How to Read a Fluke Multimeter: Field Notes from an ifm Sensor Audit

It started with a routine audit. Third week of April, 2024, a regional quality auditor came to verify our instrument calibration program. I expected the usual paperwork, a few questions, maybe a walk down to the line. Instead, I got a nonconformance that's still influences how we work today.

When the auditor asked for proof

Our documentation for the ifm flow sensors, the ifm absolute encoders, the pressure transmitters, all of it basically looked fine on paper. Manufacturers' certificates were filed. Calibration stickers were current. But then the auditor asked a question I couldn't answer well:

“How do I know your sensor actually reads what the display says? Show me the last verification record with raw readings.”

I pulled the logbook. It had checkmarks, initials, dates. No raw numbers. We'd been writing “passed” in the results column and never recorded the actual mA values. The auditor's finding said: “Inadequate metrological traceability. No independent evidence that critical measurement loops were verified.”

On one hand, he was right. On the other hand, I was frustrated. These were ifm SBN257 mechatronic flow sensors with built-in diagnostics and IO-Link. They tell you when the media temperature changes, they flag empty pipe, they self-monitor. For the ifm absolute encoder on the gantry axis, we had a certificate straight from the factory. How much more proof did he need?

Turns out, he needed proof that the whole loop worked. The sensor could be perfect in a test bench and still give a wrong signal on the machine if the cable was chafed, the input card was drifting, or the grounding was noisy. I couldn't argue with that logic. I just wished he'd said it a bit more gently.

The fight about the reference instrument

For a week, our quality manager and I went back and forth. He wanted to send every critical sensor back to the factory for recertification. I thought that was overkill and slow. We had a production schedule to keep, and the audit was about documentation, not about actual sensor failure.

“Look, no one found a faulty sensor,” I said. “We need independent, traceable verification of the installed loops. Not new certificates. We need to prove the installed system reads correctly.”

So we compromised. We bought a Fluke 117 multimeter, checked that it had a valid calibration certificate, and built a verification protocol. That's when the real story started.

How to read a Fluke multimeter when verifying a 4–20 mA loop

I've used multimeters for years, but honestly, the first time I used ours for loop verification, I still had to pause and think. If you're doing the same thing, here's how it works:

1. Plug the leads correctly. For current, red lead goes into the 10 A or 400 mA jack, black into COM. For voltage and resistance, red goes into the VΩ jack.
2. Select the right range. On a Fluke 117, turn the dial to mA if you're reading a 4–20 mA signal. For voltage, use V with the DC/AC button set correctly.
3. Put the meter in series for current. Don't clip across the loop. You break the loop, insert the meter, and let the current flow through it.
4. Read the display and record the value. 12.34 mA means roughly 52% of full scale. 4.00 mA means zero. 20.00 mA means 100%.
5. Write it down. The auditor wants to see the raw number next to the calculated value. Don't just write “OK”.

That last point was the whole lesson. We had been trusting the green LED and the local display without recording anything. The multimeter gave us a second opinion. It turned out we needed one.

The unexpected findings

We started with the ifm SBN257 flow sensor on the cooling water line. The sensor itself read 8.24 mA, which converts to about 26.5% of scale. The PLC display said 8.24 mA. Both agreed. Great.

Next, we verified the ifm absolute encoder. The encoder sends pulses to the drive; we used the multimeter to check the output signal on the A and B channels. The readings looked stable, but when we checked the wiring physically, we found a chafed cable shield near the cable carrier. That explained the intermittent position drift we'd been chasing for weeks.

We also pulled out a tabletop centrifuge in the coatings lab one afternoon—not part of the original audit scope, but the lab manager heard we had a calibrated meter and asked us to verify the speed sensor. That gave us a nice side lesson: the centrifuge's own RPM display was fine, but the imbalance switch was set so loosely that it would have let a badly loaded rotor run. A simple check caught it before it became a bigger problem.

The truth was that most of our sensors were perfectly fine. The ifm SBN257 flow sensor was fine. The absolute encoder was fine. The input cards were fine. But the systems they were part of had issues: a damaged cable, a poor mounting bracket, a borderline grounding point. We never would have found those by sending the sensors back to the lab.

The turning point: we trusted the tool, not the label

My favorite moment was when the senior tech suggested we just replace the encoder. “It's an old ifm absolute encoder,” he said. “Cheaper to swap it than to keep testing.”

I had to stop and think. Replacing a part is sometimes the right call. But the multimeter told us the encoder output was clean. The problem was the cable. Swapping the encoder would have cost us a few hundred dollars and probably still left us scratching our heads when the drift continued.

I have mixed feelings about that decision even now. Part of me wants to say “always diagnose before you replace.” Another part remembers the painful times when I didn't have a calibrated meter and a spare part was the only quick path. But the evidence was clear enough. We replaced the cable, added a strain relief, and the position drift disappeared.

That's why I'll take a Fluke in my kit over a spare encoder any day. Reading the mA value at the sensor, comparing it to the PLC word, and writing the two numbers side by side gives you a conversation no status LED can have with you.

What the 1777 power quality analyzer added

We didn't stop at loops. The audit pushed us to look at the electrical environment as well. We borrowed a 1777 power quality analyzer to record the supply to the VFD that feeds the pump.

That caught something none of us expected: occasional voltage sags and harmonic distortion that lined up with the flow instability we had been blaming on the ifm sensor. The sensor was reacting to real process changes caused by the drive acting up. We put a line reactor on the drive input and the flow readings settled down.

So the diagnosis chain ended up being: ifm sensor was telling the truth, the encoder cable was lying, and the VFD was making the pump misbehave—all of it discovered because we took the time to read the actual numbers instead of trusting the “everything's fine” lights.

Bottom line

The audit taught me something I'll keep using for the rest of my career.

You don't verify a sensor by trusting it. You verify it by measuring the signal and comparing it to a known reference, then writing both numbers down.

Now every critical loop gets the same treatment:

• Independent meter reading, recorded in mA or V
• Calculated engineering value, compared to the indicator
• Signature and date on the verification record

That binder on the shelf has saved us at least two unnecessary sensor replacements and one very expensive process restart. The auditor came back for a follow-up in June and looked at it. The finding was closed. No comments.

If you're in a similar spot, take my advice: buy a decent multimeter, learn how to read it properly, and don't trust the certificate alone. The ifm sensor certificate is nice, the internal diagnostics are helpful, but the number that comes out of your meter, with your initials next to it, is what actually proves your instrument was working that day.