MEMS reliability :
Reliability is the ability of a device or system to perform a required function for a specified amount of time. The function and level of reliability required of a device are obviously application specific. A sensor vital to an aircraft flight control system whose failure can mean loss of life is an example of a high reliability application.
Alternatively, a switch in a piece of commercial electronics may fail and be an annoyance but it is hardly a tragedy.
The reliability of a system may depend upon many things, such as the subsystems [e.g. individual electrical, mechanical, or MEMS devices], packing, power supply. software or cooling systems. The overall reliability of a system is complex because the inter relations between the different subsystems can combine in complex ways to cause a system failure. Alternatively, the system could be designed so that subsystems can compensate for failures in other aspects of the system e.g. software could be designed to monitor the system and avoid a system failure by utilizing other subsystems to perform a function.
Data to access the reliability of a device are obtained by operating a large number of devices under normal operating conditions and noting when failure occurs.
Three items need to be defined. unambiguously to formulate the reliability experiment properly.
1. Method of operation : A detailed definition of the method of device operation - for example a switch is the device to be tested for reliability. An application of voltage [lov/ov] to close or open a switch 100 Hz frequency using a square wave signal.
2. Definition of failure : Definition of what constitutes failure of this device. For example : resistance of the switch contact $\gt 1k \Omega$ when the l0V operate signal is applied or the resistance $\lt 1M \Omega$ when the 0V signal to open the switch is applied. This statement of failure denotes that contacts are fowled. So that adequate continuity when closed is not present or sufficient isolation is not present when the switch mechanism has failed so that the contacts do not open or close when the appropriate signal is applied.
3. Sample size : A large enough sample size utilized in the reliability test so that a statistically meaningful result can be obtained. Reliability data are by their nature non-deterministic because numerous factors can affect the outcome of the reliability experiment. The non deterministic factors include manufacturing variations [ e.g. deposition, etch, patterning ] material variations and environmental variations. [e.g. humidity, particulate material interactions, shock vibration, thermal].These variations cannot be predicted, therefore the reliability data must be considered random and analyzed with statistical methods. These 3 items define how to operate the device what constitutes failure and the number of items to test. The number of cycles that have elapsed when the failure occurs defines when the failure occurs, which can have units of time or cycles of operation.
and 4 others joined a min ago.
teamques10
★ 68k
