ML20077H677

Post-Test Analysis of 1153 Series a Pressure Transmitter for Westinghouse PWR Sys Div

ML20077H677
Person / Time
Site:	Waterford
Issue date:	01/18/1979
From:	ROSEMOUNT, INC.
To:
Shared Package
ML20077H671	List: ML20077H672 ML20077H676 ML20077H677 W3P83-1853, Forwards Addl Info Required to Resolve Aging & Operability Issues Re Environ Qualification of Rosemount Transmitters Model 1153A.Encls Represent Justification for Interim Operation
References
17912A, NUDOCS 8308110204
Download: ML20077H677 (16)

Text

-

h-ki. '

./

(?l.f ~:

pI 4 ose-de-2&

9. -.

.,.n.

p e-

?'.

A-

~,. -

y,

..., y..

4.-

e<

a.*2,- <

ACO*.dOUNT iMO., F OST OFr.0 hoa 3512S mI'8 S T ' fi~'- :58h # II' U; I*f fC

c..

TA/.. ; i?6. *..

.!Y 25 Oi!3

n,

f.. e.=*

t c..

. t, $

T P03T-TEST ANALYSIS

\\

0F l

1i53 Ss!Ei A PEZ3537.~. TFANSMITTEF.S I

1 FOR iESTINGOUSE Pa'E SYSTEMS,1YISION l

RMT FEPDFT NO. 17912A I

D I

j

=

t

/ 't

, 1'

.i

-1.-

Pre:sric by

.,. - s....../"

A ','.\\

% ta,' ? * T jf.iinterin/~

A:: roved by _ k, f

/

.Y o

26#3:~M-e Lite /'JN9 jQCality Assuran:e

/

/

r..

.=-..r.

h.

s.

., c. i n.

.v.

s w

1500 S s 7% t.:=:t 8308110204 830805 Aufmec y 3,i,A.lv.

ii.!23 PDR ADOCK 05000382

("+'g e.c.* * " * ' ' ' *

• A PDR

f. '9. ;,i.t.irtn No.1C215.4

~

j ' C'.1. L g*

t

__,S

?.' VISION STAT *JS i

9*

^~

[

!.- 1 1-l etaph 1 Cht.ge Descripten

A;;

21M

,Eev.,..,,.,.._-

l l

1 1

i l

l s.

9 f

I d

I i

i

- - c, t

e i

i s

8 I

l i

l j

e i

t, l

1 I

t e

s g

I e

.t I,

4 l

9 s

G 1

. r. ;

f 4

0

.,e Oa.'.*

e' M

.as==

a i...

e

.et.

l F35T-TEli A'iA'YSIS

.h 1153 SERIES A PRESSdRE TRA'iSMITTERS FOR WESTIN3H3USE PE. SYSTEMS DIVISION RMT-REPORT hD. 1 W.2A 1.0 SCOPE This report sam.arizes the results of a failure ar.alysis performed by Rose-mount ; -hree Rosemount 1153 Series A pressure transmitters wr.ich were quali,

fication testec by Westinghouse PWR, Systems Division during October and November,1975. it.ree of the four t'ransmjtters tested failed during.the post accident mor.isoring portion of tne test. Also, cne of these three transmitters exhibited irradic performance at an earlier stage in the test program.

Results of this analysis are also discussed.

2.0 BAMGD.3TQ 2.1 Westinghouse parchased four Rcsemount 1153. Series A precsure trar.smitters for the parpose'of assessing their performar.ce to Westinghouse Equipment Specificatio1 953320. Rev. 2. Sections 2.30 and 2.31. The test was per-formed te levels exceeding the pu'blished Rosemount specification for the 1153 Series A.

A comparisen "of Westinchouse and Rcsemount test levels is shown in Table 1 celow.

~

TEST ROSEMOUNT LEVELS WESTINGHOUSE LEVELS 5

7 6

Radiation 4.0x10 rads gama 8 5.Dx10 rads /hr.

5.0x10 rads gama 8 2.5x10 eads/hr.

Seismic Single Axis Sine Sweep Biaxial Random 3.5g Maximum to IEEE 344-1971 7.0g Maximum to IEEE 344-1975 Steam 350'F for 10 minutes w'#

370*F spike 300*F for 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> T [t l' 320*F for 20 minutes Ch:mical i

-' e r,-

250*F for 56 hours6.481481e-4 days <br />0.0156 hours <br />9.259259e-5 weeks <br />2.1308e-5 months <br /> Ramp to 220'T during 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> Post None d

4 g j 220*F for 3~5 hoars

. Accident TABLE 1 - QUALIFIEATION TEST LEVELS l

/,

m-4

=

i

J Transmi:ters su;. plied for the test are listes in Tatie 2 bel:,:.

2.2 h

.2:ing5:use!

~

l Calibrated I Serial Nu:ber i Delig.a-for.

j i

8 RM' Model No.

Typ

! Ranga 1

i 1153GA9 l

Giugi Fresssre j, 17C0-2500 PS*G.I, 135',27 l A%-1

___m-li53G39 l Ga ge Pre: sura l 0-3000 PSIG

{ 135540 l AW-2

' !.7 -I 1153DA5 l

Differentia *..~ressure ! 13S-52 H23 1355 5 i

1153DA5 j

Differential Presure i 475-21D" H,0 135945 i AX-2

_t TA5LE 2 - R".T TRANSMITTEF.S SUPPLIED TO WESTINGC.'SE The Westinghouse test consisted of sequ'en':ial exp:...:re of all four transmitters 2.3 to radiation, seisnic, sten? chemical and post acciden z.ging environments.

All tests were performed by Westinghouse personnel. Various phases of the test program were observed by R:semount personnel.

After corpletien of the test seqLenca, g es of the four tran'mitters were returned to Roseroer.t fcr s

analys's.

2.0 FAILURE ANAi.YSIS Tne analysis cf es:h transmitter failure is discussed in the :110 wing para-graphs after a brief sue.ary of each transmitter's performance results up to the pein of failure. The performance sumary is based cn ir.fccmation supplied by Westinghouse and/or witnessed by the observing Rosemount personnai.

3.1 Si11359t?(AW')

3.1.1 Perfcn.sr.:e This transmitter exhibited the larges: error of the four transmitters tested during radiation exposure, 50.2% of span (13.5% of upper rang limit)..

During seismic vibration, the largest error was 3.8% of span. During the steam test, the maximum error was approximately -13% of span (-3.5% cf upper range lidt),

factcring out the chamber pressure effect. -.This transmitter performed thrcugh-1 '.,','

out the test and was within 4.3% of' span (1.2% of upper range limit)' accuracy-

' * ~ "

after completion of the test.

~

'3.1.2 Analysis Results This transmitter was not returned to Rosem:unt for analysis, but the unit was Q

visually inspe:ted by Rosemount personnel at Westinghcuse.

l 1

RMTRepori179:2A.

-__p-

As car. te seer. ' rc Fi;;re 4.2 c' r.s,.estir.:*.:.;se re::rt, the ra:ia:i:n

~

• h e#fect was s.;5 isrgely to the d:5. rate and n:: the giy!-; t.p; c sage sir.: tr'iXlut tente: tc w a rec:verv wher.;.er the :si...:r. ::rce m

W1s lowered.

Rosecour.: has experienced similar but lower magnitude errors on 1153 trans-mitters during radiation testing.

The dose rates used in Rose :;e. tes-irg r

6 6

did not exceed 0.55 x 10 rads /hr. compares to 2.2 x 1C rads /nr. in the Westinghouse test.

Radiation errors are caused pre :mir.ar.tly by a change ir. the input offset voltage of the integrated circuit operatior.a1 ampli' '

fiers. This effect is'emplified by k.he circuit gain,'which is p.o;crtionsi to the range'down factor.

In this case, the range down factor was 3.75:1

.= 3.75).

Since ea:h integ*ated circuit possesses (I '*

hs - 1700 psi a different degree of radiation hardness, the large errer of tnis unit is believed to have been causec by one or mere of the three op-amps in the ur.it in co.bination with the high gain asse:iated with this calibration..The radia-tion performance of the otner three test units demonstrates that tnis large effect is nct typical f 1153 Series A performance.

3 SN135945 (t4'-3) h

~.2

.2.1 Pe-forman:e The worst case cerformance during radiation was 4.0% of span (4.M of upper rangelimit).

rior to seismic testir.g. a calibratien test it.ci:ated irracic performance of the trarsmitter.

It was found that the transmitter exhibited a linear outpat frc 0 to approximately 1500 psig. At presseras above 1500 psig, the output was high (in excess of 20 r%',.

This problem disappeared

(

wnile oeing analyNd and the sei'imic test was initiated and cc.if.c ed without.

further inciden:e.

Post. seismic testing sgeir. exhioited the problem. but good calibration data was subsequer.tially obtair.ed prior to steam testing., Durir.g j

l steam testing, the output was monitored at a set point of 2100 psig. Irracic high ouput again appeared during steam testing up un-il the fir.a1 (15th) day of tha test at which poir.t the output went to a Icw (2.19 d.) value.

Duringpost-testcalibrationtesting,theauthutrerainedatthis',cwvalue.

3.2.2 Failure Analysis Resultr.

Rosexcunt's failure annlysis began by isolating the cause of the constant low (2.19 mA) output value. The 'tilure was a short between the primary Q

, and feedback windings of the oscillator transf:mer. The trar.:ferrer was replaced and the following calibration data was taken:

Rv.T Reoort 17512A -

a

In? t Outp.:

?ressare (:sici Cu-rer:

,,h 0

4.0EE 1453 11'.S45 1503 26.050 26.054 3300 rur;her analysis cf the trar.sformer failu-e is dis: : sed i-Se: ier. 2.f.

This data indicated that the sensor was again exhibiting the probler. that,,

appeared prior te seismic testing and during the steam test. Tne circuit boards were removed and tested separately to verify that the pretiem was due to a sensor defect. Subsequent pressure cycling of the senser causeo the problem to disappear and the following calibra:ior, data was re:orded when the unit was reassembled.

Inout-Outpa:

Pressure (psi:1 Current (r.a.)

0 4.064 12.110 1500 g

3';03 -.

22.124 The irrasic sensor benavior was indicative of a intermittent shor: te: ween the sensor's center diaphragm and the deposited filn capaciter plate cn the low pressure side of the Dp cell. Tne cell was disasse:r. bled' to expose the deposited capacitor plate' surface. Visual inspe: tion of this surfa:e under a

~

microscope revea'.ed tha' the vapor deposition adhesion to the c.i' f"'.'. tJbe sarfa:e was of poor o.uality and exhibited a flaking tppearance.. No evidt:nce of free conductive particles was found otherwise. Thehighpressureside capaciter plate was 'also' examined for a similar defect.

It was fcur.d to have adequate adnesion to the fill tube surface.

~

-"I'S

~

Adhesior. of the vapor deposition to'the cell half 'is audited on each ir.-process J

2 lot by gample testing. The technique use:i is a cellophane tape pull test.

A:ceptan:e of the lot N' bas 6d on 51.0%'AQi..,It is judged that this failure

~

i could have ap;e' red in normal service and was r.ot d e to the qualification a

testinc since only radiation exposure had been completed before the symptem appeared.

Previcus radiation testing on dozens of similar transmittars has e

never brought about this type of fa'ilure.

mm

• 3.3 5ri25; 5

'4-l' 3.3.1 Perf:- a.:e The worst :ase psrformar.:e during racu-io was 13.2!. cf s:ar. 'S P '. ;,.t r range lin#:'.

..ing seisri

testing, the w:rst case errcr was 3.;'i :f s:ar..

Durin.g the stear chtni:El, test, tna w:r.t :ase errer was 7.0% cf 3ta. (4.21.

of upper range limit). The output signal wen: high af:er 8 days f post-a:cidert c:nitcrir.;. The signal returned :s nor ai periodically until after 12 days when it remained high for the rest of the test.

3.3.2 f_ailureAr.elysisResults The failure ana' lysis cor.sisted of isciating the :ause 6f the constant nigh (25.4 mA) outpet. The failure was a defe:t'ive output' stage operational anplifier. The op-aap was repla:ed ar.d the follosir.g calibraticr. ctta was taken.

~

Input Output Pressare (in H2O)

Current (S';

133 4.24' 359 12.244 O

~

s=

20.213 Further ana'ysis of the op-amp failura is dis:ussed in Se: tion 2.7 3.4 SN125945 (AX-2',

3.4.1.

Perfomence The worst case perfomance during radiaticr. was -15.9% of span (-5.K of upper-rangelicit).

During seismic testing, the wors: case error was 6.E*; of span.

During the steam chemical test, the error went up to 23% cf span (5.C% of upper range limit), but it was found that the Conax fitting had leaked, allowir.; the caustic spray solution to enter the teminal side of the electronics housir.g.

After the termir.al c mpartter.t was cleaned and resealed, the trar.s.itter cperated for three days cf post a:cident monitoring before a high output failure c: curred.

1 I

l

'~3.4.2 Failure Ar.alysis Results

'The cause of the high output failure was found to be a snort betwear. the primary l

and feedback winding of the oscillator transfomer. The transformer was replaced fs

'and the following calibration data was taker.:

y 1

~

RMT Report 17512A '

J

Inout Outoat Fressure (in N Currer.- (mM b'h 475 4.032 422 7.2'2 SE5 10.455 31E 13.078 253 16.892 210 20.103 FurtheranalysisofthetransformerfaildreisdiscussedinSection3.6...

The terminal sice of the electror.ics housing was pressurized via a special cover to determine if the Conax fitting was still leaking.

No detectable leaks were found at 13.5 ps'i pressure, indicatir.g that the extra RTV applied to the seal afte,r its ir.itial failure had held :troughost the remair. der of the test.

Additional temperature testing of this transmitter was performed tc determine the cause of the 23% errcr encour.:ered durir.; steam testir.g. After heating the trar.scitter tc 330*F in an eve., a groar.cing test was perfom.ec.

Data was taken with no ground, positive power supply,groanded and negative power supply grounded. The results cf this test indicated t..at at high teT.perature there was significant electrical leakage currant from the tranc=itter circuitry to the esse. Through a process of eliminat on, the leakage was i

isolatec to be betwee the termir.a1 block positive signal termir.a1 and the electronics housing. The cause of this electrical leakaSe was probably caustic spray soluticr. residue tripped behind the tertins1 block, as a resu*.t of the Conax fitting le.n daring the steam. test.

The transmitter was reassei.iled'with a'different electronics housing and retested at 330'F.' The error during"this test was +13%,cf span (4.6% of upper range 3

limit) and the groundi,ng effect,wns significantly lower than the prenous test

, with the original housing.

.2..

0 RMT Report 17912A.

3.5 Che-i:t1 Resida-in Electrori:s H usin:

( Insice ne cover: Of en:n of thE sie:tronics housin;s, a gray-: a:k re ic;e se: ermine if this easidee was caused by a let.; :r. tre e a:- ~ was prest. - tror.ic: tousir;... ring tne stet:: chemica". test, cr..;er E'a::r:r. Spac rc;ra '.i: ar.aiysis was pi #:r.e: at it.e Univa*sity of innes:.1 0. ; e recidae fec= the terr.dnal side cover cf SN1359'5 (N'-2). This te:hnig a is capable of detecting all elements ex:ept hydrc;en er.d helium. The residae was mixec with :olicical graphite in isepropyl al:o5ol to perform the analysis. The craphite was analyzed to establish its chemical compositien,' then ~the residue mixed with graphite:was analyzed to 3etermine tne chemical composition cf the residue. Th'e results c'i the ana'.jsis fc110w: ~ Ilements Sresent See:imen Colicidal Graphite Carbon, Oxygen Resitae mixed with Celic.:a. Scapnite 5iliccr., Sulfer, Carbon, Cxygen (i.:reased) and Boron (trace q;antities) () The rssidue dsfinitely contained sili:an, sulfer, an oxygen. Ca-bon may have been preser.t, but be:ause the residue was mixeo witn grspnite, the cart:n. coalc r.:t be cetected. Dr.ly trace gusntities cf Soren were present. The cor.:iusions that can be reached from this analy:is are: 1.- The residue was not-a corrosion product of t.se stainless steel covers. If it were,' elements such as iren, nickel and chromium would nave been ~ dete:ted. Since only traces of Bor:n and no Sodium were detected, this indtc'ates 2. that the residue was not produced from the chemi:als in the steam chemical test *,. indicating that the steam' did not leak into the ur.it. 3. The source of the silicon was probably th'e General Elect.-i; Rec F.TV used v ~ to seal the Conaf fitting. The source cf the sulfer'ecuid be ra:. dual i mechining lubricants.

• Test chemicals: 1.5 wt. percent toric acid adjusted to a n:minal pH o' 9.25 with sodium hydroxide.

[]g RMT Report 17912A.

.-=. .,.*e j 3,5 05:<'1st r Trans'or e-7a417+ :.e'. sis results ()h .~ ail;-e :f 2:se :un; transni te-s, serial r.u.ter: 125;48 'a. 21 ar.d 1355.: ', aere iso ~. ate: to t.e trar.sformer. F/t. 011E3. ; 7, 3h i'. t::rt. :

n e ec:es 774E ar.c 7745 resoec:ively.

The c:: va: failures a: seared :: be similar and there# ore e:tn d:-vices w&"e analy:et si.ltar.eousi.3 0 vice Descriotion 3.6.1 This torcict' transfor:Ier consists of a 69 tern prir.ary wir. ding .';.ifilsr) followed by three 303 turn seccr.daries (trifilar) ar.d fir. ally a 5 turr. fasdba:k winding. The concer ytre is ir.sulated with Estrasci 120 to ensure adequate isolation in 180t temperature and gemma radi2-tien envi.cor.ments. The wound core is wrapped with teficr. tape to pre-ver.t epoxy en:apular.- from perme'a-ing the wincings. The assembly is then er.capsulated in a cylindrical cup after the windings are terminated at solce-tinned copper pins. 3.6.2 Ooserved failure Mode' At the prirted wirine :::rd level, both devices exhibited thort circai.I -of less than 102 chms between their primary winding (cins i & 7) anc feedts:k windias (pins E 5 8). The s:e:ificatier control crs.:ing re-qtires insulation resistance between ail wincings at +2500 t: be 1,000 megchmt mir.ims: at 10,C VDC. ([) 3.6.3 Failure arsiveis 3.6.3.1 Failure M.cde Verificating ~ The devices were re=cved from tne circuit bcard and insulat on re-i sistan:e measurements were made with a Gen Rad :S53 megohmeter at 50 Enc -100 VDC. Reacings of les tnan 500K chms betwaen orimary and feedbacx winding were observed on bctn devices. The 100 ohm readings observed on the circuit board were not verified. 3.6.3.2 Encaosular.t Rer. oval The potting shell base and the epoxy from the top and bottom were renceed to expose the teflon tape. The original failure mode was stii'. preser.t. The tape was somewhat degraded from gamma exposure but had r.:t reatne.-. the expected powdery state..The tape was then removed revealing' ccmina-l tion leads near pins 5 through 8 twisted tcgether. Reference Figure 1. ~ ' Tape removal also revealed small amounts of scider flux. } ~ l-RKT Report 17312A ~ g .a.. 3 43 g

I hS . r ' ~ ~,- . ~ b g m.g e p u ::; % x ..s gh =e~am I P F.

• 2.:. -

p$.:. a: ??Mx * ~: j .s -2 ~=s ~ ~ u y :, k.~.. %h:A.-MI ! Idbihd,$A 44% t *%. iM-ER l ) iL 1 4 't.e:Md I.. , p;-)n..* f,' ~.~.~.? M f.:. s ~ s \\ l . ~. a.n+N1 . =.. tr2pc.-% J.- .t w;a r %.dW,' uesgr;f\\2& .3 [Q>fD. %-4 i I h+ 4... #..... s F e.t. ' %.,..% 'f~ ~... _. _..r w %w w.g.? y: i e-l Y5- %. %:' 4 ?%.WJ-n p;::;f P =:.y. f >. $ .'.'l' ' '.' LL.=t:.4/ d w. "w "..f;.d. S @u-v&- .I ^ -sI wa, w.,,.,-..e......... y t .J. h , m. .- _,n.,o m m. . :. i y m h f. Figure 1 - bin:ing Leads Twisted Together ? g 3.6.3.3 X o.ay Microc a:5s h The transfomers were sectioned in order to isolate the primary and h feedback pins from the toroid as a unit. The lead wires were cut a 's small distance from the epoxy, isolating the failure mode to tne pin temination area'. An h.P. 34702A D Y. indicated a 2 megch.T. short be- ~ j tween pins 6 an 7 of BH Electronics date ccde 7745 and a 2 megoh:a ,(. short between pins 7 ar.:i 3 of EH Ele::renics date code 7745.. Top and y side viev x-rays revealed the falicwing wcrkmanship deficiencies: L 7' Excess lead lengths D Poor lead dress I Solder balls Both primary ar.d feedback wires twisted together at one p ir.t. f, L Reference Figures 5 through E. p,. U I:. l lp, - l6 _ RMT Report 17912A - r j

\\ l 77n~ *W2Q~i5.M-.-??-?"M9""'" p _.1n ;-.=mg.u.-: i-g 4 no.v rag Sm 5g, y .r ' ' ~ W 1 ...p (, s +p,,,. ~~ Figure 5 - Date Co:ia - - - - ~. . # ;.-s. s v, 7; N'.z:.. (L.mM5 . p..-,.- s, - e.. . j_A,. 7,,,,5, '.ea :i Das i v-..=..'. or. d#49. ke,.s x s. . /hWM.W.1M*'f.N.N. 4-8 (1. to r.) S, 7, 6, 5 a... s . w w.. - ' l'nNf'NgM.,.f.: film. 1.2::xd.i-4....-~E W i. d.T. M.. E .. ~ -.:..:-

• J c

. =. .l-y$ [%.,.m.. u...; ... w - s % p,.:.. ~,.:- g p .,a.. e d ;.V '.7. T F - d. F E M % i'^ Y. s - $ N o.v/3 e G ;. - I y e m a s E M W. N E' m'.. *.* sat 2r q ......f... ;r s-p:.*.-. s $'.# $.disg 65 91& W h W 'F-W = ...m n:n. --C,v-u=:.m . i. . ntr M.? 2.~ewre e g.1s - y-4 ;. l j .-WdW-4-T.v'GMu%2,w.ys4 r.!.h. b i.' I .e >. ~..: >:m A. W~- . n. : >. m . %"e.v..s."W W eA-f.5:nbP' -M.o '. 3 .r-I Figure 5 - Date Code

-~ w,-g m

c I W.:.92 + .mu., .a y 4 4 b )

• N N N '.N N

'( or ,7 i . :mx 4,,,,

s.,

'4 ,/ M-4 /' .O< j i j P,', '. M :.; $ . M.'." d t.- .' bc .i. l i* 4 F:i 5 t-3 ~; # Sad

• 4-

.=raas ..a 1.. cgg e . *Y: V-:tWwc.r:::.h*fm. A tN~ R "*wca ~ Y.W.fs -,r,:q*tp>ac,g;p,w"MM.'.NJ&ld': . ~ nte

,s

~ D l . p.Aw.p..c - 4~4 ~ 5 a@.M e w w Zi74%.?.hw a w3. g Ed &%W? 2 %,. - g 3 R M LW-k.5: i I Oc RKT Report 17912A.

a, . IN '% ) kTi5?MMX5ERWW b:= g. s W+ a t n F T2 4 % ~ g-gw w=.emswer. .k?51'%#rWI@MihMNM2}M l b) Wl?ET W..xS5'A$'E$.h W N.% l?igure7-DateCole i .W-:.%. -. hw/s 't.*-- ..-c 5- ! 7745, Lead Designnion ar. +< @ (1. to r.) 5, 6, 7 B . : v\\,.* **=-MNM, i m t i.2 R.y.1. N['.%W. It.C.-h.Q.:&_M..? l I i . 4g.w, ' va'.".e.c. 'N 'w. 't ..":.mx.: ra..,. ,9-l nw :.~ P @ =q,M.,. - u......?alc 4 5.. Ot "'.f. e-i. h + r*.-M w W '"6 *-- 8 i .o < -s.s .....: h.%;t ;Lf-y.?.pn~=w.i.<,.. . --' ?-ft ?~"Ei.s. #.. .s,;.. :...s * -o ..~.,. /.#, l

• . E khs..8 =v=Q t.y..,=y.9.M *.G

, m. ..=~ ... M sv p.-.or. y'5 D h .ma-. t. g t.

p p.

1N'-M WE [YE** j*b. A \\/ 6 gr, w w.=m ga m.s ieg erg 4 s..q u e -

• a'. O W $ i N i.:

QA?;*iyRFESA.i$$&'* ef* E.i%.2g,.n-wdi-l ~ l

• a a = -cN.':t??2m.M.Md'..

I f. s~- 'I k Af t W h nu=ww2kF sh.ti.gM.:..*..*n.f. -@: V e.u ce .s.: .u+ :. t - b... - l , 5.[ h..:>.I M M N UE$ N M M.-.-. 5.$' fr-Ma.E. n., n.~ em:.ss.w n. mW: w~#m. ,1,~.r.-m R l .u w.ex.&pe w. .v. v g o.:u.2*$. &w.n a:$e m=em,.s., I i i s t -kon .;g h,

m. h U-.. N c-

[M i L ad D 9 r t. .w:u (1. to r.) 5, 6. 7. 8 k...s.. 4 T.EMW2 #S.4

3-

~C.% .'3 G. - i .$N.r.Ke:s.v$$..i@7.n@WdthsM;M l ~ .e 4 a 6 wWm sk."- T M % N.. @ d 3.r-w.w. % IE.-! l g w e.. ;.,a e. -s.~a.,e,m..$,w...&w x..,.g-. g -u .s ~. 4. o,4 @, W F 9 h - w 4 7. se.aWmt.'M i, .am..s.a m.w - v- ., :~, {4 RKT Report 17912A -11 4 e v 4

e-0"'I 3.6.3.4 E;_ yv Ir.5*.'2*i0n EeLi!!n* TP.e insLil: ten resisten e of a i inch square pie:e of at ::y Wai ;0r?:i#ac with that cf a ncn-irr2 disted transf r.T.er. Tr.s resista.:e of b;;t stroles was f.2 Y.10:: chts z.: 10; VD*. 3.6.4 Conclusic. The observed failares were n t reittee to a dificien:y in desi;r. or appear to be the resa'.; of cne - mere of the observed materials, but rather, A'.though the exact cause of failure could nct ba workcar. ship defects. determined, any one of the defects has the poter.tial of being tna sais cause of failure; excessively.lon5 terr.ir.aticn leads runnin; a:ross adjacent pins, terr.inatisr. les'ds tv.'isted.toge:Her and seider balls c'.cng - with solder fiux residue ir. the lead termination area. O i l i a-O 3 -~ RMT Report 17912A -

e

• s

.,6.

• ~.,. *..

3.7 0 erati:ne'. A :lifier Fa m._,e Analysis Resuitt Feilure c! Rosemoun; transmittar, serial nember 1.;5545 W.-1) 2:as isolated c an c:eraticnal am;1ifie, F/N 01153-0110. Nati,nal Seri:en:uctor date code 644. 3.7.1 Cevice Oa:-:H otion The National Semicondu: tor amplifier is an '_v.105, selec ed primarily for its icw su: ply current requirements. Tne device is hemetically packaged ir. c metal (TO-5) can. Single data code lots of 1030 pieces are,stadst - cally sampled and evaluated for radiation hardness in addition to normal 100'; functional testing at Receiving Inspec: ion. 3.7.2 Observed Failure Mode The nepiiffer output (pin 6) was pemanently :iamped to a level 0.7 volts above the minus supply voltage (pin 4) indicating that tne PNP porticn of tne complinantary output pair was fully saturate:. 3.7.3 Failure A.alvsis l 3.7.3.1 Failure Mode Verificatier. The device was removed from the printed circui: teard and retested, h confirming the failure mods des:ribed it. paragraph 3.7.2. 3.7.3.2 Internal Visual Inscetti:n ~ The oa:kage lid was removed anc the internal ctvity was enminec ur.ar a microscope. Apparent contamination was observed a; a grcwth on the aluminum :netalization and the alumir.u wires. Utiiizing vertical iiiumi-natior., tne growth appeared as a bla.ck granular substance Lnt was 2ctom-panied by a substantial in:.rease in volume over the origir.a1 Al me:a.i-zation. Whan viewed with dark field or side illuminatier, the growth had a silver-whita crystalline appearan:e Refirsace Fi5ures 9, 10 :nd 11. i _W=* dkhf} p II ihl e ^ (9iNahr filld: r.-d - c.%

• A8 9

' 'M 4 78 %;.t-h(% ' ~ ' M' '& hr h v2 5; W. : d L:: m @w @ a u.l w.-. b m, .y. h $ ~ d @ i l ~ [ T,% k $ m..._ &, e.. ~ 1 N4 1 ., w a. e pNd '. a .= I' 5.I$N. L M !fd d # h[ 9 h h h 1 N $5 $ 111uminttien ~. P.s .t lO """""! is!TfE%j&gpege3Mah l i 3

-\\ i i .5 4 e O 4 k;*:: a= - ---e, .,d. -} 3. N1 k asr - r u:sa r= y ai m p p:re d h."T. f, t l 1? b t[f)'r~ .zxcf -- .s y ..;. {..bj}-

....v-

.cp' ..s. m -- <g i aEaMaJ [.,-,,,. g. ~ f, ) pp g

• p j i

s w 'fb A 7 t .

• M' Ficure 10

$.1 h ~. :.m.M Vm MM f i Dark Field Illumination i k}}. dEC9 N k~d k ' $ W* W Q h f) [i W fl. M . t68A y

4. r... -e-

_.2 y. s &p. g L-v J.:.'tr.y.:, w = =_ .. v. e. .am%r a~.*<. I

• N

.".b er

[f,.

w e.,,. 7.,... ' e 6'-h e.SF.:.532( M A._ E Edf w, ima._. . m:~. <- i j O i @: E K(.T0 % S L M F r.i t' E E.! ge l. gy:A.,M d.t'.Das 4. ;w.e.t-W.'.... 'zch swe ...n 1 9 aes-e = ....u.+i..4 @ps < ' Y _ $ y-h $5$$M'#% l ef 3 . p( h h hfEbh.e n g . p p s ; e rs.t 4 ~~5E fe s w : N A k b I ha b ie d Illumination t m..m. e..n. m v.., y yt.m .y l ,..s- .n.a..s pkAI,h - h .+ f,. +. + i - ::w-d'*-^.S.:. ~ ~ * ' L Es:nwk...E.t-rs.:- i g..,..,...

-s. A ~. %, d e..-

c. - @.. ~ M r ? C W e-r e. ? w-kas-. 5- .. ~

nswsWa wnk:...c.

.,. M....w4xem#n&.di ~- i g,:: S Q RMT Report 17912A, .w .m e SD M C.LL m O A gg- ]

e 4 .e s 3.7.3.3 Ccerosio-Frodu:t Ane'ys's scar " p electron micros::.oe (SD') was used in the ' ert.v dis; esiv=- x-ray' 7.o 2 to determir.e the eier.en:11 cot.n:s'. i cf c.: alumind? corr:sior. product. Elements four.d in ?uded chiccine, sulfur and cal:ium. ~ 3.7.4 ' Cor.cicsion The SE9. analysis ossitively identified the con a.inates available fc-reaction witnin tne IC pa:kege. A clear.ing scivent resid;e wo id be the most li<ely sour:e of the Ohiorine and sulfur, while a dessi(ant woulo be a pcssible scurce of the calcium.

t is.possiole that the contaminates were introduced by the 10 manufacture prior to pa:kage sealing er that a defective or degraded seal. allowed contaminants to enter the package during subsequent Rosemount cleaning operaticns.

An identi:a1 amplifier of the same manufacturer's date code was ramoved fro the failes circuit board assem:1y anc helium leak rate test:c alor.g with all remaining devices in-house of the same date code. Leak rates in all cases were found to be less than I X 10-' std. c.c./sec. The sister cevice was de:a;?ed and examined revealing nc evidence of alur.inur cerrosior.. In ligh of this additional evidence and since it is highly unlikely that an IC manufacturer would use a chicrine basec solvent in his process, it is cencluded that the defe:: was most likely rancom and invclved a defec 1 tive er decraded seal which allcwed the ccnta-ir.ates to enter ne cie cavity. l k RMT Report 17912A - I l t

i (*, v 1 *. l 4.0 CO ;0!.'J5105 0F A!.: LYSIS

A'ifice

~C ~22 1 Cf The res its of ta,e 'a'estinsn:Use F.!F. 5/ste7s Civisi:E -:es A r.:.: lear pressa;. transmitters su:ptet the ru.'.3 Of. :.: the 1153 Re s er.:...: :.iiii':stion tests presente: in A!C F.e;or: 2755. Tr. ar.i 'ysi s shorts thl. :tt e r:rs er.:or:ered at: fatiares tr.at ::: rred,..i n tr.e exception of a sensor failure (Se:ti n 3.2.2), ctr oc attributed to the higher t'est levels of the '.'estinghouse procedure. For that ressor.. i-is Rosemount's.cpinion that these faileres do not constitute rep:rtatie deficien:ias ur. der the provisions of 100FR Part 21. f ~ D

e -

+ i i..- RMT Report 17512A -1E-l Gi-,- I ~' ' _}}