Professional Documents
Culture Documents
WJ 1973 07 s281
WJ 1973 07 s281
W E L D I N G R E S E A R C H S U P P L E M E N T ! 281-s
EXPANDED SCHAEFFLER CONSTITUTION DIAGRAM
FOR STAINLESS STEEL WELD METAL
18 19 20 21 22 23 24 25 26 27 28
x
CHROMIUM EQUIVALENT = £Cr+§Mo + 1.5 x-g-Si + 0.5 "rjCb
Fig. 1 — Schaeffler diagram
several variables in the measure- posit cools through a temperature distribution. However, besides being
ment and calculation of ferrite are range just below the delta ferrite c u m b e r s o m e , such a m e t h o d re-
considered, as well as the overall region. The transformation is time-de- quires considerable care to be sure
variation between them. Allowances pendent — i.e., the ferrite does not in- that the section or sections studied
for receiving inspection are recom- stantaneously transform, but does so give a true picture of the volumetric
m e n d e d to p r o v i d e adequate progressively over a short period of ferrite distribution. Also, the etchant
tolerances for variations in ferrite con- time — and theoretically the trans- selected and the degree of etch have
tent between pads. formation could be avoided almost been shown to be among the several
Measured ferrite is r e c o m m e n d e d completely if the deposit could be in- variables in procedure and technique
over calculated for specification pur- stantly quenched from the just-frozen which can influence rather s u b -
poses. state to black heat, which would re- stantially the values obtained. 1
sult in a much higher percentage of In 1949 Schaeffler published his
Background
ferrite in a Type 308 deposit, for ex- Constitution Diagram For Stainless
The Literature ample. Similarly, if the cooling rate Steel Weld Metal, 2 which provided a
The effectiveness of the presence from the molten state is rapid enough, tool for calculating weld deposit fer-
of a small but controlled amount of it is possible to supercool through the rite c o n t e n t b a s e d on c h e m i c a l
delta ferrite in preventing cracking of t e m p e r a t u r e zone in w h i c h delta analysis by graphically combining ef-
austenitic stainless steel weld d e - ferrite f o r m s and to transform d i - fects of the austenitizers nickel, car-
posits is well known. Types in which rectly into austenite. Practically bon and manganese and the fer-
this effect is taken advantage of in- speaking, however, the final amount ritizers c h r o m i u m , m o l y b d e n u m ,
clude among others 308, 308L, 309, of delta ferrite in virtually all weld silicon and columbium; this diagram
347, 318 and sometimes 316 and metal depends only to a rather minor gave a figure for ferrite content which
316L and 317 and 317L. degree upon the cooling rate. was stated to be accurate to ± 4
In cooling from the molten state The amount of ferrite in stainless volume percent (hereinafter referred
austenitic stainless weld metal of nor- weld deposits can be determined in to as %) ferrite for many of the c o m -
mal carbon content solidifies first as a any of several ways. Metallographic mon austenitic weld deposits under
mixture of delta ferrite and austenite; e x a m i n a t i o n of the d e p o s i t c a n cooling conditions present in field
most of the ferrite subsequently provide an indication of the approxi- weldments. An expanded portion of
transforms to austenite as the de- mate ferrite content in terms of area the Schaeffler Diagram is shown as
282-s I J U L Y 1 973
Fig. 1; this diagram has been widely rather accurately known. A more di- of those instruments did produce dif-
accepted and very useful because rect measuring technique is the use of ferent ferrite readings on a given
calculated ferrite is rather easily de- an appropriate magnetic measuring specimen. These readings could
termined. instrument such as an Aminco-Bren- range from 3 to 9% on samples which
In 1956 DeLong et al3 expanded on ner Magne-Gage, which measures we would rate as 5%, and from 7 to
Schaeffler's work and published a the magnetic attraction between the 19% on samples which we would rate
modified Constitution Diagram For weld metal and a permanent magnet. as 10%. The variations were trace-
Stainless Weld Metal (Fig. 2*) which Use of a properly calibrated Magne- able to two primary problems. The
added an austenitizing factor for Gage has been recommended by first was that no scientifically accu-
nitrogen and modified the location of DeLong3 and others 456 and the instru- rate method of precisely establishing
the lines for the more highly alloyed ment is widely used for measuring the quantity of ferrite present in
grades such as 309, 316 and 317 at ferrite in weld deposits. The DeLong various unknowns had been pro-
normal nitrogen levels. The modifi- Diagram was based not upon metal- posed and broadly accepted.10 The
cation increased the amount of fer- lographic examination of welds, as second was that too many in-
rite predicted over that predicted by was the Schaeffler Diagram, but upon dependently calibrated magnetic in-
the Schaeffler Diagram, and sub- magnetic determinations of ferrite struments and/or curves to be used
stantially improved the agreement be- content of over 600 weld metal pads with such instruments had evolved
tween the predicted and measured by use of a Magne-Gage calibrated as over the years, often with appreci-
values for covered electrode de- described in DeLong's 1956 paper.3 able conflict with one another in the
posits of the more highly alloyed There has in the past been signi- ferrite values reported.1-7-89
grades. The DeLong Diagram had im- ficant controversy in welding circles Standardization of magnetic
proved accuracy, stated to be ±2% as to whether what welding people measuring instrument calibration in
ferrite, especially for deposits con- call 5% ferrite is that amount or close the U.S. has recently been accom-
taining higher than normal nitrogen. to it or is significantly more or less. plished by the High Alloys Com-
Of course to use either the Schaef- Simpkinson5 has shown that different mittee of the Welding Research Coun-
fler or the DeLong Diagram the particle sizes of iron in bakelite, or dif- cil through the Advisory Sub-
chemistry of the weld deposit must be ferent orientations of iron flakes in committee on Welding Stainless
bakelite, produce substantially dif- Steels.10 This WRC calibration pro-
ferent calibration curves with a cedure includes the use of the term
Magne-Gage and its number 3 Ferrite Number (FN) instead of the
'Figure 2 as presented has been revised to magnet. Many publications78-9 show- term percent ferrite. For a given
show the WRC Ferrite Numbers and also ed that different magnetic measuring
to incorporate the revisions discussed specimen, the Ferrite Number, par-
later in this paper.
instruments or different calibrations ticularly if it is 10 or less, represents a
by William T. DeLong
CONSTITUTION DIAGRAM FOR STAINLESS STEEL WELD METAL
revised January, 1973
18 19 20 21 22 23 24
CHROMIUM EQUIVALENT = %Cr + %Mo + 1.5 X %Si + 0.5 X %Cb
284-s I J U L Y 1 973
Table 2 — Ferrite and Chemistry Variations from Electrode to Weld Deposit Spray Transfer GMAW — 1/16 In. ER 309
Electrode: 99% Argon — 1 % Oxygen Shielding Gas
Table 3 — Ferrite and Chemistry Variations from Electrode to Weld Deposit — Spray Transfer G MAW — .045 and 1/16 in. ER 308L
Electrode, Pure Argon Shielding Gas
Ferrite number,
286-s I J U L Y 1973
Table 6 - - GTAW Rod — Chemical Summary — 71 Heats
No. of
Type heats C Mn Si Cr Ni Mo N Cb
ER308 27
Range .027/.069 1.6/2.06 .35/.61 20.10/21.44 9.40/10.25 .05/.34 .020/.07 —
Avg. .046 1.79 .47 20.68 9.86 .24 .048 —
ER347 2
Range .057/.059 1.58/1.68 .52A55 19.28/19.31 9.50/9.70 .14/.20 .047/.056 .82A88
Avg. .058 1.63 .54 19.30 9.60 .17 .051 .85
ER308L 13
(a)
Range .011/.024 1.61/2.02 .34/.57 19.75/21.30 9.60/10.22 .04/.31 .019/.076 —
Avg. .018 1.80 .46 20.48 9.96 .18 .042 —
ER309 15
Range .039/.08 1.56/1.93 .39A63 23.70/24.80 12.80/13.80 .04/.26 .033/.064 82(b)
No. of
Type heats c Mn Si Cr Ni Mo N Cb
ER308 10
Range .040/.066 1.60/1.94 .37A54 20.10/21.44 9.40/TO.25 .05/.34 .037/.065 —
Avg " .046 1.78 .44 20.71 9.81 .19 .052 —
ER347 2
Range .041/.057 1.66/1.68 .52/.64 19.28/19.50 9.50/9.73 .14/.23 .047/.064 .82/.91
Avg .049 1.67 .58 19.39 9.62 .19 .056 .87
ER308L 6
Range .015/.024 1.54/1.94 .44/.57 ( a ) 19.75/20.80 9.80/10.15 .02/.31 .024/.06 —
Avg .019 1.78 .50 20.43 10.04 .16 .040 —
ER309 7
Range .039/.08 1.64/1.93 .39/.63 23.80/24.75 12.90/13.60 .06/.26 .05/.064 —
Avg .061 1.77 .48 24.18 13.26 .16 .057 —
ER316 1
Range — — — — — — — —
Avg. .040 1.67 .55 18.95 12.94 2.24 .031 —
ER316L 2
Range .019/.029 1.17/1.81 .48/.50 19.10/19.23 12.40/13.00 2.21/2.31 .028/.031 —
Avg .024 1.49 .49 19.17 12.70 2.26 .030 —
(a) Excluding one at 0.S
W E L D I N G RESEARCH S U P P L E M E N T ! 287-s
ferred across the arc, which mini- Table 9 — Ferrite Data from 53 Type 308 GTAW Pads
mizes exposure of the metal, gen-
erally provides the least nitrogen
Rod Change, Change,
pickup of the gas-shielded processes
diam, Cr DeL. Sch. DeL. Sch. M-Gl" DeL. F N t o Sch. F N t o
and hence the least variations in weld Heat in. Eq (a) Ni Eq ( b l Ni Eq |c| FN (d) FN (e) FN M-G FN M-GFN
metal ferrite content. A l t h o u g h it
would be expected that GMAW with 21.44 14.66 12.35 7.5 8.. 10.0(9> + 2.5 (9> + 1.2<9)
short circuiting transfer might show .030 9.7 + 2.4 + 1.1
.045 12.0 + 4.5 + 3.2
somewhat greater gas turbulence
.045 7.4 -0.1 -1.4
than spray transfer GMAW, it is also + 2.9 + 1.6
1/16 10.4
true that short circuiting arc transfer 3/32 9.8 + 2,3 + 1.0
does not involve the fine particles of 3/32 10.4 + 2.9 + .1.6
molten metal which cross the arc gap 1/8 9.9 + 2.4 + 1.1
in spray transfer, so the overall 5/32 10.1 + 2.6 + 1.3
nitrogen pickup could be either less 21.67 13.67 11.57 11.9 11.7 11.7 , 9 ) - 0 . 2 (3) 0.0 (9»
or greater than with spray transfer; .045 10.2 -1.7 -1.5
our data are inconclusive on this 3/32 10.8 -1.1 -0.9
point. 3/32 13.6 + 1.7 + 1.9
1/8 12.0 + 0.1 + 0.3
Gun or Torch Design. Here again 3 — 21.21 13.30 11.14 11.0 11.7 11.4 <9) + 0.4'9) + 0.7' 9 >
the effect is related to gas turbulence. 1/16 11.5 + 0.5 -0.2
Table 4 summarizes the results o b - 3/32 10.4 -0.6 -1.3
1/8 12.4 + 1.4 + 0.7
t a i n e d in G M A W tests w i t h f o u r _0.2(9)
21.29 13.80 11.70 9.6 10.1 9.9 l 9 ) + 0.3 l 9 )
different gun designs f r o m three
1/16 11.5 + 1.9 + 1.4
manufacturers; the tests covered both
1/8 9.5 -0.1 -0.6
spray and short-circuiting transfer, 1/8 8.6 -1.0 -1.5
using 99% A r - 1 % 0 2 and 90% He- 5 — 20.96 14.13 11.76 7.1 9.3 9.1(9) + 2.0 ( s ) -0.2'9>
7.5% Ar-2.5% C 0 2 shielding gases re- 1/16 9.1 + 2.0 -0.2
spectively, and both 0.045 and 1/16 3/32 9.1 + 2.0 -0.2
in. electrode diameters. Also s u m - 1/8 9.0 + 1.9 -0.3
21.01 13.92 11.52 8.0 10.0 8.2(9) (g> -1.8(9)
marized are the results of GTAW tests + 0 2
288-s I J U L Y 1 9 7 3
prepared with covered electrodes
1/16 to V* in. in diam and of three dif-
ferent coverings. Of these, the 308,
Change, Change,
diam,
308L, 309 and 347 pads averaged
Cr DeL. Sch. DeL. Sch. M-G DeL. F N t o Sch. FN to
Heat in. Eq Ni Eq Ni Eq FN FN FN M-GFN M-GFN about 6% ferrite on the old scale, and
the many 316, 316L, 317 and 317L
12 12.0*9) _0.6<9) -0.9<9>
pads ranged from 0% up to 6%. Since
3/32 22.08 13.98 11.52 12.6 12.9 11.3 -1.3 -1.6
little data were available at ferrite
3/32 12.6 0 -0.3
11.2(9) -3.6(9) -0.3<9)
levels higher than 8%, the original
13 —
3/32 21.62 12.73 11.53 14.8 11.5 12.1 -2.7 +0.6 ferrite lines at 8% and higher were
1/8 10.2 -4.6 -1.3 primarily extensions of the pattern
14 — 9.0 (8) _2.7<9> -2.2(9) shown by the data in the 0 to 8%
.045 21.77 13.84 11.74 11.7 11,2 7.7 -4.0 -3.5 range. In spite of this, our measured
1/16 10.3 -1.4 -0.9 ferrite did correlate reasonably well
15 — with the limited data which we had
1/16 22.34 13.84 12.13 14,1 11,6 13.0 -1.1 + 1.4 available for higher ferrite levels, and
16 — did support the diagram lines from
1/16 21.46 13.91 11.90 10.0 10.0 10.2 +0.2 +0.2
8% through 14% ferrite.
17 —
3/32 21.77 13.43 11.91 13.1 10.7 9.7 -3.2 -1.0 However, the recent WRC calibra-
18 — tion has increased the assigned
3/32 21.88 13.79 11.60 12.4 12.1 12.4 0 + 0.3 values for ferrite content as pre-
19 — viously described. In addition, it
1/16 21.30 14.67 11.97 7.0 9.5 9.3 + 2.3 -0.2
seemed desirable to utilize the sub-
20 — stantial body of quality assurance
.045 21.32 14.36 11.87 7.9 9.8 7.7 -0.2 -2.1
21
data which we had collected on
— GTAW and GMAW deposits
1/16 21.59 14.85 12.69 7.5 8.4 9.3 + 1.8 +0.9
22 (presented in the next portion of this
—
3/32 21.75 13.17 11.31 14.0 12.7 14.3 + 0.3 + 1.6 paper) to further improve the diagram
23 — in the high ferrite areas. This data
1/16 21.67 13.11 11.61 13.7 11.4 11.6 -2.1 +0.2 substantially outnumbered the prior
24 — data from covered electrodes in the
1/8 21.77 12.44 11.12 16.3 13.3 14.2 -2.1 + 0.9 higher ferrite area. The extent of the
25 — revisions can be s u m m a r i z e d as
1/18 21.88 13.60 12.25 13.0 10.1 13.2 +0.2 + 3.1
follows: (1) no change was made in
26 — the 0% (i.e., 0 FN) line, (2) slight
.035 22.18 13.93 11.98 13.1 11.8 8.9 -4.2 -2.9
27 changes were made in the 4 to 6 FN
— area to make a smooth transition to
3/32 21.23 13.90 11.68 9.0 10.1 9.8 +0.8 -0.3
the reduced spacing between lines
above the old 6% level and (3) at
(a) Cr Eq: Chromium equivalent calculated from filler metal chemistry (= % Cr + % Mo + 1.5 X % Si + 0.5 X % Cb); Ferrite Numbers above 8, where the
Same for both DeLong and Schaeffler Diagrams
(b) DeL. Ni Eq: DeLong nickel equivalent calculated from filler metal chemistry corrected per Table 8 ( = % Ni + 30 X %
original covered weld metal data were
C + 0.5 X % Mn•+ 30 X % N) sparse, the m a t h e m a t i c a l l y c o n -
(c) Sch. Ni Eq: Schaeffler nickel equivalent calculated from filler metal chemistry corrected per Table 8 (+ % Ni + 30 X % verted lines were shifted somewhat
C + 0.5 x % Mn)
(d) DeL. FN: Ferrite Number calculated using DeLong constitution diagram toward the 0 FN line so that the new
(e) Sch. FN: % ferrite calculated using Schaeffler constitution diagram (taken as equal to FN) diagram better reflects the ferrite
(f) M-G FN: Ferrite Number measured on pad using Magne-Gage calibrated per WRC procedure
(g) Average for heat when more than one test per heat.
levels found in GTAW and GMAW
deposits.
Table 10 — Ferrite Data from 3 Type 347 GTAW Pads Inherent Differences Between the
(For Explanation of Column Headings See Footnotes Table 9) DeLong and Schaeffler Diagrams
290-s I J U L Y 1973
Diagrams, we assumed a constant
nitrogen pickup from the filler metal Table 12 — Ferrite Data from 23 Type 309 GTAW Pads
nitrogen of 0.020% for GTAW cal- (For Explanation of Column Headings See Footnotes Table 9)
culations and 0.040% for GMAW cal-
culations (also listed in Table 8 for Rod Change, Change,
reference). These assumptions were diam, Cr DeL. Sch. DeL. Sch. M-G DeL. FN to Sch. F N t o
based on the data discussed earlier in Heat in. Eq NiEq NiEq FN FN FN M-GFN M-GFN
this paper. 43 24.64 18.45 16.17 8.9 7.0 6.8 ( a ) -2.1 , a > -0.2(a)
Tables 9 through 19 list data on the 1/16 7.1 -1.8 +0.1
129 GTAW and 32 GMAW pads in- 3/32 7.1 -1.8 +0.1
cluded in the study; these data are 3/32 6.7 -2.2 -0.3
summarized by type and process in 1/8 6.2 -2.7 -0.8
44 25.01 17.65 15.28 12.8 9.7 9.7 <a) -3.1(a) 0.0 ( a )
Tables 20 (GTAW) and 21 (GMAW).
.045 8.2 -4.6 -1.5
1/8 11.3 -1.5 + 1.6
A n Overall GTAW and GMAW 45 24.70 18.41 16.13 9.2 7.2 8.8 ( a ) -0.4<a> + 1.6 (a)
Review .045 8.8 -0.4 + 1.6
3/32 8.8 -0.4 + 1.6
Variations From Test To Test On A Single 46 25.61 17.23 15.13 16.4 11.0 14.3<a) -2.1(a) +3.3(a)
.035 13.0 -3.4 +2.0
Heat
1/8 15.6 -0.8 +4.6
47 24.96 17.47 15.37 13.3 9.4 12.0<a) -1.3(a» + 2.6 ( a )
The Magne-Gage data from Tables
1/16 10.4 -2.9 + 1.0
9 through 19 have been summarized
1/16 13.7 +0.4 +4.3
and analyzed in Table 22 to show 48 24.94 17.96 15.86 11.7 8.3 10.8 (a) -0.9(a) + 2.5 (a)
statistical data by types and in total for 1/16 13.2 + 1.5 +4.9
all heats on which multiple tests were 3/32 8.5 -3.2 +0.2
run. Each test variance shown was 49 24.86 18.10 15.58 11.0 8.7
calculated by summing the squares of 1/8 10.8 -0.2 +2.1
the differences between individual 50 25.61 18.55 16.15 12.5 9.0
test values and the average test value 3/32 12.6 +0.1 +3.6
for the heat used in each test and 51 24.67 17.58 15.18 11.9 9.2
dividing the sum so obtained by the 1/8 13.2 + 1.3 +4.0
52 25.12 16.94 14.51 15.5 11.6
figure (number of tests minus number
1/16 13.9 -1.4 + 2.3
of heats involved); this divisor c o m - 53 9.4
25.08 17.83 15.43 12.5
bines the usual (n-1) small sample 1/8 13.5 + 1.0 +4.1
divisors for the data from individual 54 24.95 16.70 15.11 15.5 9.8
heats and is the commonly accepted 1/8 9.8 -5.7 0
method of giving an overall variance 55 25.59 16.16 14.57 19.9 12.7
for several small groups of data. Tak- 3/32 18.6 -1.3 +5.9
ing t h e s q u a r e root, of t h e test 56 25.09 17.78 15.38 12.7 9.6
variance furnished the standard 3/32 13.0 + 0.3 + 3.4
57 24.83 16.84 14.38 14.5 11.3
deviation in each case.
1/16 16.3 + 1.8 +5.0
The data in Table 22, being based
on Magne-Gage measurements, are
(a) Average for heat when more than one test per heat
presented in terms of Ferrite Number.
As discussed earlier, the FN can be
r e g a r d e d as e q u i v a l e n t t o t h e
Schaeffler percent ferrite. GTAW Table 13 — Ferrite Data from 14 Type 316 GTAW Pads
data from the several types are ar-
(For Explanation o)' Column Headings See Footnotes Table 9)
ranged in order from the type show-
ing the lowest standard deviation to Rod Change, Change,
diam, Cr DeL. Sch. DeL. Sch. M-G DeL. F N t o Sch. FN to
that showing the highest. The table
Heat in. Eq Ni Eq NiEq FN FN FN M-GFN M-GFN
shows what appears to be a signifi-
cant trend, i.e., that the standard 58 22.29 16.13 14.63 6.3 5.3 5.2 (a) -1.1(a) -0.1(a)
deviation increases appreciably as 1/8 5.7 -0.6 +0.4
the average ferrite content increases. 1/8 4.1 -2.2 -1.2
The combined 316 and 316L tests 1/8 5.8 -0.5 +0.5
show an average ferrite content 5.6 59 22.14 15.96 14.43 6.3 5.4 5.6<a> -0.7(a) +0.2(a)
FN and an average standard devia- 3/32 4.8 -1.5 -0.6
tion of about 0.8 FN. The combined 3/32 6.5 + 0.2 + 1.1
308 and 308L tests average about 60 21.80 16.62 14.79 3.3 3.3 5.4'a» + 2.1(a) +2.1<a>
3/32 4.8 + 1.5 + 1.5
10.6 FN and show an average stan-
1/8 6.1 +2.8 +2.8
dard deviation of about 1.2 FN. Tests 61 22.02 16.18 14.65 5.4 4.5 6 7 (a) + 1.3 (a) + 2.2 ( a )
of the more highly alloyed 309 weld 1/16 6.6 + 1.2 +2.1
metal with about the same ferrite con- 3/32 6.8 + 1.4 +2.3
tent as the 308 show a standard 62 21.19 16.19 14.60 2.4 1.8 2.2(a> -0.2 ( a > +0.4(al
deviation of about 1.8 FN. .035 2.2 -0.2 +0.4
Although the figure for 309 may be .045 2.2 -0.2 +0.4
correct and due to that type's higher 63 21.78 16.42 14.32 3.8 4.7
.045 5.9 + 2.1 + 1.2
alloy content, it is based on only 14
64 21.59 15.81 14.34 4.6 4.0
tests and with more tests the standard
1/16 4.4 -0.2 +0.4
deviation for 309 might come down 65 21.50 16.71 14.91 2.1 1.9
toward that of the 308 family. The 3/32 4.8 + 2.7 + 2.9
overall GTAW g r o u p of 94 tests on 36
heats shows an average standard (a) Average for heat when more than one test per heat
292-s I J U L Y 19 73
Long Diagram corrected not only for
core wire nitrogen contents, which Table 16 — Ferrite Data from 6 Type 308L GMAW Pads
differed appreciably, but also for the (For Explanation of Column Headings See Footnotes Table 9)
varying nitrogen pickup f r o m differ-
ences in electrode size and covering Electr. Change, Change,
types, ranging from an average 0 . 0 1 % diam, Cr DeL. Sch. DeL. Sch, M-G DeL. FN to Sch. FN to
N pickup on c o m m o n sizes of dc lime Heat in. Eq NiEq NiEq FN FN FN M-GFN M-GFN
type electrodes to 0.05% on some
sizes of special titania types. This 35 21.34 14.08 11.59 8.9 10.6
1/16 8.6 -0.3 -2.0
s p r e a d of 0 . 0 4 % in t h e t y p i c a l
36 20.92 14.19 11.31 6.8 10.2
nitrogen pickups in itself represents a 1/8 7.1 + 0.3 -3.1
change in nickel equivalent of about 42 22.36 14.49 11.49 12.1 14.0
1.2 and a corresponding change in FN .045 9.8 -2.3 -4.2
of approximately 4.0. Appropriate 76 21.52 13.62 11.58 11.1 11.0
corrections for the varying nitrogen .045 9.1 -2.0 -1.9
p i c k u p d e p e n d i n g on e l e c t r o d e 77 21.04 13.38 11.46 10.0 10.2
covering and size plus further correc- 1/16 10.8 +0.8 +0.6
tions in the 316, 316L and 309 fami- 78 21.33 13.52 11.48 10.7 10.9
lies as described in reference 3 logi- .045 10.0 -0.7 -0.9
cally explain the better results o b -
tained with the DeLong Diagram on
covered electrodes.
Table 17 — Ferrite Data from 2 Type 347 GMAW Pads
However, the significant improve-
ment found for covered electrodes (For Explanation of Column Headings See Footnotes Table 9)
with the DeLong Diagram does not Electr. Change, Change,
carry over to the GTAW summary bar diam, Cr DeL. Sch. DeL. Sch. M-G DeL. F N t o Sch. F N t o
g r a p h s . T h e r e , the D e L o n g a n d Heat in. Eq NiEq NiEq FN FN FN M-GFN M-GFN
Schaeffler results are essentially
equivalent in total. On the DeLong 29 20.61 14.57 11.96 4.6 7.9
Diagram approximately 94% of the .045 6.1 + 1.5 -1.8
75 21.15 14.82 11.70 5.7 9.8
calculated values are w i t h i n - i 3 FN of
.045 4.1 -1.6 -5.7
the measured Ferrite Numbers, and
on the Schaeffler Diagram about 9 1 %
are within this range. These figures
are comparable to those cited above
for the DeLong covered electrode Table 18 — Ferrite Data from 8 Type 309 GMAW Pads
summary bar graph, where a ± 3 FN
(For Explanation ->t Column Headings See Footnotes Table 9)
range included about 92% of the
values, and are appreciably better Electr. Change, Change,
diam, Cr DeL. Sch. DeL. Sch. M-G DeL. FN to Sch. FN to
than the Schaeffler covered electrode
Heat in. Eq NiEq Ni Eq FN FN FN M-GFN M-G FN
summary, which showed only 78%
within ± 3 FN. 46 25.61 18.16 15.46 13.7 10.4 11.3<a> - 2 . 4 (a» +0.9(a)
.035 12.4 -1.3 + 2.0
With regard to individual types
10.2 -3.5 -0.2
welded with the GTAW process, the 45 24.70 19.34 16.46 6.5 6.6
Schaeffler Diagram understates the .045 8.5 +2.0 + 1.9
316, 316L and 309 families as it did for 47 24.96 18.40 15.70 10.3 8.7
covered electrodes 3 . The DeLong Dia- 1/16 10.7 +0.4 +2.0
gram has a better balance on these 49 24.86' 19.03 15.91 8.1 8.0
three more highly alloyed types. The 1/16 9.1 + 1.0 + 1.1
slope of the ferrite lines is greater on 50 25.61 19.48 16.48 9.8 8.3
the DeLong Diagram than on the 1/16 7.8 -2.0 -0.5
51 24.67 18.51 15.51 8.8 8.4
Schaeffler, so that the DeLong Dia-
1/8 8.6 -0.2 + 0.2
gram predicts higher ferrite values 57 24.83 17.77 14.71 11.9 10.4
than Schaeffler for types 316 and 1/16 13.0 + 1.1 +2.6
316L at normal nitrogen levels; this
difference between the two predic- (a) Average for heat when more than one test per heat
tions is even more pronounced for
type 309 weld metal. With 308 and
308L weld metal (probably also 347)
the DeLong and Schaeffler Diagrams Table 19 — Ferrite Data from 1 Type 316 and 3 Type 316L GMAW Pads
produce essentially equivalent (For Explanation of Column Headings See Footnotes Table 9)
results. Where the balance is a little Electr. Change, Change,
low as on the 308L, it is low on both diam, Cr DeL. Sch. DeL. Sch. M-G DeL. FN to Sch. FN to
diagrams. The ER308 and ER308L Heat in. Eq NiEq NiEq FN FN FN M-GFN M-GFN
rods average about 0.045% nitrogen
and an 0.020% pickup will yield a 61 22.02 17.05 14.92 3.0 '3.7
deposit average of 0.065% nitrogen, (316) .035 4.3 + 1.3 +0.6
r e a s o n a b l y c l o s e to t h e 0 . 0 6 % 69 — 22.06 15.89 14.48 6.2 5.1 e.i<a> + 2 . 1 <a) + 1.0(a>
(316L) .035 5.4 -0.8 +0.3
nitrogen content which we believe
5/64 6.7 +0.5 + 1.6
was a s s u m e d by S c h a e f f l e r for 67 17.11
covered electrode deposits. Thus the — 22.26 13.76 3.7 7.2 _ —
(316L) 1/16 3.7 0 -3.5
Schaeffler Diagram would be ex-
pected to give reasonable results for (a) Average for heat when more than one test per heat
W E L D I N G R E S E A R C H S U P P LE M E N T I 2 9 3 - s
nitrogen pickup reduces the meas-
Table 20 — Summary of GTAW Ferrite Data ured ferrite significantly and brings it
down close to the general range pre-
dicted by the diagram.
(For Explanation of Column Headings See Footnotes Table 9)
294-s I J U L Y 1973
values used to predict the ferrite from
a diagram. Analysis for C, M n , Si, Cr,
Table 21 — Summary of GMAW Ferrite Data
Ni, Mo and N is obviously subject to
error which depends upon the skill (For Explanation of Column Headings See Footnotes Table 9)
and experience of the laboratory and
the care exercised in analysis. A Change, Change,
previous paper 3 discussed this point DeL. Sch. DeL. Sch. M-G DeL. FN Sch.FN
and made assumptions which led to a Item CrEq NiEq NiEq FN FN FN to M - G F N to M-G FN
calculated s t a n d a r d d e v i a t i o n of
about 0.55 FN based upon high quali-
ER308 12 tests
ty chemistry by an experienced lab-
Range 21.01 14.13 11.47 4.8 8.0 2.1 -3.0 -6.8
oratory. to
to to to to to to to
This figure seems far too low in 22.34 15.59 12.68 11.0 10.9 11.0 + 1.7 +0.3
view of a much more recent study 11 Avg. 21.51 14.66 11.91 7.8 10.1 7.3 -0.5 -2.8
based on comparing steel mill ladle
analyses on types 308 and 308L with ER347 2 tests
Range 20.61 14.57 11.70 4.6 7.9 4.1 -1.6 -5.7
electrode producers' receiving anal-
to to to to to to to to
yses. Statistical analysis of this data 11 14.82
21.15 11.96 5.7 9.8 6.1 + 1.5 -1.8
showed a standard deviation of 0.164 Avg. 20.88 14.70 11.83 5.2 8.9 5,1 -0.1 -3.8
on the C h r o m i u m Equivalent and
0.165 on t h e N i c k e l E q u i v a l e n t ; ER308L . 6 tests
together these deviations produce a Range 20.92 13.38 11.31 6.8 10.2 7.1 -2.3 -4.2
calculated standard deviation of 0.90 to to to to to to to to
FN on the revised DeLong Diagram. 22.36 14.49 11.59 12.1 14.0 10.8 + 0.8 +0.6
The heats involved averaged around Avg. 21.42 13.88 11.49 9.9 11.2 9.2 -0.7 -1.9
10.5 FN. This later work did not incor-
ER309 8 tests
porate variations in nitrogen analysis
Range 24.67 17.77 14.71 6.5 6.6 7.8 -3.5 -0.5
but probably overstated the variations to to to to to to to to
due to carbon, since the mill analysis 25.61 19.48 16.48 13.7 10.4 13.0 + 2.0 +2.6
was often a ladle and not a check Avg. 25.11 18.61 15.71 10.4 8.9 10.0 ^0.3 + 1.1
analysis, which tends to be higher
than a ladle. The data from the receiv- ER316, 316L 4 tests
ing laboratories involved the average Range 22.02 15.89 •13.76 3.0 3.7 3.7 -0.8 -3.5
of multiple tests on a heat, and the to to to to to to to to
mill ladle analysis was usually also the 22.26 17.11 14.92 6.2 7.2 6.7 + 1.3 + 1.6
average of multiple tests. Avg. 22.10 16.49 14.41 4.8 5.3 5.0 +0.3 -0.3
riTti,^ ™LM
308L v a r i a t i o n resulting f r o m Items 1
o- o , Vrtjh , .rrfh.n, , through 4 above. Thus it is logical to
-5 0 + 5 - 5 0 +5 -5 0 + 5 - 5 0 +5
expect that the overall s t a n d a r d
296-s I J U L Y 1973
GTAW up to 0.06% electrode car- between measured and cal- Single (avg) Est. 2 sigma (95%) tol-
bon. culated ferrite contents when using measured pad erance where the avg
3. Variations in FN measurements on either diagram are greater at the FN compared ferrite content is:
multiple pads of the same filler higher ferrite levels found in GTAW to:
and G M A W deposits than when 0-8 FN 8-14 FN
metal heat.
a. The s t a n d a r d deviation for using the DeLong Diagram at the
Avg measured (see above 3a & 6)
m e a s u r e d F e r r i t e N u m b e r in lower ferrite levels of the covered
FNof5or
GTAW increases as the ferrite con- electrode deposits studied in the more pads of ±1.6 FN ±2.4 FN
tent increases over t h e range original work leading to that dia- the batch
studied. At a measured 6 FN the gram. However, the differences Calculated FN (see above 4, 5,7&
standard deviation is about 0.8 FN are slightly lower than those e n - from good de- 8 but prime data
(Types 316 and 316L) and this in- countered with covered electrode posit chemis- source is Fig. 4)
creases to a b o u t 1.2 FN at a deposits using the Schaeffler Dia- try and dia-
measured 11 FN (Type 308 and gram. With GTAW and GMAW de- gram' 3 '
308L). An average standard devia- posits a spread of up to about Schaeffler
± 3.5 FN or % ferrite can be ex- covered ± 5.0 FN ± 5.5 FN
tion of 1.8 FN was found for Type
pected between measured and bare ± 3.3 FN ± 4.0 FN
309 at about 11 measured FN, but DeLong
this deserves more study. calculated values, versus about
covered ± 3.3 FN ± 4.0 FN
± 4.5 FN or % ferrite for covered bare ± 2.7 FN ± 3.3 FN
b. Based on much less data than electrode deposits using the
were studied for GTAW, the stan- Schaeffler Diagram, (a) The estimated tolerances would
dard deviations for GMAW tests e. In types with inherently lower probably be lower if the average of multi-
were close to those of the GTAW ferrite content, such as 316 and ple measured FN determinations were
tests and the trends were similar. 316L, the range of differences be- used, and would be higher if poorer
c. To provide for test-to-test tween measured and calculated chemistry values were used.
variations in measured GTAW and ferrite seems lower than in the
GMAW ferrite, receiving accep- types with inherently higher ferrite References
tance tests should allow for vari- 1. DeLong, W. T., "Discussion of the In-
levels, such as 308, 308L and 309. ternational Testing Program for the Deter-
ations between the supplier's'tests
mination of Ferrite Content in Austenitic
and the receiving tests, e.g., if the 6. Direct m e a s u r e m e n t of ferrite,
Stainless Weld Metal", IIW Document ll-C-
supplier is required to meet 5 FN although it requires some allow- 331-70.
minimum the customer's re- ance for test variations as de- 2. Schaeffler, A. L., "Constitution Dia-
ceiving test requirement should be scribed in conclusion 3, offers the gram for Stainless Steel Weld Metal",
set at 3.5 FN m i n i m u m . Similarly, if most d e p e n d a b l e a p p r o a c h to Metal Progress, 56, 680 and 680-B (No-
a maximum of 10 or 15 FN is re- determining the ferrite content of vember 1949).
q u i r e d of the supplier the re- weld deposits. 3. DeLong, W., Ostrom, G., and Szu-
ceiving limit should be set at 12.5 7. The variables involved in calcula- machowski, E., "Measurement and Calcu-
or 17.5 FN respectively. These tion seem greater in total than the lation of Ferrite in Stainless Steel Weld
figures represent approximate two Metal", The Welding Journal, 35 (11),
variables involved in measure-
Research Suppl., 526-s to 533-s (1956).
sigma tolerance levels based on ment. 4. Simpkinson, T. V. and Lavigne, M. J.,
the data studied. 8. The reconciling of both calculated "Detection of Ferrite by Its Magnetism,"
4. Uncertainties in the determination and measured ferrite content Metal Progress, 55, 164-167 (February
of d e p o s i t c h e m i s t r y are s i g - within one frameowrk — e.g., 1949).
nificant variables in the calcula- where a specification requires a 5. Simpkinson, T. V., "Ferrite in Auste-
nitic Steels Estimated Accurately," Iron
t i o n of f e r r i t e f r o m any c o n - given ferrite range to be met both
Age, 170, 166-169 (Dec. 11, 1952).
stitution diagram. From various measured and calculated — re- 6. Fleischmann, W. L., "Determination
sources it is concluded that two quires the acceptance of rather of Ferrite in Type 347 Stainless Steel
sigma levels of about 1.6 FN at cal- broad tolerances because stan- Weld Deposit", The Welding Journal,
culated ferrite of 6 FN and less and dard deviations, both known and 33 (9), Research Suppl., 459-s to 468-s
about 2.0 FN at 10 to 15 FN cal- u n d e f i n e d , of both systems of (1954).
culated are reasonable values for rating the ferrite content must be 7. Gunia, R. B. and Ratz, G. A., "The
this variable if excellent chemistry a c c o m m o d a t e d if the r e q u i r e - Measurement of Delta Ferrite in Austenit-
is available. For average to poor ments are to be met on a practical ic Stainless Steels", WRC Bulletin 132,
c h e m i s t r y c h e c k s these values and workable basis. The data show August, 1968.
that the total tolerance in such 8. DeLong, W. T., "Measurement and
should be increased appreciably.
Effects of Ferrite in Stainless Steel Weld
5. Comparison of the Schaeffler and case may have to be as high as Metal, Subcommittee Meeting of Nov. 10
DeLong Diagrams for inert gas ± 3.4 to 4.0 FN for GTAW and in New York" (Advisory Subcommittee of
shielded deposits. GMAW deposits averaging around the High Alloys Committee of the Welding
a. The diagrams are essentially 11 FN. Research Council), IIW Doc. II-C-372-71.
e q u i v a l e n t in the d i f f e r e n c e s 9. Conclusion 8 must not be inter- 9. Rosendahl, C-H, "Ferrite in Austenit-
ic Stainless Steel Weld Metal; Round
between predicted and measured preted to mean that the ferrite con-
Robin Testing Programme 1971-1972",
ferrite in GTAW deposits of Types tent of a weld deposit cannot be IIW Doc. 11-631-72.
308, 308L and 347. measured or calculated to closer 10. "Calibration Procedure for Instru-
b. The Schaeffler Diagram under- limits than ± 4 FN. Pending the ments to Measure the Delta Ferrite Con-
states the ferrite content of the completion of the WRC coopera- tent of Austenitic Stainless Steel Weld
more highly alloyed Type 316, tive program on covered elec- Metal", Welding Journal 52 (2), Research
316L and 309 GTAW deposits. The trode deposits, which will also Suppl., 69-s to 72-s (1973).
DeLong Diagram is better on these cover the influence of pad size and 11. Hebble, T. L., "Statistical Treatment
grades, as is also the case with preparation, we are estimating the of Chemical Analysis Variability", Sup-
covered electrode deposits. following overall tolerances for plemental minutes of November '72 WRC
Advisory Subcommittee minutes and
c. The DeLong Diagram is better m e a s u r e m e n t / c a l c u l a t i o n . The
phone conversation of 1/16/73.
for GMAW deposits due to higher base is the average FN as meas- 12. E. A. Schoefer, "Welding of High
nitrogen pickup in those deposits. ured on a welded pad of the type Alloy Castings", Steel Founders' Society of
d. The differences to be expected defined in this paper. American, Publication 772-5M1.