ML20211B838

Part 21 Rept Re Misoperation of Ite 50D Overcurrent Relay Due to RCA Pnp Transistor Which Developed Abnormally High Leakage Rate Over Time.Relay Replaced.Rca Failure Analysis of Transistor Indicated Leakage Caused by Ionic Mobility

ML20211B838
Person / Time
Site:	River Bend
Issue date:	02/11/1987
From:	Kaiser A BBC BROWN BOVERI, INC. (FORMERLY BROWN BOVERI CORP.
To:	Taylor J NRC OFFICE OF INSPECTION & ENFORCEMENT (IE)
References
REF-PT21-87-037-000 PT21-87-037-000, PT21-87-37, NUDOCS 8702190573
Download: ML20211B838 (2)

Text

- ---_ - . . . . - - -. = . _ . - --. --

t

. . , . r. .

Brown Boven.

BBC Switchgear Products Group Bac arown Boveri,Inc.

4379 County Une Road Chalfont, PA 18914 Phone:(215)822 4270 February 11'1987 Telecopier;(215)822 4271 I

i Mr. J. M. Taylor, Director Office of Inspection and Enforcement U.S. Nuclear Regulatory Commission -

Washington, D.C. 20555

Subject:

BBC Brown Boveri ITE 50D Overcurrent Relay (10CFR Part 21 Report)

Dear Mr. Taylor:

On August 2,1986, BBC Brown Boveri was informed by Gulf States Utilities, River Bend i Plant of a misoperation of an ITE 50D overcurrent relay (Catalog Number 238T4575).

The misoperation was caused by a PNP transistor (RCA Type 40319) which developed an abnormally high leakage rate over time.

The relay was repaired by replacing the transistor with an Raytheon 2N2904A pending an evaluation of the failure by RCA. Replacement transistors were furnished for similar units at the site.

The nature of the misoperation is that the relay will go into its " trip" condition even though its input current is below the operating value. The target indication will not be given, which could cause confusion amoung operating personnel as to which device tripped the circuit breaker.

The abnormal transistor was submitted to RCA for failure analysis. RCA has concluded that the abnormalleakage condition was caused by ionic mobility. This condition was not detected by RCA's standard production line testing and screening. RCA has indicated, however, that the quality control sampling program now in place on the transistor production line would not permit such a processing problem to persist. Sample units from the RCA bipolar production line are periodically removed from the line and subjected to i a battery of tests including a high temperature-reverse bias test. This IITRB test will

detect the ion migration which resulted in the abnormalleakage condition. It has been concluded that while there may be a problem with a given lot of transistors there is not j an overall generic problem with the basic device, f BBC Brown Boveri's experience with this transistor supports this conclusion. Since j December 1976,43,000 units of this transistor from 50 different RCA production batches i have been applied successfully without problems. The one batch with the abnormal i

failure rate has been definitely identified. The anticipated total failure rate of transistors from this batch is greater than 10 percent.

,- 8702190 AU g Q h Se PDR FDR I S O

,o y-, .,-%,-,,-..,e---.-,,.- ww-. .-3.w y - -:-.-- y.-- , -+-.----.., - -. - - , - -y.. - , - - . -. - - - r - , ,, - - - - - -y +. , --wr - . , - . . , - - --- - - , 9.w . , , - w-,

BBC BROWN BOVERI Mr. J. M. Taylor February 11,1987 Page 2 The only other relays that were manufactured and shipped using transistors from this one batch for Nuclear applications are at GPU Nuclear Corporation, Oyster Creek (Serial Numbers 4899 and 4950 procured on their P.O. #PP025935; and at Duke Power Company, Oconee Station (Serial Number 4884 procured on their P.O. M-28723-78). GPU and Duke Power were notified of this condition by telephone on February 10, 1987 and will receive a copy of this report. Replacement transistors and instructions for installation will be furnished, or the relays may be returned to the factory for replacement parts.

A. F. KAISER, PRESIDENT Switchgear Products Group EWR/jm Attachment