You are on page 1of 17

WESLDINC RESEARCH

SUPPLEMENT TO THE WELDING JOURNAL, JULY 1973


Sponsored by the American Welding Society and the Welding Research Council

The Ferrite Content of Austenitic


Stainless Steel Weld Metal
The effect of nitrogen pickup in GTAW and
GMA W and the need for better correlation
between calculated and measured
ferrite content leads to a revised constitution
diagram and an evaluation

of experimental

and statistical error.


BY C. J. LONG AND

W. T. DeLONG

W. T. DeLONG is Vice President and


C. J. LONG is Welding Engineer,
Teledyne McKay, York, Pennsylvania.
Paper was presented at the 54th
AWS Annual Meeting held in Chicago
during April 2-6, 1973.

ABSTRACT. The possibility of


nitrogen incursion into the arc and
weld pool must be considered in any
application where austenitic stainless
weld metal ferrite content is important. This is especially true in gas
metal-arc welding (GMAW) and gas
tungsten-arc welding (GTAW) where
turbulence in the arc shielding gas
flow can admit enough nitrogen to appreciably reduce the ferrite content of
the deposit. Calculations of ferrite using filler metal chemistry and either
the DeLong or Schaeffler diagrams
will usually show a higher ferrite content than exists in the deposit, because nitrogen is picked up by the
weld metal during welding.
If the Schaeffler Diagram were the
only calculation tool available the
nitrogen-caused differences in ferrite
between filler metal and weld metal
would often be difficult to explain because Schaeffler did not include a
factor for nitrogen in his diagram.
However, use of the actual deposit
analysis and the DeLong Diagram,
which takes account of the strong
austenitizing effect of nitrogen, can
generally explain the significant
differences observed when high
nitrogen pickup has occurred.
Knowledge of the effect of nitrogen
and the magnetic measurement of
weld deposit ferrite can be of significant help in qualifying GMAW and
GTAW procedures and operators to
minimize pickup of the element and
thereby to provide consistent and acceptable weld metal ferrite.
The DeLong Diagram has been
revised to convert it to the WRC Ferrite Number system and to reflect information obtained in the study of the
GTAW and GMAW deposits described in this paper.
Comparisons between the
calculated and the measured ferrite
contents are made for both the
Schaeffler and the revised DeLong
Diagrams. On Types 308, 308L and
347 the diagrams are essentially
equal in performance except where
high nitrogen levels are involved, in
which case the DeLong Diagram is
better. For the more highly alloyed
Types 316, 316L and 309 the Schaeffler Diagram understates the ferrite
content so the DeLong Diagram is
better for these types.
The sources and magnitudes of

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

CHROMIUM EQUIVALENT = Cr+Mo + 1.5 x-g-Si + 0.5

27

28

"rjCb

Fig. 1 Schaeffler diagram


several variables in the measurement and calculation of ferrite are
considered, as well as the overall
variation between them. Allowances
for receiving inspection are recomm e n d e d to p r o v i d e
adequate
tolerances for variations in ferrite content between pads.
Measured ferrite is r e c o m m e n d e d
over calculated for specification purposes.
Background
The Literature
The effectiveness of the presence
of a small but controlled amount of
delta ferrite in preventing cracking of
austenitic stainless steel weld d e posits is well known. Types in which
this effect is taken advantage of include among others 308, 308L, 309,
347, 318 and sometimes 316 and
316L and 317 and 317L.
In cooling from the molten state
austenitic stainless weld metal of normal carbon content solidifies first as a
mixture of delta ferrite and austenite;
most of the ferrite subsequently
transforms to austenite as the de282-s I J U L Y

1 973

posit cools through a temperature


range just below the delta ferrite
region. The transformation is time-dependent i.e., the ferrite does not instantaneously transform, but does so
progressively over a short period of
time and theoretically the transformation could be avoided almost
completely if the deposit could be instantly quenched from the just-frozen
state to black heat, which would result in a much higher percentage of
ferrite in a Type 308 deposit, for example. Similarly, if the cooling rate
from the molten state is rapid enough,
it is possible to supercool through the
t e m p e r a t u r e zone in w h i c h delta
ferrite f o r m s and to transform d i rectly into austenite. Practically
speaking, however, the final amount
of delta ferrite in virtually all weld
metal depends only to a rather minor
degree upon the cooling rate.
The amount of ferrite in stainless
weld deposits can be determined in
any of several ways. Metallographic
e x a m i n a t i o n of the d e p o s i t c a n
provide an indication of the approximate ferrite content in terms of area

distribution. However, besides being


c u m b e r s o m e , such a m e t h o d requires considerable care to be sure
that the section or sections studied
give a true picture of the volumetric
ferrite distribution. Also, the etchant
selected and the degree of etch have
been shown to be among the several
variables in procedure and technique
which can influence rather s u b stantially the values obtained. 1
In 1949 Schaeffler published his
Constitution Diagram For Stainless
Steel Weld Metal, 2 which provided a
tool for calculating weld deposit ferrite c o n t e n t b a s e d on c h e m i c a l
analysis by graphically combining effects of the austenitizers nickel, carbon and manganese and the ferritizers c h r o m i u m , m o l y b d e n u m ,
silicon and columbium; this diagram
gave a figure for ferrite content which
was stated to be accurate to 4
volume percent (hereinafter referred
to as %) ferrite for many of the c o m mon austenitic weld deposits under
cooling conditions present in field
weldments. An expanded portion of
the Schaeffler Diagram is shown as

Fig. 1; this diagram has been widely


accepted and very useful because
calculated ferrite is rather easily determined.
In 1956 DeLong et al3 expanded on
Schaeffler's work and published a
modified Constitution Diagram For
Stainless Weld Metal (Fig. 2*) which
added an austenitizing factor for
nitrogen and modified the location of
the lines for the more highly alloyed
grades such as 309, 316 and 317 at
normal nitrogen levels. The modification increased the amount of ferrite predicted over that predicted by
the Schaeffler Diagram, and substantially improved the agreement between the predicted and measured
values for covered electrode deposits of the more highly alloyed
grades. The DeLong Diagram had improved accuracy, stated to be 2%
ferrite, especially for deposits containing higher than normal nitrogen.
Of course to use either the Schaeffler or the DeLong Diagram the
chemistry of the weld deposit must be

'Figure 2 as presented has been revised to


show the WRC Ferrite Numbers and also
to incorporate the revisions discussed
later in this paper.

rather accurately known. A more direct measuring technique is the use of


an appropriate magnetic measuring
instrument such as an Aminco-Brenner Magne-Gage, which measures
the magnetic attraction between the
weld metal and a permanent magnet.
Use of a properly calibrated MagneGage has been recommended by
DeLong3 and others 456 and the instrument is widely used for measuring
ferrite in weld deposits. The DeLong
Diagram was based not upon metallographic examination of welds, as
was the Schaeffler Diagram, but upon
magnetic determinations of ferrite
content of over 600 weld metal pads
by use of a Magne-Gage calibrated as
described in DeLong's 1956 paper.3
There has in the past been significant controversy in welding circles
as to whether what welding people
call 5% ferrite is that amount or close
to it or is significantly more or less.
Simpkinson5 has shown that different
particle sizes of iron in bakelite, or different orientations of iron flakes in
bakelite, produce substantially different calibration curves with a
Magne-Gage and its number 3
magnet. Many publications78-9 showed that different magnetic measuring
instruments or different calibrations

of those instruments did produce different ferrite readings on a given


specimen. These readings could
range from 3 to 9% on samples which
we would rate as 5%, and from 7 to
19% on samples which we would rate
as 10%. The variations were traceable to two primary problems. The
first was that no scientifically accurate method of precisely establishing
the quantity of ferrite present in
various unknowns had been proposed and broadly accepted.10 The
second was that too many independently calibrated magnetic instruments and/or curves to be used
with such instruments had evolved
over the years, often with appreciable conflict with one another in the
ferrite values reported.1-7-89
Standardization of magnetic
measuring instrument calibration in
the U.S. has recently been accomplished by the High Alloys Committee of the Welding Research Council through the Advisory Subcommittee on Welding Stainless
Steels.10 This WRC calibration procedure includes the use of the term
Ferrite Number (FN) instead of the
term percent ferrite. For a given
specimen, the Ferrite Number, particularly if it is 10 or less, represents a

CONSTITUTION DIAGRAM FOR STAINLESS STEEL WELD METAL

18

by William T. DeLong
revised January, 1973

19
20
21
22
23
24
CHROMIUM EQUIVALENT = %Cr + %Mo + 1.5 X %Si + 0.5 X %Cb

Fig. 2 DeLong diagram

W E L D I N G RESEARCH S U PPLE M E NT | 283-s

Table 1 Conversion Old McKay %


Ferrite (as measured with a Magne-Gage)
to WRC Ferrite Number (FN)
Old
%F

New
FN

Up to 6.0
7.0
8.0
9.0
10.0
11.0
12.0
13.0
14.0

Same as old %
7.2
8.5
9.8
11.1
12.4
13.7
15.0
16.2

g o o d average of the ferrite percentage figures which would be determined for the specimen by present
U.S. and world methods of measuring
delta ferrite, based on round robins
conducted by the WRC Advisory Subcommittee and the International Institute of Welding Subcommission IIQ 8.9.10 | n V j e w 0 f t h e WRC procedure
we have presented all experimental
data in this paper in terms of FN and
much of the discussion herein is also
in FN terms. For reference, Fig. 2 and
Table 1 both present a comparison
of the percentage figures formerly
used by our l a b o r a t o r y with the
equivalent Ferrite Number for each.
Manufacture and Use of
Covered Electrodes
Fortunately from a practical viewpoint, most major U.S. covered welding electrode producers have set up
their practices in electrode manufacture so that the differences between them have not prevented the
practical use of ferrite in the general
range of 4 to 9% to provide crack-free
weld deposits. If converted from our
prior practice, which generally corresponds to the practices of most
U.S. m a n u f a c t u r e r s , 3 t h i s is e s sentially 4 to 10 on the WRC Ferrite
Number scale. Round robin data 8
indicate that the major producers in
the U.S. who participated in the program were within approximately 0 . 5
FN of each other at a 5 Ferrite Number level and 1 FN of each other at a
10 Ferrite Number level, substantially

below both the spread found in users'


calibrations and the total world-wide
spread found in the survey. This reasonable level of agreement has been
attained in part through the interchange of weld metal specimens with
specific assigned ferrite values, and
the subsequent use of the specimens
as s t a n d a r d s for c o n s t r u c t i o n of
Magne-Gage calibration curves, and
in part through use of either the
DeLong or Schaeffler Constitution
Diagrams. It should be remembered
that despite the fact that not all producers have agreed precisely on the
percent ferrite in any given weld deposit, the agreement has been close
enough so that these widely used
tools for d e t e r m i n i n g calculated
and/or measured weld metal ferrite
have been eminently successful in
providing crack-free stainless steel
weld deposits for 20 years or more.
Many users, including the government, require in some of their purchase s p e c i f i c a t i o n s that c o v e r e d
electrodes of Types 308 and 308L and
other f e r r i t e - c o n t a i n i n g austenitic
stainless types produce undiluted
weld metal containing a particular
range of ferrite, often from 4 % to 9%
or from 5% to 15%, as determined
with either one of the abovementioned constitution diagrams or a
Magne-Gage or other specified m a g netic measuring instrument such as
a Severn Ferrite Indicator. This approach to ferrite control in covered
electrode stainless weld metal has
been successful primarily because
welding practices and techniques with
covered electrodes have over the
years become generally reasonably
uniform, so that a deposit produced
by a trained welder in the electrode
manufacturer's test lab according to
standardized procedures will be generally equivalent to one produced by a
similarly trained welder in the purchaser's facility according to the same
p r o c e d u r e s . Also, the a m o u n t of
ferrite measured in such deposits with
a Magne-Gage is reasonably close to
the amount calculated f r o m the
DeLong Diagram, for example, using
the covered electrode deposit
chemistry, because as mentioned
above the Magne-Gage calibration

FERRITE
MEASURED
IN THIS AREA
- MIN AS J
WELDEDE

-TYPE 30I, 302 OR


304 STAINLESS
PLATE.

jfMIN

-T^L

r-1 APPROX -

Fig. 3 Standard ferrite pad


284-s I J U L Y

1 973

8 WELD LAYERS
(PASSES)
MINIMUM BUILDUP-

-3"APPROX-

curves have generally been established based on the DeLong or


Schaeffler Diagrams.
Use of Bare Stainless
Electrodes and Rods
In the bare stainless area, attempts
have been made to provide deposit
ferrite control similar to that used with
covered electrodes. These attempts
have not been entirely successful,
partly because variations in GTAW
and GMAW procedures have made it
difficult to standardize in a manner
which will produce ferrite readings
agreeable to a m a j o r i t y of f a b r i cators. The result has been that few
p u r c h a s e s p e c i f i c a t i o n s for bare
stainless filler metals have in the past
included any ferrite requirement; the
apparent feeling of many users is that
when the chemical requirements are
met the deposit will be satisfactory
from the ferrite standpoint. One exception to this approach is U.S. Government Specification MIL-E-19933D,
which requires that undiluted weld
metals of the 308, 309 and 347
families contain from 4 to 9% ferrite
as measured on the top surface of a
standard pad (Fig. 3) using a calibrated Magne-Gage.
Ferrite in G T A W a n d G M A W
Deposits the Nitrogen P r o b l e m
When we first considered the MILE-19933D ferrite requirements years
ago, we used ferrite calculated from
bare electrode chemistry by the
DeLong Diagram to predict whether
the ferrite in a standard ferrite pad
would fall within the desired range; in
predicting, we assumed a minimum
change in analysis from electrode to
deposit and about 0.5 per cent loss in
ferrite, the latter assumption being
based on our feeling that cooling
rates in the low-mass GMAW standard pads were somewhat slower
than the cooling rates typical of most
production welds.
Our experience indicated that it
was difficult to produce GMAW deposits with ferrite contents within the
predicted ranges; given a heat with
10% calculated ferrite, normally skilled semiautomatic welders following
the MIL p r o c e d u r a l r e q u i r e m e n t s
could produce deposits containing
anywhere from 0 to 9% Magne-Gage
ferrite. Under such circumstances a
deposit from an ER 308L electrode
with 14% calculated ferrite, for example, might meet the MIL-E-19933D requirements when prepared by one
welder but might be outside those requirements when prepared under
nominally equivalent conditions by
another welder.
Because of the above described
difficulties in predicting and reproducing ferrite measurements, a study

Table 2 Ferrite and Chemistry Variations from Electrode to Weld Deposit


Electrode: 99% Argon 1 % Oxygen Shielding Gas
Item
1. Electrode
2. Weld deposit,
first check
3. Weld deposit,
special care

Spray Transfer GMAW 1/16 In. ER 309

Mr,

Si

Cr

Ni

Mo

.074
.063

1.63
1.52

.43
.31

24.61
24.08

13.66
13.48

.14
.12

.063
.161

.058

1.42

.38

23.81

13.48

.12

.070

Ferrite number,
calculated' 3 ' measured
11.6
2.4
0.6 W
9.1

6.0

,b)

{a) Revised DeLong Diagram (Fig. 2).


(b) Difference between measured and calculated felt at the time to be due primarily to cooling rate effects.

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,
Heat
465
465
465
055
055
055

Form
.045 electr.
Welder 1
deposit
Welder 2
deposit
1/16 electr.
Welder 1
deposit
Welder 2
deposit

Mn

calculated^) measured
14.3

C
.019

1.74

Si
.47

Cr
20.55

Ni
10.15

N
.029

.018

1.62

.54

20.60

10.10

.057

12.6

9.1

.016
.014

1.52
1.73

.52
.60

20.43
21.00

10.42
9.90

.319
.069

0
15.0

.020

1.62

.56

21.17

9.95

.112

10.5

7.8

.016

1.54

.56

21.08

10.14

.224

0.9

1.0

(a) Revised DeLong Diagram (Fig. 2)

was undertaken to determine the


effects of several variables on the
Magne-Gage ferrite of GMAW and
GTAW deposits, with primary
emphasis on changes in nitrogen
content between filler metal and deposit. Nitrogen was emphasized because we have found on covered
electrode deposits that if the ferrite of
the deposit is substantially below our
aim, say from 2 to 6 or more FN
below, the p r o b l e m is invariably
t r a c e a b l e to e x c e s s i v e n i t r o g e n
pickup in the deposit due to somewhat lax welding procedures. The
standard pad design (Fig. 3) is good
from an economics viewpoint but is
very exposed from the viewpoint of air
intrusion, and a very short arc must
be maintained.
In this study, we have used the
DeLong Constitution Diagram shown
in Fig. 2 and the Magne-Gage calibration method established by the WRC 10
as the basis for all of our ferrite
values, whether calculated from
chemistry or measured on the actual
weld deposit. As previously noted, the
ferrite contents are given in terms of
Ferrite Number (FN).
Our own direct studies were also in
part due to a customer complaint received in the early 1960's. It c o n cerned the ferrite content of semiautomatic inert GMAW spray transfer
deposits p r e p a r e d with 1/16 in.
ER309 filler metal from a heat which
showed 10 FN as calculated using the
electrode chemistry and the DeLong
Diagram. The customer, who was ex-

p e r i e n c e d and k n o w l e d g e a b l e in
stainless steel fabrication, noted that
M a g n e - G a g e c h e c k s of their d e posits showed ferrite of less than 5
FN, and in some instances less than 1
FN. Our routine handling of the c o m plaint, comprising preparation of a
standard ferrite pad with a sample of
the electrode under conditions closely
matching those used by the customer, also produced Magne-Gage
readings showing ferrite of less than 1
FN.
Based on our prior experience with
covered electrodes we would not
have been surprised at a difference of
up to 2 . 5 FN from the value calculated from bare electrode chemistry,
w h i c h w e a s s u m e d w o u l d be
equivalent to the deposit chemistry;
however, we were unable to explain a
drop from 10 FN calculated to as low
as 1 FN measured. Chemical analysis
of the undiluted weld metal from the
ferrite pad revealed that, although
average changes in other elements
were minor, the weld metal nitrogen
was more than twice that in the electrode (see Table 2, Item 2). The nitrogen level was high enough to reconcile the calculated and the measured
ferrite within the normal spread of
2.5 FN which we expected.
Because the data showed that our
welder's procedures allowed nitrogen
incursion to the arc zone, we had him
repeat the test taking special care to
hold a short arc and a short gas cupto-work distance. These precautions
decreased the nitrogen pickup to an

acceptably low level, as shown in


Table 2, Item 3; calculated ferrite on
the repeat pad was 9.1 FN and the
Magne-Gage ferrite was 6 FN. This is
acceptable from a cracking viewpoint, although not from the viewpoint of good agreement of measured
FN with the FN calculated from either
the electrode chemistry or, to a lesser
degree, the pad chemistry. Our results indicated that it was not the electrode but probably the customer's
welder's procedure which was the
problem. Thus, in spite of the customer's experience, they still had a
welder whose practices were apparently poor in GMAW of austenitic
stainless steel.
The Type 309 experience and additional more recent studies with other
types have confirmed that nitrogen
p i c k u p occurs all too readily in
GMAW and GTAW. This pickup is unfortunately not visually obvious during
or after welding because the deposit
appearance does not change significantly. The only means of detecting
nitrogen pickup are measurement of
the deposit ferrite level, analysis of
the deposit for nitrogen, or weld metal
cracking if the ferrite is too low and
the joint restrained.
It is well understood that nitrogen
pickup has a very strong influence on
weld metal ferrite content. The reason
is that nitrogen is 30 times as strong
an austenitizer as nickel, according to
the DeLong Diagram, so that pickup
of very small amounts in the weld
metal disproportionately lowers the

WELDING RESEARCH SUPPLEMENT!

285-s

Table 4 Gun or Torch Design Vs. Weld Metal Nitrogen Pickup Type 308L Using
Skilled Welder Under Excellent Laboratory Conditions
N pick up n
weld m etal %

No. of
tests
22
12
8
4
8

Gun or torch
GMAW gun A
GMAW gun B
GMAWgunC
GMAW gun D
GTAW torch E
GTAW torch F
(Gas lens)

Approx. equiv.
calculated fe r r i t e l o s s ( a )

range
.014/.069
.012/.027
.018/.026
.021/.024
.019/.034

avg.
.036
.020
.021
.022
.025

range
1.1/5.5
1.0/2.2
1.4/2.1
1.7/1.9
1.5/2.7

avg.
2.9
1.6
1.7
1.8
2.0

.012/.020

.016

1.0/1.6

1.3

(a) Assuming 0.015%N = 1.2 FN

Table 5 Changes in Calculated Ferrite Attributable to Changes of 0.2 in Chromium and


Nickel Equivalents at 21 Chromium Equivalent

Approx.
ferrite
content,
FN or %
2
6
11

Change in ferrite
content for a 0.2
change in Cr Eq
DeLongla)

Schaeffler

0.6 FN
0.8 FN
0.9 FN

0.7%
0.65%
0.6%

Change in ferrite
content for a 0.2
change in Ni Eq
D e L o n g , a ) Schaeffler
0.5 FN
0.6 FN
0.7 FN

0.65%
0.6%
0.65%

Cumul. c hange in
ferrite cc ntentfor
+ 0.2Cr Eq and - 0 . 2 Ni
Eq or vice-versa
DeLong ( a )

Schaeffler

1.1 FN
1.4 FN
1.6 FN

1.35%
1.25%
1.25%

(a) Based on revised DeLong Diagram, Fig. 2.

ferrite content; since the source of


nitrogen is air, it is evident that any
factor which decreases the effectiveness of the arc shield in GMAW or
GTAW can cause significant loss of
ferrite from that calculated using the
f i l l e r m e t a l c h e m i s t r y to t h a t
measured in the deposit. As an example, referring to Fig. 2, let it be ass u m e d t h a t e n o u g h air b r e a k s
through the protective gas shield so
that 0.08% nitrogen is picked up by
the deposit. This adds 2.4 to the
Nickel Equivalent (Ni Eq) and will
reduce the deposit ferrite level by as
much as 8 FN. Associated with this
amount of nitrogen, the deposit could
be expected to pickup about 0.02%
oxygen, which would combine with
the oxidizable elements silicon, manganese, chromium and perhaps carbon. The individual loss of any of
these would be so low that the change
in their levels in the deposit, and c o n sequently the change in the calculated ferrite level, would be negligible. Also, the small amount of additional slag or oxide so formed has
no s i g n i f i c a n t e f f e c t on t h e
appearance of the final weld. If we
assume that one fourth of the oxygen,
or 0.005% combines with each element, the losses would be as follows,
rounded off:
A. 0.005% O combines
0.005% Si to form 0 . 0 1 %
0.005% Si = 0.007 Cr Eq.
B. 0.005% O combines
0.015% Cr to form 0.02%
0.015% Cr = 0.015 Cr Eq.
286-s I J U L Y

1973

with
SiOi;
with
CrO;

C. 0.005% 0 combines
0.015% Mn to form 0.02%
.015% Mn = .007 Ni Eq.
D. 0.005.% O combines
0.004% C to form 0.009%
0.004% C = .12 Ni Eq.

with
MnO;
with
CO;

These losses in silicon, c h r o m i u m


and m a n g a n e s e w o u l d hardly be
noticeable on analysis, and their
effect on the ferrite would be quite
low. Even the assumed carbon loss
would only be equivalent to an increase of about 0.3 FN.
The solubility limit of nitrogen in
austenitic stainless steel is in excess
of 0.25% and sets a m a x i m u m on the
amount of nitrogen that can be picked
up during welding; however, an extra
0.20% nitrogen added to a 0.05% nitrogen base in the filler metal adds 6.0
to the Ni Eq on the DeLong Diagram
and represents a theoretical loss of
about 18 FN, so it can be seen that nitrogen pickup can potentially eliminate all ferrite from virtually all c o m mon austenitic deposits.
It should also be restated that exc e s s i v e n i t r o g e n p i c k u p is not
ordinarily reflected in changed bead
appearance or in the loss of oxidizable elements; one reliable and rapid
means of detecting substantial increases in nitrogen is a comparison of
the weld metal ferrite content as
m e a s u r e d using a p r o p e r l y calibrated instrument with the theoretical
ferrite content as calculated from the
filler metal chemistry with the DeLong
Diagram.

Effects of Variables on Nitrogen Pickup


Some of the most important variables affecting nitrogen pickup and
the extent of their effects are disc u s s e d in t h e f o l l o w i n g p a r a graphs.
Welder
Technique.
This is undoubtedly the major influencing variable. Poor technique, arising either
from inexperience or inattention, can
allow
aufficient
nitrogen
pickup,
either through turbulence or through
holding an excessively long gas cupto-work distance, to eliminate virtually
all ferrite from the w e l d deposit.
Table 3 illustrates the effect that
welder technique can have on deposit
ferrite. This table sets forth the
changes in nitrogen that occurred
between bare electrode and deposit
in experiments with two heats of
ER308L filler metal; semiautomatic
GMAW using spray transfer with pure
argon shielding was used to prepare
all four weld pads according to the requirements of MIL-E-19933D. Note
that except for nitrogen, the deposit
chemistry was very little changed
from the electrode chemistry. Ferrite
was measured on all pads with a
calibrated Magne-Gage and calculated from chemistry using the
DeLong Diagram.
Welder 1, who had long experience
in stainless steel GMAW, allowed far
less nitrogen pickup than Welder 2,
whose experience had been primarily
with C 0 2 - s h i e l d e d w e l d i n g using
tubular electrodes. Welder 2 did not
recognize that his t e c h n i q u e was
poor, in spite of the fact that he was a
skilled semiautomatic welder. This
emphasizes the importance of the
right kind of experience in minimizing nitrogen pickup in GMAW
stainless deposits.
In GTAW also, welder technique
can bring about loss of ferrite. For example, a heat of 1/16 in. ER309 rod
with a DeLong-calculated ferrite of 15
FN was used to prepare a standard
GTAW test pad by one of our skilled
welders who had not previously been
trained to avoid nitrogen pickup. Not
recognizing the p r o b l e m , he did not
use the proper techniques, and as a
result the pad checked at 6.7 FN by
Magne-Gage, a drop of 8.3 FN. Such
a loss is not at all typical of GTAW,
and we therefore had the welder run a
repeat pad and asked him to concentrate on maintaining optimum gas
shielding, which is our normal practice when unexpectedly low ferrite results are obtained. On the repeat pad
the Magne-Gage ferrite was 16.3 FN,
which agrees very well with the calculated 15.0 FN for the heat. We have
found that in virtually every similar
case we have e n c o u n t e r e d , t h e
Magne-Gage ferrite of GTAW deposits can be brought within the normally expected range by the use of a

Table 6 - - GTAW Rod Chemical Summary 71 Heats

Type

No. of
heats

27
ER308
Range
Avg.
ER347
2
Range
Avg.
ER308L 13
Range
Avg.
ER309
15
Range
Avg.
8
ER316
Range
Avg.
ER316L
6
Range
Avg.

Cr

Mo

Ni

Cb

Mn

Si

.027/.069
.046

1.6/2.06
1.79

.35/.61
.47

20.10/21.44
20.68

9.40/10.25
9.86

.05/.34
.24

.020/.07
.048

.057/.059
.058

1.58/1.68
1.63

.52A55
.54

19.28/19.31
19.30

9.50/9.70
9.60

.14/.20
.17

.047/.056
.051

.011/.024
.018

1.61/2.02
1.80

.34/.57
.46

19.75/21.30
20.48

9.60/10.22
9.96

.04/.31
.18

.019/.076
.042

.039/.08
.056

1.56/1.93
1.75

.39A63
.47

23.70/24.80
24.16

12.80/13.80
13.23

.04/.26
.15

.033/.064
.054

82(b)

.030/.061
.042

1.42/1.68
1.61

.31/.55
.46

18.44/19.40
18.91

12.60/13.10
12.86

2.12/2.24
2.19

.029/.05
.036

.016/.029
.020

1.17/1.81
1.58

.38/.58
.48

18.15/19.23
18.91

11.90/13.20
12.75

2.09/2.24
2.22

.025/.035
.030

(a)

.82A88
.85

(a) Excluding one at OS


(b) One heat only

Table 7 -- GMAW Electrode Chem cal Summary 28 Heats

Type
ER308
Range
Avg "
ER347
Range
Avg
ER308L
Range
Avg
ER309
Range
Avg
ER316
Range
Avg.
ER316L
Range
Avg

No. of
heats

Mn

Cr

Si

Mo

Ni

Cb

10
.040/.066
.046

1.60/1.94
1.78

.37A54
.44

20.10/21.44
20.71

9.40/TO.25
9.81

.05/.34
.19

.037/.065
.052

.041/.057
.049

1.66/1.68
1.67

.52/.64
.58

19.28/19.50
19.39

9.50/9.73
9.62

.14/.23
.19

.047/.064
.056

.82/.91
.87

.015/.024
.019

1.54/1.94
1.78

.44/.57 ( a )
.50

19.75/20.80
20.43

9.80/10.15
10.04

.02/.31
.16

.024/.06
.040

.039/.08
.061

1.64/1.93
1.77

.39/.63
.48

23.80/24.75
24.18

12.90/13.60
13.26

.06/.26
.16

.05/.064
.057

.040

1.67

.55

18.95

12.94

2.24

.019/.029
.024

1.17/1.81
1.49

.48/.50
.49

19.10/19.23
19.17

12.40/13.00
12.70

2.21/2.31
2.26

.031

.028/.031
.030

(a) Excluding one at 0.S

technique which emphasizes maximum gas shielding of the arc and


molten weld metal.
Optimum welder technique for both
GMAW and GTAW includes holding
arc length and gas cup-to-work distances as short as possible, avoiding
drafts in the welding area, and maintaining equipment, particularly gas
nozzles, clean to minimize turbulence
in the gas shield.
Transfer Mode. Insofar as the
transfer mode relates to shielding gas
turbulence, it can influence nitrogen
pickup. GTAW, because it involves
low gas flow rates and thus relatively
non-turbulent flow and because the
filler metal is melted rather than trans-

Table 8 Average Carbon Losses and C Drresponding FN Gain from Filler Metal Analysis
to Deposit Analysis

Type

No. of
heats

No. of
tests

C, avg.
heat, %

Avg.
loss,
C, %

Approx.
FN gain

GTAW
process

309
308, 347
316
308L, 316L

15
29
8
19

23
56
14
36

.059
.047
.042
.019

.018
.014
.011
.001

1.8
1.5
1.0

GMAW
process

309
308, 347
316
308L, 316L

7
12
1
8

8
14
1
9

.061
.047
.040
.020

.007
.003
.002
.000

0.7
0.3
0.2

(a) These figures were deducted from the filler metal carbon when calculating predicted ferrite content, except that no
deduction was made for 308L or 316L either GTAW or GMAW, due to the low average changes in these types. In addition, when calculating predicted ferrite using the DeLong Diagram .020% was added to the filler metal nitrogen for
GTAW pads and .040% nitrogen was added for GMAW pads. These are equivalent to approximate Ferrite Number
losses of 2.0 and 4.0 respectively.

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 minimizes exposure of the metal, generally provides the least nitrogen
pickup of the gas-shielded processes
and hence the least variations in weld
metal ferrite content. A l t h o u g h it
would be expected that GMAW with
short circuiting transfer might show
somewhat greater gas turbulence
than spray transfer GMAW, it is also
true that short circuiting arc transfer
does not involve the fine particles of
molten metal which cross the arc gap
in spray transfer, so the overall
nitrogen pickup could be either less
or greater than with spray transfer;
our data are inconclusive on this
point.
Gun or Torch Design. Here again
the effect is related to gas turbulence.
Table 4 summarizes the results o b t a i n e d in G M A W tests w i t h f o u r
different gun designs f r o m three
manufacturers; the tests covered both
spray and short-circuiting transfer,
using 99% A r - 1 % 0 2 and 90% He7.5% Ar-2.5% C 0 2 shielding gases respectively, and both 0.045 and 1/16
in. electrode diameters. Also s u m marized are the results of GTAW tests
with two c o m m e r c i a l l y available
torches, one with a gas lens and the
other without. All the tests were run by
the experienced Welder 1 of Table 3
and under c o m f o r t a b l e and c o n trolled laboratory conditions; consequently, the nitrogen increases in
these tests were all relatively low.
The G M A W tests showed very little
difference between Guns B, C and D,
all three guns providing reasonably
low weld metal nitrogen pickup; Gun
A performed somewhat less well than
the other three, possibly for the reasons that (a) it was in poorer overall
condition than the others, having
been in use in our laboratory c o n siderably longer than they, and (b) the
data shown for Gun A represent a
larger number and variety of test conditions than the data shown for the
other guns. As a general rule, it appears that when all are in equally
good (or bad) condition no significant differences will be seen
between most commercially available semiautomatic GMAW guns, except that the size and overall
geometry of the gun may affect nitrogen p i c k u p to the extent that it
prevents the welder from holding a
short gas cup-to-work distance
e.g., use of a large, heavy gun to deposit a small fillet weld in tight
quarters may result in lower deposit
ferrite than use of a smaller lightweight g u n for the s a m e j o b . In
GTAW, nitrogen pickup can be reduced by use of a torch which includes a gas lens. In the set of eight
pairs of weld deposits prepared to
compare the gas lens torch with the
ordinary torch, deposits made with
288-s I J U L Y 1 9 7 3

Table 9 Ferrite Data from 53 Type 308 GTAW Pads

Heat

Rod
diam,
in.

Change,
Cr
DeL.
Sch.
Eq (a) Ni Eq ( b l Ni Eq |c|

DeL.
FN (d)

Sch.
FN (e)

21.44

14.66

12.35

7.5

8..

21.67

13.67

11.57

11.9

11.7

21.21

13.30

11.14

11.0

11.7

21.29

13.80

11.70

9.6

10.1

20.96

14.13

11.76

7.1

9.3

21.01

13.92

11.52

8.0

10.0

.030
.045
.045
1/16
3/32
3/32
1/8
5/32

.045
3/32
3/32
1/8

1/16
3/32
1/8
1/16
1/8
1/8

1/16
3/32
1/8
1/16
3/32
1/8

1/16
1/16

21.51

13.05

11.37

13.3

11.

21.46

13.42

11.92

11.8

10.0

21.57

13.60

11.20

11.6

12.6

21.18

14.22

11.82

7.7

9.6

21.74

13.05

11.43

14.2

12.2

1/16
1/16
3/32
3/32
10
3/32
1/8
11
1/16
3/32
the former contained lower nitrogen
in all cases.
Joint
Design
and Position
of
Welding. Although no tests were run
in these areas it seems logical that
nitrogen pickup would depend upon
these variables also. Both fillet welds
and deep groove welds are obviously
physically better shielded than welds
on a flat plate. Protection against
drafts would be a variable. Gas protection could be better or poorer in
the vertical or overhead positions
than in the flat position.
Recommendations
On
Average
Nitrogen
Pickup.
From a practical
viewpoint, it is necessary to assume
some average nitrogen pickup from
filler metal to deposit in order to
supply a filler metal to a customer
which will give him a reasonable
ferrite level in the weld deposit
reasonable in the sense of providing
enough ferrite to prevent cracking or
to meet some deposit ferrite requirements or both.

M-Gl"
FN
10.0(9>
9.7
12.0
7.4
10.4
9.8
10.4
9.9
10.1
11.7 , 9 )
10.2
10.8
13.6
12.0
11.4 <9)
11.5
10.4
12.4
9.9 l 9 )
11.5
9.5
8.6
9.1(9)
9.1
9.1
9.0
8.2(9)

8.2
8.1
8.2
11.4*9)
10.7
12.1
11.6 (9 >
12.4
10.7
13.01a'
13.9
12.0
7.5 , 9 )
7.7
7.2
12.1 (9'
12.9
11.3

Change,

DeL. F N t o Sch. F N t o
M-GFN
M-G FN
+ 2.5 (9>
+ 2.4
+ 4.5
-0.1
+ 2.9
+ 2,3
+ 2.9
+ 2.4
+ 2.6
- 0 . 2 (3)
-1.7
-1.1
+ 1.7
+ 0.1
+ 0.4'9)
+ 0.5
-0.6
+ 1.4
+ 0.3 l 9 )
+ 1.9
-0.1
-1.0
+ 2.0 ( s )
+ 2.0
+ 2.0
+ 1.9
+ 0 2

(g>

+0.2
+ 0.1
+ 0.2
-1.9.(9)
-2.6
-1.2
- 0 . 2 (9)
+ 0.6
-1.1
+ 1.4 (9)
+ 2.3
+ 0.4
- 0 . 3 (9)
0
-0.5
_21(g>

-1.3
-2.9

+ 1.2<9)
+ 1.1
+ 3.2
-1.4
+ 1.6
+ 1.0
+ .1.6
+ 1.1
+ 1.3
0.0 (9
-1.5
-0.9
+ 1.9
+ 0.3
+ 0.7' 9 >
-0.2
-1.3
+ 0.7
_0.2(9)
+ 1.4
-0.6
-1.5
-0.2'9>
-0.2
-0.2
-0.3
-1.8(9)
-1.8
-1.9
-1.8
-0.4(9
-1.1
+ 0.3,
+ 1.6 <gl
+ 2.4
+ 0 7

- ,i
+ 0.4 ( 9 )
+ 1.3
-0.6
+ 2.1 <9)
-1.9
-2.4
- 0 . 1 '9I
+ 0.7
-0.9

When the variations in nitrogen


pickup shown in the data are considered, as well as the added variations which could come from variations in production welding situations
due to the influence of the joint on gas
shielding, any assumptions reached
must be accepted as being only approximations of the final results. With
this caution in mind, we assume a
GTAW pickup of 0.020% N and a
GMAW pickup of 0.040% N. The
GTAW figure of 0.020% is roughly
equal to the average data presented
and represents what we estimate
would result from good practice in
joints more protected than our test
pads. The GMAW figure of 0.040% is
above the average of the data in Table
4 but provides a factor of safety for
welders less trained in and less conscious of the nitrogen problem than
the welder who ran most of our later
tests. As our earlier tests show, the
pickup of nitrogen and the loss of
ferrite can be quite high.

Heat

M-G
FN

Change,
DeL. F N t o
M-GFN

Change,
Sch. FN to
M-GFN

12.0*9)
11.3
12.6
11.2(9)
12.1
10.2
9.0 (8)
7.7
10.3

_0.6<9)
-1.3
0
-3.6(9)
-2.7
-4.6
_2.7<9>
-4.0
-1.4

-0.9<9>
-1.6
-0.3
-0.3<9)
+0.6
-1.3
-2.2(9)
-3.5
-0.9

11,6

13.0

-1.1

+ 1.4

10.0

10.0

10.2

+0.2

+0.2

11.91

13.1

10.7

9.7

-3.2

-1.0

13.79

11.60

12.4

12.1

12.4

+ 0.3

21.30

14.67

11.97

7.0

9.5

9.3

+ 2.3

-0.2

21.32

14.36

11.87

7.9

9.8

7.7

-0.2

-2.1

21.59

14.85

12.69

7.5

8.4

9.3

+ 1.8

+0.9

21.75

13.17

11.31

14.0

12.7

14.3

+ 0.3

+ 1.6

21.67

13.11

11.61

13.7

11.4

11.6

-2.1

+0.2

21.77

12.44

11.12

16.3

13.3

14.2

-2.1

+ 0.9

21.88

13.60

12.25

13.0

10.1

13.2

+0.2

+ 3.1

22.18

13.93

11.98

13.1

11.8

8.9

-4.2

-2.9

21.23

13.90

11.68

9.0

10.1

9.8

+0.8

-0.3

diam,
in.

Cr
Eq

DeL.
Ni Eq

Sch.
Ni Eq

3/32
3/32

22.08

13.98

11.52

12.6

12.9

21.62

12.73

11.53

14.8

11.5

21.77

13.84

11.74

11.7

11,2

22.34

13.84

12.13

14,1

21.46

13.91

11.90

21.77

13.43

21.88

DeL.
FN

Sch.
FN

12

13

3/32
1/8

14

.045
1/16

15

16

17

1/16

1/16

3/32
18

19

20

3/32

1/16

.045
21

22

23

24

25

26

27

1/16
3/32
1/16
1/8
1/18
.035

3/32

(a) Cr Eq: Chromium equivalent calculated from filler metal chemistry (= % Cr + % Mo + 1.5 X % Si + 0.5 X % Cb);
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 %
C + 0.5 X % Mn+ 30 X % N)
(c) Sch. Ni Eq: Schaeffler nickel equivalent calculated from filler metal chemistry corrected per Table 8 (+ % Ni + 30 X %
C + 0.5 x % Mn)
(d) DeL. FN: Ferrite Number calculated using DeLong constitution diagram
(e) Sch. FN: % ferrite calculated using Schaeffler constitution diagram (taken as equal to FN)
(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.

Heat
28

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

20.78

14.12

11.84

6.3

8.6

20.61

13.64

11.63

7.1

8.7

1/16
3/32
28
045

M-G
FN
8.4 | a )
8.2
8.6
5.8

Change,
Change,
DeL. F N t o Sch. F N t o
M-GFN
M-G FN
+ 2.1 <a)
+ 1.9
+ 2.3

-0.2
-0.4
0

-1.3

-2.9

(a)

(a) Average for heat when more than one test per heat

DeLong Constitution
Revision

Diagram

The C o n s t i t u t i o n
Diagram
presented in 1956 3 is presented in a
revised form as Fig. 2, The major revision has been to convert the original
magnetically-determined percent ferrite values to the new WRC Ferrite

However, the recent WRC calibration has increased the assigned


values for ferrite content as previously described. In addition, it
seemed desirable to utilize the substantial body of quality assurance
data which we had collected on
GTAW and GMAW
deposits
(presented in the next portion of this
paper) to further improve the diagram
in the high ferrite areas. This data
substantially outnumbered the prior
data from covered electrodes in the
higher ferrite area. The extent of the
revisions can be s u m m a r i z e d as
follows: (1) no change was made in
the 0% (i.e., 0 FN) line, (2) slight
changes were made in the 4 to 6 FN
area to make a smooth transition to
the reduced spacing between lines
above the old 6% level and (3) at
Ferrite Numbers above 8, where the
original covered weld metal data were
sparse, the m a t h e m a t i c a l l y c o n verted lines were shifted somewhat
toward the 0 FN line so that the new
diagram better reflects the ferrite
levels found in GTAW and GMAW
deposits.
Inherent Differences Between the
DeLong and Schaeffler Diagrams

Table 10 Ferrite Data from 3 Type 347 GTAW Pads


(For Explanation of Column Headings See Footnotes Table 9)
Rod
diam,
in.

prepared with covered electrodes


1/16 to V* in. in diam and of three different coverings. Of these, the 308,
308L, 309 and 347 pads averaged
about 6% ferrite on the old scale, and
the many 316, 316L, 317 and 317L
pads ranged from 0% up to 6%. Since
little data were available at ferrite
levels higher than 8%, the original
ferrite lines at 8% and higher were
primarily extensions of the pattern
shown by the data in the 0 to 8%
range. In spite of this, our measured
ferrite did correlate reasonably well
with the limited data which we had
available for higher ferrite levels, and
did support the diagram lines from
8% through 14% ferrite.

Number scale for magnetic instruments, specifically for our MagneGages. This is a direct mathematical
conversion, as illustrated in Table 1.
Also, a slight adjustment was made
in the line locations for the higher
ferrite areas of the diagram. The
original diagram was based on data
f r o m over 600 weld metal pads

A comparison of Figs. 1 and 2


shows that the calculated Chromium
Equivalent (Cr Eq) for a given specimen will be the same for both the
DeLong and Schaeffler Diagrams, but
the calculated Nickel Equivalent (Ni
Eq) will obviously be higher for the
D e L o n g D i a g r a m b e c a u s e of
DeLong's added factor for nitrogen.
Schaeffler assumes 0.060% N for 308,
which on the DeLong Diagram represents 1.8 Ni Eq. The objective of this
section is to clarify and compare the
effects of rather small changes in
Chromium and Nickel Equivalents on
the predicted ferrite levels as calculated with the DeLong and Schaeffler
Diagrams. As base points the points
where the 2, 6 and 11 FN (DeLong) or
percent ferrite (Schaeffler) lines intersect the vertical line at 21.0 Cr Eq will
be used. The individual effects of
changes of 0 . 2 in Cr Eq and Ni Eq

WELDING RESEARCH SUPPLEMENT!

289-s

c o u n t e r e d in a g r o u p of weJd d e posits p r e p a r e d with filler m e t a l f r o m


a given heat and type.

Table 11 Ferrite Data f r o m 22 Type 308L GTAW Pads


(For Explanation

ot Column

Headings

See

Footnotes Table 9)

Rod

diam,
Heat

in.

32

DeL.
Ni Eq

21.63

30

31

Cr
Eq

1/16
3/32
5/32
.035
3/32
3/32

1/16

20.96

14.10

12.97

Sch.
Ni Eq

11.22

11.44

DeL.
FN

10.0

11.0

Sch.
FN

M-G
FN
10.8 ( a )

12.6

9.7
9.8
13.0
11.1 W
11.3
11.3
10.8
11.4<a
10.8
12.0
>(a)
9.9"
9.0
10.8
8.0 (a)
7.3
8.8
11.9 (a)
12.2
11.7
11.5 <a)
12.1
11.0

10.0

21.48

12.87

11.43

13.8

11.4

21.11

12.

11.14

12.1

11.3

1/16
33

34

3/32
1/8

21.40

14.10

11.40

9.0

11.3

.045
1/16
35

21.34

13.48

11.59

11.0

10.6

.035
1/16
36

20.92

13.59

11.31

8.7

10.2

21.85

12.16

10.99

17.7

14.3

3/32
I/8
37

21.47

13.03

11.53

13.0

11.0

21.28

12.71

11.30

13.4

11.3

21.57

13.17

11.70

13.1

10.8

20.96

13.26

11.49

10.0

10.0

22.36

13.89

11.49

14.1

14.0

11.5

3/32
39

12.8

3/32
40
1/8
41

11.5

10.8

3/32
42

+ 0.8
-0.3
-0.2
+ 3.0

(a)

-1.8
-2.9
-2.8
+0.4

a)

+ 0.1 <
+0.3
+0.3
-0.2
2.4 , a >
-3.0
-1.8
-2.2|a)

-3.1
-1.3
-1.0(a)

-1.7
-0.2
+ 0.9 ( a )

+ 1.2
+ 0.7

+ 2.8 ( a )
+ 3.4

+ 1.1<>

+ 1.3
+ 1.3
+ 0.8
0.0' a )
-0.6
+ 0.6
-1.4 ( a >
-2.3
-0.5
-3.3(a)
-4.0
-2.5
+ 1.3,a
+ 1.6
+ 1.1
+ 1.3 (a)
+ 1.9
+ 0.8

+ 2.3

14.8

1/16
38

Change,
Change,
DeL. F N t o Sch. F N t o
M-GFN
M-GFN

10.0

.035

-2.9
-1.5
-0.6
-1.6
+ 0.8
-4.1

+ 0.5
+ 0.5
+ 1.5
+ 0.7
+ 0.8
-4.0

(a) Average for heat when more than one test per heat

a n d t h e c u m u l a t i v e e f f e c t s of a
c h a n g e of + 0 . 2 Cr E q a n d - 0 . 2 Ni E q
(or - 0 . 2 C r E q a n d + 0 . 2 Ni Eq) a r e
s h o w n in T a b l e 5. T h e v a l u e s w e r e o b t a i n e d by r e a d i n g t h e d i a g r a m s of
F i g u r e s 1 a n d 2 as c l o s e l y a s p o s s i b l e .
As w o u l d be expected f r o m exa m i n i n g t h e d i a g r a m s , t h e p a t t e r n is
not a c o n s t a n t one b e c a u s e both the
s p a c i n g s a n d t h e s l o p e s of t h e c o n stant ferrite lines c h a n g e f r o m o n e
d i a g r a m t o t h e o t h e r . A t a level of 2
F N , a g i v e n c h a n g e in Cr E q a n d / o r Ni
Eq w i l l p r o d u c e l e s s c h a n g e in c a l c u lated ferrite on the D e L o n g D i a g r a m
than on the Schaeffler Diagram; howe v e r , at l e v e l s of 6 FN a n d 11 F N t h e
r e v e r s e is t r u e . A l s o n o t e t h a t t h e
c h a n g e s in c a l c u l a t e d f e r r i t e a r e
r a t h e r s u b s t a n t i a l in v i e w of t h e r e l a t i v e l y s m a l l c h a n g e of 0.2 in t h e C r E q
a n d Ni E q . A 0.2 c h a n g e in Ni E q c o u l d
b e p r o d u c e d by a c h a n g e of a b o u t
0 . 0 0 7 % in C, 0 . 0 0 7 % in N (if t h e
D e L o n g D i a g r a m is u s e d ) , 0 . 2 % in N i ,
or 0 . 4 % in M n ; a 0.2 c h a n g e in C r E q
c o u l d b e p r o d u c e d b y a c h a n g e of
0 . 2 % in C r , 0 . 1 4 % in S i , 0 . 2 % in M o o r
0 . 4 % in C b .
T a b l e 5 s h o w s t h a t a c h a n g e of 0.2
e a c h in C r E q a n d Ni E q w i l l , if t h e y r e 290-s I J U L Y

1973

i n f o r c e e a c h o t h e r , p r o d u c e an o v e r a l l
c h a n g e of a b o u t 1.1 t o 1.6 F N d e p e n d i n g o n the ferrite level a n d t h e
diagram used.
Study

of

GTAW

and

GMAW

Deposits
The following discussion covers a
s t u d y of 129 G T A W a n d 3 2 G M A W
d e p o s i t s p r e p a r e d by our Quality A s surance (QA) Department; M a g n e G a g e f e r r i t e w a s c h e c k e d o n all t h e
d e p o s i t s , a n d o n m o s t of t h e d e posits chemical analyses were run.
The study was undertaken for the
following listed p u r p o s e s :
1. T o d e v e l o p a p r a c t i c a l m e t h o d
for p r e d i c t i n g G T A W a n d G M A W
deposit
ferrite
using
chemical
a n a l y s i s of t h e f i l l e r m e t a l a n d t h e
DeLong Diagram.
2. U s i n g t h e m e t h o d d e v e l o p e d in
1, to c o m p a r e t h e D e L o n g
and
S c h a e f f l e r D i a g r a m s in r e s p e c t of h o w
closely each predicts the G T A W and
G M A W deposit ferrite for the m o r e
c o m m o n f e r r i t e - b e a r i n g g r a d e s of
austenitic stainless steels.
3. T o s t u d y t h e v a r i a t i o n s
in
M a g n e - G a g e readings typically e n -

Procedure
Several
approaches
were
cons i d e r e d in a n a l y z i n g t h e w e l d p a d s .
A m o n g t h e s e w e r e (a) t o c a l c u late f e r r i t e u s i n g o n l y t h e f i l l e r m e t a l
heat analysis a n d c o m p a r e the results
with the m e a s u r e d
(Magne-Gage)
f e r r i t e o b t a i n e d in e a c h p a d ; ( b ) s a m e
as (a) b u t u s i n g a c t u a l w e l d m e t a l
analysis including an estimate for
n i t r o g e n c o n t e n t of t h e w e l d m e t a l
(since the QA weld metal analyses did
n o t i n c l u d e n i t r o g e n ) ; a n d (c) t o
m o d i f y t h e filler m e t a l heat a n a l y s e s
s o a s t o r e f l e c t a n t i c i p a t e d c h a n g e s in
the critical elements c a r b o n and nitrog e n , a n d to c o m p a r e ferrite c a l c u lated f r o m the resulting
analyses
with M a g n e - G a g e ferrite m e a s u r e d on
the pads.
A p p r o a c h (c) w a s f i n a l l y d e c i d e d
u p o n . U s e d as t h e base w a s t h e filler
metal chemistry d e t e r m i n e d f r o m the
a v e r a g e of s e v e r a l t e s t s o n t h e h o t
rolled rod a n d / o r d r a w n wire for the
elements carbon, manganese, silicon,
c h r o m i u m , nickel, m o l y b d e n u m and
c o l u m b i u m and the supplying mill
a n a l y s i s f o r n i t r o g e n . S u m m a r i e s of
the filler m e t a l a n a l y s e s by t y p e a r e
p r e s e n t e d in T a b l e 6 f o r t h e r o d s
u s e d in G T A W a n d T a b l e 7 f o r t h e
e l e c t r o d e s u s e d in G M A W . A c o m p a r i s o n of t h e a c t u a l w e l d
metal
analyses with t h e filler metal c h e m i s tries indicated the following:
1. T h e r e w a s a d i s t i n c t p a t t e r n of
l o s s e s in c a r b o n c o n t e n t
between
f i l l e r m e t a l a n d p a d , t h e a m o u n t of
c a r b o n lost d e p e n d i n g o n w i r e c a r b o n
content and welding process; Table 8
lists t h e s e f i n d i n g s . In c a l c u l a t i n g
" p r e d i c t e d " ferrite w e modified the
filler m e t a l c a r b o n c o n t e n t in a c c o r d a n c e w i t h t h e a v e r a g e l o s s e s s h o w n in
T a b l e 8.
2. A l t h o u g h s o m e t r e n d s
were
seen indicating slight average losses
in M n , N i , C r a n d M o (in t h e M o bearing grades), which we assumed
to be primarily d u e to vaporization
rather than to o x i d a t i o n , w e i g n o r e d
t h e m b e c a u s e (a) t h e e f f e c t s w e r e
l a r g e l y s e l f - c a n c e l l i n g , (b) t h e y w o u l d
influence calculated Ferrite N u m b e r
b y n o m o r e t h a n a b o u t 0.5 F N if a t all
a n d (c) t h e r e w a s a q u e s t i o n a s t o
whether the changes might have been
d u e t o n o r m a l a n a l y t i c a l e r r o r in t h e
X-ray
spectrographic
techniques
u s e d to d e t e r m i n e t h e a n a l y s i s ( r e checks using wet chemistry techniques are being m a d e to c h e c k this
out).
In a d d i t i o n t o m o d i f y i n g t h e f i l l e r
metal c a r b o n c o n t e n t for calculation
of p r e d i c t e d w e l d m e t a l f e r r i t e , w h i c h
would apply to calculations using both
the D e L o n g and the
Schaeffler

Diagrams, we assumed a constant


nitrogen pickup from the filler metal
nitrogen of 0.020% for GTAW calculations and 0.040% for GMAW calculations (also listed in Table 8 for
reference). These assumptions were
based on the data discussed earlier in
this paper.
Tables 9 through 19 list data on the
129 GTAW and 32 GMAW pads included in the study; these data are
summarized by type and process in
Tables 20 (GTAW) and 21 (GMAW).
A n Overall
Review

GTAW

and

GMAW

Table 12 Ferrite Data from 23 Type 309 GTAW Pads


(For Explanation of Column Headings See Footnotes Table 9)

Heat

Rod
diam,
in.

43

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

24.64

18.45

16.17

8.9

7.0

25.01

17.65

15.28

12.8

9.7

24.70

18.41

16.13

9.2

7.2

25.61

17.23

15.13

16.4

11.0

24.96

17.47

15.37

13.3

9.4

24.94

17.96

15.86

11.7

8.3

24.86

18.10

15.58

11.0

8.7

25.61

18.55

16.15

12.5

9.0

24.67

17.58

15.18

11.9

9.2

25.12

16.94

14.51

15.5

11.6

25.08

17.83

15.43

12.5

9.4

24.95

16.70

15.11

15.5

9.8

25.59

16.16

14.57

19.9

12.7

25.09

17.78

15.38

12.7

9.6

24.83

16.84

14.38

14.5

11.3

M-G
FN

1/16
3/32
3/32
1/8
44
.045
1/8
45
.045
3/32

Variations From Test To Test On A Single


Heat

46

The Magne-Gage data from Tables


9 through 19 have been summarized
and analyzed in Table 22 to show
statistical data by types and in total for
all heats on which multiple tests were
run. Each test variance shown was
calculated by summing the squares of
the differences between individual
test values and the average test value
for the heat used in each test and
dividing the sum so obtained by the
figure (number of tests minus number
of heats involved); this divisor c o m bines the usual (n-1) small sample
divisors for the data from individual
heats and is the commonly accepted
method of giving an overall variance
for several small groups of data. Taking t h e s q u a r e root, of t h e test
variance furnished the standard
deviation in each case.
The data in Table 22, being based
on Magne-Gage measurements, are
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
data from the several types are arranged in order from the type showing the lowest standard deviation to
that showing the highest. The table
shows what appears to be a significant trend, i.e., that the standard
deviation increases appreciably as
the average ferrite content increases.
The combined 316 and 316L tests
show an average ferrite content 5.6
FN and an average standard deviation of about 0.8 FN. The combined
308 and 308L tests average about
10.6 FN and show an average standard deviation of about 1.2 FN. Tests
of the more highly alloyed 309 weld
metal with about the same ferrite content as the 308 show a standard
deviation of about 1.8 FN.
Although the figure for 309 may be
correct and due to that type's higher
alloy content, it is based on only 14
tests and with more tests the standard
deviation for 309 might come down
toward that of the 308 family. The
overall GTAW g r o u p of 94 tests on 36
heats shows an average standard

47

.035
1/8
1/16
1/16
48
1/16
3/32
49
1/8
50
3/32
51
1/8
52
1/16
53
1/8
54
1/8
55
3/32
56
3/32
57
1/16

Change,
Change,
DeL. FN to Sch. F N t o
M-GFN
M-GFN

6.8 ( a )
7.1
7.1
6.7
6.2
9.7 <a)
8.2
11.3
8.8 ( a )
8.8
8.8
14.3<a)
13.0
15.6
12.0<a)
10.4
13.7
10.8 (a)
13.2
8.5

-2.1 , a >
-1.8
-1.8
-2.2
-2.7
-3.1(a)
-4.6
-1.5
-0.4<a>
-0.4
-0.4
-2.1(a)
-3.4
-0.8
-1.3(a
-2.9
+0.4
-0.9(a)
+ 1.5
-3.2

-0.2(a)
+0.1
+0.1
-0.3
-0.8
0.0 ( a )
-1.5
+ 1.6
+ 1.6 (a)
+ 1.6
+ 1.6
+3.3(a)
+2.0
+4.6
+ 2.6 ( a )
+ 1.0
+4.3
+ 2.5 (a)
+4.9
+0.2

10.8

-0.2

+2.1

12.6

+0.1

+3.6

13.2

+ 1.3

+4.0

13.9

-1.4

+ 2.3

13.5

+ 1.0

+4.1

9.8

-5.7

18.6

-1.3

+5.9

13.0

+ 0.3

+ 3.4

16.3

+ 1.8

+5.0

(a) Average for heat when more than one test per heat

Table 13 Ferrite Data from 14 Type 316 GTAW Pads


(For Explanation

Heat

Rod
diam,
in.

58

o)'

Column Headings

See

Cr
Eq

DeL.
Ni Eq

Sch.
NiEq

DeL.
FN

Sch.
FN

22.29

16.13

14.63

6.3

5.3

22.14

15.96

14.43

6.3

5.4

21.80

16.62

14.79

3.3

3.3

22.02

16.18

14.65

5.4

4.5

1/8
1/8
1/8
59
3/32
3/32
60
3/32
1/8
61
1/16
3/32
62

21.19

16.19

14.60

2.4

1.8

21.78

16.42

14.32

3.8

4.7

21.59

15.81

14.34

4.6

4.0

21.50

16.71

14.91

2.1

1.9

.035
.045
63
.045
64
1/16
65
3/32

Footnotes Table 9)
Change,
Change,
DeL. F N t o Sch. FN to
M-GFN
M-GFN

M-G
FN
5.2 (a)
5.7
4.1
5.8
5.6<a>
4.8
6.5
5.4'a
4.8
6.1
6.6
6.8
2.2(a>
2.2
2.2

-1.1(a)
-0.6
-2.2
-0.5
-0.7(a)
-1.5
+ 0.2
+ 2.1(a)
+ 1.5
+2.8
+ 1.3 (a)
+ 1.2
+ 1.4
-0.2 ( a >
-0.2
-0.2

-0.1(a)
+0.4
-1.2
+0.5
+0.2(a)
-0.6
+ 1.1
+2.1<a>
+ 1.5
+2.8
+ 2.2 ( a )
+2.1
+2.3
+0.4(al
+0.4
+0.4

5.9

+ 2.1

+ 1.2

4.4

-0.2

+0.4

4.8

+ 2.7

+ 2.9

6 7

(a)

(a) Average for heat when more than one test per heat

WELDING RESEARCH SUPPLEMENT!

291-s

Table 14 Ferrite Data from 14 Type 316L GTAW Pads


(For Explanation of Column Headings See Footnotes Table 9)

Heat

Rod
diam,
in.

66

Cr
Eq

DeL.
NiEq

Sch.
Ni Eq

DeL.
FN

Sch.
FN

21.67

15.83

14.48

4.9

3.9

22.26

15.29

13.76

9.0

7.2

21.02

16.00

14.53

2.3

1.3

22.06

15.92

14.48

6.0

5.1

21.85

15.69

14.10

6.0

5.5

21.97

14.87

13.22

9.0

7.9

.045
.045
1/16
1/8
67
.035
.045
1/16
1/8
68
3/32
3/32
1/8
69
3/32
70
3/32
71
3/32

M-G
FN
7.4 ( a '
6.7
6.9
7.7
8.4
8 .2(a)
8.8
S.2
7.2
8.6
4.8(a)
3.7
4.8
5.8

Change,
Change,
DeL. F N t o Sch. F N t o
M-G FN
M-GFN
+ 2.5 ( a )
+ 1.8
+2.0
+2.8
+ 3 . 5 >a
-0.8
-0.2
-0.8
-1.8
-0.4
+2.5(a)
+ 1.4
+ 2.5
+3.5

+ 3.5 ( a |
+ 2.8
+3.0
+ 3.8
+4.5
+ 1.0 (a)
+ 1.6
+ 1.0
0
+ 1.4
+3.5 (a
+ 2.4
+ 3.5
+4.5

6.0

+0.9

6.3

+0.3

+0.8

10.2

+ 1.2

+ 2.3

(a) Average for heat when more than one test per heat

Table 15 Ferrite Data from 12 Type 308 GMAW Pads


(For Explanation of Column Headings See Footnotes Table 9)

Heat

Electr.
Diam,
in.

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

21.21

14.23

11.47

7.9

10.6

21.44

15.59

12.68

4.8

8.0

21.67

14.60

11.90

8.6

10.5

.030
.035
1/16
1
.045
2
.045
6

21.01

14.85

11.85

5.2

9.1

22.34

14.77

12.46

11.0

10.6

21.88

14.72

11.93

9.1

10.9

21.32

15.29

12.20

5.1

8.9

21.77

14.31

11.91

10.0

10.7

21.43

14.13

11.73

9.0

10.3

21.65

14.96

11.81

7.5

10.7

3/32
15
1/16
18
3/32
20
.045
72
.030
73
.045
74
.035

M-G
FN

Change,
Change,
DeL. F N t o Sch. F N t o
M-G FN
M-GFN

6.9 (a)
7.3
6.5
6.8

-1.0,a)
-0.6
-1.4
-1.1

-3.7( a >
-3.3
-4.1
-3.8

6.0

+ 1.2

-2.0

8.9

+0.3

-1.6

4.6

-0.6

-4.5

8.1

-2.9

-2.5

10.8

+ 1.7

-0.1

2.1

-3.0

-6.8

11.0

+ 1.0

+0.3

8.5

-0.5

-1.8

7.4

-0.1

-3.3

(a) Average for heat when more than one test per heat

deviation of about 1.2 FN, and an


average weld deposit ferrite content
of 9.4 FN. The GMAW data are not extensive enough to separate by type,
but the overall standard deviation of
0.96 FN at an average FN of 7.9 fits in
well with the trend indicated by the
GTAW data.
Considering that in the DeLong
Diagram a given change in chemistry
causes a greater change in calcu292-s I J U L Y 19 73

lated FN at high FN levels than at low


FN levels (as shown in Table 5 at
levels of 2, 6 and 11 FN), it could logically be expected that a similar trend
would be present in ferrite measured
with a Magne-Gage i.e., a given
change in the factors influencing
measured FN would cause a greater
change in measured FN at high FN
levels than at low FN levels. This
thesis is borne out by the strong in-

dication that the standard deviations


for measured ferrite at high FN levels
are greater than they are at low FN
levels.
If the experimental results are
assumed to produce a normal distribution curve, a range of plus or
minus one standard deviation (sigma)
from the mean should encompass
about 68% of the tests, and one of
plus or minus two sigma should encompass about 95% of the tests. It is
obviously desirable that acceptance
tests for material be set so that there
is substantially less than a 10%
chance of rejection due to testing
error or variation. Thus, if a specification for a GTAW deposit were to be
established at 5 to 12 Ferrite Number
and the assumption made that mean
values of 5.0 FN on the low side or
12.0 FN on the high side would be
acceptable, the receiving acceptance
test limits would require a test tolerance of not less than about 1.5 FN on
the low side and about 2.0 FN on the
high side in order to reduce to well
under 10% the chances of rejection at
the specification limits.
Amendments to specifications MILE-0022200/2 and MIL-E-19933D for
c o v e r e d and b a r e a u s t e n i t i c
chromium nickel stainless steel electrodes have for several years
accepted this philosophy of broader
receiving test tolerances to provide
for variations in pad preparation,
testing procedure, and the inherent
ferrite variations to be expected from
pad to pad in the weld metal itself.
Variations From Predicted FN To
Measured FN
The data from Tables 9 through 19
are presented graphically in the bar
graphs of Fig. 4, which also presents
for comparison a summary bar graph
of similar data on covered electrode
weld pads of the same types drawn
from the work described in Ref. 3. The
covered electrode graph has been
condensed to a scale similar to that of
the other bar graphs for easier comparison. As stated in the note on Fig.
4, the width of each bar representing
covered electrode data based on the
DeLong Diagram has been expanded
15% to compensate for the increase
in numerical value in going from our
old percent ferrite to the new Ferrite
Number system.
Turning first to the covered electrode summary bar graph data, it can
be seen that roughly 87% of the values
were within 2.5 FN based on the DeLong Diagram versus about 68%
within the same range based on the
Schaeffler Diagram; at 3 FN the
figures were about 92% and 78%,
respectively. The DeLong Diagram
data were significantly better than the
Schaeffler, particularly on the low
side. This was logical because the De-

Long Diagram corrected not only for


core wire nitrogen contents, which
differed appreciably, but also for the
varying nitrogen pickup f r o m differences in electrode size and covering
types, ranging from an average 0 . 0 1 %
N pickup on c o m m o n sizes of dc lime
type electrodes to 0.05% on some
sizes of special titania types. This
s p r e a d of 0 . 0 4 % in t h e t y p i c a l
nitrogen pickups in itself represents a
change in nickel equivalent of about
1.2 and a corresponding change in FN
of approximately 4.0. Appropriate
corrections for the varying nitrogen
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
covering and size plus further corrections in the 316, 316L and 309 families as described in reference 3 logically explain the better results o b tained with the DeLong Diagram on
covered electrodes.
However, the significant improvement found for covered electrodes
with the DeLong Diagram does not
carry over to the GTAW summary bar
g r a p h s . T h e r e , the D e L o n g a n d
Schaeffler results are essentially
equivalent in total. On the DeLong
Diagram approximately 94% of the
calculated values are w i t h i n - i 3 FN of
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
summary bar graph, where a 3 FN
range included about 92% of the
values, and are appreciably better
than the Schaeffler covered electrode
summary, which showed only 78%
within 3 FN.
With regard to individual types
welded with the GTAW process, the
Schaeffler Diagram understates the
316, 316L and 309 families as it did for
covered electrodes 3 . The DeLong Diagram has a better balance on these
three more highly alloyed types. The
slope of the ferrite lines is greater on
the DeLong Diagram than on the
Schaeffler, so that the DeLong Diagram predicts higher ferrite values
than Schaeffler for types 316 and
316L at normal nitrogen levels; this
difference between the two predictions is even more pronounced for
type 309 weld metal. With 308 and
308L weld metal (probably also 347)
the DeLong and Schaeffler Diagrams
produce essentially
equivalent
results. Where the balance is a little
low as on the 308L, it is low on both
diagrams. The ER308 and ER308L
rods average about 0.045% nitrogen
and an 0.020% pickup will yield a
deposit average of 0.065% nitrogen,
r e a s o n a b l y c l o s e to t h e 0 . 0 6 %
nitrogen content which we believe
was a s s u m e d by S c h a e f f l e r for
covered electrode deposits. Thus the
Schaeffler Diagram would be expected to give reasonable results for

Table 16 Ferrite Data from 6 Type 308L GMAW Pads


(For Explanation of Column Headings See Footnotes Table 9)

Heat

Electr.
diam,
in.

35

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch,
FN

21.34

14.08

11.59

8.9

10.6

20.92

14.19

11.31

6.8

10.2

22.36

14.49

11.49

12.1

14.0

21.52

13.62

11.58

11.1

11.0

21.04

13.38

11.46

10.0

10.2

21.33

13.52

11.48

10.7

10.9

1/16
36
1/8
42
.045
76
.045
77
1/16
78
.045

M-G
FN

Change,
Change,
DeL. FN to Sch. FN to
M-GFN
M-GFN

8.6

-0.3

-2.0

7.1

+ 0.3

-3.1

9.8

-2.3

-4.2

9.1

-2.0

-1.9

10.8

+0.8

+0.6

10.0

-0.7

-0.9

Table 17 Ferrite Data from 2 Type 347 GMAW Pads


(For Explanation of Column Headings See Footnotes Table 9)
Electr.
Heat

Change,

diam,
in.

29

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

20.61

14.57

11.96

4.6

7.9

21.15

14.82

11.70

5.7

9.8

.045
75
.045

M-G
FN

Change,

DeL. F N t o Sch. F N t o
M-GFN
M-GFN

6.1

+ 1.5

-1.8

4.1

-1.6

-5.7

Table 18 Ferrite Data from 8 Type 309 GMAW Pads


(For

Heat

Electr.
diam,
in.

46

Explanation ->t Column

Headings

See Footnotes

Cr
Eq

DeL.
NiEq

Sch.
Ni Eq

DeL.
FN

Sch.
FN

M-G
FN

25.61

18.16

15.46

13.7

10.4

11.3<a>
12.4
10.2

24.70

19.34

16.46

6.5

6.6

24.96

18.40

15.70

10.3

8.7

24.86'

19.03

15.91

8.1

8.0

25.61

19.48

16.48

9.8

8.3

24.67

18.51

15.51

8.8

8.4

24.83

17.77

14.71

11.9

10.4

.035
45
.045
47
1/16
49
1/16
50
1/16
51
1/8
57
1/16

Table 9)
Change,
Change,
DeL. FN to Sch. FN to
M-GFN
M-G FN
- 2 . 4 (a
-1.3
-3.5

+0.9(a)
+ 2.0
-0.2

8.5

+2.0

+ 1.9

10.7

+0.4

+2.0

9.1

+ 1.0

+ 1.1

7.8

-2.0

-0.5

8.6

-0.2

+ 0.2

13.0

+ 1.1

+2.6

(a) Average for heat when more than one test per heat

Table 19 Ferrite Data from 1 Type 316 and 3 Type 316L GMAW Pads
(For Explanation of Column Headings See Footnotes Table 9)

Heat

Electr.
diam,
in.

61
(316)
.035
69

(316L) .035
5/64
67

(316L) 1/16

Cr
Eq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

22.02

17.05

14.92

3.0

'3.7

22.06

15.89

14.48

6.2

5.1

M-G
FN

4.3

e.i<a>
5.4
6.7

22.26

17.11

13.76

3.7

7.2

3.7

Change,
Change,
DeL. FN to Sch. FN to
M-GFN
M-GFN

+ 1.3
+ 2 . 1 <a)
-0.8
+0.5

+0.6
+ 1.0(a>
+0.3
+ 1.6

-3.5

(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 measured ferrite significantly and brings it


down close to the general range predicted by the diagram.

Table 20 Summary of GTAW Ferrite Data


(For Explanation of Column Headings See Footnotes Table 9)

Item

CrEq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

M-G
FN

Change,
Change,
Sch. FN
DeL. FN
to M-G FN to M-G FN

ER308, 53 tests
Range
20.96
to
22.34
Avg.
21.50

12.44
to
14.85
13.82

11.12
to
12.69
11.75

7.0
to
16.3
10.6

8.4
to
13.3
10.7

7.2
to
14.3
10.6

-4.6
to
+4.5
0

-3.5
to
+3.2
-0.1

ER347, 3 tests
20.61
Range
to
20.78
Avg.
20.70

13.64
to
14.12
13.88

11.63
to
11.84
11.74

6.3
to
7.1
6.6

8.6
to
8.7
8.6

5.8
to
8.6
7.5

-1.3
to
+ 2.3
+ 1.0

-2.9
to
0
-1.1

ER308L , 22 tests
Range
20.92
to
22.36
Avg.
21.35

12.16
to
14.10
13.33

10.99
to
11.70
11.38

8.7
to
17.7
11.5

10.0
to
14.3
11.3

7.3
to
14.8
11.0

-4.1
to
+ 3.4
-0.5

-4.0
to
+ 1.9
-0.3

ER309, 23 tests
24.64
Range
to
25.61
Avg.
24.99

16.16
to
18.35
17.73

14.38
to
16.17
15.50

8.9
to
19.9
12.5

7.0
to
12.7
9.2

6.2
to
18.6
11.3

-5.7
to
+ 1.8
-1.2

-1.5
to
+ 5.9
+ 2.2

ER316, 14 tests
Range
21.19
to
22.29
Avg.
21.86

15.81
to
16.71
16.23

14.32
to
14.91
14.60

2.4
to
6.3
4.0

1.8
to
5.4
3.5

2.2
to
6.8
4.4

-2.2
to
+ 2.8
+0.4

-1.2
to
+ 2.9
+ 0.9

ER316L , 14 tests
21.02
Range
to
22.26
Avg.
21.76

14.87
to
16.00
15.64

13.22
to
14.53
14.17

2.3
to
9.0
6.0

1.3
tc
7.9
4.8

3.7
to
10.2
7.1

-1.8
to
+ 3.5
+ 1.1

0
to
+4.5
+ 2.3

308 and 308L GTAW deposits.


For the GMAW comparison less
data were available to work with. The
summary bar graph shows a much
better performance for the DeLong
Diagram than for the Schaeffler. All of
the values given for the DeLong Diagram are within 3.5 FN, while only
8 1 % of the Schaeffler values are
within that range. These results are
logical because the average assumed
nitrogen pickup of 0.040% for GMAW
would a p p r e c i a b l y reduce the
measured values and cause them to
fall well below the values calculated
from the Schaeffler Diagram, which
does not correct for nitrogen. Using
the Schaeffler Diagram, six of the thirty two tests run are outside 3.5 FN
and all six are negative, i.e., the
m e a s u r e d is well below the c a l culated, as expected.
The GMAW pattern by types is
reasonably normal for all types with
the D e L o n g D i a g r a m . Using the
Schaeffler D i a g r a m , however, t h e
308, 347 a n d 3 0 8 L t y p e s s h o w
m e a s u r e d ferrite well below the
calculated values. The Schaeffler Dia294-s I J U L Y

1973

g r a m w a s p r e d i c a t e d u p o n an
average 0.06% nitrogen in weld
metal of these t y p e s . W h i l e the
average electrode nitrogen is now
below this, about 0.045%, our assumed pickup of 0.040% N will produce a
typical 0.085% nitrogen in the deposit,
0.025% over the figure that Schaeffler assumed. At a multiplying factor of
30, this excess nitrogen is equivalent
to 0.75 nickel or about 2.5 Ferrite
Number. Confirming this, the data
show that on the average the measured GMAW results are about 2.5 FN
lower than the figi/res predicted using
the Schaeffler Diagram. The Schaeffler pattern with 309, 316 and 316L
does not follow the pattern of the 308
family. Although the number of tests
is not large, the measured values
average slightly higher than the calculated, but within the general 3.5
FN tolerance. This apparent agreement is due to two opposing influences: The Schaeffler Diagram u n derstates the ferrite content of these
types, as demonstrated by both the
c o v e r e d e l e c t r o d e data 3 a n d t h e
G T A W d a t a ; h o w e v e r , the h i g h

Discussion of V a r i a b l e s in
M e a s u r i n g or Predicting Ferrite
It is quite feasible to identify a n u m ber of variables or groups of variables
involved in the processes of measuring or predicting the ferrite content of
austenitic stainless steel weld metals.
Some of these can also be treated
quantitatively, as has been done previously in this paper for the variations
between individual tests with the
same heat of filler metal. The major
categories, with some background on
e a c h , a r e d i s c u s s e d in i t e m s 1
through 5 below.
1. Test variations, which include
the variables of welding technique,
pad t y p e , c o o l i n g r a t e , s u r f a c e
preparation, and measurement
errors. Many of these variables are interrelated. Earlier in this paper the
rather substantial standard deviations
found by types and overall in the
GTAW and GMAW tests have -been
d i s c u s s e d . C h a n g e s in pad t y p e ,
welding procedure or surface preparation technique could further increase these standard deviations.
A word of caution here: The standard deviations presented for GTAW
do not necessarily correspond to
those which are applicable to covered
electrodes. If the variables are similar
in magnitude, the standard deviations
for covered electrode deposits could
be e x p e c t e d to be lower s i m p l y
because the average ferrite contents
are lower. For the 308 and 309
families the average covered electrode deposits will contain about 6 FN
as against the 10.6 FN average in the
GTAW deposits.
Conclusions on the pertinent standard deviations for covered electrode deposits should await the c o n clusion of the study being conducted
by the Advisory Subcommittee of the
WRC High Alloys Committee. Unpublished data submitted to the WRC
S u b c o m m i t t e e at their N o v e m b e r
1972 meeting indicated that changes
in heat input and welding current over
a rather wide range have little influence on measured ferrite content
of covered electrode deposits, except
that a very low current for the electrode size involved can decrease the
ferrite level by about 1.5 FN, p r o b ably due to somewhat higher nitrogen
pickup than normal. A longer than
normal arc length, of course, can
decrease the covered electrode
f e r r i t e level a p p r e c i a b l y d u e to
nitrogen pickup, as is true with GTAW
and G M A W welding.
2. Variations in the
chemistry

values used to predict the ferrite from


a diagram. Analysis for C, M n , Si, Cr,
Ni, Mo and N is obviously subject to
error which depends upon the skill
and experience of the laboratory and
the care exercised in analysis. A
previous paper 3 discussed this point
and made assumptions which led to a
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 quality chemistry by an experienced laboratory.
This figure seems far too low in
view of a much more recent study 11
based on comparing steel mill ladle
analyses on types 308 and 308L with
electrode producers' receiving analyses. Statistical analysis of this data 11
showed a standard deviation of 0.164
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 ;
together these deviations produce a
calculated standard deviation of 0.90
FN on the revised DeLong Diagram.
The heats involved averaged around
10.5 FN. This later work did not incorporate variations in nitrogen analysis
but probably overstated the variations
due to carbon, since the mill analysis
was often a ladle and not a check
analysis, which tends to be higher
than a ladle. The data from the receiving laboratories involved the average
of multiple tests on a heat, and the
mill ladle analysis was usually also the
average of multiple tests.
In both cases, therefore, the standard deviations cited were lower than
those to be expected from chemical
analysis drawn from a single determination or from analyses obtained
by laboratories less familiar with and
less skilled in stainless steel analysis. Almost all typical chemistry results would have to be considered
poorer than the values cited; i.e., the
influence of errors in chemical determinations on the calculated ferrite
content w o u l d almost always be
greater than those estimated here.
As d i s c u s s e d p r e v i o u s l y w i t h
respect to Table 5, the spacing
between ferrite lines at a high level of
ferrite (about 11%) is larger on the
Schaeffler Diagram than on the DeLong Diagram. As a result, the standard deviations in Chromium and
Nickel Equivalents reported in the recent study 1 1 result in a s t a n d a r d
deviation of only 0.64% ferrite by the
Schaeffler Diagram.
Another publication12 which
describes a method of determining
the ferrite content of stainless steel
castings based upon the Schaeffler
Diagram shows a scatter band of
about 1.8% at an average 5% ferrite
content and about 2.2% at an
average 10% ferrite content "due to
uncertainty of c h e m i c a l analysis
d e t e r m i n a t i o n s for individual elements". It is assumed that the author
intended this to cover a 2 sigma value

Table 21 Summary of GMAW Ferrite Data


(For Explanation of Column Headings See Footnotes Table 9)
Change,
Change,
DeL. FN
Sch.FN
to M - G F N to M-G FN

CrEq

DeL.
NiEq

Sch.
NiEq

DeL.
FN

Sch.
FN

M-G
FN

ER308 12 tests
Range
21.01
to
22.34
Avg.
21.51

14.13
to
15.59
14.66

11.47
to
12.68
11.91

4.8
to
11.0
7.8

8.0
to
10.9
10.1

2.1
to
11.0
7.3

-3.0
to
+ 1.7
-0.5

-6.8
to
+0.3
-2.8

ER347 2 tests
Range
20.61
to
21.15
Avg.
20.88

14.57
to
14.82
14.70

11.70
to
11.96
11.83

4.6
to
5.7
5.2

7.9
to
9.8
8.9

4.1
to
6.1
5,1

-1.6
to
+ 1.5
-0.1

-5.7
to
-1.8
-3.8

ER308L . 6 tests
20.92
Range
to
22.36
Avg.
21.42

13.38
to
14.49
13.88

11.31
to
11.59
11.49

6.8
to
12.1
9.9

10.2
to
14.0
11.2

7.1
to
10.8
9.2

-2.3
to
+ 0.8
-0.7

-4.2
to
+0.6
-1.9

ER309 8 tests
24.67
Range
to
25.61
Avg.
25.11

17.77
to
19.48
18.61

14.71
to
16.48
15.71

6.5
to
13.7
10.4

6.6
to
10.4
8.9

7.8
to
13.0
10.0

-3.5
to
+ 2.0
^0.3

-0.5
to
+2.6
+ 1.1

ER316, 316L 4 tests


Range
22.02
to
22.26
Avg.
22.10

15.89
to
17.11
16.49

13.76
to
14.92
14.41

3.0
to
6.2
4.8

3.7
to
7.2
5.3

3.7
to
6.7
5.0

-0.8
to
+ 1.3
+0.3

-3.5
to
+ 1.6
-0.3

Item

Table 22 Summary of Variances and Standard Deviations from all Heats on which More
Than One Test Was Run
GTAW

Item
Type
316
No. of tests (A)
11
No. of heats (B)
5
A minus B
6
Total variance, FN
4.14
Test Variance, FN
.69
Standard deviation
FN
.83
2.2/
Range of FN on heat
5.7
Avg. FN
5.1
of heats

(a)

GMAW

316L
11
3
8
5.56
.70

All
94
36
58
82.15
1.41

308
40
14
26
36.01
1.39

308L
16
7
9
11.43
1.43

309
14
6
8
24.93
3.10

All
7
3
4
3.68
.92

.83
4.8/
8.2

1.19
2.2/
14.3

1.18
7.5/
13.0

1.20
8.1/
12.0

1.76
6.8/
14.3

.96
5.4/
12.4

6.1

9.4

10.6

10.7

10.4

7.9

(a) Includes one heat of 347. 2 tests

or better.
Overall, a standard deviation of not
less than 1.0 FN for either diagram
appears to be a good figure to use for
chemical laboratories under good
control. This standard deviation
applies to ferrite contents of 10 to 15
FN and could be lower, perhaps
around 0.8 FN, if the ferrite content
were below 6 FN.
3. Variations in diagram construction, due to i n c o r r e c t f a c t o r s for
elements, the approximations in fac-

tors for elements where they may


have different factors at certain levels
t h a n at o t h e r ( i . e . , n o n - l i n e a r
behavior), errors in line location, etc.
Also variations due to residuals not
taken into account in the calculation,
or inherent variations in response by
one heat versus another. We are not
prepared to estimate the extent of
these variations at present.
4. Instrument calibration. As disc u s s e d in the i n t r o d u c t i o n , this
variable has been demonstrated to be

WELDING RESEARCH SUPPLEMENT!

295-s

GMAW

GTAW
DELONG

SCHAEFFLER

SCHAEFFLER

DELONG

1510-

ri

0-

-5

o-

-5

M7

_n DL

riTti,^ LM
0

-5

0 + 5

0-

-5

0 + 5 - 5

5-

, rri.rVri
-5

316
ji
5

-5

0 ^ - 5 - 5

0 + 5 - 5

+5

308L
,
.rrfh.n,

, Vrtjh

-5

+5

-5

+5

309

^T+
5

,
+5

. w
0

+5

L ^
0

Caution

+5

316,

, rrftfTI ,316 L , , JlTTi,

-5

+5

^fi4 ,309, ,itfh

5 -

308,

308,

5-

, ,^n
5

+5

,3I6L, n
+5

fn

+5

40

GTAW
SUMMARY
20

GMAW

rj

The conclusions reached here are


probably not directly applicable to
stainless steel castings, a l t h o u g h
some of the principles may be. More
importantly, the conclusions cannot
be directly applied to wrought stainless steels because such factors as
soaking the billets for rolling, hot
w o r k i n g , solution a n n e a l i n g , etc.,
have very significant effects on the
quantity, size and shape of the ferrite
and therefore on its relative magnetic
response.

,_, SUMMARY

"j-.

", if:

^cf...

+ 5

3 0-

\wu J *

In
+ 5

+5

20-

10-

0
0

tw
+5

-5

+5

BAR GRAPH OF COVERED


ELECTRODE DEPOSITS FOR
COMPARISON. DATA FROM
WORK REPORTED IN REFERENCE
3, SCALED DOWN TO THE
SAME HEIGHT AS GTAW
SUMMARY BAR GRAPH. DATA
FOR DeLONG DIAGRAM
EXPANDED 15 PER-CENT HORIZONTALLY TO ROUGHLY COMPENSATE FOR THE CHANGE
TO FERRITE NUMBER.

COVERED
ELECTRODE
SUMMARY

-5

tion, and may be too severe. The


greatest d i f f e r e n c e between c a l culated and measured ferrite values
obviously occurs when the measured
value and the calculated value are on
opposite sides of the "true" value, but
Fig. 4 treats the data as though the
calculated is the "true" value, which is
obviously incorrect and tends to
overstate the deviations.
On the other hand, the difference
between calculated and measured
must obviously encompass the total
v a r i a t i o n resulting f r o m Items 1
through 4 above. Thus it is logical to
expect that the overall s t a n d a r d
deviation s h o u l d be significantly
larger than any of the standard deviations due to Items 1 through 4 individually.

+5

Fig. 4 Bar graphs showing frequency vs. change from predicted ferrite to measured
ferrite. Plus sign indicates predicted lower than measured, minus sign indicates predicted
higher than measured

a major one in the past8-9 but should


be minor after the WRC instrument
c a l i b r a t i o n p r o c e d u r e 1 0 is m o r e
generally accepted. The WRC A d visory Subcommittee program discussed under 1 will provide excellent
information on this variable.
5. O v e r a l l
variations
from
predicted FN to measured FN. This
has been discussed previously in this
paper. We are not sure at this time
296-s I J U L Y

1973

how to treat the data to derive a


reasonably valid standard deviation.
The differences presented in Tables 9
through 21 are between values determined by entirely different methods,
i.e., direct measurement of FN versus
FN calculated using chemical analysis.
The variations shown on the bar
graphs in Fig. 4 probably represent
the highest possible standard devia-

Conclusions
1. Nitrogen pickup in inert gas welding.
a. Nitrogen pickup is a significant
factor which reduces the ferrite
content of austenitic stainless steel
weld deposits.
b. In GTAW a pickup of 0.02% N
may be assumed; this is equivalent to about 2 FN.
c. In GMAW a pickup of 0.04% N
may be assumed, equivalent to
about 4 FN.
d. The amount of nitrogen picked
up can be quite variable in both
processes, with much higher
pickup possible if the arc protection is poor.
e. Magnetic tests for ferrite c o n tent are the most convenient way to
determine if excessive nitrogen
pickup has occurred.
2. Carbon loss in inert gas welding.
a. Significant carbon losses can
o c c u r w h e n w e l d i n g with filler
metal having normal carbon contents.
b. In GTAW the loss appears to increase linearly as the welding r o d
carbon level increases, from zero
loss at about 0.02% carbon to
about 0.02% loss at 0.06% carbon.
c. The pattern of carbon loss in
GMAW is similar to that in GTAW,
but the amount lost is only about
one f o u r t h the a m o u n t lost in

GTAW up to 0.06% electrode carbon.


3. Variations in FN measurements on
multiple pads of the same filler
metal heat.
a. The s t a n d a r d deviation for
m e a s u r e d F e r r i t e N u m b e r in
GTAW increases as the ferrite content increases over t h e range
studied. At a measured 6 FN the
standard deviation is about 0.8 FN
(Types 316 and 316L) and this increases to a b o u t 1.2 FN at a
measured 11 FN (Type 308 and
308L). An average standard deviation of 1.8 FN was found for Type
309 at about 11 measured FN, but
this deserves more study.
b. Based on much less data than
were studied for GTAW, the standard deviations for GMAW tests
were close to those of the GTAW
tests and the trends were similar.
c. To provide for
test-to-test
variations in measured GTAW and
GMAW ferrite, receiving acceptance tests should allow for variations between the supplier's'tests
and the receiving tests, e.g., if the
supplier is required to meet 5 FN
minimum the customer's receiving test requirement should be
set at 3.5 FN m i n i m u m . Similarly, if
a maximum of 10 or 15 FN is req u i r e d of the supplier the receiving limit should be set at 12.5
or 17.5 FN respectively. These
figures represent approximate two
sigma tolerance levels based on
the data studied.
4. Uncertainties in the determination
of d e p o s i t c h e m i s t r y are s i g nificant variables in the calculat i o n of f e r r i t e f r o m any c o n stitution diagram. From various
sources it is concluded that two
sigma levels of about 1.6 FN at calculated ferrite of 6 FN and less and
about 2.0 FN at 10 to 15 FN calculated are reasonable values for
this variable if excellent chemistry
is available. For average to poor
c h e m i s t r y c h e c k s these values
should be increased appreciably.
5. Comparison of the Schaeffler and
DeLong Diagrams for inert gas
shielded deposits.
a. The diagrams are essentially
e q u i v a l e n t in the d i f f e r e n c e s
between predicted and measured
ferrite in GTAW deposits of Types
308, 308L and 347.
b. The Schaeffler Diagram understates the ferrite content of the
more highly alloyed Type 316,
316L and 309 GTAW deposits. The
DeLong Diagram is better on these
grades, as is also the case with
covered electrode deposits.
c. The DeLong Diagram is better
for GMAW deposits due to higher
nitrogen pickup in those deposits.
d. The differences to be expected

Single (avg)
measured pad
FN compared
to:

between measured and calculated ferrite contents when using


either diagram are greater at the
higher ferrite levels found in GTAW
and G M A W deposits than when
using the DeLong Diagram at the
lower ferrite levels of the covered
electrode deposits studied in the
original work leading to that diagram. However, the differences
are slightly lower than those e n countered with covered electrode
deposits using the Schaeffler Diagram. With GTAW and GMAW deposits a spread of up to about
3.5 FN or % ferrite can be expected between measured and
calculated values, versus about
4.5 FN or % ferrite for covered
electrode deposits using the
Schaeffler Diagram,
e. In types with inherently lower
ferrite content, such as 316 and
316L, the range of differences between measured and calculated
ferrite seems lower than in the
types with inherently higher ferrite
levels, such as 308, 308L and 309.

0-8 FN
Avg measured
FNof5or
more pads of
the batch
Calculated FN
from good deposit chemistry and diagram' 3 '
Schaeffler
covered
bare
DeLong
covered
bare

8-14 FN

(see above 3a & 6)


1.6 FN

2.4 FN

(see above 4, 5,7&


8 but prime data
source is Fig. 4)

5.0 FN
3.3 FN

5.5 FN
4.0 FN

3.3 FN
2.7 FN

4.0 FN
3.3 FN

(a) The estimated tolerances would


probably be lower if the average of multiple measured FN determinations were
used, and would be higher if poorer
chemistry values were used.

6. Direct m e a s u r e m e n t of ferrite,
although it requires some allowance for test variations as described in conclusion 3, offers the
most d e p e n d a b l e a p p r o a c h to
determining the ferrite content of
weld deposits.
7. The variables involved in calculation seem greater in total than the
variables involved in measurement.
8. The reconciling of both calculated
and measured ferrite content
within one frameowrk e.g.,
where a specification requires a
given ferrite range to be met both
measured and calculated requires the acceptance of rather
broad tolerances because standard deviations, both known and
u n d e f i n e d , of both systems of
rating the ferrite content must be
a c c o m m o d a t e d if the r e q u i r e ments are to be met on a practical
and workable basis. The data show
that the total tolerance in such
case may have to be as high as
3.4 to 4.0 FN for GTAW and
GMAW deposits averaging around
11 FN.
9. Conclusion 8 must not be interpreted to mean that the ferrite content of a weld deposit cannot be
measured or calculated to closer
limits than 4 FN. Pending the
completion of the WRC cooperative program on covered electrode deposits, which will also
cover the influence of pad size and
preparation, we are estimating the
following overall tolerances for
m e a s u r e m e n t / c a l c u l a t i o n . The
base is the average FN as measured on a welded pad of the type
defined in this paper.
WELDING

Est. 2 sigma (95%) tolerance where the avg


ferrite content is:

References
1. DeLong, W. T., "Discussion of the International Testing Program for the Determination of Ferrite Content in Austenitic
Stainless Weld Metal", IIW Document ll-C331-70.
2. Schaeffler, A. L., "Constitution Diagram for Stainless Steel Weld Metal",
Metal Progress, 56, 680 and 680-B (November 1949).
3. DeLong, W., Ostrom, G., and Szumachowski, E., "Measurement and Calculation of Ferrite in Stainless Steel Weld
Metal", The Welding Journal, 35 (11),
Research Suppl., 526-s to 533-s (1956).
4. Simpkinson, T. V. and Lavigne, M. J.,
"Detection of Ferrite by Its Magnetism,"
Metal Progress, 55, 164-167 (February
1949).
5. Simpkinson, T. V., "Ferrite in Austenitic Steels Estimated Accurately," Iron
Age, 170, 166-169 (Dec. 11, 1952).
6. Fleischmann, W. L., "Determination
of Ferrite in Type 347 Stainless Steel
Weld Deposit", The Welding Journal,
33 (9), Research Suppl., 459-s to 468-s
(1954).
7. Gunia, R. B. and Ratz, G. A., "The
Measurement of Delta Ferrite in Austenitic Stainless Steels", WRC Bulletin 132,
August, 1968.
8. DeLong, W. T., "Measurement and
Effects of Ferrite in Stainless Steel Weld
Metal, Subcommittee Meeting of Nov. 10
in New York" (Advisory Subcommittee of
the High Alloys Committee of the Welding
Research Council), IIW Doc. II-C-372-71.
9. Rosendahl, C-H, "Ferrite in Austenitic Stainless Steel Weld Metal; Round
Robin Testing Programme 1971-1972",
IIW Doc. 11-631-72.
10. "Calibration Procedure for Instruments to Measure the Delta Ferrite Content of Austenitic Stainless Steel Weld
Metal", Welding Journal 52 (2), Research
Suppl., 69-s to 72-s (1973).
11. Hebble, T. L., "Statistical Treatment
of Chemical Analysis Variability", Supplemental minutes of November '72 WRC
Advisory Subcommittee minutes and
phone conversation of 1/16/73.
12. E. A. Schoefer, "Welding of High
Alloy Castings", Steel Founders' Society of
American, Publication 772-5M1.

RESEARCH S U P P L E M E N T !

297-s

You might also like