FIT data

[mvogl]

I am trying to use FIT (failures in time) data for electronic parts. The majority of the data I have found is from accelerated life testing data, which includes a temperature adjustment, but no information about the test voltage is typically listed. Does anyone know why the voltage stress/acceleration is typically neglected for FIT data?


More information about what I am doing:

The FIT data that I have received from manufacturers is calculated from accelerated life test data using the following formula:

Failure rate (FIT) = X / [2*N*h*Aft] * 10^9
(units of failures per billion hours)
X = Chi square value for given confidence level and (2n+2) degrees of freedom
n = number of failures
N = number of devices tested
h = number of test hours
Aft = temperature adjustment factor
= exp( (Ea/k)*((1/Toperating)-(1/Ttest)) )
Ea = activation energy (eV)
k = Boltzmann constant
Toperating = operating temperature (K)
Ttest = test temperature (K)

I have also seen a voltage adjustment factor included:
Failure rate = X / [2*N*h*Aft*Afv]
but it is typically left off. Additionally, the test voltage is typically not listed. Since (judging by the MIL217 model) the voltage stress can be significant even for smaller variations in voltage (e.g. for low frequency diodes, the voltage stress factor roughly doubles going from 40% to 60% of the rated voltage), I would like to account for the voltage stress.

[rwisniewski]

While it is difficult to speculate as to why voltage data is not being reported, one possible explanation is that the manufacturers are only considering the effects of temperature on operating life and are holding the voltage constant at some level during their accelerated life tests. It may be believed that temperature has the biggest impact on operating life as long as the other parameters remain within the manufacturer’s specifications. If you believe that voltage is critical to your applications, you may want to contact the component manufacturers to determine the applied voltage stress ratios during the life tests. You can then adjust your FIT data up or down based upon your application using the inverse power law for the voltage acceleration.