To determine the burn-in time, data like those given in Table 6.1 need to be obtained. This table is based, as an example, on 900 devices subjected to burn-in, and their performance is monitored in 10-hr intervals to determine the number of devices that fail in each interval. Next, the failure rate for each time interval is calculated. In Row 2 of Table 6.1 the equations for calculating the failure rate of these devices at each time interval are given. The first equation is
| (6.1) |  |
where
| | = | average estimate of the failure rate in each ? T time interval, |
| N F( ? T) | = | number of devices failing in ? T time interval, which in Table 6.1 is 10 hr, |
| N BP( T) | = | number of devices at the beginning of each ? T period involved, |
and
| ? T | = | time interval for which the |
|
 |
It may be seen that the denominator of Eq. (6.1) is the total operating device-hours during the ? T interval. But in fact there are only [ N BP( T) -N F( ? T)] devices surviving this whole ? T period, while N F( ? T) devices fail during each interval. Therefore, the actual total device hours of operation by all devices being burned-in is less than N BP( T
TABLE OF CONTENTS 
