United States Nuclear Regulatory Commission - Protecting People and the Environment

Information Notice No. 97-08: Potential Failures of General Electric Magne-Blast Circuit Breaker Subcomponents

                                 UNITED STATES
                         NUCLEAR REGULATORY COMMISSION
                     OFFICE OF NUCLEAR REACTOR REGULATION
                         WASHINGTON, D.C.  20555-0001

                                March 12, 1997


NRC INFORMATION NOTICE 97-08:  POTENTIAL FAILURES OF GENERAL ELECTRIC          
                               MAGNE-BLAST CIRCUIT BREAKER SUBCOMPONENTS


Addressees

All holders of operating licenses or construction permits for nuclear power
reactors.

Purpose

The U.S. Nuclear Regulatory Commission (NRC) is issuing this information
notice to alert addressees to potential failures of six subcomponents in
General Electric (GE) type AM or AMH 4.16-kV circuit breakers that can render
the breakers inoperable.  The subcomponents in question are (1) the trip
crank, (2) the CR2940 contact blocks that make up the power switch assembly,
(3) the manual trip lever and its supporting "L" bracket in the AMH horizontal
drawout breakers, (4) the cotter pin that holds the latch pawl hinge pin in
place, (5) the spring charging motor tie bolts, and (6) the type HMA control
relay.  It is expected that recipients will review the information for
applicability to their facilities and consider actions, as appropriate, to
avoid similar problems.  However, suggestions contained in this information
notice are not NRC requirements; therefore, no specific action or written
response is required.

Description of Circumstances

Trip Crank Failures

The NRC has learned that several plants have experienced failures of the trip
crank (GE Part No. 105C9316G1, Piece No. 28 of Figure 1 in GE ML-13 Mechanism
Renewal Parts Bulletin GEF-4379) in Magne-Blast circuit breakers.  These
failures occurred when the pin at the end of the crank broke off.  The pin may
break off the crank when the trip coil is energized.  The trip crank pin
inserts into a hole in the lower end of the link between the trip crank and
the trip coil armature.  If the pin breaks off before the trip crank can
successfully rotate the trip shaft (which has been the case in most
instances), the breaker will fail to trip electrically (although it can still
be tripped with its local manual pushbutton).


9703140142.                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 2 of 8


Discussion

Trip Crank Failures

GE has attributed the broken pins to three principal factors:  (1) lack of
adequate control of one of the critical machined dimensions on the pin during
the early 1970s, (2) lack of adequate fusion in some of the pin-to-plate
welds, and (3) grinding of the weld reinforcements flush on the back of the
trip crank plates.  Upon being informed of the first instances of pin failure
in 1988 (at Tennessee Valley Authority's Watts Bar Nuclear Plant), GE
instituted more rigorous quality control checks on the pins and finished trip
cranks.  GE also revised the pin weld detail on its trip crank fabrication
drawing (105C9316) and added "DO NOT GRIND FLUSH."  Several failures were
reported after the initial Watts Bar report,               
and many potentially susceptible breakers were found in the field, all with
their original trip cranks made in the early 1970s.  However, no instances of
failures of trip cranks manufactured after 1988 have been reported.

Also, if the remote trip signal (either from a protective relay or a manual
hand switch) is applied for more than a few seconds (which it normally is) and
the breaker fails to trip (such as it would if the trip crank pin broke),
neither the breaker-mounted auxiliary switch nor the stationary (cubicle-
mounted) auxiliary switch will signal control circuits that the breaker has
opened, and thus the trip signal will normally remain applied.  Energizing the
trip coil (which is normally energized only momentarily) for an extended
period may open-circuit the coil, thereby rendering it permanently inoperable.

Trip cranks that are potentially susceptible to this failure can be identified
without disassembly of the breaker mechanism.  With the mechanism front cover
removed, the gap between the trip crank and the right side of the mechanism
frame may be seen.  It is then possible to see whether the weld reinforcement
has been ground off.  GE is preparing a service advisory letter (SAL) on this
problem in which it intends to recommend replacement of any trip cranks that
do not have the proper thickness of pin weld reinforcement (1/32-1/16 inch). 
GE Philadelphia Operation (GE PO) can furnish replacement cranks.

Description of Circumstances

Contact Block CR2940 Contact Resistance

On February 12, 1996, the FitzPatrick licensee experienced failure of two
residual heat removal service water (RHRSW) pumps to start on demand because
their supply breakers failed to close.  RHRSW pump C failed to start on demand
during monthly surveillance testing and RHRSW pump A failed to start when
attempting to place it in service as part of a suppression pool cooling
evolution.  The licensee's investigation found that the Magne-Blast breakers
failed to close because high resistance across one of the power switch
assembly contacts prevented the closing coil from being energized.

.                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 3 of 8

Discussion

Contact Block CR2940 Contact Resistance

The power switch assembly consists of three GE type CR2940 contact blocks
stacked together so that all three sets of contacts are actuated by a single
striker.  Two of the contacts (1-2 and 3-4) are normally open and are held
closed by the striker during the spring charging operation.  When the charging
cycle is complete, the contacts spring-return to the 
open position to cut off power to the spring charging motor and the control
(anti-pump) relay (52Y).  The third set of contacts (5-6) is normally closed
and is included as an option to allow remote indication of the closing spring
status (charged/discharged), usually by means of a white indicator light in
the control room.  This third contact is often called the "white light"
contact for this reason.  This contact is wired into the breaker control
circuitry such that failure of the contact to close will prevent the breaker
closing coil (52X) from being energized and the breaker cannot be closed
electrically.

The licensee determined that the CR2940 contacts were misapplied in the Magne-
Blast breaker control circuitry because the contacts are rated for only 2.2
amps dc and are required to interrupt 6.0 amps dc (Licensee Event Report
50/333 96-002, Accession No. 960410298).  The licensee also observed that the
contacts seemed to show signs of arcing (blackened, pitted surface) after
about 2,000 operations, even though the recommended breaker service life is
10,000 operations.  General maintenance instructions in GE Technical Manual
GEI-88771D, "Magne-Blast Circuit Breaker," states that the 1,200-amp breakers
are capable of performing up to 5,000 operations and the 2,000-amp breakers
are capable of performing 3,000 operations before any replacement of parts
should be necessary.

Resistance measurements across the failed contacts varied between 200-1000
ohms.  Contacts with 1,500 operations or less did not have the arcing
indications, nor did they have high resistance readings.  The licensee also
noted that there were no recommendations to check the contact resistance
during periodic preventive maintenance in the vendor's maintenance manual. 
There was disagreement between the plant's drawings and the manufacturer's
wiring diagrams.  The manufacturer's wiring diagram indicates that the 5-6
contact should be jumpered out when not used.  One of the plant drawings shows
that when the 5-6 contact is "not furnished," it should be jumpered.  The 5-6
contact is not shown at all on the plant RHRSW pump circuit breaker elementary
drawing.

The FitzPatrick licensee has also experienced failure of CR2940 contact blocks
used as latch checking switches in Magne-Blast breakers, even though the
contacts do not experience significant "make" or "break" current.  The
licensee believes that these failures could be related to aging or the number
of operations and is evaluating whether periodic replacement may be necessary.

In a letter dated June 14, 1996, GE Nuclear Energy informed the FitzPatrick
licensee that the suitability of the CR2940 contact blocks in the ML-13
operating mechanism for the Magne-Blast breaker was confirmed by testing the
breaker in accordance with applicable American National Standards Institute
(ANSI) and National Electrical Manufacturers Association .                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 4 of 8


(NEMA) standards.  Operability of the contacts was demonstrated by breaker
life cycle testing of 10,000 operations with no failure of the contact blocks,
and there is no requirement to replace the contacts on the basis of age or the
number of operations.  However, GE stated that according to applicable NEMA
standards, the maximum number of operations between servicing is 2,000.  The
operations are listed on the basis of servicing at intervals of  6 months or
less.  GE also stated that although the published instructions do not
specifically address the contact block resistance, instructions for checking
the control power during servicing include measuring the operating voltage at
the closing coil, the trip coil, and the charging motor terminals.  GE
believes that this type of testing would reveal whether the contacts required
replacement.  GE stated that the wiring diagram clearly indicates that the 5-6
contact should be jumpered out when the "white light" function is not
utilized.  In addition, the drawing shows that another CR2940 contact used as
a latch check switch in the closing coil circuit should also be jumpered out
when this feature is not used.

GE concluded that although the contact blocks were suitable for use in the
Magne-Blast breakers, the operability demands of the nuclear power industry
and the recently reported problems from the field indicated that the contact
blocks were a weak link in the design of the
control circuitry.  GE recommended the following actions in the June 14, 1996
letter:

.    In control schemes where the "52 SM/LS" (5-6) contact is installed but
     not utilized, it should be jumpered out of the circuit.

.    In control schemes where the "52 SM/LS" (5-6) contact is installed and
     utilized for "white light" indication, but the "auto reclose" function
     is not used, the wiring should be revised to remove the contact from the
     close coil circuit.  GE can furnish a revised wiring diagram and
     nameplate.

.    For the CL/MS application, where the contact block is used to break
     charging motor current, GE is evaluating a replacement device.  The new
     switch will have a higher dc interrupting rating and will be furnished
     for those applications where breaker applications require the increased
     durability.

GE plans to issue a SAL concerning the CR2940 contact blocks in March 1997.

Description of Circumstances

Bent Manual Trip Lever and Cracked "L" Bracket

During surveillance testing in June and July 1996, the licensee for Calvert
Cliffs identified two problems with type AMH-4.76-250 (horizontal drawout)
Magne-Blast circuit breakers.  In the first case, a low-pressure safety
injection (LPSI) pump breaker failed to close.  The licensee found that the
trip lever was bent and there was no gap between the trip lever and the manual
trip rod.  Although no gap value is given in the vendor manual, there is
generally a small gap between the trip lever paddle and the manual trip rod. 
The bent trip lever prevented the trip latch from fully rotating onto the stop
pin, resulting in a less than optimal .                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 5 of 8


area of contact (wipe) between the latch and the stop pin.  As a result, the
breaker would experience intermittent failure to close.

A second LPSI pump circuit breaker failed to close during monthly testing at
Calvert Cliffs in July 1996.  Investigation found that in addition to the trip
lever's being bent, the "L" bracket support for the trip lever was also
cracked.  The "L" bracket is designed to support the trip lever and provide
additional stiffness.  A subsequent inspection of other breakers at Calvert
Cliffs found that one other breaker had a bent trip lever and two other
breakers had cracked "L" brackets.

Discussion

Bent Manual Trip Lever and Cracked "L" Bracket

GE performed extensive testing on one of the failed Calvert Cliffs breakers
and concluded that the most probable cause was insufficient trip latch reset
spring force caused by either incorrect or damaged springs originally
installed at the factory.  GE recommended a modification to the Calvert Cliffs
breakers to prevent further cases of trip lever bending and "L" bracket
failures.  The modification consists of replacing the trip paddles, the
support bracket, and the spring discharge link.  The trip lever material was
changed from American Iron and Steel Institute (AISI) 1005 carbon steel to
AISI 1018 carbon steel.  The "L" bracket was changed from AISI 1005 steel to
aluminum.  The configuration of the components was also changed.

The modification corrects for the weak spring and allows the breaker to retain
operability with the weak spring installed.  Replacement of the trip latch
reset spring is not part of the normal maintenance or overhaul activity. 
Replacement of the spring requires that a V-notch be cut into the breaker
angle support to allow removal of the trip shaft.  The Calvert Cliffs licensee
plans to replace the weak springs in the breakers during the next scheduled
overhaul.

The modification kit is available as Catalog No. 0172C8186G001.  GE plans to
issue a SAL on this issue by April 30, 1997.

Description of Circumstances

Cotter Pins for the Latch Pawl Hinge Pin and Charging Motor Tie Bolts

On September 13, 1996, the licensee for Vermont Yankee Nuclear Power Station
discovered during a tagging procedure that the "A" emergency diesel generator
(EDG) was inoperable.  The EDG output circuit breaker (GE type AM-4.16 kV
Magne-Blast) was found in its normally open position, but its closing springs
were discharged.  With the springs discharged, the breaker was incapable of
closing.

Subsequent investigation by the Vermont Yankee licensee determined that the
spring charging motor had run to failure because the cotter pin that holds the
latch pawl hinge pin in position broke.  The ears of the cotter pin had
apparently broken and allowed the cotter pin .                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 6 of 8


to fall out, thus allowing the hinge pin to work its way out of position and
prevent the latch pawls from holding the ratchet wheel in place during the
charging operation.  The charging springs were not compressed, and the
charging motor continued to run until it overheated and the motor winding
open-circuited.  Three of the four charging motor tie bolts that connect the
motor portion to the gear housing were also found lying on the floor of the
breaker cell.  Vermont Yankee personnel inspected other similar breakers and
found that 18 cotter pins were either degraded (one or both "ears" broken off)
or undersized, and in one case a cotter pin was missing from the latch pawl
hinge pin.  Three breakers were also found with one or more loose charging
motor tie bolts.

On November 25, 1996, after learning of the event at Vermont Yankee, the
licensee for FitzPatrick performed an inspection and identified 10 out of 18
safety-related Magne-Blast breakers with degraded cotter pins latch pawl hinge
pins.  Similar to the failure at Vermont Yankee, the cotter pins had one or
both ears broken off.  One undersized cotter pin was also found, but it was
not broken and the licensee determined that it had been installed by plant
personnel.

Discussion

Cotter Pins for the Latch Pawl Hinge Pin and Charging Motor Tie Bolts

The latch pawl hinge pin was originally designed in 1962 to be held in place
by cotter pins at either end.  In 1979, GE enhanced the design of the hinge
pin assembly by tapping an existing hole in the hinge pin support bracket and
installing a bolt with a washer large enough to overlap the hinge pin.  Using
the bolt and washer to hold the hinge pin in place precluded the need for
cotter pins.  According to GE, this enhancement was made only to aid in
disassembly and reassembly of the breaker during maintenance, and not because
of any perceived problem with the cotter pins.  As a result, GE did not deem
it necessary to inform                                     
customers of the change in 1979.  Testing performed by GE in 1996 demonstrated
that the cotter pins may experience damage after approximately 2,000
operations.  GE plans to issue a SAL on this issue in March 1997.

Two different styles of charging motors are used in Magne-Blast breakers. 
Initially, GE used motors manufactured by the Sioux Tool Company of Sioux
City, Iowa.  In the early 1970s, GE switched to motors made by Millers Falls
(later bought by Ingersoll/Rand).  In the late 1970s, GE went back to using
the Sioux Tool Company as the charging motor supplier for the Magne-Blast
breakers and still uses it today when customers order replacements.  

The two different types of charging motor can be easily identified.  Two black
cover plates conceal the tie bolts on the Sioux motors, and thus the bolts are
not visible from the outside.  The cover plates have to be removed to gain
access to the four bolt heads, and the tie bolts are inserted from the motor
housing into the gear housing.  In contrast, the tie bolts on the Millers
Falls (Ingersoll/Rand) motors have exposed heads and are inserted from the
gear housing into the motor housing.  The motors with the loose bolts at
Vermont Yankee were Millers Falls motors.  

.                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 7 of 8


Description of Circumstances  

Type HMA Control Relay

On December 1, 1996, a Magne-Blast breaker serving as a vital bus feed breaker
failed to close on demand during surveillance testing at Salem Nuclear
Generating Station.  The licensee determined that the HMA control relay (the
anti-pump relay [52Y]) normally closed contacts failed to reclose when the
relay was deenergized because of binding of the armature against the molded
phenolic post.  With the contacts stuck in the open position, the closing
circuit cannot be completed and the breaker cannot be closed electrically.

Discussion

Type HMA Control Relay

The relay was sent to the vendor (GE Power Management [GE PM], Malvern,
Pennsylvania) for detailed failure analysis.  The vendor found that there was
no clearance between one side of the armature tailpiece and the molded post. 
Normally, when an HMA relay is assembled at the factory, the armature is
centered between the two molded posts with a gap of  0.005 inch on each side.

The vendor recalled that a similar situation occurred in 1982 and prompted the
issuance of SAL 721-PSM No. 171.1, "HMA Relay Armature Binding," on December
17, 1982.  The original SAL stated that a tool problem at the factory in 1974
caused several relays to have improper clearance between the armature and the
molded posts.  The SAL suggested that the proper clearance could be achieved
by first removing the armature stop clamping nut and lifting the stop and
armature tailpiece from between the molded posts, and then removing some of
the phenolic post material.

The NRC discussed this issue with GE PM.  The vendor stated that the armature
could be checked for the proper clearance between the armature and the molded
posts by use of feeler gauges.  A gap of less than 0.002 inch on either side
indicates an adjustment is needed.  However, the original SAL stated that the
solution was to remove some of the phenolic material from the posts and did
not mention that customers could first try to adjust the armature to achieve
the proper clearance.  If the relay does not have the proper 
clearance, usually all that is needed is to loosen the armature stop clamping
nut, center the armature between the two posts, retighten the nut, and then
check the clearances again.  The vendor also stated that although the
recommended minimum gap given in the original SAL is 0.005 inch on each side,
a gap of 0.002 inch is considered adequate for reliable operation. 

Related Generic Communications

GE issued SAL 073-352.1, "Latest Design Configuration:  GE Type AM Circuit
Breakers and Medium Voltage Switchgear," on July 7, 1995, to alert customers
to design changes made in the circuit breakers, their operating mechanisms,
and the switchgear.  Some of the listed design changes were discussed in
previous SALs, while other changes were not originally conveyed to customers
because the changes were made to facilitate assembly, maintenance, .                                                            IN 97-08
                                                            March 12, 1997
                                                            Page 8 of 8


or operation of the equipment.  The SAL states that customers should evaluate
each item listed and consider the applicability to their particular equipment.

Recent NRC information notices (Ins) concerning Magne-Blast circuit breakers
are as follows:

IN 90-41, "Potential Failure of General Electric Magne-Blast Circuit Breakers
and AK Circuit Breakers," issued June 12, 1990.

IN 93-91, "Misadjustment Between General Electric 4.16-kV Circuit Breakers and
Their Associated Cubicles," issued December 3, 1993.

IN 94-54, "Failure of General Electric Magne-Blast Circuit Breakers to Latch
Closed," issued August 1, 1994.

IN 96-43, "Failures of General Electric Magne-Blast Circuit Breakers," issued
August 12, 1996.

IN 96-46, "Zinc Plating of Hardened Metal Parts and Removal of Protective
Coatings in Refurbished Circuit Breakers," issued August 12, 1996.

This information notice requires no specific action or written response.  If
you have any questions about the information in this notice, please contact
one of the technical contacts listed below or the appropriate Office of
Nuclear Reactor Regulation (NRR) project manager.


                                       original signed by M.M. Slosson

                                       Thomas T. Martin, Director
                                       Division of Reactor Program Management
                                       Office of Nuclear Reactor Regulation

Technical contacts:  Kamalaka Naidu, NRR
                     (301) 415-2980
                     E-mail:  krn@nrc.gov

                     Stephen Alexander, NRR
                     (301) 415-2995
                     E-mail:  sda@nrc.gov

                     David Skeen, NRR
                     (301) 415-1174
                     E-mail:  dls@nrc.gov

Page Last Reviewed/Updated Tuesday, December 03, 2013