ORNL-1955 rT'

C-84 — Reactors=Special Features‘
Aircraft Reactors

VELOPMENT REPORT s

FABRICATION OF HEAT EXCHANGERS AND
RADIATORS FOR HIGH TEMPERATURE '

 

 

  
     
    

& \ REACTOR APPLICATIONS
B ol | =
'i-“" - (‘,J i P. Patriarca
=3 N e G. M. Slaughter
L.\"Erl h,; T W. D, Munly
oS S8 - R. L. Heestand
o= v—! Na hl 5
= | P .
g8 35 W 5 DECLAS Gl5tn vp 2wy v
3 2 N 44| DECLASSIRES gy
i A\ e -’Zé oy
25 ‘ = -

 

ot Z /‘.g /1 _-
For: M. 1. Bray, lw

  
 

== l.-_; iy N Fo |
A
e AT -
m‘r\ ™ :‘
fln.i"'!" *"" ““A“
v X% B ‘ o R i J
FUutie < E 74 RH.T. Uy ‘ UL—

= Y

PRl 1}{,1.;: g _ 2. 10-10
? £ Q /3 J15¢
P. 5. BAKER, (mHLjCO = 11114,1,3 DLTE

Easidadinilad

OAK RIDGE NATIONAL LABORATORY
OPERATED BY
UNION CARBIDE NUCLEAR COMPANY 2

A Division of Union Carbide and Carbon Corporation

POST OFFICE BOX P * OAK RIDGE, TENNESSEE

L S

NSy 3

se y s g = B

BV d‘-s W fi"mmfi:‘?’.hwfl “““ - Fan i 4 - 1w §
the Atk i, .

TR e L e Aha e o a ofits cTEmWS
SVl

  
  
 

  

 
 

 

 

 

LEGAL NOTICE

This report was prepored os on oceount of Government sponsered work. Neither the United Stotes,

nor the Commission, nor any person acting on behalf of the Commission:

A. Mokes any warranty or representotion, express or implied, with respect to the accuracy,
completeness, o usefulness of the information contained in this report, or thot the use of
any informotion, apparatus, method, or process disclosed in this report may net infringe
privately owned rights; or

B. Assumes any liabilities with respect to the use of, or for domoges resulting from the use of
any Information, apporatus, method, or process diselesed in this report.

As used in the cbove, *'person acting on behalf of the Commission® includes any employee or

cortractor of the Commission to the extent thot such employee or contractor prepares, hondles

or distributes, or providas access to, any information pursuant to his employment or contract
with the Commission,

 

 
 

ORISR V) 4 o 10 ST 35, ik
RN 5765 A L R P PSS
TIPS G, ™ i e TR e, Vfifi?fi”‘?’“‘” 2 m
DO S R MATIE S SR ORNL-1955
FRATRMRN C-84 Reactors - Special Features of Aircraft
Reactors

   
 

This document consists of T4 pages.
CoPy/ S of 260 copies. Series A.

FABRICATION OF HEAT EXCHANGERS AND RADIATORS
FOR HIGH TEMPERATURE REACTOR APPLICATIONS

P. Patriarca, G. M. Slaughter,
and W. D. Manly
Metallurgy Division
Oak Ridge National Laboratory

R. L. Heestand
Fox Project
Pratt and Whitney Aircraft

Period Covered by Work: February, 1955 - April, 1955

Work Performed By

R. E. Clausing R. G. Shooster
0. E. Conner C. E. Shubert
BR. L. Heestand L. C. Williams
B. McDowell

Metallurgy Division
A. G. Towns
Aircraft Reactor Engineering Division
Welding Under Direction Of
T. R. Housley
Engineering and Mechanical Division
Photographic and Metallographic Work By

M. D. Allen R. J. Gray
J. C. Gower E. P. Griggs

Contract No. W-Th05-eng-26
JUN 14 1355

  

T
Herdyiy o

 
 

 

1.

. Biology Library

. Health Physics Library

. Metallurgy Library

. Central Research Library
. Reactor Experimental

. Laboratory Records Department
. Laboratory Records, ORNL R.C.

p:bfacq?i?:?:b*#tnDuujc.orz'fitytdc4t+t'b

. Air Materiel Area
. Air Research and Development Command (RDGN)
. Air Research and Development Command (RDZPA)
. Air Technical Intelligence Center

. Aircraft laboratory Design Branch (WADC)

-1i-

ORNL-1955 )
C-84 - Reactors-Speclal

. :,_’/‘;'
©

Features of Aircraft Reactors

INTERNAL DISTRIBUTION

 

C. E. Center

Engineering Library

M. Weinberg L5,
B. Emlet (K-25) L6,
P. Murray (Y-12) h.
A. Swartout L8.
H. Taylor h9-50.
D. Shipley 51.
C. Vonderlage 52.
H. Jordan 53.
P. Keim Sk .
H. Frye, Jr. 55.
S. Livingston 56.
R. Dickison 57 .

. C. Lind 58.
L. Culler 59.
H. Snell 60.
Hollaender 61.
T. Kelley 62.
Z. Morgan 63.
A. Lane 6h.
A. Lincoln 65.
S. Householder 66-67.
S. Harrill ‘

EXTERNAL DISTRIBUTION

 

Baltimore
Burbank
Marietta
Santa Monica
Seattle
Wood-Ridge

Plant Representative,
Plant Representative,
Plant Representative,
Plant Representative,
Plant Representative,
Plant Representative,

HEEEEE

v

36.
37.
38.
39.
40.
L1,
Lo,
43.
L,

Winters
Billington
Cardwell

. Manly

King
Miller
Cowen
Charpie
Picklesimer
Skinner
Boyd

ender
Boudreau
Clarke
Fraas
Schultheiss
Yarosh
Gray

. Savage
Taboada
Patriarca

H. Coobs

H. Cook
Hoffman
Housley
DeVan
Douglas, Jr.
Heestand
Savolainen
- Y-12 Technical Library,
Document Reference Section

TQRuHduuHdgdEHagBHPODqgRUO=Z00H

PHUC IR TP NI P IERQEEIOPEIIYO

O
% .
B

 

 

;v b
 

 

107-109.
110.
111-11k,
115.
116.
117.
118.
119.
120.
121.
122.
123.
124,
125.
126.
127.
128.
129-131.
132-135.
136.
137.
138.
139.
140.
141.
1h2-1h4).
145-259.
260.

-1ii-

. ANP ProJject Office, Fort Worth

. Argonne National Laboratory

. Armed Forces Special Weapons Project, Sandia
. Assistant Secretary of the Air Force, R&D

. Atomic Energy Commission, Washington

. Battelle Memorial Institute

. Bettis Plant

. Bureau of Aeronautics

. Bureau of Aeronautics (Code 34)

. Bureau of Aeronautics General Representative
. Chicago Operations Office

. Chicago Patent Group

. Chief of Naval Research

. Convalr-General Dynamics Corporation

. Director of Laboratories (WCL)

. Director of Requirements (AFDRQ)

. Director of Research and Development (AFDRD-ANP)
. Directorate of Systems Management (RDZ~lSN;

Directorate of Systems Management (RDZ-1SS

Equipment Laboratory (WADC)

General Electric Company (ANFD)

Hartford Area Office

Headquarters, Air Force Special Weapons Center

Idaho Operations Office

Knolls Atomic Power Laboratory

Lockland Area Office

Los Alamos Scientific Laboratory

Materials Laboratory Plans Office (WADC)

National Advisory Committee for Aeronautics, Cleveland
National Advisory Committee for Aeronautics, Washington
Naval Air Development Center

New York Operations Office

North American Aviation, Inc. (Aerophysics Division)
Nuclear Development Corporation

Patent Branch, Washington

Powerplant Laboratory (WADC)

Pratt & Whitney Aircraft Division (Fox Project)

San Francisco Operations Office

Sandia Corporation

School of Aviation Medicine

Sylvania Electric Products, Inc.

USAF ProJject Rand

University of California Radiation Iaboratory, Livermore
Wright Air Development Center (WCOSI-3)

Technical Information Extension, Oak Ridge

Division of Research and Development, AEC, ORO
 

 

-iv- e "*“

TABLE OF CONTENTS

AB O T RAC T sttt iit v tssoiennoacsetononeeasasonsooesnasesssossonssesossensss 1
INTRODUCTION et connionsconsesacassasesoanonnssoessanesnssossnssescennsoas 2
FABRICATION OF 500~KW HEAT EXCHANGERS « v v v trevscnsioensooconcacenessss L
INtroduction .cooeeerveenerenoeeencaconanns f ettt e iearet et . L
Fabrication Procedure Gt e e e e i s eaea et etae s et erorste s atnanansns 7
SUMMATY oeseveessnvosses e rreesececnacence Gt eaceer ettt eescasecs st enens 19
FABRICATION OF 500-KW NaK-TO-AIR RADIATORS ocevo.. fh e e e creeneeanaeenen e 21
I OdUC i On ottt ittt ottt ittt erenneeeetsasosesssesesaennseanesns 21
Fabrication Procedure coocieererrenieeeenioeenneneeseoseansennnsonnoas 23
SUMIMBITY 2 50 0 o0 s o 0 s et aesonsensennssnonsarenessonnsanssssssosnnssosesss 37
FABRICATION OF 20-TUBE HIGH-VELOCITY HEAT EXCHANGER ..... ceeevesnresoeas 37
TNETOAUCHION oo v e veonnnennenennsnnsnseneen e e . 37
Fabricalion Procedure oeeieeiiiiiiniionieeeeneneeeeenosesesesnonnnnns 39
Sumary ....... C e e e e e st e e e e e ae et e e e e ane et reaeeenseaaaeeben0ns L0
CONCLUSIONS o0t vvcotconeososneonnseesoasssssscsnosossosonneseses seescacne U2
LIST OF FIGURES -+ 00 vt eesesoessssanasesonsenssonsssasseeensnnesnsnssennss L3
BB OGRAPHY & it iiiointoneeenennenoeeesoneeesoeaeneassnnenenennennnnens L5

 
 

ABSTRACT

Two 500-kw fused~fluoride~to-Nak heat exchangers, two 500-kw
NaK-to~air radiators, and a 20-tube high~velocity heat exchanger
were fabricated for a heat~exchanger development program. A con-
struction procedure, utilizing both inert-arc-welding and high-
temperature dry-hydrogen brazing, was used successfully on all of
the units. The tube-to-header joints were welded and back-brazed;
the manifold joints were inert-arc-welded with full penetration;
and the tube-to-fin joinis were brazed. A detailed description of
the fabrication of each type of component is discussed and a cost

analysis of the 500-kw units is presented.
 

- -

INTRODUCTION

The heat exchangers and radiators to be used in conjunction with high-tempera-
ture nuclear reactors which utilize highly corrosive and radioactive fluids must
necessarily be the ultimate in integrity. Precise control of the procedures used
in their construction must therefore be constantly maintained. It is well recog-
nized that faulty workmanship or the improper selection of a joining technique in
one location on a unit may result in a catastrophe or, at least, a costly shut-
down during repair.

The component designs under consideration for the Aircraft Reactor Test in-
stallation incorporate multitudes of thin-walled small-diameter tubes in extremely
close-packed configurations. The fabrication of these units poses a difficult
problem; highly specialized equipment and procedures, which have been proven sat-
isfactory in rigid tests under simulated service conditions, are required. Since
the complexity of design of the heat exchangers and radiators is, in many cases,
unigue to the atomic energy field, much of the developmental work on these join-
ing techniques has been done by the personnel who are actually confronted with the
problems of component construction.

Most of the fabrication problems associated with these units may be classified
into three general categories: (1) the production of sound tube-to-header joints,
(2) the production of high-quality manifold joints, and (3) the attainment of sat-
isfactory tube-to-fin joints. The development of procedures and techniques for the
solution of these problems has been under way at Oak Ridge National Laboratory for
several years. Numerous successful test assemblies have been fabricated during this

time, and refinements in construction procedure have been continuously introduced.
 

-3

A heat-exchanger test loopl has now been designed and set up in the Aircraft
Reactor Engineering Division of ORNL to provide data on corrosion, mass transfer,
and reliability of a fuel-to-NaK-to-air system operating under conditions compara-
ble to those expected in the Aircraft Reactor Test. A small-scale heat-exchanger
test was also operated for the purpose of investigating heat-transfer character-
istics, through the Reynolds-number range of 0 to 5500, on the fluoride-mixture
side of the fuel-to-NaK heat exchanger.

The Welding and Brazing Group of the Metallurgy Division was assigned the job
of fabricating the units used in these test loops. The loops consisted of two 500-
kw fused-fluoride-to-NaK heat exchangers; two 500-kw NaK-to-air radiators; and a
20-tube high-velocity, fused-fluoride-to-NaK heat exchanger. Services of the Engi-
neering and Mechanical Division of ORNL were also used extensively in the construc-
tion of these components. This report describes in detail the fabrication of these
heat exchangers and radiators and contains information pertaining directly to the
construction of similar units for the ART. It should also provide assistance to
other groups interested in the production of equipment for high~temperature high-

corrosion applications.
 

i SRS,

-

A. Fabrication of 500-kw Heat Exchangers.

 

1. Introduction.

Two 500-kw fuel-to-NaK heat exchanger tube bundles were required for use in
a heat-exchanger test loop under investigation in the Aircraft Reactor Engineer-
ing Division. The test loop is part of a long-range heat-exchanger development
program designed to obtain information on the operating characteristics, corro-
sion, and reliability of fuel-to~-NaK-to-air systems under conditions similar to
those stipulated for the ART.

The test loop incorporates these heat exchangers in the regenerative-type

1 'The NaK flows from a 1-Mw gas-fired

circuit shown schematically in Fig. 1.
heater through one tube bundle to a radiator. The stream then goes to a liquid-
metal pump and back through the second tube bundle to the heater. The fuel mix-
ture NaF-Zth-UFh (50-46-4 mole %) will be circulated outside the tubes counter-
current to the NaK flow and will be alternately heated and cooled by the NakK
stream.

Each heat exchanger was composed of 100 Inconel tubes, 3/16-in. OD, 0.017-in.
wall thickness, and approximately 6 ft. in length. The tubes were incorporated
into the heat exchanger as shown in Fig. 2, an assembly drawing of a typical tube
bundle. It is essential that the completed heat exchangers be leaktight and that
they be fabricated in such a way as to withstand the severe conditions of tempera-
ture and pressure indicated in Fig. 1.

Another heat-exchanger test unit, of a smallér but somewhat similar design,
has been fabricated,2 tested,‘3 and examined.LL The tube-to-header joints in this
early unit were manually inert-arc-welded. The unit performed successfully for

1680 hr before being terminated due to a failure in a tube-to-header weld. Me-

tallographic examination of several of the welded joints revealed the presence

 
 

-5-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ORNL-LR-DWG 5673

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8O° F 700° F
AIR IN | — e o e e e i AR QUT
15.4 psia - Mw 15 psia
NaK-TO-AIR
1099° F RADIATOR 1279°F
20 psia 30 psia ‘
|
| E
x5! °
zZ5 : %
|
FUEL
PUMP
U —————
u .=
% '3_ 37 gpm w .g wio
Mo wl.g o 2 Om 8.
2le og 8 @ < RNlo
Y = e e
L [ i |
{ J | L w | }<—I |
LoE 1,2 Zi 122
wl
| hég I gI gI | 5 |
| = | |7 3 Q1 1 22
g - < X = |
I 2O | fmY mu | oF |
o2y AR
e Loy wl | g%
| & | | 2 |
] L
wil@ Ll o
o w o @
9 o O; é 8 Q OLL g
IR 3 Q1% ole
njN |
2 2 (&
1590°F
74 psi 67 psi
psia 1~ Mw c)psncx
GAS-FIRED 80" F AIR IN
F
cggflg&sjgosw 1000° £ HEATEH‘/// 30 psia 128 Ib/sec
4——————-———— ——————
T 15 psia >~ o
ov NS 8O F GAS IN
30 psia  0.071 Ib/sec

 

 

 

Fig. 1. Flow Diagram of Loop for Testing Intermediate Heat Exchanger No. 2.
 

 

of extensive microfissuring in the fusion zone, as evidenced in Fig. 3. It is
thought that the differential thermal expansion between the tubes and the casing
caused stress concentrations at the roots of the tube-to-header welds. These
stress concentrations tended to propagate cracks through the welds in the course
of thermal cycling; particularly since the columnar dendrites, which are typical
of a weld structure, were aligned in such a way as to aid the formation of paral-
lel fractures. A photomicrograph of a tube-to-header joint exhibiting a similar
crack in the early stages of propagation is shown in Fig. L.

The results of the metallographic investigation emphasized the extreme desir-
ability of utilizing the advantages afforded by back brazing. As can be seen in
Fig. 5, a photomicrograph of a welded and back-brazed tube-to-header joint, this
process (1) eliminates the "notch effect" resulting from incomplete weld penetra-
tion and (2) insures against the development of leaks in the event of corrosion
through an area of shallow weld penetration. This duplex fabrication technique
was therefore employed in the production of the critical tube-to-header welds.

2. Fabrication Procedure.

a. Tube-to-header welding. The 200, 3/16-in. OD by 0.017-in. wall, Inconel
tubes for these fuel-to-NaK heat exchangers were formed into the desired configu-
rations and delivered, along with the drilled headers and other necessary compo-
nents, to the Welding and Brazing Group. All parts were meticulously degreased
prior to assembling, and the headers were deburred to facilitate entry of the
tubes.

The inert-arc-welding of the LOO tube-to-header joints was performed on the
semiautomatic rotating-arc equipment shown in Fig. 6. The equipment incorporates
a commercially available inert-arc-welding torch which is attached to a drive mech-

anism originally designed for contour-cutting with an oxygen cutting torch. An

B S IR e ©

 
 

 

 

 

7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Y 15604
FARTS LIST =y
':.h' e mAME Bl AT AL :\_'__
! - WOITLE 2% BEM'D 40 FUPE i Ot |
2 |n HOUS WG N0 PLATE | 14" PLATE ™COMEL
3|4 SHELT 14" PLATE O
4 | ® HOUSIME SiDE FLATE /4" BLATE WCONEL
3 | DRAIN COMMECTION [172° SCH'D 40 PIFE INCOMEL
L ELBOw A" BCH'D 40 mPE IMCONE L
T (aE BPACER BREET S0 SHEET IO L
8 |200| TUuBE BUNMDLE 3800 K O WALL TUBS INCOREL ]
"2 WPER SMELL 2" FLATE iNCOmMEL
o | 4 HOUSING COVER 3" SCND 40 FIFE N CONEL ‘
wle ADAPTER 14" LA INCOMEL
e |z PN IFB" AL FIN BHCOMEL | i
ial|lz LOWER SMELL B° PLATL MOONEL =
4 (234 SFACER DM A WIRE MOCNEL L =
i~ (2
— — 2
A
( :J'I (8) 9 = (4
3 I I X
5
: [ | .
_— 11}
- Py 2
PSR
= }
b S |
"5;6\"-1-’——/ Il
= |
; — —_— i —_—
e
I2
Y-13004
v
p— m—
I
o
=z
0.02
0,03
<
- O
o

 

 

 

Fig. 3. Tube-to-Header Weld Exhibiting Extensive Microfissuring. Etch:

electrolytic; oxalic acid.

 
 

 

= e 4
N e AR
™ R - f.,fi 4L

'
o) -

 

R e
AT R

 

 

 

Fig. 4. Tube-to-Header Weld Exhibiting Crack in Early Stages of Propagation. Etch: electrolytic;
oxalic acid.

 

 

Fig. 5. Tube-to-Header Weld ofter Bock Brazing. Etch: 10% oxalic acid; electrolytic. 12X.
Ve

L
L e

A=
&b '

B
..-.—m-_ SE= B ¥

=1 —}-——

LA R &

‘r ";“_m_

4

Fig. 6. Semiautomatic Tube-to-Header Welding Equipment.

Y-13504

 
 

 

 

 

 

~10-

of fset-cam mechanism was developed and installed, which permitted the torch to
travel around a tube periphery of any desired diameter. A d~c welding generator
served as the power supply, and the argon shielding gas was fed through the torch
at the top. The modified metal lathe shown in the picture was not used in this
fabrication, but it is useful for such special applications as circumferential
inert-arc-welding of thin-walled tubing.

The arc is initiated by a superimposed high-frequency current, and the weld-
ing current is applied for slightly longer than one revolution. Extinguishing of
the arc is carried out by means of a commercially available foot-controlled arc-
decay attaclment. This mechanism minimized the tendency to form arc craters by
permitting a gradual reduction of the output voltage of the generator. Thus,
"feathering" of the arc current is made possible. <Extreme1y low welding currents
can be attained on this equipment, when desired, by the utilization of a bank of
variable high-wattage resistances. Good arc-striking conditions can be simulta-
neously attained since a high open~circuit voltage is available.

Prior to starting the actual fabrication, a set of experiments was conducted
to determine the optimum combination of welding conditions. In the experiments
two Inconel test headers, of the same size and physical shape as the one shown in
Fig. 2, were machined, and sample tube-to-header welds were made under a variety
of controlled conditions. ©Several of the sample joints were examined under high-
power magnification to determine the presence of weld microfissuring, but this type
of cracking was not found and, therefore, presented no problem. Small imperfec-
tions showed that the preparation of the header surfaces for welding, after inser-
tion of the tubes, should definitely not be done by abrasive grinding. Entrapped
abrasive in the joint resulted in severe arc instability and in inconsistent welds.

A more reliable method consisted of preshaping the tubes to conform to the curvature

 
 

-]]-

of the header before assembly. The tube can then be expanded with a special tool
before final welding. By the use of this method of header preparation, the magni-
tudes of the welding variables were found which gave the most consistent weld pene-
tration and which were least likely to result in excessive hole constriction or in
undesirable preferential melting of the tube wall.

Previous experience had shown that an arc distance of 0.050 in. and a weld
time of approximately 6 sec produced consistently good welds with this tubing with
a 3/16-in. QD and a 0.017-in. wall thickness when it was joined to relatively thick
headers. These values were, therefore, used for determining the optimum diameter
of electrode rotation and the proper welding current. A rotation diameter between
0.21 in. and 0.22 in. was found to be desirable, since values less than this often
resulted in preferential melting of the tube wall, and values greater than this
produced maximum weld penetration in the header plate and not at the joint, where
it is of most importance. An arc current of 60 amp at an arc voltage of 10 v pro-
duced consistently satisfactory welds with penetrations of approximately twice the
tube wall thickness. A photomicrograph of a typical weld produced under these
optimum conditions is shown in Fig. 7. No weld porosity or cracks are evident and
excellent penetration was achieved. The two-pass effect present in the nugget re-
sults because a weld overlap of approximately one-half revolution is used after the
complete peripheral weld has been made. During this period the weld current is
gradually decreased to prevent the formation of undesirable arc craters. The uni-
formity of welds made under these conditions can be seen in Fig. 8, a photograph
of a 100-weld header section of this heat exchanger. Helium backing-gas was used
throughout to minimize oxidatibn at the root of the weld. Leak testing of each
header section was performed, and all welds were leaktight before initiation of

the following stages of fabrication.

N L
19

o
] INCHES I;

E

jo.c8
x
2

 

 

 

|o.08

Fig. 7. Tube-to-Header Weld Produced with the Semiautomatic Welding Equipment. Etch: electro-
lytic; oxalic acid. 75X. Reduced 13%.

 

Fig. 8. One End of Heat Exchanger ofter Welding Tube-to-Header Joints.
 

~13-

b. Header welding. The side plates, top plates, and nozzles were joined to
the Inconel headers entirely by manual inert-arc-welding. The welding procedures
are stipulated in Appendix 1 and the qualifications for welding operators are
stipulated in Appendix 2. The joint designs used were those recommended for high-
pressure, high-corrosion applications (cf. Appendix 3).

To minimize cracking of the brittle brazing alloy and to prevent the areas to
be welded from becoming contaminated with such elements as boron and silicon from
the brazing alloy, the welding of the components was performed before the tube-to-
header joints were back-brazed. However, to permit accessibility to the joints,
the preplacement of the brazing alloy was completed first.

The welding in an early stage of fabrication is shown in Fig. 9, and it can
be seen that strongbacks were used to minimize distortion of the header and con-
sequent danger of tube-to-header weld fracture. A photograph of a completed leak-
tight header section is shown in Fig. 10.

¢. Back brazing of tube-to-header joints. The back brazing of the tube-to-
header welds in this heat exchanger was desirable since it afforded a means of
minimizing the stress concentrations formed at the root of the weld during service.
The brazing alloy would, of course, also minimize the chances for leaks to occur
from weld corrosion or microfissuring during operation of the unit. Because the
bragzing alloy would be in intimate contact with the fused fluorides and might
come in localized contact with the metal inside the tubes, it was desirable that
the alloy possess good resistance to corrosion in both media.

A long-range program for the development of brazing alloys has been conducted
by the Welding and Brazing Group of the Metallurgy Division at ORNL for several

5

years. A thorough evaluation” of the corrosion resistance of brazing alloys has
been conducted by the General Corrosion Group and, as a result of their work, a

suitable alloy was found.
Y-14480

 

 

Fig. 9. Header in Early Stages of Fabrication.

Fig. 10.

 

A Completed Leaktight Header Section.

-Pl-

 
 

-15-

Low-melting Nicrobrazé, an alloy of the nominal composition 80Ni-5Cr-5Si-3B-
6Fe-1Cr, was found to be compatible with both liquids. A photomicrograph show-
ing the resistance of this alloy to fluoride attack is presented in Fig. 11, and
a photomicrograph of a sample after testing in molten sodium is presented in Fig.
12. This alloy possesses good flowability at 1040°C, a temperature at which boron
diffusion into the Inconel base metal is not a serious problem. Therefore, only
minor embrittlement of the tubes would be expected to occur during brazing.

The low-melting Nicrobraz alloy was preplaced on the headers as a dry powder
and was secured firmly in position with Nicrobraz cement. Since the Globar-heated
brazing furnace contained a heating zone of limited length, only one end of the
tube bundle could be brazed at a time. As a result, the welding and the brazing
of one header of a tube bundle were completely performed before the brazing alloy
preplacement and before the welding of the second end was initiated. This was
necessary since there was danger of spalling of preplaced alloy from the under-
side of the second header during the brazing of the first header. It was thought
that hot hydrogen gas might volatilize the binder and leave the brazing alloy in-
securely positioned in its overhead locations. This important consideration in-
creased the fabrication time to some extent because the welding of both heads
could not be performed in succession.

The brazing operation was performed in a dry hydrogen atmosphere of -70°F
dew point or below. Good flowability is obtained without the addition of a flux
because the highly reducing nature of this hydrogen prevents the formation of
most metal oxides. Accelerated corrosion, common to joints containing entrapped
brazing flux, is, therefore, avoided. Conventional canning procedures were used
during brazing; that is, the heat exchanger was placed in a stainless steel retort

in which hydrogen inlet and hydrogen exit tubes were the only openings. The pure
      

5 ¥-15090
-
¥
'
A
iy
:—:"‘:‘J: ;.I -'.-' .
oo . =
0 -
Ll T g
o el e R R S s R et e =
*-'?:"'*-"."r'.r"‘ el I A o '?"::-"-&J'a"&.'.""h L

B g A e v g e 3
£ i 4 b T . o

- -

  

ERE]

o
o o
-~

 

5

T
NC:%ES

[TTRITTTRI

 

o
2
o

L[]}

 

G
P
o

"

150

 

 

Bl |

Fig. 11. An Inconel T-Joint Brazed with Low-Melting Nicrobraz and Tested in Fluoride Bath
for 100 hr at 1500°F. Only very slight attack is present. Etchant: Glyceria regia. 150X. Reduced 14%.

Y-15091
-
” i
12
8 i
" ‘ . 1 --*. q‘
S8
Bt Dl
e

 

TTTEL

333113

1
NCIHE 5

 

HEEE:

o
~
o

<l

150

 

 

 

&
o

 

No. 44

Fig. 12. An Inconel T-Joint Brazed with Low-Melting Nicrobraz and Tested in Sodium for 100 hr at

1500°F. Negligible attack is present., Etch: None. 150X. Reduced 14%.

 
 

-17-

dry-hydrogen atmosphere was then maintained in this retort while it was heated in
a large Globar high-temperature furnace.

An elaborate purification and drying train was used to remove oxygen and water
from the inlet hydrogen gas. The oxygen was converted to water by means of Deoxo
(palladium catalyst) manufactured by Baker & Co., Inc., Newark, New Jersey, and the
water was removed by an activated alumina drier obtained from Pittsburgh Lectro-
dryer Corp., Pittsburgh, Pennsylvania.

The dry-hydrogen furnace brazing was performed in a horizontal position in a
large Globar-heated furnace. The heat-exchanger tube bundle was securely jigged
in a long rectangular stainless steel retort, and a dry-hydrogen atmosphere was
maintained in the retort during brazing. A time-temperature record of every braz-
ing thermal cycle was measured by means of a chromel-alumel thermocouple firmly
attached to the assembly. A plot of a typical cycle is shown in Fig. 13. A pre-
heated brazing furnace was used to shorten the time required for each operation,
and no visible warpage or distortion of the heat exchanger resulted.

d. Fabrication and assembly of "comb spacers”. The use of wire comb spacers
was considered to be the most practical means of holding the tube bundle rigid
throughout its free~span length and of keeping the tubes separated to permit the
required mode of fluid flow between them. The assemblies were to be placed at
5-in. intervals along the bundle with alternate spacers positioned at an angle of
90° to each other. It was required, for a heat-transfer evaluation, to have the
spacers on one bundle placed perpendicular to the axis of the tubing, whereas they
were placed in a plane at a 30° angle to the tubing axis on the second bundle. The
comb assemblies were fabricated and attached according to these requirements; and
a photograph of both types of spacers on a sample tube bundle is given in Fig. 1l.

It can be seen that the fabrication and the assembling of these spacers required
1200

oo

1000

900

BoQ

700

8 o
8 &

TEMPERATURE - °C.

w
Q
o

P
o
o

100

 

400 |

 

 

 

 

Y 15841
/(_-\
\
i# \
//’ \
= /
7
s/
i |
|
e FURNACE | AR
[I: COOLING T COOLING
. e —— HEATING ——II-"—"'!-‘ERAIING
| | | 1 | | 1 | | | | |
0 05 LD L5 20 25 30 35 40 45 50 55 6.0 €5 70 15 B0 B85 9.0

TIME- HOURS

Fig. 13. Typical Time-Temperature Thermal Cycle Used in the Brazing of Heat Exchangers.

 

Fig. 14. Sample Tube Bundle Showing Two Types of ‘“Comb Spacers."”’

 
 

-19-~

the use of precision jigging facilities and the careful determination and control
of welding conditions. Meticulous care also was maintained when punching holes in
the strip headers because extremely close tolerances were specified.

e. Installation of tube bundle into pressure shell. The installation of the
tube bundles into their pressure shells required a light machining of the shell to
permit proper fit. Each pressure shell had to be precision machined to adjust for
the very small variations in dimensions between it and its corresponding tube bun-
dle.

The heavy-walled pressure shell was beveled according to the recommended joint
design, and the root pass was inert-arc-welded as prescribed for high-corrosion ap-
plications. The remainder of the welding was performed by use of the metallic-arc
process in order to minimize the heat input and consequent distortion of the unit.
Fig. 15 is a photograph showing the heat exchanger after completion of the root
pass. The large I-beam was used as a strongback for the system in order to pre-
vent distortion.

f. Joining of the two tube bundles. The two 500-kw tube bundles were joined
together after being welded into their respective pressure shells, and a photograph
of the completed unit is presented in Fig. 16. It can be seen that the operation
required several manual inert-arc-welds to connect the sodium circuits and to at-
tach the inlet and exit nozzles.

3. Summary.

Leak testing of this unit with a mass spectrometer type of helium leak detec-
tor indicated no flaws, and the heat exchanger was subsequently delivered to the
Aircraft Reactor Engineering Division for installation into their testing rig.

A step-by-step record of the time required for each operation in the fabrica-
tion of the two tube bundles was maintained, and the compilation of these times is

presented in Appendix L.
-20-

¥-14589

¥-14810

 

Fig. 16. A Completed 1-Mw Heat Exchanger.

 
 

-21-

B. Fabrication of 500-kw NaK-to-Air Radiators.

 

1. Introduction.

The Intermediate Heat Exchanger No. 2 test loop described in the preceding
section required the fabrication of two 500-kw air-cooled radiators for installa-
tion as shown in Fig. 1. An assembly drawing of one radiator is shown in Fig. 17.

Several test radiators, containing very compact tube-to-~fin matrices and lim-
ited in size and heat transfer capacity, have been fabricated and tested.! These
early units, which contained austenitic stainless steel or nickel fins, possessed
good structural integrity But only fair performance characteristics at the high-
operating temperatures. A high-conductivity fin developmental program was there-
fore initiated8 to determine the most suitable high-thermal-conductivity materials
for high-~-temperature services. Type 310 stainless-steel-clad copper was found to
possess excellent resistance to high-temperature oxidation, and the good thermal
conductivity of the material was not materially reduced after extended periods at
the service temperatures. A fin material of 0.006-in. OFHC copper, clad on each
side with 0.002 in. of type 310 stainless steel, was therefore selected for this
application. The desired spacing between successive stacked fins was obtained by
proper design of the punch-and-die equipment, that is, by the depth of lip protrud-
ing from the flat sheet after punching.

Experience with the early fabrication procedures, in which the tube-to-fin,
tube-to~-header, and manifold joints were all brazed in one operation, indicated
that every critical joint could not consistently be made leaktigfii in one braz-
ing operation. Minor variations in joint spacing, base~metal composition, hydrogen
coverage, hydrogen purity, and brazing technique influenced the brazing-alloy flow-
ability to such an extent that several rebrazing operations were often required to

obtain a radiator that was leaktight when tested with a helium mass spectrometer.
-29-

PARTS LIST
—_—e e ——
iy | nIt

RAGIATOR Fin
G PLATE

I

_SuPPORT
BCTI0M ELAMSED FIATE
MARATON TURE

—SEL DETRIL "W

Yg) & o =
¥ _AFb - ;

      

 

 

 

Fig. 18. Sodium-to-Air Radiator with Stainless Steel Fins Fabricated by the Combination Welding

and Brozing Technique.

 
 

-23-

These rebrazed sites were considered undesirable because of their adverse effect
upon the mechanical properties of the base matérialse Increased grain growth of
the parent material resulted; and embrittlement, resulting from the diffusion of
constituents from the brazing alloys,‘was promoted. Also, the temperature gradi-
ents across the radiator during rebrazing induced localized stresses in the rigid
structure, which condition might.initiate the formation of microfissures in the
braze or base metal.

In view of these difficulties, a combination welding and brazing procedure was
developed whereby all manifold welds (butt welds, nozzle welds, and end-cap welds)
were inert-arc-welded before subsequent dry-hydrogen brazing. Also, all tube-to-
header joints were welded with the intention of back brazing as discussed in the
section, "Back Brazing of Tube-to-Header Joints". As was explained in this pre-
vious section, back brazing eliminates the notch effect and prevents leaks in the
event of corrosion through localized areas of shallow weld penetration. Back braz-
ing can also be used to seal occasional welds containing small cracks or other de-
fects. This procedure proved to be successful, and two leaktight liquid-metal-to-
alr test radiators were fabricated after only one brazing operation on each.h’9
Photographs of these two units are shown in Figs. 18 and 19. In view of the satis-
factory results obtained by this duplex technique, it was chosen as the best method
for use in the construction of the 500-kw units.

2. Fabrication Procedure.

a. Tube bending. The Inconel tubes were bent to the desired configurations
before being delivered to the Welding and Brazing Group. It was specified that
these tubes should have an outside diameter between 0.187 in. and 0.188 in. to per-
mit proper fitup for flow of the brazing alloy in the joint. The use of tubing

with an outside diameter of 0.185 in. resulted in such a loose tube-to-fin fitup
T

 

Fig. 19. Sodium-to-Air Radiator with High-Conductivity Fins Fabricated by the Combination Welding
and Brazing Technique.

 
 

—25-

that bonding was extremely inconsistent, whereas tubing with an outside diameter
of 0.190 in. was so large that assembling the fins on the tubes was very difficult.
The tubing was radiographically inspected and vacuum-tested for flaws before the
radiators were assembled, but no apparent defects were found.

b. Fin shearing and inspection. Approximately 1500 ft of type 310 stainless-
steel-clad-copper high~conductivity fin material was obtained in coiled spcols from
the General Plate Div. of Metals & Controls Corp. The spools were purchased in
1 15/16-in. widths because this was one of the specified fin dimensions. The
0.010-in.~-thick material was composed of 0.006 in. of OFHC copper roll-clad on each
side with 0.002 in. of stainless steel for oxidation resistance. Twenty-two hun-
dred fins were individually sheared from these coils to the desired 1 15/16-in. by
8-in. final dimensions. Precision shearing was required since it was important to
have the copper edge completely exposed for the aluminizing edge treatment (to be
described in the next section). The fins were then carefully inspected for such
surface imperfections as blistering and exposed copper. Thorough degreasing in
polychlorethylene was alsc performed prior to aluminizing.

c. Aluminum-bronze edge-protection process. The edge protection of exposed
copper on the sheared fin edges 1is extremely desirable in order to minimize oxida-
tion and in order to overcome the severe fin distortion resulting from the volume
changes occurring on account of the formation ¢f the oxide. An aluminizing process
was developed whereby edge protection is obtained through the formation of a highly
oxidation-resistant copper-aluminum alloy. A procedure was devised to permit the
similtaneous edge protection of large numbers of type 310 stainless-steel-clad
copper-sheet fins. This procedure, for which the precision fin-shearing was re-

quired, is presented below.
 

26~

The fins were stacked in groups of approximately 200 and securely clamped with
1/l4-in.-thick stainless steel side plates. The exposed copper edges were then
sprayed with three coats of Kestron acrylic spray produced by Aerosol Products
Company, South Hampton, Pennsylvania. This sealed the cracks between fins and pre-
vented the flow of an aluminum-bearing slurry onto the stainless steel cladding.

A coat of slurry, consisting of 100 cc of acrylic resin and 4O g of atomized alu-
minum powder (-325 mesh), was then applied evenly to the fin edges. After drying
sufficiently, the clamped fins were heated in helium at 750°C for 2 1/2 hr in order
to accelerate the formation of the oxidation-resistant aluminum bronze. 4 high
helium flow rate (80 ft3/hr) was maintained for the first hour but was reduced to
1O ft3/hr for the remaining heat treatment. After cooling, the excess aluminum

was easily removed from the fin edges with the aid of a brass wire brush.

Test specimens of fins edge-protected by this method have been metallograph-
ically examined after exposure for 100, 200, 500, and 1000 hr in air at 1500°F.
Fig. 20 illustrates the excellent protection provided by this technique after 1000
hr at this temperature. The depth of oxide penetration into an unprotected fin is
shown on the same print for comparison purposes. OCyclic tests on these aluminized
fins were also conducted to evaluate their stability under simulated service con-
ditions. After 500 hr at 1500°F and after having been intermittently cooled to
room temperature 200 times during this period, no adverse effects wefe evident.

After edge protection was completed, the fins were punched to accommodate the
Inconel tubing. It was found that the use of a lead-oxide die lubricant should be
omitted if possible, since the subsequent removal of this substance is difficult.
However, if this lubricant is used, it is essential that every trace of the lead

oxide must be removed.

 

e e RO ANk AT A TR R S St e ko
 

Fig. 20. Photomicrograph Exhibiting the Excellent Fin Edge-Protection to High-Temperature Oxi-
dation Afforded by the Aluminum-Bronzing Operation. Fin on left was unprotected. Etch: None. 60X.

 
 

 

 

~28-

d. Preplacement of brazing alloy. Since this radiator was to have stainless-
steel-clad copper, high~conductivity fins, a brazing alloy was required which pos-
sessed a flow point below 1083°C. The alloy,lO Coast Metals No. 52, exhibits good
flowability at 1020°C and its mechanical strength at the service temperature of
1500°F is sufficient to warrant its use in this high-temperature application. Its
resistance to oxidation is excellent as can be seen in Fig. 21, a photomicrograph
of a brazed Inconel T-joint that was heated at 1500°F for 500 hr in static air.

The compatibility of this metal with liquid sodium is also good, as can be noted in
Fig. 22, which shows a brazed "A" nickel T-joint tested for 100 hr in sodium at
1500°F. Stability in sodium is an important prerequisite for two reasons: (1) the
brazing alloy may actually come in contact with the liquid metal during service, if
it 1s used to seal small pin-hole leaks, or if sufficient weld metal corrosion oc-
curs to expose a new braze-metal surface; (2) severe tube-wall dilution during braz-
ing or solid-state diffusion in service may produce a high concentration of brazing-
alloy constituents near the circulating liquid.

An analysis of the design of the radiator shown in Fig. 19 indicated that crit-
ical control of the quantity of brazing alloy placed on each tube-to-fin Jjoint must
be maintained. An amount of braze metal only sufficient to protect the exposed cop-
per on the punched lips of the high-conductivity fins was used. Applying alloy in
excess of the amount required could result in "puddling®, whereby the excess is
merely concentrated on the bottom fins. This latter situation is findesirable since
the air passages between these fins may be sealed and localized fin distortion can
occur.

The use of extruded wires, as a means of obtaining controlled quantities of
brazing alloy, has been investigated extensively; but the lack of a suitable binder

makes their use unattractive. With the use of acrylic-binder materials, wires are

 
.29-

INCHES

 

 

 

Fig. 21. Inconel T-Joint Brazed with Coast Metals No. 52 Alloy and Oxidized for 500 hr at 1500°F.
No attack is present. Etch: electrolytic; oxalic acid. 100X. Reduced 12.5%.

¥-14542

INCHES

 

 

 

Fig. 22. A" Nickel T-Joint Brazed with Coast Metals No. 52 Alloy and Tested in Sodium for 100 hr
ot 1500°F. Negligible attack is present. Etch: electrolytic; oxalic acid. 100X. Reduced 12.5%.

 
 

~30-

produced that become brittle after aging for a few hours at room temperature. With
other binders that were tested, wires were produced that had insufficient strength
to permit ease of handling. Since Coast Metals No. 52 brazing alloy can be obtained
as a fine powder, a dny*powder'technique of brazing alloy preplacement was an obvi-
ous alternative.

A process was developed whereby a controlled amount of alloy could be preplaced
on each tube-to-fin joint. A stainless steel template, containing holes that were
precision drilled, was placed over a sheared fin leaving an annulus between each
fin lip and the template hole peripheries. After application of the brazing-alloy
powder and subsequent removal of the excess, a ring of alloy of controlled quantity
remained. Upon careful removal of the template, the powder was made secure to the
fifi with Nicrobraz cement and allowed to dry. A photograph showing a fin, a tem-
plate, and a fin with preplaced alloy is presented in Fig. 23. This 1/16-in.-
thick template, which contains holes 0.246 in. in diameter, was found to provide
the optimum quantity of brazing alloy for each joint and was, therefore, used to
preplace alloy on the 2200 fins in preparation for assembly on the tubes.

e. Assembling of fins. The assembling of the fins on the radiator was con-
ducted with the tube-bends down, as is shown in Fig. 24, a photograph of the ini-
tial stages of fabrication of one radiator. The assembly of the first L in. of
fins was extremely time-consuming since the tubes were not held rigidly in place.

A heavy metal template was necessary to force each individual fin in place. As
some of this difficulty was also in consequence of the slight curvature of the
straight lengths of tubing, it is thought that improvements in the bending techni-
ques might substantially assist the assembling of future units.

A set of lucite "finger" jigs was designed and built, which simplified the

problem of assembling the fins. These jigs, which can be seen in Fig. 24, enabled

 
 

Fig. 24. Fin Assembly in Early Stages of Fabrication. It can be seen that the process was difficult
even with the use of precision jigs.
 

 

the tubes to be aligned in their desired positions; namely, to fit the punched geom-
etry of the fin. Supporting sheet-metal channels were placed at L-in. intervals to
impart over-all structural stability to the radiator. These channels were inert-
arc welded to form sheet-metal side plates. Two thin sheets of Inconel were also
placed tightly together against the channels at these L-in. intervals. This modi-
fication was used to aveid the undesirable accumulation of excess alloy which could
result from normal variations in the quantity deposited. The capillary joint be-
tween these two Inconel sheets acts as a sump to accommodate any excess brazing
alloy.

f. Assembling of headers. The original design of these radiators required
that all tubes enter the cylindrical headers normal to the curvature at the point
of entrance. Sample specimens were prepared to determine the optimum tube-bending
and fitting techniques, but at the time, no simple method was devised. The complex
tube design resulted in extreme difficulties in placing the headers on the tubes.

In view of these complications, a new header, utilizing a 3-in., schedule LO
pipe, was designed and accepted. All tubes, except the two outside rows on each
radiator half-section, entered the headers without being bent. Several test sam-
ples were then prepared for the determination of the proper bending techniques and
for use in the operators' qualification test.

Careful hand-polishing of each of the tubes was required to permit assembling
the split headers without the use of additional force from a hand press. Precision
drilling and deburring of the headers was also employed to prevent bending and dis-
torting the thin-walled Inconel tubes. After assembling the headers, the tubes
were meticulously hand-filed to conform to the exact curvature of the interior sur-
faces. Abrasive grinding or wet-milling was avoided since entrapped abrasive or

lubricant would lead to difficulty during tube-to-header welding.

 

 
 

=33~

g- Welding. All tube~to-header welding on the two radiators was performed
using the manual inert-arc~process. Test samples were prepared to determine the
optimum welding conditions. A welding current of approximately 35 amp. was found
to give consistently sound welds with good penetration. Helium backup on the under-
side of the joint was used throughout the welding process.

The manual inert-arc-welding of the split headers, nozzles, and endplates was
accomplished by using qualified operators (Appendix 2) and procedures (Appendix 1).
Extreme care was taken to insure complete penetration and good coverage by the back-
ing gas.

After the manifold welds were completed, the integrity of the system was de-
termined by helium leak testing. Although it was assumed that pinhole leaks could
be sealed during the subsequent brazing operation, large leaks were not permitted.
It should be pointed out that headers containing questionable welds can be removed
by simply cutting off the tubes at the underside of the header plates. New head-
ers can then be attached by rebending the tubes and repolishing as discussed above.
This procedure was actually employed to obtain leaktight units before brazing was
done on these radiators.

h. Brazing. The back brazing of the tube-to-header welds and the brazing of
the tube-to-fin joints were accomplished by use of a conventional canning techni-
que. The brazing alloy had already been preplaced on the fins, as described pre-
viously, and a medicine-dropper technique was fiséd to place the braze slurry on the
tube-to-header joihts.

A special baffle design was used for the brazing of these radiators because it
was very important to flush out the copious amounts of methacrylate binder present
in the tube-fié»fin matrix. A sheet-metal cover was attached to the front of the

radiator, and the clean, dry inlet gas was forced between the fins and then was
 

T SR A AR e A S AT kit

 

 

 

=3l-

exhausted to the outside of the can. The inside of the sheet-metal cover and the
inlet tube were designed to permit relatively even flow throughout the entire fron-
tal area. A good circulation bf hydrogen was also obtained around the tube-to-
header joints.

The radiator was securely supported on a special heavy framework to prevent
sagging or distortion during the high-temperature brazing cycle. The support bars
were made from stainless steel which had been previously aluminized at 1500°F.

This aluminizing prevented the support bars from being brazed to the radiator be-
cause the aluminum oxide film prevents wetting by the Coast Metals No. 52 alloy.

The radiator was placed in the brazing can with four thermocouples attached
to various positions around the periphery. This temperature measurement around
the radiator was desirable from a control standpoint, even though previous experi-
ments had indicated that all parts of a unit of this size could be held within 25°C
of the desired brazing temperature. In the preliminary experiments, 12 thermocou-
ples were attached at various points on the test specimen. This close control was
necessary since the alloy possessed a flow temperature of 102000, and the copper in
the fins melts at 1083°C.

The thermal cycle used in brazing is shown in Fig. 25. So that distortion
would be minimized, the unit was placed in the furnace at room temperature and was
gradually raised to the brazing temperature. Fig. 26, a photograph of the two com-
pleted radiators, shows that negligible distortion is present. No excess brazing
alloy was encountered, which verifies the practicality of the "sump" technique for
large assemblies. Good flowability of the brazing alloy was obtained, and a photo-
micrograph of a section of a typical tube-to-fin joint is given in Fig. 27. This
Joint was brazed under conditions sirmulating those used in the fabrication of the

500-kw radiators. It can be seen that good edge protection of the exposed copper

 

 
TEMPERATURE—°C

1200

Y -15072

 

100+

000 |—

900}—

800|—

700—

600 |—

S00—

400

300

200

100

 

        

  

HEATING | BRAZING FURNACE
COOLING

 

 

 

-— AR COOLING

 

l I I | l | | | I l | I | |

.98-

 

 

1.0 1.5 2.0 2.5 30 35 4.0 4.5 50 5.5 6.0 6.5 7.0 7.5
TIME - HOURS

Fig. 25. Typical Time-Temperature Thermal Cycle Used When Brazing Radiators.

8.0

 
-36-

: ~r-15mr_

 

Fig. 27. A Typical Tube-to-Fin Joint Showing Edge-Protection of Exposed Copper on Punched Fin
Lips. Etch: None. 50X. Reduced 14%.

 
 

 

-37-
e
was obtained in the tube-to-fin joints and that good filleting was present in the
tube-to-header joints.
3. Summary.
Leak testing with a mass-spectrometer helium leak detector indicated no flaws
in the integrity of the closed sodium circuits. The two radiators were submitted

to the Aircraft Reactor Engineering Division for installation in the test rig. An;_,.___

analysis of the fabrication cost of these two radiators is presented in Afl.@fifigx L.

C. Fabrication of 20-Tube High-Velocity Heat Exchanger.

 

1. Introduction.

A small-scale heat-exchanger testl is being conducted by the Aircraft Reactor
Engineering Division to determine the heat-transfer characteristics of fluorides
through a Reynolds number range of O to 5500. For this test, the fabrication of
the fuel-to-NaK heat exchanger, shown in Fig. 28, was required. The heat exchanger
is installed in the rig shown schematically in Fig. 29. The design of this unit
specified that the 3/16-in. 0D, 0.017-in.-wall Inconel tubes were to be joined to
a dished header at the NaK inlet end and to a radial header at the NaK outlet end.
A thick-walled Inconel pressure vessel was used to confine the molten fuels located
outside the tubes. All welds in this assembly were performed by using the manual
inert-arc-process to ensure the utmost in soundness and corrosion resistance. With
the exception of the tube-to-header joints, which were later back brazed, complete
weld penetration was utilized throughout. 411 welding procedures (Appendix 1) and
operators (Appendix 2) were qualified according to rigid specifications. The root
of the welds was completely protected by an inert gas at all times in order to min-
imize scaling and oxidation.

Back brazing of the tube-to-header joints was employed for two reasons: (1) to

eliminate the "notch effect” resulting from incomplete weld penetration, and (2) to
 

   

 

-38-

 

           
   

 

 

PARTS LIST MAK QUTLET
pant] wo.
nAME nEMAREY

e Y 15605

1 1 (NLET, TUBE SHELCT

2 {2 {meEcar 3" ST'0 WY BUTT WELD

31 Tus- 1 118" PIFE SCWO 40

4 1) | mOuSINE-TURE INLET 3 PIPE SCHO 40

8 | HOUSING -TUSE OFFSET ¥ MPEL SCW'D a0

& | 2 | FLUORIDE BAFFLE

T L STUB-MAK QUTLET (172" MPE 3CHD 40

| FILLER SUDASSEMOLY

* BPACER WIN 032 pia

0|1 | STUN- FLUDAIDE INLET 11727 PIPE SEW'D 40

1|1 | GUTLET TUBE HEADER 3T PIPE SCH'D 40

12 | v | HEADER INNER Ring

15 | met §* 3CH'D 40

140 2 1 HEADER OUTER Ribd

19 | | TRAMSITION SLEEVE

18 |20 | WAk Tuse 1878 00X OF7 WALL

17| 1 | MEAY EXCHANGER SMELL | (/27 PIPE SCHD 40

18| 1§ STUB-FLUDAIDE QUTLET | 41/2" FIPE ACHD 40

 

 

 

 

 

 

 

 

SECTION "a-a"

SECTION '8-8"

 

Fig. 28. Assembly Drawing of 20-Tube High-Velocity Heat Exchanger.

SECRET
ORNL-LR-DWG 5674

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

° 200-kw o
134°F
R NaK-TO- AIR 3 _
23 psio RADIATCR 20 psio
2
n
a
o
o
Y
NaK
PUMP
o
a
w0
¥
144°F ~ 3aF Y
. . ‘45 .
24 psio FUEL-TO-NaK psia
AT ANGER
1500°F HEAT EXCHANGE 1400° F
95 psia - 24 psio
2
a
o
o~
FUEL
PUMP
2
o
Q
0
4500° F 200-kw 1400° F 1
- ELECTRIC —t _
96 psia HEATER 142 psia

 

 

 

Fig. 29. Flow Diagram for Loop Containing 20-Tube High-Velocity Heat Exchanger.

 
 

..39..

ensure against the development of leaks in the event of corrosion through an area
of shallow weld penetration. OSince the qorrosive environment present in this heat
exchanger was to be the same as that in the iOO-tube,heat exchangers, low-melting
Nicrobraz was chosen as the most desirable brazing alloy. A discussion of the pro-
perties of the alloy was presented in the section "Back Brazing of Tube-to-Header
Joints".

2. Fabrication Procedure.

a. Welding sequence. The welding operations and the leak testing were per-
formed in the order listed below:

1. welding the tubes to NaK inlet header,

2. welding the tubes to NaK outlet header and then leak testing the welds,

3. welding the tube-inlet housing to offset housing,

,. welding the fluoride outlet stub to offset housing,

5. splitting of offset housing and welding a baffle to it,
6. welding the outer ring headers to a é-in. pipe,

7. welding the inner ring header to fluoride inlet stub,

8. welding the above part (Step 7) to the NaK outlet header,
9. welding the assembly (Step 8) to the outer ring headers,
10. welding the NaK inlet stub to a 3-in. pipe cap,
11. welding the assembly (Step 10) to the NaK-inlet-tube sheet,
12. welding the housing (Step 5) to the NaK-inlet-tube sheet and to itself,
13. welding the pipe cap tc the assembly (Step 12),
1l;. assembling the wire spacers and filler plates,
15. welding the pressure shell,

16. welding the transition sleeve onto the pressure shell and the NaK outlet header,
17. leak testing the entire unit,

18. welding the "window" after back brazing.
 

 

=10~

b. Brazing. The back-brazing process was performed in a dry-hydrogen atmos-
phere and conventional canning procedures were used. Two brazing operations were
required because the Globar pit furnace available in the Welding and Brazing Lab-
oratory did not possess a heating zone of sufficient length to heat both ends of
the unit up to the brazing temperature. A "window" was removed from the pressure
shell to permit preplacement of the brazing alloy on the NaK inlet end of the tube
bundle. A photograph of the window in the NaK inlet end is presented in Fig. 30.
The window was welded shut and helium leak testing indicated that all the welded
and brazed joints were leaktight. A photograph of the completed heat exchanger is
shown in Fig. 31.

3. Summary.
The unit was installed in the test loop, and operated satisfactorily in ser-

vice for approximately 1500 hr before termination of the test.

 
-41-

Y-14799

 

Fig. 30. The NaK Inlet End of a 20-Tube Heat Exchanger. Note the “‘window’’ through which the

brazing alloy was preplaced and inspected.

Y=14796

 

T L"'u‘_:\.‘ .

 

Fig. 31. The Completed 20-Tube High-Yelocity Heat Exchanger.

 
 

 

=2~
CONCLUSIONS

On the basis of experience obtained in the fabrication of two 500-kw heat ex-
changers, two 500-kw radiators, and a 20-tube high-velocity heat exchanger, the fol-
lowing conclusions can be drawn:

(1) The fabrication of heat exchangers and radiators containing thin-walled
small-diameter tubes in closely packed configurations can be performed successfully
if proper construction procedures and equipment are employed.

(2) A combination fabrication procedure, in which the tube-to-header joints
are welded and back brazed, the manifold joints are welded, and the tube-to-header
joints are brazed, is a very satisfactory method of construction.

(3) High quality, leaktight tube-to-header welds can be produced semiauto-
matically once the optimum conditions and qualification of procedure have been de-
termined.

(L) High quality, leaktight tube-to-header and manifold welds can be made
manually if suitable joint designs and if qualified procedures and operators are
employed.

(5) Dry-hydrogen furnace brazing can be used successfully in the fabrication
of large, complex assemblies if suitable gas-purification equipment is used and if

proper manifolding and jigging techniques are employed.

 
N Ul B W

10.

11.

12.

13.

15.
16.
17.
18.

19.

20.

2l.

 

-1i3-
LIST OF FIGURES

Flow Diagram of Loop for Testing Intermediate Heat Exchanger No. 2.
Assembly Drawing of 500-kw Fuel-to-NaK Heat Exchanger.

Tube-to~Header Weld Exhibiting Extensive Microfissuring.
Tube-to-Header Weld Exhibiting Crack in Early Stages of Propagation.
Tube-to-Header Weld After Back Brazing.

Semiautomatic Tube-to-Header Welding Equipment.

Tube-to-Header Weld Produced with the Semiautomatic Welding Equipment.
One End of Heat Exchanger After Welding Tube-to-Header Joints.

Header in Early Stages of Fabrication.

A Completed Leaktight Header Sectiom.

An Inconel T-Joint Brazed with Low-Melting Nicrobraz and Tested in Fluoride
Bath No. Ll for 100 hr at 1500°F. Only very slight attack is present.

An Inconel T-Joint Brazed with Low-Melting Nicrobraz and Tested in Sodium for
100 hr at 1500°F. Negligible attack is present.

Typical Time-Temperature Thermal Cycle Used in the Brazing of Heat Exchangers.
Sample Tube Bundle Showing Two Types of "Comb Spacers®.

One Tube Bundle After Completion of Root‘Wéld Pass on Pressure Shell.

A Completed 1-Mw Heat Exchanger.

Assembly Drawing of 500-kw NaK-to-Air Radiators.

Sodiun-to-Air Radiator with Stainless Steel Fins Fabricated by the Combination
Welding and Brazing Technique.

Sodium=-to-Air Radiator with High-Conductivity Fins Fabricated by the Combina-
tion Welding and Brazing Technique.

Photomicrograph Exhibiting the Excellent Fin Edge-Protection to High-Temperature
Oxidation Afforded by the Aluminum-Bronzing Operation. Fin on left was unpro-
tected.

Inconel T-Joint Brazed with Coast Metals No. 52 Alloy and Oxidized for 500 hr
at 1500°F. No attack is present.
 

 

22.

23.
2l.

25.
26.

27.

28.
29.
30.

31.

Ll

"A" Nickel T-Joint Brazed with Coast Metals No. 52 Alloy and Tested in Sodium
for 100 hr at 1500°F. Negligible attack is present. .

A Fin, a Template, and a Fin with Preplaced Alloy.

Fin Assembly in Earlj Stages of Fabrication. It can be seen that the process
was difficult even with the use of precision jigs.

Typical Time-Temperature Thermal Cycle Used When Brazing Radiators.
Completed 1/2-Mw and 1-Mw Sodium-to-Air Radiators.

A Typical Tube-to-Fin Joint Showing Edge=Protection of Exposed Copper on
Punched Fin lips. '

Assembiy Drawing of 20-Tube High-Velocity Heat Exchanger.
Flow Diagram for Loop Containing 20-Tube High-Velocity Heat Exchanger.

The NaK Inlet End of a 20-Tube Heat Exchanger. Note "window" through which
the brazing alloy was preplaced and inspected. ‘

The Completed 20-Tube High Velocity Heat Exchanger.

 
10.

 

~L5-

BIBLIOGRAPHY

. E. MacPherson, ANP Quar. Prog. Rep. March 10, 1955, ORNL-186L, p 36.

 

. Patriarca et al., ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, p 72.

 

 

R
P
A. P. Fraas et al., ANP Quar. Prog. Rep. March 10, 195L, ORNL-1692, p 18.
R

. J. Gray, P. Patriarca, and G. M. Slaughter, ANP Quar. Prog. Rep. Sept. 10,
1954, ORNL-1771, p 12L.

 

E. B. Hoffman, C. F. Leitten, W. D. Manly, P. Patriarca, and G. M. Slaughter,
An Evaluation of the Corrosion and Oxidation Resistance of High-Temperature
Brazing Alloys, to be published.

 

A developmental brazing alloy obtained from the Wall Colmonoy Corp. Detroit,
Michigan.

W. S. Farmer et al., Preliminary Design and Performance Studies of Sodium-to-
Air Radiators, ORNL-1509 (Aug. 26, 1953).

 

H. Inouye, High-Conductivity Fin Materials for Radiators, to be published.

 

P. Patriarca, G. M. Slaughter, J. Cisar, and K. W. Reber, Met. Semiann. Prog.
Rep. April 10, 1954, ORNL-1727, p 6.

 

C. V. Foerster and R. C. Kopituk, New High-Nickel Brazing Alloys, Publication
of Coast Metals, Inc., Little Ferry, New Jersey.

 
 

 

46
APPENDTX 1

PROCEDURE SPECIFICATION P.S.-1 FOR D. C. INERT ARC WEIDING OF
INCONEL PIPE, PLATE, AND FITTINGS

First Revision
5-1-55
SCOPE:

 

1. This specification covers the procedure for welding of Inconel pipe,
plate and fittings in the range of thickness .109" through .500"
using the inert-gas-shielded tungsten arc process (D.C.).

2. All welding performed by this procedure shall be done by welders
qualified in accordance with operators Qualification Test Specifica-
tion GTS-1. If the operator has not welded in accordance with this
procedure specification during the preceding 30 day period, he shall
be required to pass tests in the 2G and 5G positions on pipe or plate
in sizes similar to that being used in the construction before he
will be eligible to perform work in accordance with this procedure.

REF'ERENCES:

Procedure Specification Figure PS-1-A.
Qualification Test Specification QTS-1.
Qualification Test Specification Figure QTS-1-A.
ASTM B 166-L9T (Rods and Bars).

ASTM B 167-49T (Seamless pipe and tubing).

ASTM B 168-49T (Plate, sheet and strip).

ASME Boiler & Pressure Vessel Code Section IX.
INCO Technical Bulletin T-2.

PROCESS::

The welding shall be done by the inert-gas-shielded tungsten arc
process (D.C.).

BASE METAL:

The base metal shall conform to the specifications ASTM B 166-49T,
ASTM B 167-49T and ASTM B 163-L9T for Inconel rods or bars, seamless pipe
or tubing, plate, sheet or strip.

FILLER METAL:
The filler metal shall be Inconel INCO # 62 or equal.

POSTITIONS:

The welding may be done with the axis of the plpe in the horizontal
rolled, horizontal fixed and vertical fixed positions. The welding of the
plate shall be done in the flat, horizontal, vertical or overhead positions.

 
 

~47-

PREPARATION OF BASE METAL:

 

1. The pipe and plate to be joined by welding shall be beveled and
fitted in accordance with the joint design shown in Figure P.S.-1-A.

If machine preparation is impractical, grinding, filing or other
means may be employed to obtain the same results as machining.
Before meking the welds, all filing, grease, etc., shall be re-
moved and the plpe cleaned.

2. Plpe or plate to be welded shall be supported to prevent excessive
stress during the welding operation.

CLEANTNG:

Before assembling any jolnts to be welded, the mating surfaces and ad-
Jacent areas shall be thoroughly cleaned by wiping with clean cloths satu-
rated with trichloroethylene.

Caution:

After cleaning with trichloroethylene, allow time for fluid to
evaporate before welding is started.

NATURE OF WELDING CURRENT:

 

The welding current shall be direct current (D.C.). The base metal
shall be on the positive side and the electrode on the negative side of the
line (straight polarity).

INERT GAS (MONATOMIC):

 

1. Argon of 99.8% purity shall be used as the shielding gas through the
welding torch at the rate and with the proper size gas cup as specified
on Figure PS-1-A and shall be the minimum applied in field welding.

2., For the inert blanket inside the pipe, helium of minimum purity of
09.99% shall be used at all times welding is being performed.

WELDING TECHNIQUE:

 

1. The welding technique, electrode size, amperage, and size of filler
metal shall be as shown on Figure PS-1-A. A maximum variation of
plus or minus 10% is permissible for welding currents listed.

2. The basic principle of the welding technique employed 1s to insure
that all welding is done in a constant atmosphere of inert gas.
Therefore it is absolutely necessary to perform all welding in the
field or shop in a qulet atmosphere where no air current or draft
exists in the immediate vicinity where a joint is being welded. In
a majority of cases where welding is being performed 1t is necessary
to shield the location to assure a constant inert gas atmosphere
free from air disturbances and drafts in the immediate welding area.
 

10.

 

~LE-

. The inside or all internal surfaces of the Joint to be welded for a

minimum of 10 inches on each side of the weld shall be purged with
inert gas (helium) before the welding starts and shall remain
blanketed until the temperature of the metal has returned below
LOO°F.

. The tungsten electrode shall protrude approximately l/h inch to

3/8 inch beyond the edge of the gas cup, and in order to have best
arc control, the tungsten electrode shall be dressed to a sharp
point or a bevel having an included angle of approximately 90°,
The tungsten size used shall be as specified in Flgure PS-1-A.

A copper band approximately 1 inch wide shall be made availlable
near the joint to be welded. The arc 1s struck by swinging the
torch in a pendulum like motion immediately over the copper band
toward that point on the band where the welding 1s to start,
allowing the tungsten tip to become red. The operator shall then
go to the welding groove without breaking the arc. This entire
operation shall be repeated each time the arc is broken.

. The torch shall be moved uniformly at all times in order to control

uniform and accurate bead width. A short arc shall be maintalned

so that the pool of molten metal will appear bright, free of internal
porosity and/or oxides and to assure that maximum penetration and
fusion is obtalned.

. The filler wire should be added or kept in front of the heat at an

an angle of from 20 to 30°. The wire should be fed in short jabs
into the very front edge of the weld pool and as near as where re-
quired, drawing the arc back along the weld if necessary to prevent
the arc from playing on the end of the filler metal.

. A uniform quantity of filler wire should be added or melted off with

each jab and the filler wire withdrawn about 1/4 inch out of the
welding zone with the frequency of adding wire to be adjusted in re-
lation to the forward welding speed to give the required welding
bead. Extreme care should be taken at all times to keep the end of
the welding rod in the atmosphere of argon.

. Precautionary measures are to be followed to obtaln continuous and

uniform penetration for the entire root of the weldment. Root pene-
tration will not be permitted in excess of 1/16 inch. Caution shall
be exercised at all times with particular emphasis on the first pass
when starting and stopping to prevent voids and pinholes at the root
of the weld.

The flow of argon shall be maintained on the electrode and the weld
until returned to normal color.

Care should be exercised to prevent blowing the pool of molten metal
with the torech or stirring it with the filler wire. If at any time
the tungsten electrode becomes contaminated with flller or weld metal,
welding should stop at once and the electrode cleaned or broken off
and reshaped before welding is continued. This is to keep the weld
free of tungsten contamination.

 

 
“Lg-

DEFECTS:

Any cracks or blowholes that appear on the surface of any tack weld or
bead of welding shall be removed by chipping, grinding, or filing.

All surface oxides must be removed by wire brushing with a stainless
steel wire brush before depositing the next successive bead of welding.

PREHEATTING:
In general, preheating is not necessary except when metal temperature
is 32°F or below, then the joint shall be preheated to TO°F or over for 6

inches on each side of the welding groove.

IDENTIFICATION OF WELDS:

 

Each weldment shall be identified by the welding operator's stencil
number adjacent to the weld, placed at the most convenient point of ex-
posure for ready examination.

GENERAL:

1. At all times during welding of Jolnts in accordance with this speci-
fication the authorized inspector shall evaluate the weld quality in
terms of the soundness requirements of Qualification Test Specification
QT8-1. This shall include the right to remove and test questionable
welds. If the joint is substandard to the soundness requirements of
QTS-1 the operator shall be disqualified until a satisfactory re-test
is made.

2. It is of utmost importance that the operator welding under this pro-
cedure check all fit ups and keep the proper amount of argon gas
flowing on the work to assure that the welding being performed will
be of the highest quality possible.

3. It shall be the duty of the welding operator using this procedure to
report to his foreman any irregularities which might impair the quality
of workmanship desired by this procedure.

 

Engineering & Mechanical Divisio

-

 

P.S.-1 First Revision Pa ?a'f,riarca, Metallurgisf
Metallurgy Division

Date: 5-1-55
 

 

 

 

-50-

FIG. PS

DESIGN FOR GROOVE WELDS

b= A

IN PIPE OR PLATE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T
=z
ES
c 0 W o o| o 0 o o 0 0 o O o
RM < < o ® 0 ™~ o © ~ o w0 ® M
>
O
2 dnosvo e |~| o ol o | o|® | oo |0 |0 o
H4dD 0 o o wl| o w | w w0 0 w0 o 0 w0
< dN-MIvg-9H | e | s | ¢ S| | ||| || T | <=
2 Hao-gmo-2H 8|2/ el 8lslels8glsla]gls]®
T
5 3Z1S HITUG [0 -2 S| mS| ]S | wlm [ole | lm | -0 |~ oy | o |
m 3Z1S 9NN1L —@ L —j@| oY mIN| o | —j© | o =0 | o |—l© | 4@ | ~|o | J©
"
ON SSvd ||| =| 1| -] ao| T |l a|®|-]a]|?
o m %) ")
S3SSVd ON o M < 0 ©
T
<
MP w [ w | O Ol o | w | © o o | w ) o | o
RM < < 0 © 0 ~ o 0 ~ o O ® <
A - — -
S
O
O dnN) Sv9O ~ |~ ® o | ® © | © © ® ® ® © ©
A
H4D 0 0 W
> 1% dnyovE-oH | S| S| S| S| ||| |F ||| ¥
-+
© _ H30-dnd-v | S| 8| R”| R| 3|3 |8 |R|8|8|8 |8~
E © 3Z1S 437774 || 1| M| @IS 0N | miS [l [mih | 1o |-l [ mig | —jo | o
wn
Mau 3ZIS ONNL —C | 2| m$ Q| —l® | o |—jo |-I® [ ~jo |~|® | o | —|o Ao
| " 0 "
ONSsYd |- |l -1 -1Y =T -]
o o o < N < o~ <
M D
S3SSVd ON _ b © ~ o
o
7p)
n
- L =
— ARWT o O © O —|y
©w FCaov = S ~ P
0 »

 

 

 

 

 

 

 

 

 

REVISION ¥

DATE: 5/1/55

L Bl

 
 

~51=

APPENDIX 2
OPERATOR'S QUALIFICATION TEST SPECIFICATION QTS-1 FOR INERT
ARC WELDING OF INCONEL PIPE, PLATE, AND FITTINGS

First Revision

5-1-55

SCOPE:

This specification covers the qualification of operators for approval

to heliarc weld Inconel pipe, plate, and fittings in accordance with Pro-
cedure Specification P.S.-1.

REFERENCES:

Procedure Specification PS-1.

Procedure Specification Figure PS-1-A.
Qualification Test Specification Figure QTS-1-A.
ASTM B 167-LoT.

INCO Technical Bulletin T-2.

ASME Boller & Pressure Vessel Code Section IX.

MATERIAL REQUIRED:

 

Test weldments shall be made using ASTM B 167-4OT Inconel pipe as noted

below:

TEST

For groove weld tests,

 

Four pieces of 3" diameter schedule 4O pipe are rquired. Each piece
shall be approximately 4" long and beveled as shown on Figure QTS-1-A.

For fillet weld tests.

 

Two pieces of 2" diameter schedule 80 pipe 3" long and one piece of
1-1/2" diameter schedule 4O pipe 1-1/2" long are required. All pieces
of pipe shall have both ends sguare cut.

Filler Metal.

The filler metal shall be INCO # 62 welding wire or approved equal,

POSITIQNS:

l. Grove welds.

Test # 1 - Position 2G - A groove weld shall be made between two
pieces of 3 inch pipe placed with the axis in the vertical position and
the welding groove in a horizontal plane as shown on Figure QTS-1-A
test # 1. After welding, the pipe shall be stamped with numbers 1, 2,
3, and 4, clockwise and approximately 90° apart.
 

 

-50-

Test # 2 - Position 5G - A groove weld shall be made between two
pleces of 3 inch pipe, placed with the axis in a horizantal position
and welding groove in a vertical plane as shown on Figure QTS-1-A test
# 2. Before welding, the pipe shall be stamped with numbers 1, 2, 3,
and U clockwise, starting with number 1 on the top when arranged for
welding.

2. Fillet welds.

Test # 3, part A, a full fillet shall be made joining the 2 and
1-1/2" pipes as shown in Figure QTS-1-A test #3 part A. This joint
shall be welded with the axls of the pipe in the horizontal fixed
position and the weld in a vertical plane.

Test 7 3, part B, a full fillet shall be made on the other side of
the Joint, Joining the 2" and 1-1/2" pipes as shown in Figure QTS-1-A
test # 3 part B. This jolnt shall be welded with the axis of the pipe
in a vertical fixed position and the welding plane horizontal.

Make the close-in passes as shown on Figure QTS-1-A test # 3 using
any convenient welding position. The pipe shall then be stamped with
the numbers 1, 2, 3, and 4 at 90° intervals around the weldment.

WELDING REQUIREMENTS:

 

1. The welding operator shall be required to follow procedure specifi-
cation PS-1 in making the test welds and shall not be allowed to
rotate or turn the pipe during welding.

2. An Inspector shall be present at all times while the gqualification
test is in progress. The Iinspector may refuse acceptance of a test
weldment if the operator does not comply with the standard procedure
in all respects.

NON-DESTRUCTIVE INSPECTION OF WELDMENTS:

 

The finished weldments shall be inspected for deviation from the pro-
cedure specification and for the points listed below:

Groove and fillet weld specimens:

1. The outer surface of the weld bead reinforcement shall be 1/16 inch
(+1/32 inch) and shall be uniform in size and contour.

2. There shall be no undercut, overlap, or lack of fusion.
Groove weld specimens:
1. There shall be complete, uniform penetration. Any weldment having

weld metal protruding inside the pipe or tube more than 1/16 inch,
or having pinholes at the root will not be accepted.

Gl rABMET 35 | B b £l ™t B oy AR IR SR 2 2 7o B i i e e, i e s s e
 

-53-

TEST SPECIMENS AND DESTRUCTIVE TESTS:

 

1. Groove weld specimens.

(a)

(o)

(4)

The weldments shall be machine cut longitudinally so as to
form four coupons, each approximately 1 inch wide and bearing
a stamped identification number. Additional specimen cutting
is required in paragraph (e) below.

Weld reinforcement on the specimens shall be removed flush with
the surface of the pipe by machining, filing or grinding and it
will not be permissible to remove undercutting or other defects
below the surface of the base metal.

Neither will it be permissible to remove any base metal from
inside of the pipe in order to conceal any evidence of lack of
penetration or fusion at the root of the weld. The edges of
all groove weld specimens shall be rounded by removal of the
burr with a file.

A guided bend jig proportioned for 0.216" material shall be
used to conduct bend tests®as follows:

Specimens 2G-1, 2G-2, 5G-1 and 5G-2 shall be given a guided face
bend test. Specimens 2G-3, 2G-4, 5G-3 and 5G-4 shall be given a
guided root bend test.

Two sample pleces approximately 3/4 inch wide shall be removed
as welded from each weldment from positions approximately 180°
apart and these shall be stamped with the proper identification
number of the weldment, position and operator. These sample
pleces shall be radiographed and the radiographs shall be ex-
amined for evlidence of cracks, porosity, and tungsten inclusions.
The welds shall then be prepared for metallographlic evaluation
of the transverse section and examined in the as polished and
etched condition for evidence of flaws,

2. Fillet weld specimens.

(2)

()

the weldment shall be machine cut longitudinally so as to form
four coupons, each approximately 1 inch wide and bearing a stamped
ldentification number. Additional specimen cutting is required in
paragraph (d) below.

Weld reinforcement and the backing ring on the specimens shall be
removed flush with the surface of the pipe by machining, filing
or grinding and it will not be permissible to remove undercutting
or other defects below the surface of the base metal.

Each specimen shall be subjected to a gulded root bend test using
a gulded bend jog proporticned for 0.216" material.
 

(a)

-5k4-

Two weld sample pieces approximately 1/2 inch wide shall be re-
moved as welded from the weldment from positions approximately
180° epart and these shall be stamped with the proper identifi-
cation number of the weldment, position and operator. The welds
shall be prepared for metallographic evaluation of the transverse
section and examined in the as polished and etched condition for
evidence of flaws.

TEST RESULTS REQUIRED:

 

1. Groove weld specimens:

(a)

The convex surface of the bend specimens shall be free of all
cracks or other open defects. Cracks occurring at corners of
specimens during testing shall not be considered unless it is
Indicated that the origination was from a welding defect.

The convex surface of the weld shall show complete penetration
with no evidence of lack of fusion at the root of the weld.

The radiographic and metallographic examinations shall show no
evidence of cracking or incomplete fusion. Gas pockets or in-
clusions shall not exceed one per specimen and none shall
exceed 1/32 inch in its greatest dimension.

2. Fillet weld specimens.

RETESTS:

(a) The convex surface of the bend specimens shall be free of gll

cracks or other open defects. Cracks occurring at corners of
specimens during testing shall not be considered unless it is
Indicated that the origination was from a welding defect.

(b) The metallographic examination shall show no evidence of crack-

ing or incomplete fusion. Gas pockets or inclusions shall not
exceed one per transverse section and none shall exceed 1/32
inch in its greatest dimension. Penetration into the base metal
shall be 1/16" + 1/32".

In case a welding operator fails to meet the requirements as stated,
a retest may be allowed under the following conditions:

1. An immediate retest may be made which shall consist of two test welds
of each type and test position that has been failed, all of which shall
meet the requirements specified for such welds, or;

A complete retest may be made at the end of a minimum period of one
week providing there is evidence that the operator has had further
training and/or practice.
 

-55-

ASSIGNMENT OF CODE UPON PASSING QUALIFICATION TEST:

Welding operators passing the above test will be qualified and his
operator's card so marked for weldlng by the hellarc process as specified
under Procedure Speclfication PS-1.

RECORD OF TEST:

A record shall be kept of all pertinent test data with the results
thereof for each operator meeting these requirements. This record shall
be originated by the inspector.

Tested specimens shall be identified and made avallable for examination
by interested parties until all fabrication requiring the use of this speci-
fication has been completed and the system has been accepted.

   

   

 

  

 

T. R. Housley, Chief Welding InSpector
Engineering & Mechanical Divisign

 
 

. Patriarca, urgls
Metallurgy Division

  

QTS-1 First Revision

Date: 5-1-55
 

 

_56...
FIG QTS—-1-A
DETAILS FOR GROOVE AND FILLET TEST WELDMENTS

| . \( | OO°>/
XN ZZ7 A

Iy

 

‘r

 

A
ol
ng
O
>
»

 

 

 

 

 

POSITION 5 G

(Pipe axis hortzontal fixed)

TEST NO 2

POSITION 2G

(Pipe axis vertical fixed)

TEST NO. |

 

 

JOINT FIT-UP AND PASS SEQUENCE FOR TEST NO 3
] % / 2" Sch. BO pipe, 3" long
I

- N\ -

Close-in welds
(6 passes)

 

 

 

 

 

"

l% S5ch.40 pipe
(Backing ring) et |

 

Y

L
2

Axis vertical

  

Weld completely
around this side
(Passes 1,2)

Weld oround
this side
(Passes 3, 4)

   

AXxis
horizontal

 

PART "A"
(Pipe axis horizontal fixed) __PART B
( Pipe axis vertical fixed)

REVISION ¥ TEST NO. 3
DATE:5/1/55

 
 

-57-
APPENDIX 3

JOINT DESIGN FCOR INERT ARC WELDED VESSELS

 

Since stainless steel and Inconel vessels fabricated at Oak Ridge National
ILaboratory are generally intended for service in a severe corrosive environment,
and oftentimes at an elevated temperature, the choice of Jjoint design for this
fabrication becomes important.

The valid assumption that incomplete penetration in welded Jjoints will
give rise to the possibility of accelerated corrosive attack, generally re-
ferred to as "crevice corrosion,” in itself justifies the joint designs pre-
sented herein.

Although strength requirements are generally considered secondary, the
combination of pressure and elevated-temperature service would suggest that
designs for optimum strength as set forth by the A.S.M.E. Boiler Construction
Code should be given serious consideration.

Inert arc welding has, in recent years, proved its superiority to con-
ventional metallic arc welding for high-corrosion applications. Slag in-
clusions in weld metal and unavoidable irregularity in penetration suggest
that metallic arc welds be avoided, at least in those areas where corrosive
media will come into direct contact with the weld deposit.

The critical nature of the majority of our work has given rise to the
necessity of training and qualifying our welding operators for the highest
quality welding possible.

Welding procedure and qualification test specifications are available
which meet these requirements. These are Carbide and Carbon Chemical Com-
pany's Engineering Specifications ESP 4-3-26 and ESP 4-4-26 for the inert
arc welding of the austenitic stainless steels and the Oak Ridge National
Laboratory's Specification PS-1 and QTS-1 for the inert arc welding of Inconel.
The joint designs as presented require operators qualified under the specifi-
cations mentioned above.

Fabrication of any standard vessel will involve the use of relatively

few basic joint designs:
/ | 00‘3’

I. Butt Welds.

A. Bingle bevel:

 

 
 

 

-58-

The above design is applicable to butt welds in pipe or fittings where
wall thicknesses are 3/8 inch or less. The dimensions as shown are fixed
with the exception of the root spacing 8. This dimension should be 1/16
inch for 1/8 inch diameter pipe gradually increasing to 5/32 inch for 3 inch
diameter pipe and above. Details are available in Figure ES L4-3-26 of the
specifications mentioned previously.

This joint design is also applicable for girth and seam welds 1in plate
up to 3/8 inch in thickness used for fabrication of the shell. When plate
thicker than 3/8 inch is used a straight bevel as shown in A will recuire an
excessive amount of weld deposit. Therefore, a double bevel or a single U
is recommended.

B. Double bevel:

 

 

 

 

 

 

This joint is applicable to girth and seam welds in plate up to 3/4 inch
in thickness where the underside of the weld is accessible. When the corrosive
environment will have access to only one side of the weld deposit, metallic arc
welding may be substituted for that half of the weld not subjected to corrosion.

The operators must be qualified according to ESP 4-L4-20 for metallic arc welding
of stalnless steel.

C. Single U:

~3pa

/
/.

I

 

 

 
 

 

5 TR ] SR A L

_59..

Where the weld joint is accessible from one side only, a single U
bevel is applicable for plate thicker than 3/% inch. The radius R will
vary from 3/32 inch to 3/16 inch depending on the plate thickness. 1In
this case as in the case of the double bevel, metallic arc welding may
be used by a qualified operator to complete the weld after two or three
passes have been deposited in the root using inert arc welding.

It must be remembered, that a back-up gas is to be applied at all
times during welding in order to protect the opposite face from oxidation.
This also applies when completing the weldment with metallic arc.
II. Attaching a Nozzle to a Shell,

Al

 

i |
s Z

 

 

 

el

 

\
§ WA

 

 

¢

This joint is applicable when nozzles of 1 inch pipe or under are
attached to a shell. The root spacing S will vary from 1/16 inch for
1/8 inch pipe to 1/2 inch for 1 inch pipe.
 

 

-60-

 

 

%\:
AT o
A A A 3

 

 

 

 

e
e

h_} "

L
)

 

4
%)

 

 

When the nozzle is l—l/h inch pipe or larger and the shell thickness
is 3/8 inch or less, the above design may be used. The root spacing S is
varied from 1/8 inch for nozzles of 1-1/4 inch pipe to 5/32 inch for pipe
3 inches or over.

c.

A

<
>

<

 

 

L
S R

 

 

 

 

 

 

R -

¥
w

 

When the shell thickness is greater than 3/9 inch and the nozzle is
l-l/h inch pipe or larger, a single J is applicable. The root spacing S
will vary from 1/8 inch for 1-1/L inch pipe to 5/32 inch for pipe 3 inches
or over. The radius R will vary from 3/32 inch to 3/16 inch depending on
the plate thickness.

As indicated previously, metallic arc welding by a qualified operator
may be used to complete the weld after two or three root passes have been
deposited by inert arc welding.

 

 
-61-

When it is desirable to have plipe of all sizes extend into a vessel
the joints applicable are either IIB or IIC above depending on the shell

thickness.

ITI. Attaching Heads.

Since dished heads are commercially available a simple butt joint as
described in part I of this memorandum can be used. Occasionally, it may
be necessary to machine dished type heads when specific sizes are not avail-
able. As a last resort, attachment of a flat head may be necessary. 1In
this event, the following design 1s recommended:

e \P;BD_' 4

 

 

 

 

 

 

 

| »
be—S G

 

 

 

 

 

 

 

 

 

As can be seen, the attachment of a flat head consists primarily of
the use of the basic joints previously described i1n part IIB or IIC de-
pending on the head thickness. It may be noted however, that the shell
as shown extends above the head surface approximately 1-1/4 times the
shell thickness to accommiodate a fillet.

As previously indicated, metallic arc welding may be used to complete
the weldment after two or three root passes have been deposited by inert

arc welding.

IV. Flanges.

It is recommended that weld neck flanges which are commercially
available are used wherever possible. The simple butt welds described
in part I are then applicable. Occasionally a slip-on flange must be
used, however. In thls event, the following joint 1s recommended.
 

 

 

 

 

 

 

 

it e e e —— -

 

 

 

 

 

 

 

It is hoped that the contents of this memorandum will be useful to
design men concerned with fabrication of critical assemblies.

 

Maintenance Division
Cak Ridge National Laboratory

 

 

   

. ratriarca
Metallurgy Division
Oak Ridge National Labeoratory

 
 

-63-
-

Hepmmry L >

INTER-COMPANY CORRESPONDENCE
Oak Ridge National Laboratory

To W. F. Boudreau Date April 27, 1955

Location 9201-3, Y-12

Attention Subject Estimate of Time Required to

Copy to Construct Radiators and Heat Exchangers
for the 1 MW Intermediate Heat Exchanger
Test

CONFIDENTIAL | UNDOCUMENTED

The two-one hundred tube heat exchangers and the two radiators to be used in the inter-
mediate heat exchanger test were fabricated by the Welding Laboratory of the Metallurgy
Division. A record has been maintained of the number of man-hours that were required
to construct the four components. A detailed breakdown of the time required to com-
plete the fabrication is presented.

Since the record includes all experimental development necessary to the successful con-
struction of the four units, two time estimates are given. One estimate (Column 4) in-
cludes all time spent on developments including changing procedures, repetition of work
because the original design was inadequate, and repairing units that were not leaktight.
The second estimate (Column B) includes only the time that would be required to build
additional units if no exigencies arise during the fabrication. The two estimates
therefore might be considered to include the minimum effort required to fabricate an-
other assembly in one case, and the maximum as encountered in fabricating the first
complete experimental intermediate heat exchanger test assembly.

The time spent by P. Patriarca and G. Slaughter in supervising the fabrication is not
included.

Estimated Cost of Materials
Radiators

 

Reference ORNL Dwg. D-SK-16590

 

1200 lineal feet 310 stainless steel clad copper radiator fin material $ 234.00
252 feet 3/16-in OD, 0.025-in wall Inconel tubing 100.00
Cost of hydrogen at $10.00/brazing operation 20,00
Inert welding gas, $25.00/unit 50.00
$ LOL.oO

Heat Exchangers
1000 feet 3/16-in 0D, 0.017-in wall Inconel tubing $ L00.00
150 pounds Inconel at $2.50/pound 375.00
Cost of hydrogen at;i&fia@?/bxazing operation 10.00
Inert welding gas, $25.00/unit 50.00
$ 865.00
 

 

_6l-

Estimate of Time Required to Construct Radiators
and Heat Exchangers for the 1 MW Intermediate
Heat Exchanger Test.

 

If a cost per man-hour of $5.00, including overhead, is taken as a reasonable figure,
the following cost estimates are obtained:

Actual fabrication cost of 2 radiators: 998 hr x $5.00 $ L,990.00

Cost of dies for punching holes in fins 600.00
Cost of materials for radiators LOL .00

$ 5,99L.00

Actual fabrication of 2-100 tube heat

exchangers: 579 hr x $5.00 $ 2,895.00
Cost of materials for heat exchangers 865.00
$ 3,760.00

Total cost of fabricating radiators and heat
exchangers for the 1 MW Intermediate BHeat
Exchanger Test $ 9,754.00

Predicted fabrication cost of two radiators:

82l hr x §5.00 $ 1,120.00
Cost of materials for two radiators L0L.00
Estimated minimum cost of constructing two

additional radiators $ L,524.00
Predicted cost of fabrication of two heat

exchangers: 459 hr x $5.00 $ 2,295.00
Cost of materials for two radiators 865.00

Estimated minimum cost of constructing two
additional 100 tube heat exchangers $ 3,160.00

 

 

M. J. Whitman G. M. Slaughter

 
2 . - . .

Aty

TIME EXPENDED IN THE FABRICATION OF TWO-ONE HUNDRED TUBE HEAT EXCHANGERS FOR INTERMEDIATE HEAT EXCHANGER TEST
Metallurgy Divis‘.z.q_r_l » Dak ;Riflg"fi National Laboratory

 

 

 

A B
Total Minimum
Man-Hours Man-Hours
Date | Expended Expended
Prior to
Feb. 10 Fabrication of a sample header with welded tubes to determine
optimum welding conditions.
Machining 16 -
Welding 2 -
Metallography 10 -
Construction of jigs 10 -
Feb. 1-10 Expanding tubes, Jjigging, degreasing, welding tube-to-header 25 25
joints and leak testing Tube Bundle #1.
Feb. 10-18 Expanding tubes, jigging, degreasing, welding and leak testing 25 25
tube-to-header joints of Tube Bundle #2.
Feb. 14-15 Welding side plates and nozzles on Header #1 of Bundle #1 40 Lo
Feb. 18 First attempt to back braze Header #1 of Bundle #1. Two men 16 16
required to control atmosphere in hydrogen furnace, weld
bundle into can containing hydrogen atmosphere, load bundle
into furnace and remove bundle from furnace.
Feb. 19 Repeat back brazing of Header #1 of Bundle #1 because brazing 16 -
was unsatisfactory.
Feb. 19 Welding side plates and nozzles on Header #1 of Bundle #2. 16 16
Feb. 21 Repeat back brazing of Header #l1 of Bundle #1 with dry brazing 10 -—
alloy powder added to seal last few leaking joints.
Feb. 22 Leak check and inspection of Header #1 of Bundle #1. 5 5
Feb. 21-22 Welding side plates and nozzles on Header #l1 of Bundle #2. 2l 2l
Feb. 23 Fourth and last back brazing of Header #l of Bundle #1l. 10 -
Feb. 24-25 Welding side plates and nozzles of Header #2 of Bundle #1. 32 32
Feb. 2L Successfully back brazed Header #1 of Bundle #2. . 10 10
Feb. 25-26 Fabrication of comb spacers. 12 12
Feb. 26 Completed welding of side plates and nozzles of Header #2 of 16 16
Bundle #1 and began welding side plates and nozzles on
Header #2 of Bundle #2.
Feb. 28 Completed welding side plates and nozzles of Header #2 of 16 16
Bundle #2. v “5 ‘_; Y

Sunmny

 

...g9..
Time Expended in the Fabrication of Two-One Hundred Tube Heat Exchangers for Intermediate Heat Exchanger Test
Metallurgy Division, Qak Ridge National Laboratory

 

 

 

A B
Total Minimum
Man-Hours Man-Hours
Date Expended Expended
Feb. 28 Successfully back brazed Header #2 of Bundle #l. 10 10
Feb. 28 Preparation and construction of test specimen to check comb 2l -
spacer installation and placement of spacers on Bundle #1.
March 1-2 Installation of comb spacers on Bundle #1. 2L 24
March 1-2 Fitting Bundle #1 into shell channel. 2L 2L
March 4,5,7 Welding Bundle #l1 into shell channel. L0 L0
March 5 Back brazing Header #2 of Bundle #2. 16 16
March 7-8 Installing comb spacers on Bundle #2. 32 32
March 9 Fitting Bundle #2 into shell channel. 16 16
March 10-11 Welding Bundle #2 into shell chamnel. 2l 2L
March 12-1L Welding assembled Bundles #1 and #2 together and 32 32
welding on nozzles.
March 15 Final leak test of complete assembly. L L
Total man-hours expended 579 459

in fabricating the heat
exchanger assembly.

n99.—
 

Metallurgy Division, Oak Ridge National Laboratory

TIME EXPENDED IN CONSTRUCTING THE TWO RADIATORS FOR THE INTERMEDIATE HEAT EXCHANGER TEST

 

 

 

A B
Total Minimum
Man-Hours Total
Date Man-Hours
Jan. 10-2) Determining optimum amount of brazing alloy to be preplaced 32 --
and making templates for preplacing alloy.
Jan. 12 Shearing 2200 high conductivity fins to 2-in x 8-in. 16 16
Jan. 1l Inspecting fin surfaces for blisters and exposed copper. L L
Jan. 17-18 Degreasing fins. 8 8
Application of aluminum edge protection to 2200 fins. Batch
process with 1500 in first batch and 700 in the second.
Clamping fins into stacks. L L
Jan. 27-28 Spraying to seal stack edges. 2 2
Jan. 31 Application of aluminum by painting. 3 3
Feb. 1 Canning stacks for furnace treatment. 2 2
Furnace heat treatment. 7 7
Removing fins from cans. 2 2
Wire brush cleaning fins to remove excess aluminum. 6 6
Inspection of fins. 8 8
Feb., 2-11 Punching fins. 100 100
Machining jigs for tube bending. Done at Y-12 50 -
Bending tubes. 40 Lo
Feb. 11 Inspecting fins. 3 3
Feb. 10-11 Washing and degreasing punched fins. 15 15
Feb. 1L4-17 Preplacing brazing alloy and binder on fins. 146 146
Inspection of preplaced alloy on fins. 17 17
Assembling fins and tubes to Radiator #1. L2 42
Feb. 18 Assembling Radiator #1. 16 16
Jig manufacture for radiator assembly. L -
Feb. 19 Assembling Radiator #1. 16 16
Feb. 21 Assembling Radiator #l. 16 16
Assembling Radiator #2. 16 16
Feb. 22 Assembling Radiator #1. 12 12
Assembling Radiator #2. 12 12
Design of jigs and fabrication of alignment pins. 9 -

-

   

-L9_
Time expended in Constructing the Two Radiators for the Intermediate Heat Exchanger Test
Metallurgy Division, QOak Ridge National Laboratory

 

 

 

A B
Total Minimum
Man-Hours Total
Date Man-Hours
Feb. 23 Completing assembly of fins and spacers to 16 16
tubes of Radiator #1.
Assembling Radiator #2. 16 16
Feb. 1-24 Fabrication of radiator spacer plates. 12 12
Feb. 24 Completing assembly of fins and spacers to tubes, Radiator #2. 16 16
Feb. 24 Welding side plates on Radiator #1. 3 3
Feb. 25 Welding side plates on Radiator #2. 3 3
Feb. 25 Design of jigs for bending tubes to fit headers. 3 -~
Feb. 25 Preparing header plates for both radiators. 10 10
Feb. 26 Construction of test specimen to check techniques 8 -
for bending ends of radiator tubes.
Feb. 28 Fitting and clamping tube ends prior to bencing for header L L
installation.
March 1 Fitting and clamping of tube ends prior to bending. 11 11
March 2 Completed construction of tube bending test specimen. 8 -
March 3-9 Radiators sent to ¥-12 for assembly of headers to tubes.

This was not accomplished at Y-12 since decision was made to
redesign header to eliminate much of the tube bending to facil-
itate assembly of headers to radiator tubes. HRadiators were
returned to X~10.
March 10 Preparation of specimens for use in optimizing bending of 16 -
radiator tube ends and assembling and weldirig headers to
radiator tubes.

March 12 Bending ends of radiator tubes. 3 3
March 14 Bending and trimming tubes to proper length. 8 8
March 15 Bending and trimming tubes to proper length. 8 8
March 16 Polishing ends of radiator tubes to make it easier to fit headers. 8 8
March 17 Deburring holes on the four headers. 8 8
March 18,21,22 Fitting one of the two headers on each radiztor. 19 19
March 26 Welding Header #1 on Radiator #1 and Header #1 on Radiator #2. 8 8
(tube-to-header welds)
March 28-29 Welding end plates, covers, and nozzles on Header #l of 32 32

Radiator #1 and Header #1 of Radiator #2.

 

-89_
Time Expended in Constructing the Two Radiators for the Intermediate Heat Exchanger Test
Metallurgy Division, Oak Ridge National Laboratory

 

 

 

A B
Total Minimum
Man-Hours Total
Date Man-~-Hours
March 30 Fitting tubes into Header #2 of Radiator #1l. 7 7
March 30-31 Tube-to-header welding of Header #2, and welding end plating 16 16
cover and nozzle on Header #2 of Radiator #1.
April 1-2 Leak tested Radiator #1. Leak discovered in tube-to-header weld. 2l
Cut all tubes to remove header that leaked. Installed new header
and made tube-to-header welds, welded on end plates, cover, and
nozzle. Repeated leak test.
April 5-6 Preparing Radiator #1 for brazing which included: placing braz- 16 16
ing alloy on tube-to-header joints, jigging, attaching thermo-
couples, adding baffles to produce even hydrogen flow through
fins; and canning the radiator.
April 7 Furnace brazed Radiator #1. 16 16
April 8 Leak tested Radiator #1. L L
April 11 Fitted second header to Radiator #2. 7 7
April 11-12 Made tube-to-header welds, and welded end plates, cover, and 16 16
nozzle on Header #2 of Radiator #2.
April 13-14 Leak tested Radiator #2. Found large leak in tube-to-header 2L L
weld. Cut tubes below tube-to-header welds to remove header
and replaced header. Made tube-to-header welds and welded
end plates, cover, and nozzle on new header.
April 15 Leak tested Radiator #2. L -
April 18-19 Prepared Radiator #2 for brazing (See April 5-6). 16 16
April 20 Furnace brazed Radiator #2. 16 16
April 21 Leak tested Radiator #2. L L
Total man-hours required to 998 824

fabricate the two radiators

Sy

_69...