TMElectronics,
Inc.
Specialists in Leak, Flow and Package Testing
TM
Electronics, Inc.
45 Main Street
Boylston, MA 01505
For
Immediate Release
6/20/05
Contact: Karen Franks
Tel: (508) 869-6400
Fax: (508) 869-9955
Email: karen.franks@tmelectronics.com
Website: www.tmelectronics.com
TME
Statistics Packages:
For Quality and Process Control in Leak and Package
Testing
Boylston,
MA: Leak identification and measurement are essential
in the manufacturing and packaging of items from
all industries. Quantification of information
obtained in leak tests or package tests allows
the data obtained in the tests to be used in several
ways, including:
- Documentation
that your product meets the required specifications
for leakage;
- Validation
of your testing protocol, particularly important
in the medical device manufacturing/packaging
industry;
- Process
control.
TM
Electronics' leak testers and package testers,
including the TME Solution high-resolution, multi-port
leak tester and the TME BT-1000 Automated Package
Tester, contain a standard statistical package
that provides not only quality documentation but
also process control tools such as control charts,
histograms and graphic presentation of each individual
test.
Control
charts are commonly used to aid in manufacturing
process control. The objective of control charts
is to monitor the process in real time so if something
goes wrong, it can be noted and corrected with
the minimum of lost product. With regard to a
package sealing process that you are monitoring
using burst testing with a TME BT-1000 Package
Tester, for example, the concept behind control
charts is as follows:
- A
sealing process "in control" will
result in burst test values that fall consistently
in a predictable range around the average. In
addition, the average burst value will not change
appreciably over time when the sealing process
is "in control".
- A
single burst test value is referred to as "x".
The average of burst values over a period of
time is referred to as the mean, or "X-bar".
The range of burst test values is the difference
between the maximum and the minimum values.
- Because
processes always vary slightly due to manufacturing
and material variations, "good" product
burst test values will go up and down within
a range around the mean burst value. That range
can be statistically predicted using the mean
burst test value plus and minus three standard
deviations (a measure of the variation inherent
in the process). The "acceptable"
range is the set of burst test values that fall
between the upper and lower control limits.
These control limits are automatically calculated
in the TME test instrument from the previous
test results in the Datalog.
- In
the BT-1000, the data points on the control
chart consist of subgroups of burst test results.
These subgroups can be as small as two tests,
or as many as 20 tests. Subgroups are used to
minimize the effect of a testing error or a
single bad part.
Control
charts for the mean (X-bar) can help the manufacturer
in several ways:
- If,
for example, a temperature problem in the sealing
equipment is causing weaker than usual seals,
the downward trend in burst values will be obvious
on the control chart even before the product
reaches the point of failures. This gives the
machine operator an opportunity to correct the
temperature problem with little or no loss of
product.
- Several
data points outside of the control limits may
give the machine operator an indication that
an instability is developing in the process
that needs to be investigated before a large
quantity of bad product is produced.
Control
charts for range (the difference between the maximum
burst value and the minimum burst value within
a subgroup) also have a place in identifying when
the sealing process is becoming erratic and inconsistent.
For
more information on the statistical analysis packages
that are standard in TM Electronics' Leak and
Package Testers, contact us at 800-370-0501 or
visit www.tmelectronics.com.
Back
to List |
