ML20102A440
| ML20102A440 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 07/15/1992 |
| From: | ELGAR CORP. |
| To: | Tomlinson R TEXAS UTILITIES ELECTRIC CO. (TU ELECTRIC) |
| References | |
| REF-PT21-92 NUDOCS 9207270046 | |
| Download: ML20102A440 (10) | |
Text
I~
t
[
9250 Brown Deer Roco a San Diego.Confomra 02121 m 610 450-0035 m Te.x 211063 p
via fax 817-897-6482 15 July 1992 i
Texas Utilities Electric Conanche Peak Steam Electric Station
{
B M. Road 56 J
Glen Rose, Texas 76043 T
Attn:
Ron Tomlinson
Subject:
Deviation in Elgar Inverters
.:=5 d
Reference:
- 1) Docket Nos. 50-445 and 50-446 d
d
- 2) ' tone & WebstP lener %BS-BFr5C, oated 08 January 1992 g
- 3) Elgar letter dated 20 March 1992 x
- 4) Elgar letter dated 13 January 1992
- 5) Elgar letter dated 25. March 1992
- 6) Elgar ECN 8208, dated 10 April 1992
- 7) Elgar Quality Alert
Dear Mr. Tomlinson,
After extensive review of the Elgar participation in the bring up of the Elgar Inverters, Model 103-1-132 for Unit 2, the following is the summation of causes and corrective actions taken for the specific deviations noted.
It should be first noted that the Inverters were first supplied to Texas Utilities in the 1981 time frame.
All of the noted deviations of operution wers recorded in the fall of 1991 and subsequent.
It should also be noted that none of the deviations (ref. 1) caused a safety related concern, as the units are not currently in operation, in a
safety related application.
The first not.ed deviation was the blowing of the F1, fuse during start up of the Inverters.
This was due to improper operation of the timing circuit for SCR firing on card J7, Elgar Part Nurter 628-126-42.
This s first reported in Sep* ember 199L, and the
[
root cause not dett c 9 ined until Elgar went tc tUE in February 1992.
It was determined, oat the U103, C101/C102 circuit had a tolerance band issue.
It was only evident on spare PCB's supplied after delivery of the Inverters, presumably to replace PCB's taken out to assist in the start up of Unit 1.
This was corrected by Elgar ECN 8162, bringing the PCB up to revision L.
The failure analysis was submitted ano approved by Stone & Webster on behalf of TUE (ref.
2).
This appears to be a generic problem with current production of these PCB's, so all affected sites have been notified by letter from Elgar (ref.
3).
All PCB's made available to Elgar were reworked and fully tested 7.t Elgar enen returned to TUE.
The second noted deviation was lack of record of notification of TUE of a chance to card J3 Elgar P/N 643-119-40.
No record is retrievable at Elgar or TUE of the notification, and it is unclear 9207270046 920715
/({l PDR ADOCK 05000445 I
S PDR Fox Numoers i
^10-4584267(Sales) e 610 458-0257 (Acm:n ) a 610-458-0225 (Puren ) e e10-J53 0231 (Rnonce)
t whether notification was required.
Elgar did issue a notification
-to all affected sites currently holding the subject PCB (ref. 4).
Current events and notifications show the systems are in place to inform all required sites of any deviations.
All PCB's were reworked on sit.e.
'The next noted deviation was the high inrush current to the Inverter.
This was traced to the Charger Gate Drive PCB.
The PCB in question was supplied as a spare board.
All the originally supplied boards of this type functioned correctly.
The root cause was the transformers on the PCB were not wound in a fashion which eliminated the noise from one secondary winding to the other.
The secondaries were wound bifilar (on top of each other, at the same time) rather than concentric (one after the other, separated by insulation).
This caused noise between the secondaries and allowed the SCR's to fire prematurely.
A quality Alert (ref 7) was issued to the fransformer Department reminding them not to wind transformers bifilar, unless specifically required by the drawing.
Also, all affected sites were notified to have any PCB's which were not origina1' equipment be verified by test or returned to Elgar, for test and rework if necessary.
All PCB's on site which were made available to Elgar were reworked to the latest revision.
The next noted deficiency was that the J2'PCB, Elgar part number 628-134-40, was noted as missing a resistor, R144.
An Elgar ECN number 1899 added the resistor, but when the ECN was incorporated into the drawing the R144 was left off the parts list.
This allowed Elgar to build and inspect the PCB without noticin, "he missing. resistor.
This again was the result of an ECN after shipment of the original units.
This has been corrected by a later ECN bringing the revision of a correct PCB to revision H.
All applicable sites were notified of the error (example ref. 5).
All PCB's on site available to Elgar were reworked to the latest revision.
A workmanship error was noted on the crimps of four (4) pins in
' Inverter CP2-ECIVEC-04.
It is unclear when the discrepancy happened, as_the units had been delivered in 1981 and many parts were removed by TUE personnel to get Unit 1 power supplies up and running.
There was no documentation available to Elgar to determine what was removed or replaced from the time of delivery from Elgar.
This was verified to be the only unit, of all the Inverters on site, inspected by Elgar, where crimps were found to be discrepant, Proper tooling and new parts were provided to the Elgar representatives on site and the crimps properly were reworked.
The next noted deficiency was the requirement to ' field select' a value for resistcrs R3 and R4 on Elgar assemblies 628-432-41, 64i-
j 305-40 and 644-311-40.
The resistor installed at Elgar worked for the_ load:used to test _the operation of the circuit at Elgar, but with the:TUE load, a different-value.was. required.
The correct values were calculated and added by Elgar ECN-8208__(ref 6)'to the.
noted assemblies, to be the Elgar installed values.
This'will-
-assure that any spare-PCB's supplied to TUE will have the correct value for the TUE's_ load installed, at Elgar prior to delivery.
During current start up testing, a wiring error to transformer A2-T3 was noted.
Again Elgar believes that the wiring error may have been caused by reworked performed at TUE af ter delivery from Elgar.
The units passed all tests at Elgar, including _a 100 hour0.00116 days <br />0.0278 hours <br />1.653439e-4 weeks <br />3.805e-5 months <br /> burn in test prior to delivery to TUE in 1981.
This testing would have caught this type of wiring error.
This wiring error was corrected and the unit _ operated per specification.
The last noted deficiency was cracked solder joint on a terminal block on PCB 634-105-40.
Aftgr further investigation two other PCB'_s of a_similar design had cracked solder joints, Elgar part
~
numbers. 643-124-41 and 643-125-40.
These PCB's are single sided without plated thru holes.
When the terminal block screws are tightened beyond a certain torque, the solder is subjected to a
~
an adverse stress.
This over time will crack the solder joint.
It was recommended to
- TUE, and accomplished by the Elgar representatives on site, to check dll three PCB's and resolder any
-joints, showing sign of stress or cracking after the screws have L
been tightened.
This would alleviate the residual stress left on the solder joint due to the tightening.
This was the_ design which went through all the environmental' and seismic testing without failure. -Due to this information, a letter has been generated to all sites-which havs this_ type of PCB design, requesting them to inspect all of the noted types of PCB's.
Any sites with concerns have been told that the current Elgar design utilizes a feed thru hole with pads on both sides, and they can procure new PCB's, if they desire.
This will provide additional support to the terminal
-block pin.
While Elgar-was on site assisting TUE/Bechtel in the start up effort from December 1991 to February 1992,'many deficiencies were L
noted in the trouble _ shooting techniques utilized during start up u
of the Unit-2 Inverters.
L l
One of-the specific cases was the use of a piece of aluminum as a fuse, to replace a fuse which kept _ blowing.
This allowed the continued trouble shooting, while possibly advers ly af fecting the unit.
The: _ fuse blowing problem should have been addressed technically, not by replacing the fuse with the aluminum block and possibly damaging the unit.
Other areas of concern were the lack of proper ESD practices, and the inability of the Elgar Technicians to-contribute'to the trouble shooting, due to tne inability to be
- _ - - - - - -. - ~ ~ - - -. ~
h at n d s : o n.
This was all brought up to the attention of TUE management.
This resulted in a much more cooperative effort between Elgar and:TUE, resulting in the completion of returning the units to an operational _ state.
Also the lack of documentation of what was removed and replaced-in the units'over the 10 years TUE had the units prior to this start up effort is of concern.
It is therefore Elgar's position that we will only warrant the work we performed and documented, and not the units as a whole.- All of Elgar's work is documented and available on site at TUE, and at Elgar.
Elgar will continue to support Texas Utilities in their effort to complete the start-up of Unit 2 in any way-we can.
Feel free to contact me at (619) 458-0247 if we can be of assistance.
Sincerely, di'
)
Timothy Roth Manager Quality Assurance / Customer Service attachments cc R. Danie1<
NRC - Document Control Desk
1
. REFERENCE 2 -
l STONE 6 WEBSTER ENGINEERING CORPORATION COMANCHE PEAK-STEAM ELECTRIC STATioH-P.O. BOX 1002, FARM ROAD $6 GLEN ROSE, TEXAS 76043
.u....
"!"J* * ".
CPSES-9200592
~,7,".,",..,",,*r:o *.*,*' *-
25WEC 9200025
. =
VBS-BF05C Date: January 8, 1992 No Response Required
,1
-Mr. Timothy Roth
'Elgar Corp.
9h0 Brown Dear Road California 92121
SUBJECT:
NOTIFICATION OF DOCUMENT STATUS
~
REFERENCE:
1.
Purchase Order No. S0026853 7S2 2.
Specification No. 2323-ES-009, Rev. 1 3.
VL No. 7594, vendor letter for failure analysis of PCB
- 628 126-42 dated 01-06-92 4
SWEC Letter =CPSES-9130470, dated 12-02-91 5.
SWEC Letter =CFSES-9200457, dated 01-06-92
Dear Mr. Roch:
By copy of this letter please be advised of the status of the following document which was submitted in accordance with the referenced purchase order.
' VENDOR DOCUMENT TITLE ITEM DOCUMENT NUMBER SHEET REV.
DESCRIPTION STATUS
.1 Elgar letter N/A N/A Vendor letter for _
1 dated Jan. 06, 92 Failure analysis of PCB 628-126-42.
Key to the status of documents:
1.
Approved 2.
Approved with Comments 3.
Not Approved 4
'For Information only l,
inv STONE 1%EBsTER-I
CPSES. 9200023 IMPORTANT It is understood that the approval of the Contractor's documents whether general or detailed is a general approval only.
Such approval shall not relieve the Contractor of errors, discrepancies, or omissions of detailed requirements; nor shall such approval relieve the contractor of any responsibility for the proper execution of the work or performance in accordance with the Contract Specification and Contractor Drawings.
If you have any quantions, please contact Bharat Tailor rc (817) 897-8500, eat.
7574.
All erittan correspondence including document resubmittal must be addressed to:
DCC Vendor Document Group TV Electric P.O. Box 1002 Glen Rose, Texas 76043-9990 ftY$
J.E. Woods C28 SWEC Project Engineer Unit 2 JLil/vv cc:
E. Lavigne C28 F.A. Garcia C07A F.D. Stobaugh C07A W. Lieneck AP3 B.T. Thompson C07A B. Tailor AP3 CCS E06 VETIP Coordinator F04 Job Book #10-1 C28 Chrono File C28 A
,... sto m m s m.....
l REFERENCE 3 ELG6R 9250 Broe Deer Rood
)
kn Diego. Cohfanio 02121 TJephone (619) 4500085 T: lex; 211063 2G March 1992 via fax 817-897-6482 1
Texas Utilities Comanche Peak Steam Electric Station F.M. Road 56 Glen Rose, Texas 76043
Subject:
10CFR21 Notification for Elgar PCB 628-126-42
Reference:
Start Up of Unit 2 at Comanche Peak SES
Dear Ronnie Tomlinson:
During the Start Up sequence of a unit at the referenced subject site it was determined that the subject PCB did not operate ps require.
This problem is only evident during Start-up sequence of the unit from an Inverter off status.
If the Equipment is operating the problem does not occur.
After conclusion of a Failure Analysis on the subject PCB, it has been determined that three components on the PCB were inadequately toleranced, capacitors C101, 102 and Integrated Circuit (IC)
Z103.
This could allow pulse widths outside the original design intent, thus not allowing sufficient time for capacitor precharge.
This could cause the Inverter fuse to fail at turn on.
Elgar has initiated ECN 8162 to change the above noted components.
Elgar is recommending that all PCB's delivered after original delivery and start up be returned to Elgar for verification to Elgar Acceptance Test Procedure T28-126-42.
The PCB's installed in operating units should also be verified.
If they do not meet the units timing requirement, we recommend that the PCB's be brought up to the current revision level, revision L, per Elgar ECN 8162.
1 If you require any further information feel free to contact me at (619) 458-0247.
l l
l Sincerely, f J.aa 9
Timothy Koth Manager Quality Assurance / Customer Service
e
~
i-REFERENCE 4
' 0250 Brcwn Deer RocJ
- Son Diego. Conferno 92121 Tewonone (619)t.504085
- Teler 211063 TECHNICAL BULLETIN 13 January 1992 Mr. Mike Cook TU Electric Company 5 Miles N.W. of Glen Rose TX off FM 56 Glen Rose, TX 76043
Subject:
Technical Bulletin rio.
2, PCB. P/N 643-119-40 D' ear Mr. Cook:
The subject PCB had a change done to, resistor R124 wh$ch decreased the sensitivity of the 'out of sinc' detect circuit.
This change reduces the amount of false 'out of sinc' alarms.
Integrated Circuit gIC) Z107 was also changed to another IC which is less sensitive to small frequency changes, thus eliminating some static transfers for units which use pin 15 of this board.
The-resistor' changed from Elgar part number 801-473-05 to 801-153-05.
The IC went from Elgar part number 849-C40-11 to 340-C40-93.
Elgar recot. mends that during the next service of the unit, these resistor and IC be changed.
This will bring t';.e PCB revision up to revision H or higher.
Any spare PCB's should be returned to Elgar for changing of the cepacitors.
This can be
-done at a minimal charge of $450.00 per PCB.
If you require any further information feel free to contact me at (619) 458-0247.
L Sincerely, Y=
l Timothy Ro}th l
Manager Quality Assurance l
REFERENCE 5 0250 Brown Deer Road Son Dego. Col:fornia 92121 tieohone (619)4504085 T & r 211063 25 March 1992 via fax 817-897-6482 Texas Utilities Comanche Peak Steam Electric Station F.M. Rocd 56 Glen Rose, Texas 76043
Subject:
TUE 10CFR21 notification of assembly error on Elgar PCB 628-134-40
Reference:
Elgar trip to TUE Comanche Peak SES, 9 March 1992
Dear Mr. Tomli'ason,
After revjuw of the trip report concerning the referenced tripf Elgar has evaluated the missing R144 and determined that it is reportable under 10CFR21.
It was found that some of the subject PCB's were missing the resistor R144.
It was found that ECN 1899 to the subject print was not fully incorporated.
This change added resistor R144.
Tne resistor was never added to the parts list or the assembly print.
Therefore the resistor was not put on the last PCB's delivered to TUE.
Testing at Elgar utilizes a voltage source, like the actual unit, to set the low AC current set point.
This will not verify the presence of R144.
The error in the process was the incorrect incorporation of the above noted ECN.
The print has been corrected, and forwarded to TUE site personnel.
Elgar recoma.
's that all of the subject PCB's be verified for the presence
' this resistor.
Any found without the resistor need to have tnem added.
If you require any further information feel free to contact me at (619) 458-0247 Sincerely, W
- /&
/h-t.e Timothy Roth Manager Quality Assurance / Customer Service
REFERENCE 6 C %)./.
~
NsdN ENGINEERING CHANGE NOTICE g
PREPARED SY [MMEE/g,v' OATE //#7/ft SH T. / OF /y Ecu riru zw An& coxeevr dr.Au mak.sra.e 01v nacxx
~
eMMKJCT3 AFFECTED PROOUCTS AFFECTED MODEL 3salAL MODEL SEMIAL R EvisiON iNcomponATED TYPE DOCUMONT NUNCER v/.6 /48*/-dII. /7d, /'j '7
- 8
'8
'N'T O'T8 I
/
' ~
hM d.28 - /.32-sf/
0 '0
(~;n_
ggf.gog,gg
_ C h CHANGE TYPE ECRS RESOLVED d'
O pm00,iMPmovtMENT M-3// 4(-d C h O aELiAsiLiTV O cost ME DvCTiON Q sA7ETY/UL CHANGE CATEGORY dCJy/'
.fjS-#26-(,f/
- f. 6 g
Q AEFLECT As SulLT Q MANDATomY p
gg. y/g-Q g
{
'Q Q 00CERMOR O luenOVEMENT C DESIGN ERROM DOCUMENT ONLY
[Mt'-gog go
'h f OTHER CW RONLY NUCLEAR g
f 1
STOCK O!SPOSITION
[USE O aEwOak O cCaAr G SEE ATTACHQ,,O2 agGut.sTEO sv OAfg map A
DEPT.s*EAO CA
,f. Q oAtt dxrwse ArAr dl.gN.iwo/> Wo TEAM r 4-*L.9&rt& cont. f(4-;y, CdNTROLLED/-fH sum,suisem. soma ne < r m.a>r NUCLEAR gDUCT DESCaimOn OF FaOsuu
~
4'.0wfnu. arac y dwsT' &Adur ? k&f1446 eMl? MP3 A2;ccu Ado Tb arf 2:vsa Ap.asrYD /~Xtu1 doAo OESCf "lON OF CH ANGE ffWSd A"ff/dfbA VAlthS~.dn'O toAffA4if' M.:Af :
2,,3 80/- /$0-65 WS42, $W 22. st?O Y s,ZC/- 04~
M J:2,.6)V Rd 4 Osrm'd 72, : JV7-Ma-as 1: s,q Sw A 'a r c. 4 ? e A M w ty re A' @ A w.4 c m- " Ws r d Ps a r.-c p t g.o;; g o x g
.4 ACAf 4/d 7' 4*tamf,Jk:w1liSv'7'::fut$.ott~C7* 70 4;' War M48 4 6 42~CO 'WO.
26MA
~
Wry.S And Penl'w'fdf'7'yrscAi)
CVA LWATJON Aase.u6 ll']l-;*)],
f
/10A /d " r/'YdA 4k./7*
/000.'/
.$ M7'5
~
."p.g. asps, J4' / Mu7444"d? /.6~> o' s r.g.ptf sw, =%z9 23-A'4x' /
7bPd 2"Mo.w ab' 238.frt ~ (!@4% DF) M, 294'.,w/$ diiDE rbers = /. 2w
~
M-au <~~re xsse,.u:<< rsi-,.,,,,,ee a m,s a a.e
~7A*'#2 /4.6% 4 c. ~ o %. //0 sM/S 6.6~yngg a,fy m
CCS APPAOVALS
@ APPROVE O aE;ECT 70s Hsif O peelNG MAMP.
DATE QA DATE i CON P1'1-C T CAfg E P P E CTiv e TY
= $O
- Of y
9y z
. A,E. -
fM../L.A. -r 2. A 7s *4.%. -., sk
REFERENCE 7 Quality Alert u.
- COP.RECTED COPY **
f internal h
Corre;ponde nce To:
N. Nguyen, S. boa b
From:
T. Roth
Subject:
991-052-90 transformer Date 03/19/92 Problem : Some of these transformers have been found to have the secondaries wound bifilar. The print requires them to be wound concentric.
It was found out that the practice in the magnetics department was to wind multiple secondaries bifilar, if the' turns ratio / inductance are the-same.
This may induce noise in the output from the secondary.
Action:
Production: Build in accordance with the print, and do not wind any multiple secondary windings bifilar.
In-Process Inspection: Verify all subject and other trans-former secondaries are wound concentric.
This can be accomplished by a simple capacitance check between the secen-dary windings.
cc: P. Kelly (Supervisor Magnetics)
J.
Yee (Manager Production)
QA Personnel
- -.