Home > Reliability Solutions > Reliability & The Coffin-Manson Model…

Reliability & The Coffin-Manson Model…

Reliability Solutions News and PR from Reliability Solutions - Published 30 March 2016 The Coffin-Manson model was originally developed to model thermally induced metal fatigue failures.
This of course has since been used to effectively model thermal induced stress failures in electronic packaged devices and complex electronic PCBAÂ’s.

It is well established that repeated Thermal Cycling stresses PCBAÂ’s and enables quick stimulation of latent defects.

The number of thermal cycles to stimulate weaknesses to the point of failure is however dependent on the strength of the latent defect which will only fail when the cumulative stress level exceeds the underlying strength of the latent issue.

Many engineers have followed the ‘HALT theory’ that massively quick temperature transition for only a few thermal cycles will detect all the latent defects. This however can lead to fracture of some solder alloys rather than realistic failure of solder joints over time in stress.

The modified Coffin-Manson model has been shown repeatedly to provide a reasonable acceleration model for qualifying PCBA reliability, though it is very limited in providing a solution for full product with other electronic devices, hence the reliability engineer must be careful not to ‘over assume’.

When Thermal Cycling is combined sequentially with Random Vibration (preferably Thermal Cycling performed first) , some excellent results can be achieved in detecting range of PCBA failure types, though there is no reliable method to add any acceleration factor for Random Vibration portion of the stress test.

The moral of the story for the Reliability Engineer is always to focus on maximising Test Strength with variety of stress types performed sequentially on same devices / products to maximise defect detectability, then worry about the simulated time period later. Never over assume the acceleration factors to simply meet a schedule and assume the end product reliability will be as predicted from test data combined with an ‘estimated’ simulation period.

For more information feel free to check out the rest of the site and if you want to discuss this further, feel free to contact me either by phone or email. IÂ’ll get back to you with a personal response.
Many engineers have followed the ‘HALT theory’

Other announcements from Reliability Solutions