Last spring, on a Tuesday morning, I almost signed off on 4,800 Lemforder control arm kits. The paperwork was clean. Batch number matched, the CMM report was green, and a truck was already booked for Thursday afternoon. Then I picked one kit from the top pallet and ran my thumb over the rear bushing.
It was not flush. Close, but not flush.
That one difference cost us a full day of rework, one tense call with a customer, and a rule I still use: measure the physical part, not just the certificate.
What the drawing said
For context, I’m a quality/compliance manager at an automotive stamping and CNC machining plant. I don’t design suspension arms, and I can’t speak to exactly how a bushing gap becomes a clunk at highway speed. What I can tell you from the quality side is how the gap got there in the first place.
The Lemforder control arm kit in question called for the rear bushing flange to sit 2.0 mm above the arm face, plus or minus 0.3 mm. The sample I measured was 2.9 mm. A few other kits were 2.6 and 2.7 mm. The CMM report on the paperwork said 2.2 mm, which is inside spec. That turned out to be the clue: the bad parts came later in the run, after a locator pin in the press fixture had worn down.
The set-up technician told me it was within the industry ballpark. But we didn’t approve parts based on an industry ballpark. We approved them based on the drawing. Put another way, the part might work in someone else’s application, but it was not the part we had promised to build.
We quarantined the batch and checked every kit by hand. Almost 700 needed to be pressed again. That was the boring, expensive version of prevention. It would have been far more expensive to ship the problem and let the customer find it.
I still kick myself for not adding the bushing-height check earlier. The fixture was on the shelf. The set-up operator used it for first-piece checks. We trusted the press’s force sensor to tell us the bushing was seated, but the force sensor didn’t measure position.
The same pattern on a differential wheel assembly
Two weeks earlier, the same lesson appeared in a differential wheel assembly for a powertrain customer. The runout spec was 0.10 mm; the finished part measured 0.13 mm. The machinist said 0.03 mm was invisible. He was right. At road speed, though, that 0.03 mm repeats and turns into a small vibration. We caught it before shipping.
The control arm kit and the differential wheel assembly were different products, but the failure pattern was the same: the measurement that people skipped was the measurement that mattered.
The engine mount follow-up
That same release also included Lemforder engine mounts. The first few looked fine. But after the bushing surprise, I wasn’t going to rely on looking. A Lemforder engine mount can look perfect on the outside and still have a mounting stud a few tenths of a millimeter out of position. That is enough to put the mount into the subframe at an angle, which makes the hydraulic bushing work against its design axis.
So now each Lemforder engine mount gets placed in a fixture before I sign the release. The fixture costs less than one field return. It also reminds everyone that prevention is not about being dramatic. It’s about measuring the right thing before the part is packed.
The thermostat explanation
When I stopped the line, a manufacturing engineer asked why I was making a fuss over a gap he could barely see. The best answer I had was a house. How does a thermostat work in a house? It measures the air temperature, compares it to the setpoint on the wall, and turns on the heating before the room feels cold. It acts on a threshold, not on a feeling.
Quality control works the same way. The threshold on the control arm drawing is 2.0 ± 0.3 mm. The threshold on the differential wheel assembly is 0.10 mm. We don’t check those numbers because they’re nice. We check them because they predict when a part will stop doing its job.
Why I changed the air filter
That same week, home reminded me again. The furnace was cycling longer than normal, and the service technician asked when I last changed the 18x14x1 air filter. I couldn’t remember. The filter was not the dramatic part of the system. It just sat there collecting dust, making the blower work harder every time the thermostat called for heat.
An 18x14x1 air filter and a bushing height check are both unglamorous. They don’t produce a product feature. They prevent a failure.
I don’t want to overstate it. Not every out-of-spec part should automatically be scrapped. Sometimes an engineer can review the actual installation conditions and accept a deviation. That review should come from an engineer, not from a production schedule. On that Tuesday, nobody had made that case. They just wanted me to sign.
Bottom line: five minutes of verification beats five days of correction. Trust me on this one. I have the rework ticket and the phone call to prove it.