EN
Title: We Used the Wrong AWG Die for 3 Months – Here's What Happened to 1,200 Crimps
By our quality & failure analysis team
Three years ago, we had a job that still keeps me up at night. 1,200 crimps on 14 AWG wire, all made with a die marked "16-14 AWG." The tool was in spec. The operator was certified. Every crimp passed visual inspection.
Then the field failures started.
The .003 inch that cost us $4,700.
We didn't catch the problem during production. The crimps looked fine. The pull tester showed acceptable results – just barely. We shipped the harnesses to a packaging equipment manufacturer.
Three months later, their service team started reporting intermittent faults. Machines would run fine for weeks, then suddenly stop. No pattern. No warning.
We flew out. We cut open 20 field-returned crimps. Under magnification, every single one had a tiny gap between the wire strands and the terminal barrel – about 0.002 to 0.003 inches. The die was nominally correct for 14 AWG, but the actual stripped wire diameter on that specific batch was on the low side of the tolerance band. The die cavity was 0.003 inches too wide for that particular wire.
That 0.003 inches – less than a human hair – had prevented a true cold weld from forming. The connection was held together by friction, not fusion.
We replaced all 1,200 crimps. Cost to us: $4,700 in materials, labor, and travel. Plus the customer trust we had to rebuild.
The physics of a good crimp is a narrow window.
A proper crimp isn't just squeezing metal until it stays. It's deforming the terminal barrel and the wire strands together into a unified structure – what metallurgists call a cold weld. The barrel wall has to collapse enough to eliminate all air gaps, but not so much that it fractures or extrudes.
That window is about 0.003 inches wide. If the die cavity is too small, the barrel splits. If it's too large, air gaps remain. If the wire diameter varies too much from what the die was designed for, you get one of two failure modes. Both end badly.
We measured everything after that incident.
Now we measure every reel of wire before we crimp it. Not just the AWG marking on the label – we use a calibrated micrometer and check the actual stripped conductor diameter against the die's tolerance band.
For a 16 AWG die with ±0.003" tolerance, we accept wire diameters from 0.0478" to 0.0538". If a reel measures 0.0465" – just 0.0013" below the band – we don't run it. We move it to a different job or change the die.
It takes 30 seconds per reel. In the last two years, that 30-second check has caught seven reels that would have produced substandard crimps. Seven batches we didn't have to rework. Seven times we didn't have to call a customer and explain why we were late.
The pull test numbers that changed my mind about "good enough."
We used to do pull tests randomly – maybe one per 100 crimps. If it passed, we moved on. Then I started tracking the actual numbers, not just pass/fail.
A UL 486A-compliant 16 AWG crimp needs to hold at least 135N (30 lbf). Our acceptable crimps averaged around 160N. The marginal ones – the ones that had those tiny air gaps – averaged 140N. Still "passing," but barely.
Over 12 months, we tracked every field failure against the pull test result at time of production. Every single failure had come from a crimp that was at the low end of the pull test range – 135-145N. None of the 160N+ crimps failed.
We raised our internal pass threshold to 150N. Anything below that gets cut out and re-done. We reject about 3% more crimps than we used to, but our field failure rate has dropped to near zero.
The die wear trap: AWG tolerance changes over time.
A die that starts at ±0.003" tolerance doesn't stay there. After about 5,000 cycles, we measured our 16 AWG die cavity at 0.004" over nominal. By 8,000 cycles, it was 0.006" over.
That die was still producing "passing" crimps on the pull tester – just barely. But we knew from our tracking that those marginal crimps were the ones that failed in the field. We now replace dies at 5,000 cycles, not 8,000 or 10,000. The extra $40 per die has saved us at least $15,000 in rework over the last 18 months.
The visual inspection lie.
Here's something I learned the hard way: a bad AWG match often looks fine to the naked eye. The terminal barrel doesn't crack visibly. The wire doesn't look crushed. Everything seems normal until you pull it or put it under load.
We had an operator who had been doing crimps for six years – excellent tech, very careful. He inspected every crimp before putting it in the harness. He missed a whole batch of under-crimped 18 AWG terminals because the die was worn and the wire was on the small side of tolerance.
We bought a bench microscope for final inspection after that. Not expensive – about $250. Now every operator inspects every crimp at 10x magnification before it leaves the workbench. We catch about 1 in 500 crimps that we wouldn't have caught with the naked eye.
What we do now, and why:
-
Measure actual wire diameter before every job – not just trust the label. 30 seconds. Prevents 90% of our AWG-related issues.
-
Check die cavity tolerance with a gauge pin weekly. Replace at 5,000 cycles, not 10,000.
-
Pull test minimum raised from 135N to 150N for 16 AWG. We have internal thresholds that exceed spec.
-
10x magnification inspection on every crimp. $250 microscope. Paid for itself in two months.
-
Track every field failure back to production pull test data. We know exactly what passes – and what doesn't.
The bottom line, from someone who's paid the rework bill:
AWG compatibility isn't about following the rules. It's about the fact that a 0.003-inch mismatch can take a functional connection and turn it into a field failure. I've seen it happen. I've reworked it. I've flown to customer sites to explain it.
We don't do that anymore.