The $4,800 Flat Coil Spring Mistake: What I Learned About 'Identical' Specifications

In September 2022, I sat at my desk staring at a purchase order that looked completely clean. It was for a batch of flat coil springs — about 2,400 pieces — destined for a customer's aftermarket suspension kit. We were also bundling a serpentine belt tensioner bracket into the same shipment. Total order value: roughly $2,400.

I'd been handling automotive components orders at beru for six years at that point. Five of those years had been relatively uneventful, which — looking back — made me a little too comfortable. I'm a procurement and quality manager, and I've personally made (and documented) three significant mistakes over the past eight years, totaling roughly $12,000 in wasted budget. This flat coil spring story is mistake number two, and it's the one that stings the most.

Here's how it happened.

The Setup: A New Supplier and a 'Simple' Part

We were expanding our aftermarket product line that year. beru glow plugs and beru coil pack orders were growing steadily, and customers kept asking for more metal components to bundle with them — springs, brackets, tensioner parts. The flat coil spring wasn't a complex stamping. The drawing had about a dozen key characteristics: wire diameter, coil diameter, free length, spring rate, material spec (60Si2Mn), and a tight tolerance on the fillet radius at the end coil.

Our regular stamping supplier was booked solid through Q4. So I tried a new shop in Guangdong. Their quote came back 18% lower than the incumbent. Their sales rep sent a folder of certificates, a few sample photos, and the line I've heard a hundred times since: "We manufacture exactly to your drawing. No problem at all."

I didn't push further. I approved the order. No pre-production meeting, no process audit, no first-article inspection requirement in the PO. That's the first thing I'd do differently today.

The Arrival: Everything Looked Fine (Until It Didn't)

Three weeks later, the shipment arrived. Visual inspection was clean. Dimensional spot-check on 10 pieces passed — 8 were dead-on, 2 were off by about 0.05mm on free length, but that seemed within the general tolerance band for a spring of that size. I signed off on incoming inspection.

Then came the fatigue test.

The customer required flat coil springs to maintain at least 90% of their original spring rate after 500,000 compression cycles. Standard stuff for suspension components. The supplier claimed their 60Si2Mn material met the spec.

At 180,000 cycles, two samples showed permanent deformation. By 250,000 cycles, three out of five test pieces had visibly failed. The spring rate had dropped 15–22% — way beyond acceptable.

My first assumption was material. We sent samples to a lab for composition analysis. Turned out the steel chemistry was actually fine. The problem was elsewhere.

The Turn: It Wasn't the Material — It Was the Tooling

I flew down to the supplier's factory with our quality engineer. We walked the floor, looked at the setup, asked for the die maintenance log. That's when we found it.

The drawing specified a fillet radius of R0.8 at the transition zone on the end coil. The supplier's die had been built at R1.2. When I asked why, the tooling manager shrugged and said: "R0.8 is harder to form without cracking. We adjusted it slightly. It doesn't affect function."

He was wrong. For a flat coil spring under cyclic loading, that fillet radius controls stress concentration. R0.8 distributes the load across a tighter geometry that the material was specifically heat-treated to handle. R1.2 changed the stress profile in a way that passed static testing but failed under repeated compression. The difference was invisible until we ran the fatigue test.

When I compared the failed batch side by side with samples from our old supplier — same drawing, same material spec, same test conditions — the contrast was obvious. The old supplier's springs held up past 800,000 cycles with minimal rate loss. The new supplier's failed at a third of that. That's when I finally understood that "same specification" on paper does not mean "same manufacturing process" in reality.

Unfortunately, this batch was bundled with serpentine belt tensioner brackets that were fine. But without the springs, the whole kit couldn't ship. One wrong component stalled the entire order.

The Cost: $4,800 and a Bruised Reputation

Here's the damage breakdown:

• Rework and replacement springs: $2,800
• Expedited air freight to meet the original delivery window: $600
• Customer discount to compensate for the two-week delay: $1,400

Total: $4,800. On a $2,400 order. We lost money and, more importantly, we lost credibility. That customer kept ordering from us, but their vendor scorecard dropped our delivery rating from A to B. It took three clean shipments to get back to A.

Around the same time, that same customer emailed asking if we could add a "when to replace car air filter" guide to their aftermarket catalog. They'd seen our beru glow plugs and beru coil pack product pages and liked the content. I remember thinking: we can't even get our own spring spec right, and now we're supposed to teach people about air filter intervals? It was a humbling week.

What I Changed: The Checklist That Exists Because of This Mistake

I spent two weekends after that incident rebuilding our supplier onboarding process. The current checklist has 17 items. These five came directly from the flat coil spring failure:

1. Written tooling change notification. Any die modification — even a "minor" fillet adjustment — requires written notice and a first-article inspection report before production continues.
2. Process audit before first order. Not just sample approval. We walk the floor, check die maintenance logs, and confirm the actual tooling matches the drawing.
3. Fatigue testing on production samples. Not prototypes. Samples pulled from the actual production run, tested to the customer's real-world cycle count.
4. Material traceability per batch. Mill certificates with heat numbers, not just a generic "meets spec" letter.
5. Side-by-side comparison against incumbent. When switching suppliers, we now run both parts through identical test protocols before committing to volume.

Some of my colleagues thought I was overcomplicating things. But the bottom line is simple: $4,800 in rework taught me that a half-day of verification is cheaper than a two-week delay.

The Bigger Lesson: The Industry Has Moved On From 'Looks Fine'

Five years ago, evaluating a stamping or spring supplier mostly came down to price, lead time, and whether the sample looked right. That approach worked in 2020. In 2025, it doesn't.

Customers now expect documented fatigue life, full material traceability, and process capability data — not just a certificate and a handshake. The tools have improved too. CMM scanning, automated spring testing, and digital die-maintenance tracking are accessible to mid-size suppliers now, not just the big Tier 1s. But the tooling change I mentioned? The supplier's die was built five years ago and never updated. They were still operating on 2020 logic. I was too, until the test results forced me to catch up.

I'm not a metallurgist. I can't tell you the exact grain structure difference between R0.8 and R1.2 fillets under cyclic load — that's beyond my expertise. What I can tell you is that from a procurement and quality management standpoint, the process controls matter as much as the material spec. Maybe more.

I learned this in late 2022. The specific test data in this story reflects that period. Fatigue standards and testing methods may have evolved since then, so verify current requirements with your customer before finalizing any spring or stamping specification.

If You're Sourcing Flat Coil Springs or Similar Components

Trust me on this one: don't assume "same drawing" means "same part." It doesn't. The drawing is the starting point, not the guarantee.

Before you place an order — especially with a new supplier — ask three questions:

1. How are you making this? Get the process parameters, not just the material spec.
2. What could change during production? Tooling wear, batch differences, heat-treat variation — make them name it.
3. How will you prove nothing changed? First-article inspection, in-process checks, final test data. Get it in writing.

Those three questions would have saved me $4,800 and two weeks of damage control. They're now item numbers 3, 7, and 11 on our checklist. Hopefully they save you the trouble.