The Load Cell That Wasn't Broken: A Costly Weighing Mistake on a Potato Chip Line

The Tuesday Everything Looked Fine

It was August 2019. I was standing on a catwalk above a potato chip packing line, watching 50-gram bags fill one after another. The scale under the packing head was drifting. Not a lot—maybe two grams—but enough that the in-house checkweigher flagged it twice already. The line was running at 120 bags per minute. The checkweigher was rejecting about one bag every two minutes. Not a storm, but enough to keep a QC guy standing there with a clipboard.

The plant manager, Dan, was standing next to me. 'It started Monday,' he said. 'Yesterday it was fine. Today it's not.'

I'd been doing industrial scale service for about five years then. I was confident. Probably too confident.

'It's the load cell,' I said.

That sentence cost $4,800.

What I Did First

I ordered a Rice Lake load cell—the brand the plant spec'd for that scale. I checked the part numbers, pulled the old one, put the new one in, and ran a quick calibration. The readout held steady for about an hour. Dan shook my hand. I drove off feeling like a hero.

Three days later, Dan called. 'It's back.'

So I did what a lot of techs do: I threw more parts at it. New junction box. New cable. I even replaced the digital indicator's power supply. Another $1,100 in parts, plus my labor. The drift still came back, usually in the afternoon.

I called Rice Lake technical support somewhere in the middle of that. They walked me through the standard diagnostics: check the millivolt output, check corner load, check for binding. Everything looked good on paper. The load cell was within spec. I still swapped it anyway, because I'd already told Dan it was the load cell, and backing down felt harder than replacing it. That's a terrible reason to spend $620.

I won't guess how many times this pattern repeats in a typical year. I also don't have hard data on how many scale drift cases are actually air-related, but based on the 20-plus service calls I've been on, it's now the first thing I check.

The T865 Thermal Camera Changed the Diagnosis

A senior electrician named Mike was in the plant that week for a different job. He carried a Fluke T865 thermal camera—not our brand, but it's what he had. He walked up to the scale and spent about 30 seconds looking at the camera screen.

'You're chasing the wrong thing,' he said.

He pointed at the display. There was a clear temperature difference between the left and right sides of the scale base. I still didn't get it.

'So the base is warmer on one side. So what?'

'Follow the air,' he said.

He crouched down and held his hand under the hopper. There it was: a constant stream of air coming from a wall vent. The air was pushing up on one side of the hopper, creating enough differential pressure to change the tension on the load cell by a fraction of a percent. In a 50-gram bag, that's a couple of grams. Enough to fail a checkweigh.

Afternoons were worse, which should have been a clue. The HVAC system ramped up in the afternoon heat, so the airflow got stronger. I didn't connect it until Mike showed me.

The final fix cost about $40: a flexible duct to redirect the vent airflow, plus a small baffle. Took us 20 minutes.

The 'Class F' Mystery

Here's the part that still embarrasses me. The original specification sheet had a line that read: 'Class F Rice Lake weighing systems.' I had never asked what Class F meant. I assumed it was a product grade or maybe a customer's internal code. It wasn't until after the air fix that I sat down with the plant's maintenance engineer and asked him.

'Class F is our food-grade classification,' he said. 'It means all parts in contact with the product must be food-safe and easy to sanitize.'

I had replaced a perfectly good load cell because I didn't understand what a label on a drawing meant. The scale's accuracy class had nothing to do with Class F. The load cell was never broken. The junction box was never broken. The cable was never broken.

What most people don't realize is that load cells don't fail as often as they get blamed. They're the only electronic part in the force path, so they're the natural suspect. But the real culprit is usually moisture, overload, or—in this case—air movement.

The Bottom Line

Here's my gross math for that one job:

• New load cell and shipping: $620
• Junction box and cable: $1,100
• Power supply: $180
• My labor across three visits: roughly $2,900
• Plant downtime during peak production: six days

I don't count the lost supermarket order Dan mentioned later. That was between him and his customer. But it wasn't a small amount.

The old rule of thumb—'if the scale drifts, it's the load cell'—was true 20 years ago. Back then, most plants didn't have high-VAV heating and cooling systems blowing directly at scales. Today, a packing line can have six vents, two exhaust fans, and a compressed-air nozzle all within three feet of a weighing station. The conditions changed. The rule didn't.

What I Do Now

My checklist includes three steps, and I don't skip the first two:

1. Check the environment first. Airflow, vibration, temperature gradients, static pressure. Use a thermal camera if you have one. A Fluke T865 or any equivalent with a decent resolution is a no-brainer for a plant that runs multiple scales.
2. Look at the mechanical path. Check for binding, debris, or anything touching the scale platform.
3. Only then suspect the electronics. And when you do, record the conditions—temperature, humidity, airflow—so you can prove the measurement was valid.

According to FTC guidance (ftc.gov), advertising claims about performance should be backed by evidence. That's not just a legal thing. It's good engineering. If I'm going to say a scale is reading correctly, I need to know what else was happening in the room at that moment.

I'd rather spend 10 minutes explaining options than deal with mismatched expectations later. That's why I now ask for the full spec sheet before I quote a job. An informed customer asks better questions and makes faster decisions.

One More Thing: A Caliper Question

A week after the air fix, one of the QC guys asked me: 'Where can I buy a Mitutoyo 8 inch digital caliper?' He wanted to order one quickly, and he'd seen cheap listings online.

I'm not going to name any auction sites, but I'll say this: calipers are precision tools. If a deal looks too good to be true, it's a red flag. Buy from an authorized distributor. The cheap ones fail in ways that make you trust bad data, and bad data in a QC department is worse than no data.

That advice is the same as what I tell plant managers about load cells: don't buy based on price alone. Buy based on the support you'll get after the purchase.

What I'd Tell a New Tech

If you're newer in this trade, learn from my mistake. When a spec says Class F, Ex rated, IP69K, or anything you don't understand, stop and ask. Ask the customer, ask the manufacturer, ask the original equipment designer. Do not assume.

And when a scale drifts, don't start by swapping the load cell. Start by watching the air around it. A 30-second walk around the scale with a thermal camera or even a piece of string will save you hours and thousands in parts.

A load cell is a piece of steel with strain gauges glued to it. It doesn't have mood swings. If the readings are inconsistent, something around it is lying to you.

The scale on that chip line ran fine for the next year and a half. No drift. No recalibration. The only thing we changed was the environment around it.

I've made that $4,800 mistake so you don't have to.