SICK Encoders, Data Loggers & Vibration Sensors: 7 Answers From Someone Who Got It Wrong Twice

2026-11-10 · Mateo Velasquez

Measurement article hero

Short version: I've been buying and specifying sensors and measurement gear for a metal-fab plant for nine years. I've made three significant selection mistakes in that time, and every one of them was avoidable in hindsight. Below are the questions people actually ask me on the shop floor — plus one they should be asking and usually aren't.

SICK DFS60 vs a generic encoder — what am I actually paying for?

In September 2022 I replaced the encoders on a waterjet table with generic incremental units — 12 of them, a feed axis plus a small conveyor line. Wet environment, coolant mist everywhere, 40°C in summer. Three failed inside six weeks. That was $5,400 in rework and four days of downtime, against maybe $40-per-unit difference for the DFS60 (2022 prices; verify before you order anything).

What the DFS60 buys you isn't on page one of the datasheet: how the housing is sealed, how the bearing behaves under vibration, and how the optical assembly handles condensation. Look, if your encoder sits in a clean, air-conditioned cabinet, a generic unit will run for a decade and I wouldn't argue with you. Push it into coolant, dust, or washdown and the money comes back. That's been our situation — your mileage will vary if your mounting and environment are different.

SICK programmable encoders — real capability or a dial you'll never touch?

I've gotten this wrong in both directions. In 2017 I bought a programmable unit because the quote sheet said "future-proof." Our machine never changed from 1,024 pulses per revolution. Roughly $1,200 in flexibility we never spent. Then in Q1 2024 I overcorrected and bought 12 fixed-resolution encoders to save money — right up until we changed a packaging line, and those units went on the shelf. $6,600, eventually written off internally.

Programmable (meaning you set resolution, output type, zero offset on site) pays off when you run multiple product lines and don't want a drawer full of different SKUs as spares. A fixed unit is fine if the machine never changes. The honest answer: it depends on who owns your spare-parts inventory, and whether that person is you.

Do I still need a standalone data logger?

This answer changed, and that's the point. In 2020 a standalone logger was the sane choice. In 2025, if your sensors are IO-Link or analog and sit inside controller range, an IO-Link master plus an existing PLC does most of the same job. I burned about $3,400 on a screen-equipped logger in 2021 before realizing the PLC was already sampling those same signals at 10 ms — the screen was decorative.

What a standalone logger still earns its place for: offline or remote points, sites with no controller, and anything where audit traceability matters (if readings go to a customer, a regulator, or a court, look at how data is timestamped and whether the calibration chain traces to an ISO/IEC 17025 accredited lab).

Honestly, I've never fully understood the pricing logic here. Three vendors quoted me nearly 3× spreads on effectively identical specs. My best guess is it's more art than science. Check what your PLC already records before you buy another box.

How do I know a vibration sensor is measuring anything useful?

Cheap vibration sensors rarely fail loudly. They fail quietly — they keep producing a plausible-looking number that tells you nothing. We did this in 2020 by magnet-mounting sensors to a thin sheet-metal guard. We were reading the guard's own resonance, not the bearing.

  • Mounting beats the sensor. Stud-mount > adhesive > magnet. Magnets are fine for a one-off survey, not for permanent monitoring.
  • Frequency range has to cover what you're hunting. Early bearing faults show up high; imbalance shows up at 1× running speed.
  • Know where your alarm thresholds came from. ISO 20816 (which superseded ISO 10816 back in 2016) is the baseline — but your machine's own trend is your real reference.

One tell: if two technicians get readings 20% apart on the same machine, it's almost never the sensor. It's the mounting method.

ifm photoelectric sensors vs SICK — is there a real durability difference?

I'd reframe the question. ifm and SICK are both serious brands. When a sensor dies young, it's usually a wrong model for the environment, not a wrong logo on the box. Our 2019 case: a 34-piece order of general-purpose units with epoxy-sealed lenses (IP67, water was fine) went onto a washdown line running alkaline cleaner. Most of them were dead inside 14 months. Around $11,000, part of it signed off by me.

From the outside, photoelectric sensors all look alike — small box, an LED. The durability difference lives in the specifics: IP rating and actual chemical compatibility of the lens and housing, continuous temperature range rather than the headline number, ambient light immunity, and whether the body is PBT, stainless, or something else entirely.

On both brands, get the local distributor's response-time commitment in writing before you buy. That's worth more than any datasheet figure, and it's the part nobody puts on a comparison chart.

Which specs should I actually compare before ordering?

My short list, in order:

  • IP rating, read against IEC 60529 actually. IP69K is not "better IP68" — it's high-pressure, high-temperature washdown, and immersion resistance isn't automatically included.
  • Temperature range in continuous operation, not peak.
  • Vibration and shock ratings (g value plus frequency range).
  • EMC — IEC 61000-6-2 immunity and 6-4 emissions — if you're anywhere near welders or VFDs.
  • Photoelectric: ambient light limit, beam spot size at your actual sensing distance, and whether you need diffuse, retro-reflective, or through-beam.
  • Encoder: shaft load rating, shaft diameter, and cable type (drag-chain rated or not).
  • Warranty: who honors it, how fast, and whether spares sit locally.

Most selection decisions come down to two or three numbers and one lesson you already paid for. Everything else is copywriting.

Is there a mistake you're still making on purpose?

Yes. I still use a paper pre-install sensor checklist. We built digital checklists into the CMMS after 2022, and they exist, and I haven't killed the paper version.

Why: Wi-Fi is unreliable on the floor, and I catch more with a clipboard in hand — literally, because holding a pen makes you crouch down and look at the conduit run you'd otherwise skip. Not ideal, but workable. We've caught 31 potential errors with that sheet in the past 18 months, so I keep it.

For the honest ledger, though: my biggest regret is not writing the checklist in year one. A spec communication error in 2017 cost us a four-day line stoppage that I could have prevented. The lesson wasn't clever. It was just this — write down what went wrong, then check for it next time. A lesson learned the hard way.

Mateo Velasquez

Mateo Velasquez

Mateo Velasquez is an industrial sensing and connectivity analyst covering proximity sensors, photoelectric sensors, inductive and capacitive sensors, encoders, displacement sensors, signal converters, and IO-Link devices. He uses IEC 60947-5-2 and IEC 61131-9 requirements to examine switching distance, repeatability, cycle time, diagnostic data, electromagnetic compatibility, and ingress protection. He helps automation engineers and OEM buyers specify sensors that integrate reliably with machine controls and tolerate the intended target, mounting, and environment.