Why Emerson Temperature Sensors Can Still Give You Wrong Readings (And How I Fixed It)

I learned this the hard way: your Emerson temperature sensor is probably fine. It's everything around it that's not.

I'll never forget the phone call. It was a Tuesday in October 2023. The shift supervisor was furious because a $500,000 batch of specialty chemicals had been scrapped. The culprit, according to his report, was a faulty Emerson temperature sensor.

I drove to the plant, expecting to find a dead probe. Instead, I found a perfectly good Emerson 644H temperature transmitter paired with a Rosemount 68 thermowell. The electronics were fine. The RTD was fine. But the reading was off by 1.8°C—enough to ruin the reaction kinetics.

That batch cost $50,000 in raw materials alone. Plus a full shift of downtime. All because of something that had nothing to do with the sensor itself. (I've made this mistake myself, more than once, so I know the feeling.)

The problem you think you have (and the one you actually have)

When a process temperature reading goes south, everyone blames the sensor. I get it. It's the most obvious component. You look at the HART communicator, see a weird value, and immediately think: "Sensor's dead. Order a replacement."

Here's what I've learned in nearly a decade of calibrating and troubleshooting Emerson instrumentation: roughly 80% of the time, the sensor is fine. The issue is in the installation, the wiring, the transmitter configuration, or something as simple as a thermowell with the wrong insertion depth.

I used to swap sensors first and ask questions later. Looking back, I should have started with the basics. At the time, the pressure of keeping a production line running made me jump to conclusions. (We've all been there, right?)

The real reasons your Emerson temperature measurement is drifting

1. Your thermowell is creating a heat sink you never accounted for

This is the #1 mistake I see. You install a heavy-duty thermowell to protect your Emerson sensor from harsh process conditions. But that same thermowell acts like a heat sink, wicking heat away from the tip. If the thermowell is too short, or if it's not fully immersed in the flow, the sensor reads a mix of process temperature and ambient temperature.

In that October 2023 incident, the thermowell was only inserted 150mm into a 300mm pipe. The tip was in the flow, but the stem was exposed to the pipe wall temperature, which was about 5°C cooler. The sensor averaged the two.

I once ordered a batch of Emerson 68 thermowells with standard lengths without checking the pipe diameter. Every single one was too short. $1,200 in hardware, straight to scrap. (Ugh.)

2. Your wiring is picking up noise, and you don't even know it

Emerson temperature transmitters are designed to reject noise, but they're not magic. If you run your 4-20mA loop alongside high-voltage power cables, or if you use unshielded twisted pair in a noisy environment, you'll see drift.

I once spent three days troubleshooting an Emerson 644H that would read 22.3°C for hours, then suddenly jump to 28.5°C for no apparent reason. The transmitter was fine. The sensor was fine. The problem was a corroded ground lug on a nearby VFD. The noise was coupling into the thermocouple extension wire.

We said 'standard shielded cable.' The electrician heard 'any two wires will do.' Discovered this when the order arrived and the cable had no shield at all. (Saved $80 on the cable, wasted $400 on troubleshooting labor.)

3. Your transmitter configuration doesn't match your sensor type

This one is embarrassingly common. You order a Emerson temperature transmitter configured for a Pt100 RTD, but you connect it to a Pt1000. The reading will be off by roughly 10x.

Or worse: you use the default sensor curve in the transmitter, but your RTD has a different alpha (temperature coefficient). Emerson transmitters are smart, but they can't read your mind. They apply the curve you tell them to apply.

In 2022, a startup plant commissioned their entire reactor train with Emerson 3144P transmitters. The configuration file had the wrong sensor type for Zone C. It took two weeks and about $15,000 in process re-validation to fix. (The upside was a new checklist for the configuration team. The risk was repeating the same mistake on the next project. I kept asking myself: is the time saved worth potentially losing the client's trust?)

What happens when you ignore these issues (I've seen it)

Let's talk consequences, because I've paid for most of them with real budget dollars.

Impact on process control: your PID loop goes crazy

A drifting temperature reading means your controller thinks the temperature is lower than it actually is. It calls for more heat. The heater runs harder. The overshoot gets worse. The system oscillates. You end up with a loop that can't hold setpoint.

I once watched a jacketed reactor cycle between 180°C and 210°C because the Emerson sensor was reading 5°C low. The controller was pushing the heater to 100% output, trying to reach a target that had already been overshot. The batch was ruined. (This was back in 2020; the cost was $3,200 in redo plus a 1-week delay.)

Impact on quality: specs get missed

In many processes, temperature tolerance is tight. A sensor that's off by 2°C might mean the product is out of spec. You might not even detect it until the lab analysis comes back 24 hours later.

That's what happened at a food processing plant I worked with. Their Emerson temperature transmitter was installed in a pipe with poor flow mixing. The sensor read the cold spot, not the average temperature. The pasteurization step never reached the required holding time for the coldest particle. The entire batch had to be reprocessed. The client lost a day of production and nearly lost a major retail contract.

Impact on safety: worst-case scenario

I don't like talking about safety close calls, but they happen. A runaway exothermic reaction in a 10,000-liter reactor is terrifying. A sensor that reads low will delay the alarm. By the time the operator sees the high temperature, it's too late.

Calculated the worst case: a $500,000 vessel rupture. Best case: a $50,000 emergency shutdown. The expected value said fix the sensor issues, but the downside felt catastrophic. We fixed them.

So what should you actually do? (Spoiler: it's not 'buy a better sensor')

By now, you probably realize that replacing your Emerson temperature sensor with another Emerson sensor isn't going to fix a thermowell or wiring problem. Here's what I've learned to do, and what I now teach our new technicians.

1. Check the mechanical installation first

Before you touch the electronics, verify the thermowell insertion depth. The rule of thumb: the tip should be in the middle third of the pipe cross-section. For pipes under 4 inches, use a thermowell that extends to the centerline. For larger pipes, at least 100mm into the flow.

Also check for steam or condensate in the thermowell. A wet well will read erratically. (I've seen this on steam lines where the thermowell wasn't sealed properly.)

2. Validate your wiring and shield grounding

Use shielded twisted-pair cable for your Emerson temperature transmitter. Ground the shield at one point only—usually at the transmitter end. Grounding at both ends creates a ground loop that can introduce noise.

If you're using thermocouples, make sure you're using the correct extension wire. Type K extension wire for Type K sensors, Type J for Type J, and so on. Don't guess—check the spool label.

3. Confirm the transmitter configuration

Use an Emerson HART communicator or AMS Device Manager to read the sensor type, the connection type (2-wire, 3-wire, 4-wire), and the input range. Make sure they match the physical sensor.

A quick sanity check: measure the resistance of your RTD at room temperature with a multimeter. For a Pt100, you should see about 108-110 ohms at 20°C. If you see 20 ohms, you've got a 10-ohm copper sensor configured as a Pt100. (This happened to a colleague in 2021. The transmitter was reading -40°C instead of 20°C. Took her a day to figure out.)

4. Perform a loop calibration and a sensor simulation

Calibrate the full loop: disconnect the sensor, connect a precision resistor or a thermocouple simulator, and verify the 4-20mA output. Then reconnect the sensor and compare the reading to a known reference (like a calibrated handheld thermometer placed in the same location).

This is the only way to isolate whether the error is in the sensor, the transmitter, or the wiring. (I've missed this step before and wasted hours chasing a problem that didn't exist.)

5. Document everything (trust me)

After the third rejection in Q1 2024, I created our pre-check list. It covers: insertion depth, thermowell condition, wiring type, shield grounding, transmitter configuration, and calibration status. We've caught 47 potential errors using this checklist in the past 18 months. Each one would have caused a shutdown.

This is the kind of thing that seems tedious until the day it saves your batch. (And your budget.)

Additional resources that can help

Getting a handle on your temperature measurement doesn't stop at field troubleshooting. The broader Emerson ecosystem offers tools for systematic improvement. For a comprehensive overview of the portfolio designed to solve these exact issues, including the latest in sensor technology and digital diagnostics, you should explore the full emerson products lineup available.

For those looking to master the most common measurement parameter, specifically focusing on thermal processes, reference materials on the emerson temperature sensor range provide deeper dives into RTD and thermocouple best practices.

Final thought (and some self-disclosure)

I'm not saying Emerson sensors are perfect. No equipment is. But the sensor is the most tested and most reliable component in the loop. The weak links are the things we assume are fine: the thermowell, the wiring, the configuration, the installation.

I learned this in 2018 when I blamed an Emerson 644H for a bad reading, only to discover a loose terminal screw on the head assembly. (The sensor was fine. The transmitter was fine. The screw was loose.) That lesson cost me a half-day of troubleshooting and a bruised ego.

If I could redo that decision, I'd invest in better upfront specifications. But given what I knew then—nothing about the subtleties of thermowell immersion depth—my choice was reasonable at the time. Now I know better.

This was accurate as of March 2025. Technology and Emerson's product line change fast, so verify current specifications, calibrations, and support documents on the Emerson website before budgeting for new installations or replacements.