Automotive exhaust system is supported by the hanger rods welded to it. Tenneco validates the structural integrity of these joints to customer specific reliability and confidence targets. As part of that validation, it is common to hold part of the exhaust system and apply constant amplitude, fully-reversed load on the hanger rod and perform a fatigue test till there is a failure somewhere in the toe of the weld or in the hanger rod.The exhaust system is designed, engineered and manufactured per system design specification of the customer. In an exhaust system program life cycle, the same fatigue test is performed on Prototype and Production parts. The same test is also performed on an additional batch of Pre-Production parts in the current case. Prototype, pre-production and production parts are referred here as Design Verification (DV), Pre-Production Verification (PPV) and Process Verification (PV) tests and parts. A significant difference in results between DV and PPV parts required an investigation of the root cause of the difference.This paper is about this root cause investigation. Once a section is cut at the crack initiation location of the tested parts, it is found that, difference in securing the part for welding created a gap between the rod and exhaust muffler. This gap makes the results for PPV parts worse than DV. It is also found that the curl between the muffler outlet head and muffler body is worse from DV to PPV parts. It is proved with further tests that it is indeed the gap that caused the difference in results. To manufacture PV parts, new fixture is designed and made to eliminate the gap. Parts made with the new fixture are tested to prove it once again.