Research On Reliability Assessment Method Of LED Driver Power Supply

Rather than paying attention to process and mechanism, Tranditional reliabilityassessment only forcused on the failure time and results. For long life and highreliable complex electronic products, the assessments had great disadvantages.Because of the high failure rate of driver power supply, the LED street lamp lightingdevices had meet big difficults in the development. This paper put forward thereliability assessment based on performance degradation. This mehtod revealed theweak parts in the LED driver power supply for further optimization andperformance improvement, which had great significant meaning.Firstly, failure mechanisms of discrete components in LED driver power supplywere studied, providing falure threshold. Taking environmental conditions, materialand work stress into account, degradation model was derived. According failuremechanism, Accelerate Degradation Test was designed and used for the follow-upreliability research in electronic system.The main power LED drive power components model was simulated in finiteelement software, verifying by thermal infrared imager. For reducing thetemperature of device and electrical connection, the position of components, heatsink size and material should be changed.Electrical connection is a critical part in addition to electronic components;consequently the research of solder joints had a great significance. Then thermalstress analysis and stress and strain under temperature cycling analysis was carriedout the though-hole solder joints of LED driver power supply, verifying by test. Wecan obtain that thermal stress firstly occurred in the solder and the substrateinterface and the crack initiation mainly took place between solder and lead pin bymetallographic testStudied degradation models were applied in the Monte Carlo method. A rank,which was key components, was achieved when device degradation impact onsystem. By the correlation analysis of the key components, the system reliabilitywas calcucated. Then the reliability results were compared between commondegradation and correlation analysis. Small difference between the two analysis, andthis method can reduce the simulaiton time and test cost.