C01_5E3.pdf part of Fluke C01 5E3 C01_5E3.pdf, Fluke 5700a doc C01

Text preview for : C01_5E3.pdf part of Fluke C01 5E3 Fluke 5700a doc C01_5E3.pdf

Back to : C01_5E3.pdf | Home

5700NAN LESSONS LEARNED

Kevin R. Abercrombie
Quality Manager, Metrology Branch
Naval Air Warfare Center
22 113 Fortin Circle, Unit 7
Patuxent River, MD 20670-1 1 18
(301) 342-1654, DSN 342-1654
FAX (301) 342-0920
[email protected]

Abstract

Since incorporating the Fluke 5700A/AN into the NAVAIR Metrology and Calibration Program,
NAVAIR incurred unacceptably high repair costs and down time. This paper examines the root
causes of the problem and the lessons learned as aresult of the investigation.

Introduction

The Fluke 5700A/AN meter calibrator is a precision instrument designed to calibrate
general purpose measuring and test equipment. It provides a source of direct and
alternating voltage, direct and alternating current, and resistance. This unit is highly
accurate and very user friendly, and because of these reasons, it was selected as the
standard meter calibrator for use throughout the NAVAIR Metrology and Calibration
(METCAL) program. Furthermore, because of its versatility, NAVAIR does not limit the
use of the 5700A/AN to solely calibrating meters, it is also used in the calibration of
other Ground Support Equipment (GSE) and General Purpose Electronic Test Equipment
(GPETE). In all, the 5700A/AN meter calibrator is used in approximately 480 Instrument
Calibration Procedures (ICPs).

Unfortunately, NAVAIR has incurred high repair-costs and significant down-time since
incorporating the 5700A/AN meter calibrator as astandard in theNAVAIR METCAL Program.
Because of these unanticipated problems, an investigative effort was undertaken in an attempt to
identify the primary cause(s) so that effective solutions could be implemented.

This paper describes the data collection method used to acquire failure information, the
investigation into the failures of seven 5700MAN meter calibrators, and the lessons learned as a
result of the investigation. Finally, this paper contains recommendations for utilizing the

2001 NCSL International Workshop & Symposium
5700NAN, which includes revisions to ICPs, appropriate safeguards, and good measurement
practices.

Survey

The method employed to gather the data pertaining to the failures of the 5700AIAN was a survey
form printed in the Metrology Bulletin (METBUL - July, 1997 and December, 1998). When
completed, the survey form provided pertinent information about the use of the 5700NAN at the
time of failure, such as: the model number of the Test Instrument (TI), the ICP that was being
performed, and the environmental conditions. This survey data was then reviewed to determine
if there was any commonality such as an ICP or type of TI.

During the data collection phase of the project, the survey form was modified to provide
additional information to help characterize the failures. The information added to the form
included diagnostic error-codes to aid in troubleshooting, the procedure revision date to ensure
that the most current ICP was in use, and characterization of the failure as either electronic or
physical in order to separate 5700AIAN failures caused by physical abuse; e.g., damage
sustained in shipping vice failures caused by normal use in accordance with approved NAVAIR
ICPs. Additionally, the contact information was
updated to provide both a mailing address as well as a FAX number to facilitate
submitting the survey to the proper point-of-contact.

Review of 28 surveys revealed that 12 of the 5700NANs failed at power up, 8 failed during the
calibration of theunit, and the remaining 8 failed while being used to perform the calibration of a
TI. Of these eight 5700A/ANs, 7 involved ICPs used to calibrate GSE. This indicated a possible
problem with how the 5700A/ANs are being employed in those procedures. Based on the survey
results, it was decided to further investigate these 7 ICPs.

5700AAN Failure Characterization

The 5700NAN failures can be caused by several factors. First, the units are beginning to age,
and therefore the failure rate is naturally increasing. Second, they are being damaged in use, but
the problem is not discovered until later. And third, the 570OAIANs are being subjected to
unexpected conditions, such as the application of reverse voltage, which can either result in a
catastrophic or a soft' failure. For the purposes of this paper, the term "reverse voltage" refers to
an unexpected voltage present at the TI test points that is fed back into the 5700AIAN when it is
connected to the TI.

According to the Fluke Manufacturing Company, "the CMOS technology used in the 5700A/AN
could sustain damage from a reverse voltage being applied to the unit, and the resulting fault
may not show up for several months.'' Furthermore, the exact amount of reverse voltage
required to damage the 5700NAN is not known; however, voltages of less than 3 volts may be
enough to damage the 5700NAN. This supports the hypothesis that the 5700A/ANs could be
subjected to reverse voltages that result in a soft failure.

' A soft failure is the result of a reverse voltage exposure that is not sufficient to cause an immediate catastrophic failure, but
rather eventually results in &failure.
200 1 NCSL International Workshop & Symposium
Of the seven 570OMANs that were reported as damaged while calibrating GSE, all 7 had
catastrophic failures.

In order to determine the most likely instances when the 5700A/AN was subjected to reverse
voltages, it is necessary to examine the reported failures and the ICPs in use at the time the
5700NAN failed. The 3 most likely factors that contribute to the 5700A/AN being subjected to
reverse voltage are: the TI being defective, operator error, and procedural errors. Table 1 shows
these probable contributing factors as they relate to the reported 5700NAN failures. For
instance, if the survey indicated that the TI was found to be defective, then TI Failure was
included as a probable contributing factor. Likewise, if the operator indicated that he had made a
mistake in the setup or connections, or the 5700MAN was substituted incorrectly, then Operator
Error was selected as a probable contributing factor. And finally, if the review of the procedure
showed that the procedure lacked safeguards or contained illogical sequences, then the
Procedural Error was listed as aprobable contributing factor.

Operator Procedural
TI Failure2 Undetermined'
Error3 Error4
2 3 5 1

Table 1. 5700NAN Failure Characterization

In 4 of the 7 cases, more than one factor may have contributed to the failure of the 5700NAN.
In 5 cases, the procedure may have contributed to the failure. In 1 case, the survey data did not
indicate any abnormal situations or problems with the TI, nor was there any indication of
operator error. Additionally, a review of the procedure showed the procedure incorporated
safeguards and good measurement practices; however, the repair data revealed that the A3
Motherboard was damaged, which is astrong indicator that the 5700 was subjected to a reverse
voltage. In this case, there was not enough information provided to determine the probable
cause.

Procedure Analysis

Each ICP was examined to determine: the workload supported; the frequency of its
performance; the relationship between Section 1 Performance Specifications and Section
2 Minimum Use Specifications regarding the use of the 57OOAIAN; how the .5700A/AN
was used in the ICP; if a substitute, either methodology or instrument, for the 5700A/AN
should have been used; if the construction of the ICP contributed to the failure of the

TI Failures: These are instances when the survey respondent indicated that the TI was defective, resulting in damage to the5700AfAN.

' Operator Error: These are instances where the operators actions contributed to the 5700AfAN failure, such as improperly substituting the
5700MAN for another standard or improperly connecting the 5700AfAN to the TI.

`Procedural Error: These are instances where the procedure may have contributed to the failure due to a lack of safeguards.
support any conclusion.
Undetermined: These are instances where thereis no data to

200 1 NCSL International Workshop & Symposium
5700A/AN; and, how the procedure could be improved to avoid damaging the meter
calibrator. An example of suggested revisions to improve each ICP is provided.

A number of conclusions were reached during the analysis of the procedures. Many of
the conclusions were the same or very similar between the ICPs examined. A summary
of these conclusions are:

Ensure that the technician places the meter calibrator in standby mode any time
the test connections or functions are changed.

Verify that test points are not energized before connecting the meter calibrator to
the TI.

0 If the TI performance specifications do not require the accuracy of the
5700A/AN, use another instrument or methodology in place of the 5700AIAN.

Ensure that WARNINGS or CAUTIONS appear in the ICP and are appropriately
placed in the ICP.

Conclusion

The most likely cause of damage to NAVAIR 5700A/AN meter calibrators is the
unintentional application of voltage to its output terminals from an external source. It
was determined that this condition is most likely to occur during the calibration of GSE.
Several of the GSE ICPs examined during this project lacked fundamental measurement
practices and safeguards, and therefore could have contributed to damaging the
5700A/AN. This indicates that greater caution must be exercised when using the
5700A/AN to perform these procedures, such as verifying that no voltages are present at
the test points prior to connecting the 5700A/AN to the TI, and ensuring that the
5700A/AN is in standby mode any time the test configuration or output function is
changed.

200 1 NCSL International Workshop & Symposium