SiliconFarEast.com FAQ

- questions most frequently asked by visitors of www.siliconfareast.com

Equivalent Life of Temperature Cycle Testing

What is the equivalent life in the field of a Temperature Cycle Test (TCT)?

If a device
passes TCT, what will its expected life be in the field?

There
seems to be no single absolute answer to these questions. The answer that we most frequently encounter for
this question, however, is what will be discussed here.

The temperature cycle test
is performed to determine the resistance of a device to alternating
temperature extremes. It consists of exposing the device to many
cycles (usually 500 to 1000 cycles) of high and low temperatures, and primarily tests the device's resistance to fatigue failure ('fatigue'
is defined as the failure due to cyclical loads). Needless to say, the
higher the number of temperature cycles that a device passes, the longer
is its equivalent life in the field. The question, however, is how
long is long.

When finding the equivalence
of temperature cycling to field lifetime, one has to consider two components of
the equivalence: 1) the equivalence of the number of TCT
cycles performed on the device in the lab to the number of temperature
cycles that the device will see
in the field; and 2) the equivalence of the number of temp cycles that
the device sees in the field to the device lifetime.

The first step, which is
correlating the number of 'lab test' temp cycles to 'field' temp cycles,
is based on an acceleration factor (AF) equation. One
commonly used AF equation is given as follows:

**
AF = (∆Taccel/∆Tuse)**^{m}
where: **∆**Taccel =
Tmin(accel)
- Tneutral;
**
∆**Tuse =
Tmin(use)
- Tneutral;
and Tneutral =
zero stress temperature (approx. 175
deg C for plastic packages - the temperature at which the package
is molded)). Tmin(accel) is the minimum temperature used for lab
temperature cycling while Tmin(use) is the 'average' minimum temperature
that the device will see in the field. The value of the
fracture-property dependent constant *m* depends on the failure
mechanism, and is usually set to 20
for plastic package
cracking.

AF is the ratio of the
number of 'field' temp cycles to the number of 'lab test' temp cycles.
As an example, if the 'lab test' TCT employs a low temperature of -65
deg C and the device will see (on the average) a minimum temperature of
0 deg C in the field, then AF = (-240/-175)^20, or AF = 554.
This means that a single cycle of the lab TCT (where the minimum
temperature is -65 deg C) is equivalent to the device experiencing 554
cycles in the field (where the minimum temperature is 0 deg C) as far as
packaging cracking is concerned.

Another
common AF equation used for temp cycling is as follows:

AF = Nuse/Ntest
= C(**∆**Tuse)^{-n}/C(**∆**Ttest)^{-n
}where
∆Tuse
is the difference between the maximum and minimum temperatures that the
device will see in the field within one 'cycle', and
∆Ttest is the
difference between the maximum and minimum temperatures used in temp
cycling.

Once the
equivalent number of 'field use' cycles is determined, one can proceed
to the next step, which involves estimating how
many 'field use' cycles are equivalent to every year of field use.
This is the more tricky (and more subjective) part of the process, since
one has to make several assumptions in order to accomplish this.
For example, a device used under the hood of a car may be assumed to
undergo X temp cycles a day if the car is expected to be
parked and driven X times a day. This is equivalent to 365 X 'field use' cycles
in a year.

Finally, one
has to combine: 1) the equivalence of 1 lab cycle to 1 field cycle, and
2) the estimated # of field cycles in a year, in order to obtain the
estimated equivalence between 1 lab cycle and the expected life of the
device. To illustrate this, assume that the car above will be
driven and parked 10 times a day, or 3,650 times a year. Also
assuming as above that each lab cycle equals 554 'field' use cycles,
then a device that passes 1000 lab TCT cycles will pass 554,000 'field'
cycles. As such, it is not
expected to fail due to 'field use' temp cycling-induced package
cracking within 150 years! Note that the numbers used in the
example discussed here are purely hypothetical to simplify the
discussion, and may not be reflective of what the actual figures would
be.

See Also:
Temperature Cycle Test;
Reliability Modeling

Back to the FAQ Page...

HOME

Copyright
©
2006
**SiliconFarEast.com**.
All Rights Reserved.