ML20084T663

Advises of Safety Sys Relay Failures.Caused by Misalignment Resulting from Carrier (Armature) Not Being Properly Guided. Paint Removed from Face of Pole Pieces for Relays in Field

ML20084T663
Person / Time
Site:	Palisades
Issue date:	07/29/1970
From:	Stuart I GENERAL ELECTRIC CO.
To:	Boyd R US ATOMIC ENERGY COMMISSION (AEC)
Shared Package
ML20084T657	List: ML20084T663
References
NUDOCS 8306230406
Download: ML20084T663 (3)

Text

.

• ^

' ' ; l} ,

h.3),I((g4kk, l [ f,h h5 h l h NUCLE AR ENERGY C 0 M P A'N Y "" "

J m

in CUE 1 Nit Avf . SAN JO5f, CAUF 93125 AatA CODE 408, TEL. 297-3000 TWXNO 408 287 6484 ATOMIC POWER fQUIPMENT DEPARTMN July 29, 1970 [

Mr. R. Doyd c.

Assistant Director, Licensing 2 R'FQbi.gh ,

USAEC Ab3 797Ca Washington, D. C. 20545 ( u.s m ., __

glg. g[pr .

SUBJECT:

\ j SAFETY SYSTEM RELAY FAILURES I ,W Q '!! -'

arm Jw 9 At a recent meeting with some of your technical staff a request was made to provide -

information on reported ralay failures. -

evaluation of the cause and corrective action taken.Following is a description of the failures, '_

request. I trust this will fulfill the <

I. Descrip_ tion of Failure In April of this year our Nuclear Instrumentation Department (NIC) was infomed that three canrgi nd. separate safety system logic AC relays had failed to open un n de-It was further reported that ene relay cperated normally

" armture was opered manually. The relay was removed from the panel 6.. er the examined Spring tension appeared normal and all parts appeared clean upon disn albly. 4 Residual nagnetism caused by a DC fault was Easpected as a possible f.ilure mechanism, but no such fault could be found, II.

gg Aps ' f ,/ a Evaluation of_ Cause /W t

/

3 One of the relays was returned to the vendor for further camination. The vendet discovered that portions of the moldcd plastic contact i rrice were chy.ptd and -

the conclusion was reached that the failure mechanism was due to misaligment  ;

resulting from the carrier (arnature) not being properly guided. All sites were i instructed to inspect and replace any chipped. carrier guides. At th nue time, the vendor initiated a design improvement program to find a stronger guide material and to reduce the shipping damage by improved packaging.

/

Continued testing at HID and the vendor revealed that the cause of sticking was i not the carrier guides, but rather the paint on the pole piece which adhered to -

the steel armature plate of the energized relay. To prove this mechanism, P.9 _

AC relays and 32 DC relays had the paint removed from the pole piece and re-placed with a thin teflon coating. Twenty-nine (29) additional AC relays and 32 DC relays were used as received from the vendor. All 122 relays were mounted _-m in panels similar to those in which the reported failures occurred and were ener-gized at rated voltage for a period of 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br />, de-energized and observed for proper operation.

Meanwhile, the vendor wa's performing a similar test using -

clamps to provide added pressure between the armatures and the pole pieces.

Five relays were untreated and 10 had teflon applied to the armature. _

8306230406 700010 PDR ADOCK 05000255 0.%M S -

PDR ,

.: DM . 0 :

b[llER AL $ ELEC C July 29, 1970 .

R. Boyd PAGE ,

TheresubsofthetestsareshowninTableI(vendorresultsinparenthesis) and support the conclusion that the mechanism of failure is adherence of the paint on the pole piece to the armature plate while the relay is energized. _

The fact that the DC relays did not stick was thought to be due to their having -

E a lower operating temperature,(no eddy current losses).

i 1ABLE I E

l Relay Type Quantity Number Stuck

?

f DC 32 0 g DC-Treated 32 0 AC 29 (5) 7 (4) J F

F AC-Treated 29(10) 0(0) '

The seven inoperative relays were manually opened and re-energized 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> to .-

g determine if heat curing the paint might be a poss-ible curb ageinst recurrance of the failure. All seven operated correctly, at the end of 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> (as did .-

e the other 112). On the basis of these remlts and the recommendation of the  %

vendor all reactor sites with flID-supplied relays of the same. type and vendor -

r were instructed to energize the relays for 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> unless it'could be posi-tively established that a particular relay had been energized for that period ^

or longer. In the latter case, the relay was exempt from the 48 hour5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> run if it could further be proven to have operated properly upon being de-energized. P K

• In addition, all flID-equipped sites were instructed to de-energize the normally .

energized relays once every 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> for a week following the 48 hour5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> run. A .;

[ similar instruction was issued to flID I4anufacturing for relays in the shop. -

k In June, word was received from one of the overseas sites that ralays were still

_ sticking after five cycles of operation. -

E III. Corrective Action ,

t Discussions between 111D and the vendor resulted in the conclusion that the paint -

g ' was most likely thermoplastic, and the only solutions to the problem were to p either remove the paint from the pole piece or to coat the armature with a -

material to which the paint would not adhere. The vendor recommended a teflon i paint which had been tested and proved satisfactory. When requested to provide --

i proof that the taflon pa. int would last for the life of the relay, the vendor de- #

f murred. The material was too new to have real time life data and the relay life s was too long to allow Arrhenius curves to be generated since the necessary extra-

polation would not be defensible.

5 Further investigation by the vendor revealed that relays made by others used the -

same pole piece material, uncoated, met MIL-C-2212, and had not experienced any k

corrosion difficulties. Also, another relay type made by the vendor of the il problem relay was using the'same pole piece material uncoated.

' " - - --= Y

&__g% _ .

s l \'

, , ~..'

~

, DE!! Fila L !2) E LE COC O July 29, 1970 Mr. R. Boyd l PAGE On the basis of this information,111D and the vendor decided to remove the ij paint from the face of the pole pieces for relays in the field, in flID's 3 i

shop and in the vendor's shop. The vendor will leave the pole face unpainted for future production. A Field Memo was issued by flID to all NID-supplied i sites instructing in the paint removal process, precautions to be observed and tests to be run following reassembly.

The vender reporteo on July 1,1970 on the results of .a corrosion test on a relay in which a pole piece with an uncoated face was submerged in salt water for a period long enough to build up a corrosion layer. The pole piece was reassembled into its relay and tests were run which gave no indication of degradation of performance. The vendor was requested to devise and run a test which would prove that unremoved paint from surfaces other than the face of the pole piece would not migrate to the face and cause stici.ing. It is felt that since only a small percentage of relays with paint on the pole face stuck in the field and in the tests described above then it most likely required an optimum thickness to provide sufficient adherence. The small amount which might migrate to the face is highly unlikely to be such an amount and, further, if the paint has sufficient strength to bond it most likely is too viscous to run. Results of these tests will be made available to you as soon as they are available.

IV. Conclusions The relay in question has been inanufactured and used for 30 years witnout prob-lems and with a record of high reliability. For this reason, it was e b en for this critical application. Discussions with the vendor indica'te that rior to 111D4 purchase of the relays which failed, the vendor moved his manufacturing plant and was in the process of training new personnel. This change in person-nel and lack of control of paint thickness resulted in the failures.

I feel the steps taken as described above in paragraph III, plus incrased usting and surveillance by llID Quality Control adequately preclude recurrance of this failure.

If I can be of further assistance in this matter, please let me know.

Sincerely,

,s ] ..G

• ' ' I. F. STUART, Manager Licensing Unit M/C 632 IFS:jb l cc: S. Levy A. P. Dray

• J. Barnard H. Hendon

, 4

.