SICK Absolute Encoder AFM60A, Rotary Encoder, and Other Sensors: Field Notes on Reliability and RTN

What you'll find here

I've been on the receiving end of "the production line just stopped" calls more times than I can count. When a critical sensor fails and you've got a machine down, the last thing you want is a datasheet. You want someone who's seen it fail, fixed it, and knows what works when it matters.

This isn't a textbook. It's a collection of questions I get asked regularly about SICK sensors—and a few others you probably didn't know to ask.

1. What's the deal with the SICK AFM60A absolute encoder? Is it worth the premium?

If you're asking whether the AFM60A is worth more than a standard incremental encoder, the answer depends on one thing: do you need to know the exact position after a power loss?

I've used the AFM60A (part number series 1036611, if you're ordering) on a few large-scale projects. After a power cycle—and in industrial environments, that happens more often than you'd think—it remembers the position. No homing cycle. No scrap. No recalibration.

In March 2024, I had a client whose whole packaging line went down after a brownout. They were using incremental encoders on the rotary cutters. Every splice was off by 3mm. Took them 2 days to re-home everything. With an AFM60A? It would have been a 15-minute restart.

I don't have hard data on ROI for every application, but based on the two dozen installs I've been involved with, the breakeven is usually around 3-4 unplanned outages. If your line loses power once a quarter, it pays for itself in a year.

2. How do I choose between a SICK rotary encoder and something cheaper?

I get this question a lot. The short answer: match the encoder to the environment, not just the price tag.

SICK rotary encoders—like the DFS60 or the DBS50—tend to have better ingress protection (IP67 is common) and better shock/vibration specs. I've pulled apart a failed "budget" encoder that had corrosion on the bearings after 6 months in a washdown environment. A SICK encoder in the same spot ran for almost 3 years before needing replacement.

My experience is based on about 80 installations across food processing, packaging, and material handling. If you're working in a clean, temperature-controlled lab, a cheaper option might work fine. But the 'cheap' encoder cost us $120 in downtime per replacement. The SICK cost $240 upfront but saved us $600 in labor and lost production over its lifespan.

That said, I should note: not every application needs IP67. A simple dust cover on a budget encoder can sometimes bridge the gap if budgets are tight.

3. Can I use my iPhone as a thermal camera for checking sensor heat?

There's a reason this keeps popping up in my search terms. People want a cheap way to diagnose sensor issues—and an iPhone with a thermal attachment is way cheaper than a professional unit.

I'm not 100% sure about the latest models, but as of early 2025, the FLIR One Pro (around $300) works well with iPhones. It's fine for identifying hot spots on panels, motors, or even a flow sensor that's running hot. But here's the catch: its resolution is roughly 160x120. For precise temperature measurement on a small component—like the electronics inside an encoder—you're likely to get a reading that's off by ±3°C.

Take this with a grain of salt: I've used an iPhone thermal camera to find a failing bearing on a conveyor motor. It worked. But when I needed exact temps for a quality audit, I borrowed a proper Fluke Ti401 Pro ($8,000 class). The iPhone camera said "around 65°C." The Fluke said 61.2°C on the bearing face and 68.7°C on the housing.

So: yes, you can use it. Just don't rely on it for precision. It's a diagnostic hint, not a measurement.

4. How do I test a flow sensor that seems to be giving bad readings?

I still kick myself for the time I swapped a $600 flow sensor without doing the basics first. Turned out it was an air bubble in the line, not the sensor itself.

When I'm triaging a suspect flow sensor—SICK's range includes the FDK and FTM series for gases and the FQS for liquids—I run through a quick checklist:

1. Visual check: Is the display showing an error code? SICK sensors usually flash a specific pattern. Check the manual, but a common one is two flashes = power issue, three = signal out of range.
2. Clean the sensing element: Flow sensors I've pulled from coolant lines often have buildup. A gentle wipe with isopropyl can fix a false low reading.
3. Verify the pipe diameter: I once found a 1-inch sensor installed in a 1.5-inch line. The flow rate was correct, but the cross-sectional area was wrong—so the sensor's calculated flow was off by a factor of 2.25.
4. Test with a known good sensor: We keep a spare SICK flow sensor on the shelf. Swapping them takes 10 minutes and proves whether the sensor is the problem or the system is.

If the sensor passes all those, then look at the PLC program. I've seen more than one "bad sensor" that was actually a scaling mistake in the logic.

5. How to test a Rice Lake load cell (and what to watch for)?

I know this seems off-topic for a SICK-focused article, but I've had enough engineers ask about Rice Lake that it's worth addressing. Load cells fail differently than encoders or photoelectric sensors.

Rice Lake is a solid brand—their RL series is common in bench scales. The most common issue I see is a zero drift. If your scale reads 0.5 kg with nothing on it, that's a load cell issue, not a display issue.

Here's the test I use:

1. Check the power supply: Load cells need stable excitation voltage (usually 10V DC). A bad power supply causes erratic readings.
2. Measure the output: With no load, the signal should be near 0 mV. With a known weight (we use a 10 kg certified weight), the output should match the spec. For a Rice Lake with 2 mV/V sensitivity at 10V, that's 20 mV at full load. A 10 kg weight on a 50 kg cell should give about 4 mV.
3. Inspect the cable: I've traced intermittent failures to a kinked cable. The shield was broken inside the jacket.
4. Look at the mounting: If the cell is bolted down unevenly, it introduces a preload that shifts the zero point.

If you're doing this in a rush—because a scale went down during shipping—the first step is always power. I'd say 30% of the "load cell failed" calls I've taken resolved with a simple power cycle or cable reseat. Don't order a replacement until you've checked that.

6. Why do SICK sensors seem to have a higher failure rate in some applications?

I've heard this complaint. And it's usually not a quality issue—it's a selection issue.

Here's a real example from last year. A client had SICK photoelectric sensors failing on a case packer. They'd last maybe 2 months. The first thing I asked: "What model are you using?" Turns out they'd bought a WTB4-3P (a standard diffuse sensor) with a sensing range of 1.3 meters. But their case packer had high-dust levels and the cases were dark brown.

A diffuse sensor in a dusty environment loses range. A better choice would have been a retro-reflective sensor (WL4-3) or a background-suppression model (WTT4). We swapped them, and those sensors are still running fine 10 months later.

In my opinion, most "premature failures" I've seen with SICK sensors come down to either wrong model selection or incorrect mounting (vibration loosening a bracket, then the cable gets stressed). The hardware itself is solid.

7. When does a smaller order for SICK sensors make sense—and when doesn't it?

When I was starting out in maintenance, the vendors who took my small orders seriously are the ones I still call for my big projects.

I've ordered single SICK encoders ($200-400) for emergency replacements. Some distributors treat a single-unit order like a hassle. The good ones don't. A distributor who's willing to process a one-off order for a $250 encoder at list price is a distributor I trust with a $5,000 annual contract.

That said, small orders have a hidden cost: shipping. A single M12 connector might be $8, but with overnight shipping, you're paying $30 to deliver it. If you know you'll need sensors down the line, a small buffer stock—even 1 or 2 units of your most common models—saves you the rush fee.

Small doesn't mean unimportant, it means potential. The client with a $200 encoder order today might be the client with a $20,000 line upgrade next year. Treating small orders with respect is just smart business.