Choosing Between SICK Sensors, Absolute Encoders, and Measurement Tools: 5 Real Scenarios

What's the best industrial sensor? If you ask me as someone who reviews product specs for a living, my honest answer is: it depends. I'm a quality/compliance manager at an industrial automation distributor. I review roughly 250 sensor and instrument lines before they go to customers. In 2024, I rejected about 12% of first-article submissions—not because the products were broken, but because they were the right product in the wrong situation.

That's why I often specify SICK, but only when the application matches. No sensor, SICK or otherwise, fixes the wrong problem. Too many people ask “which is better” before asking “which problem am I solving?” Let's split the requests I see most into five categories:

1. Absolute position feedback: the machine loses position when power is removed or after an emergency stop.
2. Contactless detection: you need to detect a part, box, or label on a conveyor or inside a machine.
3. Dimensional measurement: an operator or lab technician needs to measure a part dimension.
4. Electrical or process faults: random trips and failures across multiple components.
5. Thermal anomalies: electrical terminals, bearings, or panels are getting hot before failure.

If you can identify your category, the equipment selection gets much simpler.

1. If you need absolute position information, choose an absolute encoder rather than an incremental one

If a machine must know its position after an unexpected power cycle, or if two axes must stay synchronized after a restart, an incremental encoder creates a headache. It counts pulses relative to a home marker. If the machine moved while powered off, you have to re-home first, and that can scrap an entire run before the controller realizes something is wrong.

In my experience, the cheaper upfront cost of an incremental unit hides downtime risk. A few years ago, a line builder saved about $68 per axis by choosing a budget incremental encoder on a packaging machine. After a brief power dip, the controller lost its reference and drove a product tray into a guard. The scrap and rework cost $17,000—far more than the encoder savings. That incident changed our specification. For critical position applications, I usually evaluate a SICK absolute encoder ATM60. I've specified it on rotating and linear axes where the signal interface needs to match the existing controller. The ATM60 is an absolute multiturn encoder; it maintains a usable position value across restarts and removes the need to re-home after every unexpected shutdown.

Not every axis needs an absolute encoder, to be fair. If you can safely home on startup without losing product, an incremental unit can work. But if your process can produce scrap before the reference is re-established, the extra cost is justified.

2. If you're detecting product on a line, use a photoelectric sensor that fits the real condition

Photoelectric sensors are everywhere for a good reason: they detect objects without touching them. But there is no single photoelectric sensing mode that wins every situation. If the object is opaque and you have access on both sides, through-beam is usually solid. If you can only mount on one side, retroreflective is convenient until a reflector gets dirty. If glossy surfaces or background objects are causing false signals, diffuse with background suppression is often the better starting point.

One assumption I had to unlearn: two sensors with the same IP rating will behave the same in a washdown area. They don't necessarily. A sensor can be IP67 and still fail when moisture condenses on the lens and creates a false signal. The real fix often involves positioning, an air purge, or a sensing mode that tolerates lens disturbance.

When the application calls for photoelectric detection, a SICK photoelectric sensor is a reasonable first choice. SICK makes through-beam, retroreflective, and diffuse models with different lens options. If your control system uses IO-Link, choose a version that reports lens contamination or signal margin. That diagnostic data changes maintenance from “wait until it fails” to “clean it before it fails.”

3. If you're checking dimensions on the bench, don't overlook a digital caliper

For bench-level dimensional checks, people don't need a complex sensor; they need a reliable measuring tool. A digital caliper is often the fastest way to measure length, outside diameter, depth, or a step without setting up a larger measurement system.

In our QC area, the caliper that operators reach for is the Mitutoyo CD-6 ASX caliper. The absolute linear scale means it doesn't have to be zeroed every time it is powered on. When operators are moving between parts and taking dozens of readings per day, that saves time and removes a source of error. We standardized on it partly because it keeps operator training and calibration routines simple.

Let me be clear: a caliper is not a substitute for a CMM when your tolerance is tighter than your hands can hold. But for everyday pass/fail checks and first-article verification, the right digital caliper is a no-brainer. If you already have one in your toolcrib, send it out for calibration regularly and check the jaw for wear before assuming the reading is correct.

4. If components keep failing randomly, use a power quality analyzer before replacing more parts

One of the most common scenarios I deal with is this: “We replaced the sensor, then the drive, then the PLC card, and the fault still comes back.” Many of those mysteries are power quality problems hiding behind component failures.

A few years ago, a customer lost a dozen sensors over two months. Our lab tested them and found most were still functioning normally. Then we connected a power quality analyzer to their incoming supply for ten days. It recorded voltage sags below the sensor's minimum operating level every time a large forklift charger kicked in. The sensors were not defective. They were undervolted.

A power quality analyzer seems expensive until it finds a problem that no one else could see. Look for a unit that logs sags, swells, transients, and harmonics over days or weeks, so you can capture intermittent events. If you don't own one, rent one or hire an electrical contractor who does. Replacing parts randomly will cost more than renting an analyzer for a week. To be fair, the analyzer doesn't fix the problem by itself—it tells you where to look. That's exactly what you need.

5. If you're hunting for hot spots, use a thermal camera and understand its limitations

Thermal cameras are one of the most requested inspection tools I see, and they come with a popular myth. Can thermal cameras see through walls? No. A thermal camera reads infrared radiation from the surface it is pointed at. It does not give you X-ray vision through drywall, insulation, concrete, or metal. Even if the camera is a FLIR model, that doesn't change the physics.

What a thermal camera can do is show temperature patterns on a surface. If an electrical terminal is loose, it will produce a hot spot on the panel before it fails. If a steam pipe is overheating a wall, you may see a warm area on the wall surface. But the camera is seeing the wall surface, not the pipe behind it.

Where thermal cameras shine in industrial maintenance: checking electrical cabinets for loose connections, scanning motor bearings, finding steam trap failures, and locating missing insulation. If this is your scenario, a compact thermal camera from a known brand like FLIR is a practical addition. Take baseline images first, then inspect the same equipment regularly so you can catch changes early.

How to tell which scenario you're in before you buy

If you're on the fence between an encoder, photoelectric sensor, caliper, power quality analyzer, or thermal camera, answer these questions first:

• Do failures happen after a restart or power loss? Consider an absolute encoder such as the SICK absolute encoder ATM60.
• Do false signals depend on the product, background, or lens contamination? Test a photoelectric sensor with your actual product samples, and include SICK photoelectric sensor options in that trial.
• Do you need a dimension value rather than just presence? Use a caliper with good ergonomics and calibration support, like the Mitutoyo CD-6 ASX caliper.
• Do failures cross multiple vendors and components? Stop replacing parts and start with a power quality analyzer.
• Are terminals, bearings, or panels getting hot? Use a thermal camera—but don't expect it to see through walls, FLIR or not.

Notice I didn't make every answer SICK. That's intentional. SICK makes excellent position and detection products, and I trust them in those roles. But quality is a system: a sensor can be perfect and still fail because of poor power, bad mounting, or missing calibration. The selection process has to start with the physical condition, not just the brand name.

What was best practice in 2020 was often “buy the same sensor we bought last time.” In 2025, the execution has changed: IO-Link diagnostics, easier fieldbus integration, cloud-enabled power analyzers, and smarter thermal cameras mean you can solve problems faster than ever. But the fundamentals haven't changed. Know your environment, verify your measurement, and test before you trust. Take it from someone who has rejected expensive first articles for not matching the actual scenario: define the problem clearly before you choose the solution. That one habit will save you more money than any brand shortcut.