Instant Reliability Test Report Frequently Asked Questions
What is an “Instant Reliability Test Report”?
A: An "Instant Reliability Test Report" uses your
equipment test or field experience data to calculate the Mean Time
Between Failure (MTBF) of your equipment. The report is prepared in
accordance with MIL-STD-781D, “Reliability Testing For Engineering
Development, Qualification And Production” and MIL-HDBK-338B, “Military
Handbook, Electronic Reliability Design Handbook”. All you need to do
is enter 3 pieces of information: quantity of test operating hours or
field operating hours, quantity of failures, and a confidence level.
It's that simple. We process your data to calculate the equipment MTBF,
enter it into a Reliability Test Report, and e-mail it to you within
What is the mathematical basis for the MTBF calculation you will provide
in the Reliability Test Report?
reliability engineering it is well known that electronic parts fail at a
constant rate. Because of this concept, the exponential (Poisson)
distribution model is the appropriate model for calculating the MTBF of
equipment from test or field experience data.
MIL-HDBK-338B, paragraph 5.3.4,
“Exponential Distribution” and paragraph 5.3.8, “Poisson
Distribution”, explain the mathematical basis for the MTBF calculation.
What is a simple source for Reliability Test Data that I can use when
ordering a Reliability Test Report?
A: Many engineers don’t realize that reliability testing
is just putting operating time on the equipment under test. Also, they
don’t realize that this information already exists in their facility.
For example, most equipment is operated for a fixed number of hours
before shipment to the Customers. This is called “Post Production
Burn-In” time. The purpose of Post Production Burn-In is to detect
birth defects as required by MIL-HDBK-781D, “Reliability Testing
Plans/Procedures for Electrical Equipment”. Additionally, other sources
for Reliability Test Data may be found by reviewing field maintenance
records or service center records.
What Confidence Level should I select?
Confidence Level is the likelihood, expressed as a percentage, that the
results of a test are real and repeatable, and not just random. For
example, an equipment having an MTBF of 100,000 hours at 95% confidence
level means the probability the equipment will operate for at least
100,000 hours is 95%. Typical MTBF confidence levels
are 60%, 90% and 95%. As the confidence level increases, so too should
the sample size. If you
don't know what confidence level to select, we suggest using 60%
For the same operating hours, the MTBF gets higher as the
confidence level is set lower. With 300,000 hours of equipment
operation and 0 failures:
MTBF = 100,143 hours at 95% CL
MTBF = 130,288 hours at 90% CL
MTBF = 327,407 hours at 60% CL
Reliability engineers typically use 60%
Some statisticians consider 90% to be the minimum
confidence level for statistically significant results. In medical
research, for obvious reasons, there's a strong preference for even
higher levels of confidence.
How should I gather field experience data?
The reliability test methodology is to operate the equipment, Unit Under
Test (UUT), for a specified period of time. If the equipment fails,
record the number of hours of operation to failure. The UUT should be
operated at its normal environment (ambient temperature, humidity and
vibration) and electrical stress.
For example: Estimate the equipment MTBF at 60%
confidence level based on the following life test data. 50 units were
operated for 4 weeks. 4 weeks is 672 hours. One unit failed after 372
hours of operation. A second unit failed after 424 hours of operation.
A third unit failed after 473 hours of operation. The total unit test
hours is 32,853 hours, (47 units x 672 hours) + 372 hours + 424 hours +
473 hours = 32,853 hours. Enter confidence level of 60%, demonstrated
operating time of 32,853 hours, and 3 observed failures into the data
input screen above. The equipment MTBF is calculated to be 7,868 hours.
What is the “Bathtub Curve”?
A: The “Bathtub Curve” is a time versus failure rate
curve that is widely accepted by engineers in the reliability field.
The characteristic pattern is a period of decreasing failure rate known
as infant mortality, followed by a long period of constant failure rate
known as useful life, followed by a period of increasing failure rate
known as wear-out. The exponential distribution is the only model that
applies to the constant failure rate portion of the Bathtub Curve, so it
is the appropriate distribution model for calculating the MTBF of
What is an efficient way to perform a Reliability Life Test?
A: An efficient way to perform reliability life testing
is to put many identical units under test and operate them for a shorter
number of hours. For example, a test of 10 units operating for 100
hours each is mathematically equivalent to a test of 1 unit operating
for 1000 hours.
Probabilistic Software, Inc. (PSI) ▪ PSI Building Suite 101
▪ 4536 Indianola Way ▪ La Canada Flintridge, CA 91011
▪ Telephone: (818)
790-6412 ▪ e-mail:
Copyright © 2002 Probabilistic Software, Inc. All