MIL-PRF-38535K
TABLE IA. Screening procedure for hermetic classes Q, V and non-hermetic class Y microcircuits - Continued.
13/ Burn-in shall be performed on all QML microcircuits, except as modified in accordance with SMD section 4.2, or above their
maximum rated operating temperature (for devices to be delivered as wafer or die, burn-in of packaged samples from the wafer
lot shall be performed to a quantity accept level of 10(0)). For microcircuits whose maximum operating temperature is stated in
terms of ambient temperature (TA), table I of TM 1015 of MIL-STD-883 applies. For microcircuits whose maximum operating
temperature is stated in terms of case temperature (TC), and where the ambient temperature would cause TJ to exceed +175°C,
the ambient operating temperature may be reduced during burn-in from +125°C to a value that will demonstrate a TJ between
+175°C and +200°C and TC equal to or greater than +125°C without changing the test duration. Data supporting this reduction
shall be documented in the QM plan and shall be available to the acquiring and qualifying activities upon request. For devices
with solder terminations, burn-in test may be performed before solder balls/columns have been attached to the packages.
14/ When test condition F of method 1015 for temperature accelerated screening is used for either burn-in or reverse bias burn-in , it
shall be used for both. Also, when devices have aluminum/gold metallurgical systems (at either the die pad or package post),
the constant acceleration test shall be performed after burn-in and before completion of the final electrical tests (e.g, to allow
completion of time limited tests but that sufficient 100 percent electrical testing to verify continuity of all bonds is accomplished
subsequent to constant acceleration).
15/ Where applicable, dynamic burn-in test shall be performed, and test condition F of method 1015 and temperature accelerated
test requirement shall not apply. For class V or class Y (class level S), burn-in test shall be performed in accordance with
TM 1015 of MIL-STD-883, on each device for 240 total hours at +125°C. For a specific device type, the burn-in duration may be
reduced from 240 to 160 hours if three consecutive production lots of identical parts, from three different wafer lots pass percent
defective allowable (PDA) requirements after completing 240 hours of burn-in. Sufficient analysis (not necessarily failure
analysis) of all failures occurring during the run of the three consecutive burn-in lots shall not reveal a systematic pattern of
failure indicating an inherent reliability problem which would require that burn-in be performed for a longer time. The
manufacturer's burn-in procedures shall contain corrective action plans, approved by the qualifying activities for dealing with lot
failures.
16/ The reverse bias burn-in is a requirement only when specified in the applicable device specification and is recommended only for
a certain MOS, linear or other microcircuits where surface sensitivity may be a concern. When reverse bias burn-in is not
required, interim post burn-in electrical parameter measurements shall be omitted. The order of performing the burn-in test and
the reverse bias burn-in test may be inverted. Static burn-in may be substituted for high temperature reverse bias burn-in based
on device technology and must be approved by the QA. Moreover, burn-in time-temperature regression table I of TM 1015 of
MIL-STD-883 can be used for determination of reverse bias burn-in time and temperature.
17/ The percent defective allowable (PDA) shall be 5 percent or one device, whichever is greater. This PDA shall be based, as a
minimum, on failures from group A, subgroup 1 plus deltas (in all cases where delta parameters are specified) with the
parameters, deltas and any additional subgroups (or subgroups tested in lieu of A-1) subject to the PDA as specified in the
applicable device specification or drawing. If no device specification or drawing exists, subgroups tested shall at least meet those
of the most similar device specification or Standard Microcircuit Drawing. In addition, for class V or class Y (class level S) the
PDA shall be 3 percent (or one device, whichever is greater) based on failures from functional parameters measured at room
temperature. For class level S screening where an additional reverse bias burn-in is required, the PDA shall be based on the
results of both burn-in tests combined. The verified failures after burn-in divided by the total number of devices submitted in the
lot or sublot for burn-in shall be used to determine the percent defective for that lot, or sublot and the lot or sublot shall be
accepted or rejected based on the PDA for the applicable device class. Lots and sublots may be resubmitted for burn-in one
time only and may be resubmitted only when the percent defective does not exceed twice the specified PDA (10 percent) or
2 devices, whichever is greater. This test need not include all specified device parameters, but shall include those
measurements that are most sensitive to and effective in removing electrically defective devices (see A.4.6.1.1 and A.4.6.1.2 of
MIL-PRF-38535).
21
ACTEL [ Actel Corporation ]
