7 Rules for Sensor Testing: Lessons from $15k in Mistakes

Who This Checklist Is For

You're troubleshooting a production line where a Turck proximity sensor keeps failing falsely. Or you're onboarding new sensors for a CMM integration and need to verify they're working before installation. Or maybe you're just tired of swapping out "bad" sensors only to find the real problem was a loose wire or a misconfigured tester.

I've been handling sensor orders for about 6 years now. In that time, I've personally scrapped roughly $15,000 worth of product due to bad testing decisions. The checklist below is the result of those mistakes. Use it before you swap a sensor, before you approve a batch, or after you get a weird reading on a laser tracker setup.

There are 7 steps. Skip none of them.

Step 1: Confirm the Basics—Power and Output Type

This sounds obvious, but I'd say 40% of the "bad sensor" returns I've seen were just DC power issues. Most buyers focus on part numbers and completely miss the voltage requirements. The question everyone asks is "is it NPN or PNP?" The question they should ask is "can my system supply 24V DC stable enough for this sensor's tolerance?"

Per the Turck proximity sensor manual (as of January 2025): most inductive sensors require 10–30V DC. But if your power supply drifts below 12V under load, some sensors will behave intermittently. That's not a sensor fault — that's a system fault.

Check this: Use a Fluke multimeter (the same one you'd use to test a capacitor) to measure voltage at the sensor's connection point while the system is running. If it dips below spec, fix the power rail first.

Step 2: Inspect the Sensor Itself (Physical Condition)

I once ordered 200 Turck uprox sensors for a project. Checked the shipping receipt, approved the order, installed them. The result: 15 sensors failed within a month. We caught the issue when we noticed the threaded barrels had micro-cracks from transit damage. $2,700 in labor + the cost of replacement sensors, straight to the budget waste column.

That's when I learned: look at the sensor before you install it. Run your finger along the barrel. Check for dents. Verify the cable jacket is clean—no cuts, no kinks. If you're using a sensor tester (like the Turck tester), make sure the connection pins aren't bent before you plug it in. Bent pins inside a tester can cost you $200+ to replace the tester. Ask me how I know.

Step 3: Use the Right Tester (and Use It Right)

The Turck sensor tester is a solid tool. But "solid" doesn't mean idiot-proof. I was teaching a new technician how to test a radar sensor last year. I said, "plug it into the tester." They heard, "plug it into the tester with the power on." Result: we shorted a pin, and the tester displayed an error code we'd never seen before. The repair cost $450 plus downtime.

Correct procedure: Power down the tester. Connect the sensor. Power on. Select the correct output type (NPN/PNP/NAMUR). Once you get a reading, move the target slowly toward the sensor. Look for the switch point. If the LED turns on before the target reaches the rated sensing distance—or doesn't turn on at all—the sensor may be defective. But also check if your target material is correct: steel vs. aluminum vs. plastic makes a huge difference.

Roughly speaking, 1 out of every 50 new sensors has a manufacturing defect. The tester catches most of those.

Step 4: Simulate the Real Installation Environment

This is the step most people skip. Testing a sensor on a bench with no metal nearby is fine for basic functionality. But an inductive sensor installed inside a machine with 50mm of metal on three sides behaves differently. We discovered this when a CMM project failed because the sensor we tested on a wooden bench couldn't detect our target through vibration and ambient metal interference.

If possible, cable the sensor to the tester, then mount it in the actual bracket location. Run the target past it at real speed. If the signal glitches, you've caught a problem before it gets installed. If it works, you're gold.

Step 5: Validate the Cable and Connector

We were using the same words but meaning different things. I said, "check the cable." The technician heard, "look at it." I meant: continuity test. Discovered this when a $3,200 order of sensors was returned because half of them didn't work. The issue: the connector we ordered separately wasn't pinned out correctly. One wrong pin assignment caused the sensor to output a false signal. That mistake cost $890 in redo plus a 1-week delay.

Use a multimeter to check continuity from the sensor's pins to the controller's input. If you're using a Turck cordset with a pre-wired connector, that's usually reliable—but double-check if you've cut and spliced anything.

Step 6: Test the Full Signal Chain (Including the Controller)

The most frustrating part of sensor troubleshooting: swapping a sensor three times and still getting false readings. You'd think a new sensor would fix a bad signal, but the problem might be in your PLC input card or a loose wire in the junction box. After the third time this happened, I was ready to replace the entire panel. What finally helped: using the sensor tester to bypass everything and confirm the sensor itself was good, then testing each segment of the cable run.

Here's a practical test: set your multimeter to measure voltage at the controller input. Have someone trigger the sensor manually. If you see voltage changes at the controller, the sensor and cable are working. If not, move the meter closer to the sensor and repeat. You'll isolate the failure point in minutes instead of hours.

Step 7: Document Everything and Keep a Log

There's something satisfying about finally having a testing log. After the endless back-and-forth with vendors over whether a sensor died or was murdered, we started keeping a simple log: date, sensor part number, tester result, environment notes, and resolution. In the past 18 months, we've caught 47 potential errors using this checklist and the log. That's 47 times we avoided a $500+ rework.

What to record: sensor serial number, the actual sensing distance measured (not just pass/fail), the ambient temperature (temperature drift affects some sensors), and any unusual behavior (like the LED flickering before failing). If you ever need to file a warranty claim, this log is your best friend.

Things to Avoid

Don't rely on the LED indicator alone. LEDs lie—especially on older sensors where the driver may fail while the emitter still works.

Don't test a sensor at 24V if your system runs on 12V. The sensor may work on the bench but fail in the actual application. Test at the voltage you'll use.

Don't assume a "coordinate measuring machine" or "laser tracker" setup is immune to sensor interference. I've seen laser trackers give bad data because a nearby sensor was radiating electrical noise. Check for EMC compatibility before you finalize the installation.