The 47 Load Cells That Weren't Broken: Rice Lake Load Cell Wiring and Weighing Indicator Calculation Lessons

The Load Cell That Wasn't Broken

Last spring, a returned load cell reached my bench with an RMA note that said faulty—erratic reading. By the time that cell got to me, it had already cost someone around $1,850 in replacement parts, freight, and service time. It passed every check we ran: zero balance, bridge resistance, insulation resistance, output response. It wasn't faulty. And the scale it came from was still not fixed.

I'm on the quality team at Rice Lake Weighing Systems. My job includes deciding whether a returned load cell is genuinely defective or whether something else caused the symptom. Roughly 300+ assemblies and returned units cross my bench each year. The sign above my bench says check everything else first. That's not a slogan.

In our Q1 2024 audit, we went through 142 load cells returned from the field as defective or erratic. Forty-seven of them, 33%, met factory specifications. The telling part was that they didn't come from random sites. They came from installations where something had changed recently: a junction box opened, an indicator replaced, a cable repaired. The load cell got the blame because it was the part with wires going into it.

Why the Diagnosis Goes Wrong

A load cell that genuinely fails leaves a signature. An overload usually shifts zero and changes bridge resistance. Moisture intrusion shows up as low insulation resistance. If the cell reads clean when you disconnect it and measure it directly, the sensor is probably fine. The fault is elsewhere in the weighing circuit.

Wiring mistakes that mimic a dying sensor

The clearest pattern in that batch was intermittent contact. It's the most aggravating problem in weighing because it usually disappears while you touch test leads and comes back once vibration starts. A loose terminal, a corroded splice, a wire flexed one time too many: all of them can pass a static resistance check and still make a scale erratic.

Ground the cable shield at one point. Just one. If a shield is grounded at both ends, it becomes a path for small stray currents, and those currents show up as an offset voltage. In a plant full of motor drives, that offset doesn't stay still.

Next, the sense wires. They are easy to ignore and easy to jumper out. On a six-wire load cell, sense leads let the indicator measure voltage where the cell is, not just where the indicator is. When you defeat them, the indicator controls a clean 10 V at its terminals and assumes the cable isn't dropping any of it. On a short cable, that's a fair assumption. On a long run, or a cable routed through a hot plant, the resistance of copper changes with temperature. The scale drifts. The load cell didn't drift. The wire did.

Water is the third suspect. It doesn't have to rain on the scale. Condensation inside a junction box, washdown water creeping past a gland, or even high humidity can lower insulation resistance enough to make a good load cell look intermittent. It will pass a bench test after it dries out, which makes it even easier to misdiagnose.

The indicator calculation that gets skipped

The second pattern is less physical. It's a number in the indicator setup menu that doesn't match the load cell bolted to the scale. An indicator doesn't know what's attached to it. It relies on setup values and calibration. If you put a 3.0 mV/V cell where a 2.0 mV/V cell used to be, same capacity, and the indicator was set up for the old cell, the display will read roughly 50% high. The sensor is fine. The math is wrong, and it will be wrong consistently. Consistent errors are dangerous because they look trustworthy.

The basic weighing indicator calculation hasn't changed in decades:

Full-scale signal (mV) = excitation voltage (V) x rated output (mV/V).

At 10 V, a 3.0 mV/V cell gives 30 mV at full capacity. If the indicator was calibrated for a 2.0 mV/V cell, it is expecting only 20 mV at that capacity. The extra 10 mV doesn't mean the sensor is overperforming. It means the displayed weight is wrong. Setup errors are easy to dismiss until a known test weight says otherwise. That's why we always do the math before swapping hardware.

What a Wrong Diagnosis Really Costs

Why does this matter? Because replacing a load cell doesn't fix a wiring fault or a bad setup value. It just adds cost to an unsolved problem.

One site we worked with last year chased an erratic scale for six weeks. The contractor replaced the junction board, then the indicator, and finally returned the load cell to us. The load cell passed. When we asked about the cable, the maintenance lead remembered it had been spliced after a forklift hit a guardrail. The splice was inside a conduit where washdown water collected. The fix was a new cable and four hours of labor. They had spent nearly $9,000 getting to that point.

Then there's the quiet version. A 1% reading error doesn't stop a line, so it can run for months. On a 500 lb batch scale running 100 batches per day, 1% low means 5 lb of extra material per batch—500 lb per day in giveaway. Across 200 operating days, that's 50 tons from one line. Not from a failed load cell. From a number nobody checked.

If the scale is in legal-for-trade service, the same hidden error is an audit and liability risk, not just a material loss.

Before You Replace Anything, Run a 15-Minute Check

Here's what we ask customers to do before we authorize a return. It takes a multimeter, a datasheet, and fifteen minutes.

1. Read the cell directly. Disconnect it at the junction box and compare input and output resistance with the datasheet. While the meter is connected, wiggle the cable. If the reading jumps, the problem is in the cable or the connection, not the strain gauge.
2. Measure excitation at the cell. If the indicator is delivering 10 V at its terminals but the cell sees something meaningfully lower, the cable is eating voltage. That's a wiring issue, not a load cell issue.
3. Open the junction box. Look for water, corrosion, loose screws, and wires terminated on insulation. Ask what changed recently. In our experience, the answer to what changed is usually the answer to what broke.
4. Do the calculation. Full-scale mV = excitation voltage x rated output. Compare that with the indicator setup. If they don't agree, stop and fix the setup before you order anything.

When our support team started using this sequence in 2022, the number of no-trouble-found returns dropped noticeably. Not because load cells got better, but because the rest of the system got checked.

The Rice Lake load cell wiring diagram for your model and the Rice Lake weighing indicator calculation example are in the product manual. Use those references before you use a return authorization.

The Fundamentals Haven't Changed; the Hardware Has

Load cell wiring is still about a complete bridge, one shield ground, no moisture, and honest numbers in the indicator. What changed is what we expect from the readout. Modern indicators can resolve tiny signals that older analog meters simply hid. That's not a problem with the equipment. It means that old compromises—a splice that was good enough, a ground that was close enough, a setup value that was never verified—now show up in the weight.

Can a load cell fail? Absolutely. Overload, lightning, fatigue, or contamination can kill one. But when one in three returned cells passes factory specs, the industry standard diagnostic habit deserves a second look. The sign above my bench says check everything else first. It's there because, more often than not, the load cell is the last thing to fail.