Information Notice No. 96-52: Cracked Insertion Rods on Troxler Model 3400 Series Portable Moisture Density Gauges

                                 UNITED STATES
                         NUCLEAR REGULATORY COMMISSION
               OFFICE OF NUCLEAR MATERIAL SAFETY AND SAFEGUARDS
                            WASHINGTON, D.C.  20555

                              September 26, 1996


NRC INFORMATION NOTICE 96-52:  CRACKED INSERTION RODS ON TROXLER MODEL 3400 
                               SERIES PORTABLE MOISTURE DENSITY GAUGES


Addressees

All U.S. Nuclear Regulatory Commission portable gauge licensees and vendors.

Purpose

NRC is issuing this information notice to alert addressees to the potential
for cracks to develop in the insertion rod of Troxler model 3400 portable
moisture density gauges.  If not detected early, the cracks may propagate,
eventually leading to complete failure of the insertion rod and release of the
contained radioactive material.  This notice provides information that users
of these gauges may consider to avoid crack initiation and to detect existing
cracks.  It is expected that recipients will review the information for
applicability to their licensed activities and consider actions, as
appropriate, to avoid similar problems.  However, suggestions contained in
this information notice are not NRC requirements; therefore, no specific
action nor written response is required.

Description of Circumstances

NRC was notified of several incidents where the source cup of a Troxler 
model 3401 portable moisture density gauge separated from its insertion rod,
releasing the contained, nominal 333 MBq (9.0 mCi), cesium-137, sealed source.

One of the users of a failed gauge contacted the manufacturer and requested an
examination of its failed gauge and all other Troxler gauges it possessed. 
The examination determined that the failure occurred in the vicinity of the
weld joining the stainless steel insertion rod and source cup and that several
other gauges in the user�s possession were suspected of containing cracked
rods/cups, as well.  The manufacturer recommended that the user discontinue
use of all gauges suspected of being damaged until further inspection.

Following examination of the user�s gauges, the manufacturer initiated an
inspection and testing program on model 3400 series gauges returned to it for
disposal or repair that contained �unthreaded� (not joined with a threaded
connection) source cups and insertion rods.  This included only gauges with
serial numbers less than 13301.  The manufacturer only examined gauges with
�unthreaded� rods/cups because all gauges that were reported as having cracks
were this configuration.  The manufacturer provided the following information
and results from its investigation:

   �  235 gauges that had been returned for repair and 75 gauges that had been
      returned for disposal were inspected.


9609200181.                                                            IN 96-52
                                                            September 26, 1996
                                                            Page 2 of 7


   �  Drop and bending tests were performed on the 75 gauges that had been
      returned for disposal to determine the force necessary to cause failure
      of the weld joint.  Fatigue was not considered during the testing.

   �  Two of the 75 gauges returned for disposal were determined to contain a
      transverse crack on the insertion rod in the weld region.

   �  A number of the rod/cup welds of the 235 gauges returned for repair were
      determined to be �unusual,� but none were determined to contain cracks.

   �  The manufacturer received a report of another gauge, not included in the
      investigation, whose rod/cup joint had also failed completely in the
      field, releasing the contained sealed source.  The manufacturer
      suspected that the gauge had been abused during use, but the gauge was
      not available for the manufacturer to perform a full examination.

Based on the results of its investigation, the manufacturer concluded that
weld failure is an unlikely event when the gauge is subjected to normal
conditions of use and typical drop accident conditions.  It also concluded
that fatigue and thermal stresses would not be expected to have an adverse
effect on the weld joint under normal conditions.  However, there was no
fatigue nor thermal testing during its investigation.

Concurrent with the manufacturer�s investigation, NRC contracted an
independent third-party testing facility to examine, in detail, the failed
gauge, previously examined by the manufacturer, and four other of the user�s
model 3400 series gauges (three of which were suspected of also being
damaged), and determine the root cause of the failure, if possible.  This
examination resulted in a number of conclusions and observations, including:

   �  The failure resulted from an initial crack in the weld region of the
      insertion rod that grew progressively until it failed completely in a
      brittle manner, releasing the contained sealed source.

   �  The initial transverse crack likely resulted from the insertion rod
      being subjected to a severe bending load (possibly an impact) not
      sufficient to cause catastrophic brittle failure.

   �  The initial crack propagated slowly due to repeated loads during use. 
      The crack extended approximately 290� before the final brittle fracture.

   �  The region of crack initiation (the heat-affected zone around the weld)
      was very hard, but the weld was very soft.  This differential in
      hardness was caused by material selection, joint configuration, and the
      welding procedure, and provided a stress concentration in the region
      around the weld.  The stress concentration made the weld area more
      susceptible to crack initiation in a high-bending load condition.

.                                                            IN 96-52
                                                            September 26, 1996
                                                            Page 3 of 7


   �  One of the other gauges received from the licensee was confirmed to
      contain a transverse crack in the region of the weld.  Hardness testing
      indicated that this gauge also contained a soft weld and a high-hardness
      region surrounding the weld.  In addition, the rod/cup was constructed
      using the same materials and processes as the failed gauge.

   �  The remaining three other gauges were undamaged, but had similar
      hardness characteristics as the failed gauge.

   �  All gauges received for testing contained �unthreaded� rods/cups and
      were constructed using the same materials and processes.  The source
      cups on newer model 3400 series gauges (serial numbers above 13300) are
      threaded to the insertion rod before joining by welding.

   �  The weld configuration of the rod/cup connection is not optimum for its
      intended application, and provides a high stress concentration in the
      weld area.  An alternate selection of materials and joint configuration,
      and a post-heat treatment may have reduced the high-hardness region,
      resulting in a lower stress concentration in the weld area.

Based on the above observations and testing performed, the third-party
contractor concluded that all gauges manufactured in the same manner as the
failed gauges would be susceptible to crack initiation when subjected to a
severe bending load.  In addition, the contractor concluded that a crack on
this type of gauge would be expected to propagate slowly over time and could
be easily detected visually, either by using a dye penetrant evaluation
technique or by observing the weld area under magnification.  The contractor�s
visual examination without dye penetrant or magnification indicated a
potential crack in an insertion rod, that was subsequently determined to only
be a surface feature.  Therefore, visual examination should not be considered
a reliable way to identify cracks.

The figure below shows the typical configuration of a Troxler model 3400
series gauge and the locations of the insertion rod, source cup, and weld:

















.                                                            IN 96-52
                                                            September 26, 1996
                                                            Page 4 of 7























































                                                            IN 96-52
                                                            September 26, 1996
                                                            Page 5 of 7


In June 1996, Troxler issued a bulletin (attached) discussing the above
incidents and the results of its investigation.  The bulletin provided gauge
users with three recommendations on how to reduce the likelihood of future
source cup detachment incidents.  These included:

   1. Gauges should be used in accordance with the manufacturer�s instructions
      found in the operator�s manual.  Under no circumstances should the
      insertion rod ever be driven or forced into the test material.  A drill
      rod accessory is provided with the gauge to prepare a hole in the test
      material into which the rod is inserted.  If the user purchases a used
      gauge that does not include the drill rod accessory, the user should
      contact the manufacturer to obtain a replacement drill rod.

   2. Users should have the gauge and insertion rod inspected at least once
      every 5 years for unusual wear and tear.

   3. Companies purchasing used gauges should require an inspection before
      purchase.  Gauges involved in accidents should be inspected before
      further use.

Discussion

The aforementioned observations and investigations indicate that a gauge
manufactured using the same materials, welding process, and rod/cup
configuration as those identified as containing cracks would be expected to be
susceptible to crack initiation in the weld region when placed in a high-
bending load.  Cracks in the weld region of the insertion rods would be
expected to propagate over time, possibly leading to complete failure and
release of the contained sealed source.  However, before complete failure, a
crack of this type should be identifiable by macroscopic visual examination or
dye-penetrant testing.  It is unknown whether crack initiation could occur
under other conditions not discussed here nor considered during the testing
performed by the manufacturer or the third-party contractor.

When used according to the manufacturer�s instructions and recommended
conditions of use, crack initiation would not be expected.  Users of Troxler
model 3400 series gauges, and other similar gauges, should always operate
their devices in accordance with the manufacturer�s instructions and under the
recommended conditions of use, and should never place the insertion rod in a
bending-load condition.  The bending-load force necessary to initiate a crack
was not determined during either the manufacturer�s or the third-party
contractor�s investigations.  Therefore, bending-loads on the insertion rod,
other than those encountered during normal use, should be avoided.

The manufacturer has changed the design, materials, and manufacturing process
for model 3400 series gauges several times since their introduction.  The
manufacturer�s and third-party contractor�s investigations were limited to
model 3400 series gauges with serial numbers less than 13301 (gauges with
�unthreaded� rods/cups and the same materials and methods of construction). 
Therefore, the susceptibility to crack initiation of gauges with �threaded�
rods/cups or with other materials of construction and welding processes .
                                                            IN 96-52
                                                            September 26, 1996
                                                            Page 6 of 7


(models with serial numbers greater than 13300) is not known.  However, all
failed gauges to date have had serial numbers less than 13301.  A threaded
rod/cup connection may provide an additional measure of safety even if a crack
began and led to complete failure, as the source cup would likely still need
to be unthreaded from the rod to release the contained sealed source.

Based on the results of the manufacturer's and the third-party contractor�s
investigations, and the recommendations made in Troxler�s bulletin, users of
Troxler model 3400 series gauges with serial numbers less than 13301 are
encouraged to have their gauges inspected for cracks in the weld area of the
insertion rod.  In addition, it is recommended that these gauges be rechecked
for cracks on at least a 5 year basis (more or less frequently depending on
use level and conditions).  It is further recommended that users of model 3400
series gauges with serial numbers 13301 and above also have their insertion
rods checked for cracks in the weld area and consider periodic re-inspection,
as well.

Inspections of insertion rod weld areas should only be performed by persons
qualified to conduct an appropriate nondestructive examination (NDE) for
macroscopic cracks (e.g., visual examination with magnification or dye
penetrant) and licensed to handle the contained radioactive material.  Visual
examinations without magnification and examinations by unqualified persons are
not recommended, as these could lead to inaccurate results and unnecessary
doses to persons performing the examination.  Persons who possess Troxler
model 3400 series gauges who wish to have their insertion rods inspected
should contact either the manufacturer or an NDE inspector qualified to
perform the inspection.  For guidance on finding qualified NDE inspectors,
licensees may contact the Technical Services Division of the American Society
for Nondestructive Testing, Inc., at (800) 222-2768 or the Non-Destructive
Testing Management Association at (800) 317-0822.

Users who determine that their gauge�s insertion rod contains a crack should
immediately stop using the gauge and notify the manufacturer.  Repair or
replacement of cracked insertion rods should only be performed by qualified
persons, authorized by license to perform these activities.  The manufacturer
or other licensee authorized to performed the repair or replacement will
indicate the manner in which the gauge should be returned for service.

The incidents and investigations discussed in this notice also demonstrate the
need to determine the appropriate weld material, configuration, and processes
for a particular use application when designing gauges of this type.  The
choice of weld specifications should be based on the requirements of the gauge
during normal and accident conditions of use that it is likely to encounter
during its working life.  It is imperative that manufacturers identify these
likely conditions as best as possible and design their gauges accordingly. 
Although the choice of weld material, configuration, and processes for model
3400 series gauges with serial numbers below 13301 appears to be sufficient
under normal conditions of use, it appears to provide a means by which a crack
could begin in the weld area if subjected to higher than normal bending loads..         
			                                                   IN 96-52
                                                            September 26, 1996
                                                            Page 7 of 7


This information notice requires no specific action nor written response.  If
you have any questions about the information in this notice, please contact
the technical contact listed below or the appropriate regional office.




                                    Donald A. Cool, Director
                                    Division of Industrial and
                                      Medical Nuclear Safety
                                    Office of Nuclear Material Safety 
                                      and Safeguards

Technical contact:  Douglas Broaddus, NMSS
                    (301) 415-5847
                    E-mail:  dab@nrc.gov

Attachments:
1.  Troxler Bulletin Issued June 1996 (Text Only of Bulletin Attached to       
    Electronic Copy)