Three months ago, I almost caused a $4,000 utility billing dispute because I couldn't read a Sensus digital water meter correctly. Not because the meter was broken, and not because I'm technically incompetent—but because I skipped a 30-second step we now have written into our team checklist.
That mistake got me thinking about every equipment failure I've documented over the past six years. I've personally made (and recorded) 23 significant errors in that time, totaling roughly $18,000 in wasted budget, scrapped parts, and lost credibility. Now I maintain our team's pre-check list to keep others from repeating them.
Here's the uncomfortable truth I've landed on: most of those mistakes weren't equipment failures—they were expectation failures. And they were preventable.
The Surface Problem: What I Initially Blamed
For a long time, I thought my problems were unique. A drive system that randomly lost position feedback. A sensitive pipetting protocol that kept ruining batches. A temperature reading that looked reasonable but was 14°C off reality.
Let me walk you through the specific disasters, because the pattern only becomes obvious when you see them side by side.
Disaster #1: The SICK DBS60 Encoder That 'Failed' on Day One
In March 2024, I spec'd a SICK DBS60 incremental encoder for a conveyor drive retrofitted at a packaging plant. SICK encoders are supposed to be workhorses. The DBS60 with its solid shaft and high IP rating should have handled a dusty, vibrating environment without complaint.
Within a week, the PLC was showing erratic speed readings. Production stopped twice. My first call was to the vendor: “Your encoder is garbage.”
Turns out, the encoder wasn't garbage. The issue was that I'd mounted it onto a shaft that had excessive axial play in the coupling. The encoder's internal bearing was being hammered every cycle. The sensor side was fine—my mechanical design was the weak link. I'd focused so hard on electrical specs (resolution, outputs, voltage) that I'd ignored the mounting conditions I could see right in front of me.
That mistake cost $890 in labor for reinstallation plus a 3-day production delay. And the root cause? I didn't read the mechanical section of the datasheet. I stopped at the electrical specs.
Disaster #2: The 'SICK Thermometer' Search Rabbit Hole
Now, let's address the elephant in the room. If you clicked on this article because you searched “thermometer sick fever temperature”—you're probably looking for a medical thermometer, not an industrial automation company. That's not what this SICK is about. This SICK is the German sensor manufacturer, SICK AG. I'm not a doctor, and I wouldn't want to play one for you. But I have learned a thing or two about industrial temperature measurement.
In my world, temperature isn't about fevers—it's about bearing temperatures, motor windings, and process fluids. On a $1,200 hot-melt adhesive line, I once trusted an uncalibrated infrared gun and set the heater profile 15°C too high. The resulting charred adhesive clogged two nozzles and cost us half a shift. I'm not a thermography expert, so I can't speak to emissivity correction across different surfaces. What I can tell you from a maintenance perspective is: an unverified temperature reading is worse than no reading at all. It gives you false confidence.
Disaster #3: The Research Plus Pipette Calibration Gap
In a lab setting, measurement errors are quieter but no less expensive. A colleague of mine had a set of Eppendorf Research Plus pipettes that were passed on through a year of heavy use. They never physically dropped them, never noticed visible damage. But when they sent the pipettes for a routine gravimetric check, two of the four units were far out of spec for the low-volume range—consistently delivering ~12% less than the target.
It's easy to assume that a quality pipette stays accurate until you break it. Actually, seals degrade, springs fatigue, and contamination builds up in the shaft. Skipping annual calibration saved maybe $200 per pipette. But it ended up invalidating a month of assay data that had to be repeated—not to mention the cost of the reagents themselves.
That's when I learned: instruments don't fail loudly. They drift quietly.
Disaster #4: The LCR Meter and the 3,000-Capacitor Blunder
I'll own up to this one. In my first year (2017), I was using an LCR meter to verify a batch of 3,000 capacitors before a production run. I set the measurement frequency to 100 kHz because that's what the datasheet spec was quoted at. Simple enough.
Except I didn't account for the test fixture's stray capacitance. At 100 kHz, the fixture added a few picofarads that pushed the readings outside tolerance. I rejected the entire batch, demanded a replacement order, and set the project back by a week. The vendor's technical rep—rightly annoyed—asked me to re-test using a proper Kelvin connection with a short-circuit compensation.
Let me tell you, that conversation was humbling. The capacitors were fine. My measurement methodology was flawed. The $3,200 batch was scrapped for no reason.
The Deeper Cause: It's Never the Instrument (at First)
If you look at those four disasters side by side, the pattern is obvious: every single failure involved me skipping the same verification step before blaming the tool.
Here's what I mean. We all carry a mental model—a simplified version of how a device works. SICK encoders are robust, so I assume any malfunction is a defective unit. Research Plus pipettes are quality instruments, so I assume they hold calibration. An LCR meter gives me a number, so I assume the number is true.
The moment I stop testing those assumptions is the moment I start burning money.
Why This Keeps Happening in Industrial and Lab Settings
In an ideal world, you'd verify everything. But the modern technician is under pressure to move fast, and field instruments are designed to look self-explanatory. A digital display implies certainty. A rugged IP67 housing implies indestructibility. That combination—trusting the tool without checking the environment—is a recipe for the exact failures I keep documenting.
There's a second, subtler layer too: specification confusion between 'resolution' and 'accuracy.'
My SICK DBS60 had a fine number of pulses per revolution. High resolution. But the coupling play introduced shaft wind-up that no amount of encoder resolution could compensate for. The LCR meter displayed five digits, but the fixture's stray capacitance limited the real accuracy. The pipette can deliver tiny volumes, but a worn seal makes it inaccurate before it ever becomes imprecise. High resolution just means you measure the error more precisely, not that you avoid it.
I'm not a metrologist, so I can't speak to the full theory of measurement uncertainty. What I can tell you from a practical perspective is that the datasheet specs apply to controlled conditions. The floor of a packaging plant is not a controlled condition.
The Cost of Ignoring the Pattern
Let's put some numbers on this. Over six years, I've tracked 23 significant measurement and instrumentation errors. Here's the breakdown:
- Direct replacement labor: ~$6,400 (re-installation, re-calibration, troubleshooting time)
- Scrapped materials: ~$7,100 (the capacitor batch, the pipette assay data, the charred adhesive)
- Expedited shipping for rushed replacements: ~$1,200
- Customer-side credibility damage: hard to quantify, but it's real
The frustrating part isn't the total dollar amount. It's that every single dollar was avoidable. None of these errors required special expertise to catch. They required a willingness to say, “Before I blame the equipment, let me verify my own setup.”
After the Sensus water meter incident in October 2024, I finally hit my limit.
The Fix: A Simple Verification Routine
I want to be clear about something: I don't have a brilliant solution. There's no magic software or fancy tool that solved this. What finally helped was building a mandatory pre-check checklist for our team of five technicians. That's it. Paper and pen.
Here's what's on it:
- Define what “good” looks like before you measure. Know the reference value, the tolerance band, and the specific test conditions.
- Check the mounting or connection first. For a sensor: shaft coupling, alignment, torque. For an electrical measurement: test leads, fixture compensation, connection quality. For a pipette: is it recently calibrated?
- Do a sanity-check measurement on a known reference. A spare encoder with a known-good signal. A fixed resistor for an LCR meter. A gravimetric test with distilled water for a pipette.
- Document the result. Not just PASS/FAIL—but the actual reading, the time, the operator. We've caught 31 potential problems using this checklist in the past 18 months.
- Ask the vendor when in doubt. This is the hardest for proud technicians. But the vendor who said “this isn't our strength—here's who does it better” earned my trust for everything else. I'd rather work with a specialist who knows their limits than a generalist who overpromises.
One more thing that changed my perspective: knowing what you don't know is a professional asset. I'm not a hydrologist, so I can't explain the internal flow measurement logic of a Sensus meter. When our billing dispute came up, I didn't guess—I called the utility company and asked them to walk me through the display cycle. It took ten minutes, and the problem turned out to be a misinterpreted screen that showed cumulative flow, not instantaneous rate.
“The vendor who said 'this isn't our strength—here's who does it better' earned my trust for everything else.”
Look, I'm not saying instruments don't fail. They do. But in my six years of documentation, the instrument failed first roughly 20% of the time. The other 80%—the failure was in my assumptions, my setup, or my skipped step.
That's the lesson I'd pass on to anyone dealing with SICK sensors, lab pipettes, LCR meters, or utility meters: trust the tool, but verify the environment. Your checklist won't be as exciting as a new instrument purchase, but I'd bet it saves you more money in the long run.
