The automotive industry takes durability testing seriously. They cannot afford to do otherwise. If defective parts enter the assembly stream undetected, a lot of vehicles may be built and sold before the ultimate durability testers, car buyers, expose the problem.
Engines undergo extensive dynamometer testing lasting weeks or even months. This testing continues in order to identify further reliability and performance improvements, even after an engine design has gone into production. In a recent extended dyno testing program, an engine manufacturer in Michigan encountered a rash of connecting rod bolt failures in multiple test engines as they reached the equivalent of 30,000 miles in dyno time. Engine production immediately stopped while an urgent investigation was implemented. Thread “laps” – seams in the bolt threads resulting from defective manufacturing – were identified as the cause of the fractures. The bolt vendor brought the manufacturing process back to specified parameters, eliminating the lap defect. A new lot of bolts was quickly brought in, tested to confirm that no defects were present, and engine production resumed.
A major problem remained, however. The auto manufacturer had no idea how many defective lots of connecting rod bolts had been introduced into the system and assembled into engines prior to their discovery during dynamometer testing. To further complicate the issue, it appeared that only one in every 200 or 300 bolts was defective. Each V-6 engine contained twelve connecting rod bolts. This meant that on average, between one-in-16 to one-in-25 previously installed engines could have been assembled with a defective bolt. Depending on how far back the problem went, cars with defective engines could be waiting delivery to dealerships, at dealerships awaiting sale or sold to customers and on the road. Fortunately, the manufacturer had set aside a percentage of each lot of connecting rod bolts for possible “post assembly analysis” in case just such a situation occurred. Unfortunately, the set asides from previously assembled lots constituted just over 5000 bolts.
Immediate analysis of these bolts and identification of when defective bolts first entered the supply stream was required. The manufacturer didn’t have the staff to perform these analyses in the required time, and called Metallurgical Associates Inc. We received the bolts at our Milwaukee materials analysis lab at 4PM on a Friday afternoon. Our metallurgical engineers and technicians went into immediate “round the clock” shifts, visually examining each of the 5000 bolts by stereomicroscopy to sort potentially defective bolts. These were then examined by Scanning Electron Microscopy to confirm the presence of thread laps. Bolts with confirmed laps were then sectioned, mounted, polished, and metallographically examined to document the extent and depth of the thread laps to critical “pass/fail” criteria. By Monday morning at 8AM, 72 hours after receiving the first lot, all 5000 bolts had been examined, the percentage of defects in each lot calculated, and the date on which the first defective lot had entered the production stream, identified.
With the test results provided by Metallurgical Associates Inc. materials analysis engineers, the auto manufacturer was able to quickly identify the production dates of engines with potentially defective connecting rod bolts. Several thousand had already been installed in cars and had left the assembly plants. These were diverted from the supply chain before delivery to dealerships and retro-fitted with replacement bolts, an expensive and time consuming process. This expense, however, was a fraction of the cost of a nation wide recall both financially and to the manufacturer’s public image.