When a scale starts drifting or the indicator flashes a load cell error, everyone wants to blame the cell. I used to think the same way. When I first started reviewing weighing systems, I assumed a bad reading meant a bad load cell. After a few hundred hours inside junction boxes and cable trays, I learned the cell itself is usually the last thing to fail—not the first.
I'm a quality manager at an industrial instrumentation company. We integrate load cells, pressure transmitters, and control systems for process plants, and I sign off on roughly 150 to 200 weighing systems a year before they ship. When a Rice Lake load cell comes back with questionable readings, this is the checklist we walk through. It has six steps, in order. Most of the time, you'll find the problem before step four.
Step 1: Visual and physical inspection
You can skip this step and still find the problem eventually, but it's the cheapest test you'll ever run. Look at the cell before you touch it with any meter.
- Check for dents, bent overload stops, or rust on the cell body.
- Make sure the load point is actually aligned. A cell that's side-loaded because the scale deck shifted produces exactly the kind of drifting readings that look like an electrical fault.
- Inspect cables for cuts, crush marks, or a strain relief that's pulled out of the connector.
- Look at the connector pins on the cell and the junction box. Green or black corrosion is a short answer.
- Check that nothing is touching the cell that shouldn't be—a snagged cable, a stray piece of process debris, a coat of cured paint.
We had a batch of load cells pulled from a skid scale a few years ago, all reading erratically. Turns out one cell was sitting on a scrap of angle iron left under the deck during installation. We leveled the base, put the same cells back, and the readings settled immediately. No replacement needed.
Step 2: Open the junction box
Moisture is the number one killer of load cell systems, and the junction box is where it hides. Even a sealed enclosure can draw in condensation through the conduit. Before you start measuring resistance, open the box and look at what's inside.
- Check for water droplets or residue on the terminals.
- Look for white or green corrosion on the screw terminals.
- Verify that every screw is tight. Loose terminal screws are embarrassingly common.
- Check the cable between the junction box and the indicator, especially at the entry points. We've found more than one cable crushed by a conduit fitting.
This is also where a thermal camera earns its keep. A FLIR One Edge Pro, the wireless model, is what I keep in my bag. Scan the junction box and the cable run while the system is powered. A warm spot at a terminal usually means a high-resistance connection. It's the kind of intermittent problem that shows up at 2 a.m. but reads fine at 9 a.m. There's something satisfying about catching one of those early, instead of getting the phone call after the line stops.
(Should mention: if the system is in a hazardous area, check your site's hot-work and ex-zone rules before opening anything powered. I won't pretend to know your site's safety requirements—that's a separate discipline.)
Step 3: Electrical tests with a multimeter
Now you're measuring. Disconnect the suspect cell from the junction box first. You're testing the cell, not the wiring, and you want a clean reading. My daily carry is a 114 electrical multimeter—it's compact, CAT III rated, and good enough for the field. Honestly, any meter with the right rating will do the job. The 114 is just what fits in my bag.
Here are the numbers we look for, typical for a Rice Lake load cell:
- Input resistance (excitation to excitation): usually 350 Ω, though some models are 700 Ω or 1000 Ω. Check the data sheet.
- Output resistance (signal to signal): should match the input within a few ohms. If it's off by 20 Ω or more, the strain gauge bridge is damaged.
- Insulation resistance (each wire to the cable shield and cell body): this should be in the megaohm range. If you're reading less than 1 MΩ, moisture has gotten inside the cell. If I remember correctly, our incoming inspection rejects anything under 20 MΩ, but the exact threshold should come from your spec.
Two notes. First, test the insulation resistance at the cell cable itself, not through an extension. Second, never use a high-voltage insulation tester—a megger—on a load cell. It will destroy the strain gauge. A standard multimeter's resistance range is the right tool.
While you have the meter out, check the excitation voltage at the junction box. If the indicator is supposed to supply 10 V and you're reading 8.2 V, that alone explains low output and drift. It's a simple check that people skip because they assume the indicator is fine.
Step 4: Zero balance with no load
Reconnect the cell, apply no load, and power up the indicator. Measure the DC millivolt output between the signal leads. For a 2 mV/V cell with 10 V excitation, a no-load output near 0 mV is healthy. If the zero is off by 10% or more of rated output, something is mechanically pre-loading the cell, or the cell has been damaged.
One thing I've learned: a zero balance that changes every time you read it points to the wiring or connector, not the cell. A consistent offset is more likely mechanical. Temperature drift is real too—if the zero shifts as the sun moves across the scale, suspect uneven thermal expansion of the mounting structure before you fault the cell.
Step 5: Live load and corner tests
This is where the cell proves itself under real conditions. If you have test weights, apply them at 25%, 50%, 75%, and 100% of the cell's rated capacity, and plot the readings. If the response isn't linear—50% reads 48%, 100% reads 96%—the cell is suspect.
For multi-cell systems, run a corner test: put the same weight on each corner of the deck and compare the readings. Check tolerances against the manufacturer's documentation. Rice Lake publishes theirs, and we refer to NIST Handbook 44 for legal-for-trade applications in the U.S. If you're selling equipment internationally, OIML R60 is the recommendation to reference.
This step needs a calibrated indicator. If the indicator is lying, the cell will look bad no matter how good it is. When we see the same error across all corners, the indicator is usually the culprit, not the load cells.
Step 6: Cross-check against process instruments (the step everyone skips)
Here's the one that saves us the most time, and it's almost always overlooked. If the scale reading conflicts with what the process says it should be, don't automatically trust one over the other. Verify with a known-good reference.
For a tank that uses weight for inventory, measure the head pressure with a dwyer pressure sensor or a dwyer digital differential pressure gauge and convert it to level. If the level says the tank is half full but the load cell says it's 500 kg heavier than it should be, one of them is lying. The disagreement is the information.
We had a skid scale showing a 15% weight drop every night. Everyone assumed the load cells were shot. The electrical tests passed, so we put a dwyer digital differential pressure gauge on the tank—and the level was dropping too. A leaking discharge valve was draining the tank, not a failed load cell. We dodged a $15,000 unnecessary replacement.
I should add that this cross-check works the other way as well. If the load cells are healthy but the pressure transmitter has drifted, you've found the real problem. In my opinion, every weighing system should have at least one independent way to sanity-check the reading. It's cheap insurance.
Common mistakes and final precautions
- Don't trust the indicator's error message blindly. Its diagnostics usually describe the symptom, not the cause.
- Never hi-pot or megger test a load cell. Repeating this because it costs people money.
- Don't mix cable types when replacing a damaged cable. Different cables have different resistances and shield constructions, and that shifts the bridge balance.
- Torque the mounting bolts to spec. An over-tightened bolt introduces preload that reads as a permanent weight offset.
- If the junction box has visible corrosion, replace it. Cleaning the terminals doesn't repair corroded traces inside the board.
- After a load cell replacement, schedule a full calibration. Zeroing the indicator is not calibration.
- If nothing in this checklist catches the problem, call Rice Lake's technical support with your readings in hand. They know their cells, and they've seen every weird failure mode you can imagine.
This checklist reflects what works in our installations as of early 2025. We're a mid-size integrator working mostly in climate-controlled process plants. If you're dealing with marine environments, high humidity, or ATEX zones, your failure modes and safety rules will be different—some of these steps will apply, some won't.
And one more thing: know your boundary. I test load cells all day, but I don't pretend to know every indicator's programming menu. When the problem is on the indicator side, we call in the vendor's specialist. A supplier who says "this isn't our strength—here's who does it better" earns more trust than one who claims to handle everything. That applies to us too.
