Protecting People and the EnvironmentUNITED STATES NUCLEAR REGULATORY COMMISSION
ACCESSION #: 9803250033
GE Nuclear Energy
General Electric Company
175 Curtner Ave. San Jose, CA 95125
March 19, 1998
98-01NR2.DOC
MFN 012-98
Document Control Desk
United States Nuclear Regulatory Commission
Washington, DC 20555
Subject: Spring Return Binding in GE Type SBM Control Switches
This letter supersedes my previous letter (same subject) dated January
23, 1998. Information is provided concerning the possible failure of
certain GE Type SBM Control Switches with the spring return function to
reset properly after operation. The failure mode has been identified as
binding due to an interference between the rear brass bearing and the
phenolic rear bearing support. GE Nuclear Energy (GE-NE) is
conservatively assuming that SSM Control Switches with the spring return
function manufactured since March 1996 may be subject to this failure
mode.
GE Type SBM Control Switches are manufactured by GE Electrical
Distribution and Control (GE-ED&C) Power Management, Malvern, PA as
commercial grade items. GENE has dedicated these switches and supplied
them to several licensees as basic components for unspecified, safety
related applications. Since the specific applications and associated
safety functions of the switches are not known to GE-NE, we have informed
known affected licensees pursuant to 10 CFR Part 21.21 (b) so that they
may evaluate the condition. However, since additional licensees may have
obtained these devices through other dedicating entities, we cannot
assure ourselves that all end-users have been notified. We are therefore
providing tills information to the NRC for appropriate action.
On January 6, 1998, GE-NE was advised by a licensee that a safety related
SBM Control Switch with the spring return function had failed to reset
(return to normal position) properly. The switch (model
16SBMB3A02S1S2P1) was returned for evaluation by GE-NE with the support
of GE-ED&C Relay Product Engineering. Additional failed switches were
subsequently identified and also returned for evaluation.
The failure mode has been identified as binding due to an interference
between the rear brass bearing and the phenolic rear bearing support The
root cause has been determined to be "post mold cure" shrinkage of the
phenolic material. Post mold cure is normal for this material, and takes
place over a period of one-to-two years after molding.
GE's investigation has found that the mold used to make the phenolic rear
bearing support has worn over the years to a point where the diameter of
a part being produced, while still within specification, is near the
minimum allowable value. It has also been determined that the existing
design does not address normal post mold cure shrinkage. Thus, switches
that arc initially functional can exhibit sluggish return or binding
several months after assembly. The switches identified by the licensee
as failing to reset properly had been operating correctly for
approximately one year.
The failed switches have been examined and, in each instance, the bearing
support hole was found to be undersized and the bearing was at, or near,
its maximum allowable diameter. Testing has confirmed that switches with
bearings that are at their nominal diameter function properly, even when
the support hole is undersized. Consequently, the failure is not seen in
every switch. The failure is most likely preceded by a gradual increase
in the force required to operate the device, and a sluggish return.
Although the interference can exist in any SBM Control Switch, GE has
determined that this is only an issue in those switches that have the
spring return function. In spring return switches, the reset spring does
not generate enough torque on the operating shaft to overcome the
additional friction resulting from the interference. Thus, the operating
handle, and consequently the contacts, may not reset when the handle is
released from the momentary positions The switch contacts will function
when the switch is operated and, if the switch is manually returned to
the reset (normal) position, the contacts will also return to their
normal configuration. Measurements indicate that the torque required to
return a switch to its normal reset position is less than two inch-
pounds. During functional testing of several switches that exhibited
spring return binding, a licensee's operator was able to place the
switches in the reset (normal) position in all cases.
"Maintain position" SBM Control Switches may also have the interference,
but the additional force required to overcome the friction resulting from
the interference is not great enough to prevent proper function.
Nonetheless, it is considered good practice to assure that these switches
are in the proper position, as indicated by the switch pointer, before
releasing the handle.
A loss of torque in the tie bolts that secure the switch components has
been observed in conjunction with the spring return binding. The loss of
bolt torque is also the result of post mold cure shrinkage and, although
it has been noted in those devices that have failed, it is not the cause
of the binding and does not indicate impending failure. Loss of tie bolt
torque due to shrinkage of the phenolic materials was considered in the
switch qualification, and requires no additional action. There is no
need for inspection or retorquing of SBM Control Switch tie bolts.
All of the failed switches identified to date were manufactured in the
period from November 1996 through February 1997. There have been no
reports of spring return binding in earlier SBM Control Switches, despite
the fact that these switches have been widely used in a variety of
applications for many years.
2
Although test results demonstrate that most of the post mold cure
shrinkage occurs in the first twelve-to-eighteen months after molding,
GE-NE is conservatively assuming that SBM Control Switches with the
spring return function manufactured since March 1996 may be subject to
this failure mode. These switches can be identified from the last two
characters (date code) shown on the green "QC Acceptance" sticker. The
affected date codes are PL, RL, SL., TL, UL, VL, WL, XL, YL, ZL, NM, OM,
PM, RM, SM, TM, UM, VM, WM, XM, YM, ZM, NN and ON.
The SBM Control Switch design has been revised to address this
phenomenon, and switches manufactured beginning March 1, 1998 (date code
PN) are no longer susceptible to the failure mode.
There are two safety concerns with regard to spring return binding: (1)
possible damage to control circuitry caused by the circuit being
maintained in the momentary position for a prolonged period; and (2) the
possibility that control circuits will be prevented from performing their
proper function by the failure of the switch contacts to return to their
reset (normal) configuration.
The loss of the spring return action does not prevent the SBM Control
Switch from performing its function of providing manual circuit control
switching, and GE-NE's evaluation has determined that neither the need to
manually assist the reset action nor the loss of tie bolt torque will
prevent proper switch function or degrade the device qualification.
If SBM Control Switches with the spring return function, manufactured
since March 1996, have been installed in safety related applications, it
is suggested that station personnel who operate these switches be advised
to return the operating handle to the normal (reset) position after
operation. This will preclude any control circuit difficulties.
It should be noted that all of the switches identified to date as failing
to reset properly had previously operated correctly for approximately one
year. Control switches that are at least two years old are not subject
to the failure mode (binding due to post mold cure of the phenolic
material). Any corrective action or preventive measure undertaken in
response to this Notification should not be necessary after the subject
switch is at least two years old.
If you have any questions, please call me at (408) 925-1019.
Sincerely,
Michael A. Smith, Program Manager
Safety Evaluations
3
cc. S. D. Alexander (NRC-NRR/DISP/PSIB)
G. C. Cwalina (NRC-NRR/DISP/PSIB)
J. F. Quirk (GE-NE)
H. J. Neems (GE-NE)
G. W. Sanders (GE-NE)
J. A. Steininger (GE-NE)
J. Teague (GE-ED&C/Malvern)
GE-NE PRC File
4
*** END OF DOCUMENT ***
