I'm going to tell you about a mystery that took two years to solve.

I'm a quality engineer at an industrial robot factory in Shanghai. We buy RV reducers from a specialized supplier. Each reducer goes into a robot joint. The reducer is the heart of the joint—it transmits the motor's power to the arm.

In 2024, we noticed a pattern. Some robots were quiet. Some were loud. The quiet ones measured 65 decibels at full speed. The loud ones measured 85 decibels. Same reducer, same motor, same assembly.

We tested the reducers before assembly. All passed. We tested the motors. All passed. We tested the assembly process. Identical.

But 20% of the joints were loud. And we couldn't figure out why.

The First Clue

I started by measuring everything.

The reducers came with a test report. Backlash: within spec. Efficiency: within spec. Runout: within spec. Nothing explained the noise.

I measured the housing temperature. The loud ones ran 5°C hotter than the quiet ones. Heat is energy. Energy is noise. The loud ones were converting more energy into heat instead of motion.

But why?

The Second Clue

I disassembled a loud reducer and a quiet reducer. I laid the parts side by side. Same gears. Same bearings. Same housings. Same bolts. I measured each part with a micrometer. Everything was within spec.

But one thing was different. The grease. The loud reducer had been assembled with a slightly different amount of grease. Not much. Just a few grams.

I called the reducer supplier. "Do you measure the grease quantity for each reducer?"

"Yes," they said. "We use an automated dispenser. It dispenses exactly 12 grams per reducer."

"Can I see the calibration records?"

They sent them. The dispenser was calibrated every six months. The records showed it was within spec.

The Third Clue

I visited the supplier's factory. I watched the grease dispensing process. The automated dispenser was accurate. It dispensed 12 grams every time.

But there was a problem. The reducers arrived at the dispensing station at different temperatures. Some were cold from the previous shift. Some were warm from the production line.

The grease viscosity changes with temperature. Cold grease flows slower. Warm grease flows faster. The dispenser dispensed by volume, not by weight. A warm reducer received 12 grams of warm grease. A cold reducer received 12 grams of cold grease. But the cold grease was denser. There was more grease mass in the cold reducer.

The difference was about 10%—from 12 grams to 13.2 grams.

The Fourth Clue

I asked: "Why does the grease quantity matter?"

The design engineer explained: "The reducer has a specific grease volume. If there's too much grease, it churns. The churning creates heat and noise. If there's too little, the gears don't lubricate properly and create noise."

The sweet spot was narrow. 12 grams was perfect. 13.2 grams was too much. The cold reducers were getting too much grease.

The Fifth Clue

I checked the temperature at the dispensing station. The morning shift started with cold reducers from the overnight storage. The afternoon shift started with warm reducers from the production line.

The morning shift had more noise complaints. The afternoon shift had fewer.

The noise problem was not random. It was temperature-based. The cold reducers got too much grease. The warm ones got the right amount.

The Sixth Clue

I asked the supplier: "Why don't you control the temperature before dispensing?"

"We didn't think it mattered," they said.

That was the answer. They didn't think it mattered. So they didn't control it.

The Solution

The solution was simple and cheap.

We installed a warming station before the grease dispenser. The reducers would sit in the warming station for 30 minutes before dispensing. The temperature would stabilize at 25°C.

The grease dispenser was recalibrated to dispense by weight, not volume. It would measure the grease mass before dispensing, not the volume.

The noise problem disappeared. 100% of the reducers now operated at 65-68 decibels.

What I Learned

Here's what I learned from this.

First: Temperature matters. A small temperature difference in the production process can cause a large difference in the final product. Nobody thought to check it. They assumed the room temperature was consistent. It wasn't.

Second: Volume vs. weight is a real difference. The dispenser measured volume. The spec was written in weight. Nobody noticed the mismatch. When you measure by volume, temperature affects the result. When you measure by weight, it doesn't.

Third: The root cause was not the reducer. It was the grease. The reducer was fine. The grease was fine. The process was the problem.

Fourth: The most obvious things are often the things we don't check. We checked the reducer. We checked the grease. We didn't check the temperature of the reducer during assembly.

What I'd Tell Every Buyer

If you're buying reducers or any precision mechanical component from China, ask these questions:

Question 1: "How do you control temperature during assembly?"

If they don't control temperature, they're introducing variation. Temperature matters for grease, for adhesives, for thermal expansion. It's not a small thing.

Question 2: "Do you measure by volume or by weight?"

If they measure by volume, ask if they compensate for temperature. If they don't, you're getting inconsistent quantities.

Question 3: "What's your batch tracking process? Can you trace a specific reducer to a specific production condition?"

If they can't, they're not managing variation. They're hoping for consistency.

Question 4: "Have you ever had a noise or vibration issue? How did you solve it?"

The answer tells you about their problem-solving capability. If they've solved it before, they know what to look for. If they haven't, you're the first.

The Aftermath

We started buying reducers from the same supplier. But we added a new specification to our purchase order: "Grease dispensing temperature shall be controlled to 25°C ± 2°C. Grease quantity shall be measured by weight, not volume."

The supplier accepted the new spec. They didn't even complain. It cost them almost nothing to add the warming station.

The noise problem is gone. The reducers are consistent. The robots are quiet.

The fix was simple. Finding it took two years.

The Question I Still Have

I still wonder: how many other small temperature differences are causing problems in other factories? How many assembled parts are getting too much or too little lubricant because nobody checked the temperature?

I don't know the answer. But now I always check the temperature.

We assembled 100 robot joints. 80 were quiet. 20 were loud. The parts were identical. The assembly was identical. The re

100 identical RV reducers. 80 quiet. 20 loud. Same parts, same assembly line, same operators. The difference was so small nobody had noticed it for two years.

Jangan direproduksi tanpa izin：Blog-Pulviso » The Robot Joint Noise Mystery: Same Parts, Same Assembly, 20-Decibel Difference