wasn't a depth issue. It was a perpendicularity issue.

For Eppendorf pipettes

This is the one that stings. When someone asks me how to calibrate an Eppendorf pipette, I do not start with a checklist. I ask what volume they actually use and what liquid. ISO 8655 and the manufacturers recommend a gravimetric calibration procedure, using a balance, water temperature, and vapor pressure correction. You do not just do a few drops at full volume. Eppendorf's approach—and any sensible lab procedure—covers the usable volume range, typically 100%, 50%, and 10% of nominal volume. The balance resolution and water temperature can shift the result enough to matter at small volumes.

Tip choice matters too. An off-brand tip can leak or change the air seal enough to push you outside tolerance. And if your balance only reads to 0.1 mg, you cannot confidently verify a 5 µL volume. The calibration method is simple; the application is not.

The price of ignoring the system

The numbers from my own files are more honest than a generic warning.

In 2019, I approved a sensor setup for a packaging line. The Turck proximity sensor was new, labeled as calibrated, and the output was correct. But the mounting bracket let the sensor drift 2 mm during warm-up. Forty-two bottles collided before someone watched the sensor while the line ran. That was $1,900 in rework plus a nearly lost client. A sensor can be perfect and still lie to you if its operator is careless.

In September 2022, a technician used the depth rod of an 8 inch caliper to measure a recess depth. He held it at a slight angle and got 5.78 mm. The part was actually 5.62 mm. The 120-piece order was already in assembly. $3,200, straight to scrap. A proper depth caliper—or the tiny discipline of asking is this rod perpendicular?—would have caught the error.

One of my biggest regrets: not documenting the liquid medium when I sized a Turck level sensor. We set it for a cleaning solution with one dielectric constant. A line flush swapped in a different chemical two weeks later. The level reading went from accurate to useless overnight. That mistake didn't just cost money; it cost credibility. I still kick myself for not asking the one question I was supposed to ask: What else might be in this tank?

So glad I dodged a bullet on another one. I went back and forth between replacing an expensive sensor and re-machining the bracket for two weeks. On paper, replacing the sensor made sense; the sensor was old and the reading was jumpy. But my gut said the bracket had too much play. I re-machined the bracket for $210. The sensor worked fine. A little embarrassing, but much cheaper than a replacement.

What I do now, briefly

I didn't become a calibration expert. I became a better setup questioner. Before replacing, recalibrating, or calling a vendor, I open with three questions:

1. What exactly am I measuring? Presence, distance, level, depth, or volume? These sound obvious, but I've watched people replace a level sensor when the real problem was a density change in the liquid.
2. What are the reference conditions? Target material, liquid dielectric constant, temperature, tip type, mounting position. If the real conditions don't match the datasheet assumptions, the reading will be wrong no matter how new the calibration sticker is.
3. Did I verify the setup? Is the bracket tight? Is the depth caliper base sitting flat? Is the pipette straight and using the correct tip? Is the sensor wired to the right output? I've seen a bad sensor turn out to be a loose mounting screw more than once.

I also learned to admit my own limits. If someone asks about a vibration-rated mount or a chemical compatibility question outside my experience, I say so. I'd rather work with a specialist who knows their limits than a generalist who overpromises. A vendor who says this isn't our strength—here's who does it better earned my trust for everything else.

Now I maintain the team's checklist with those three questions. It has caught 47 potential errors in the past 18 months. That's not a brag; it's proof that most measurement mistakes were never about the measurement itself. If you found this because you were about to replace a Turck proximity sensor, or you were searching how to calibrate Eppendorf pipette, or you were wondering whether your 8 inch caliper is the right tool for a depth check—stop. The instrument might be fine. The question is whether the rest of the system is telling it the truth.