2003 Keith Millis Symposium on Ductile Cast Iron

Casting Defect Analysis Procedure and a Case History

A. Alagarsamy
Citation Corporation, Birmingham, Alabama
ABSTRACT

Foundries are still using trial and error methods to solve The Ductile Iron Society undertook a study to do just
casting problems. There are benefits to using a more that. It was determined that silicon level has a significant
disciplined approach to define, identify and determine effect on the internal porosity in unfed sections of the
the root cause of a defect. Use of international standard castings.
defect codes for classifying the defects is illustrated.
Powerful techniques such as defect mapping, CASTING DEFECT CODES
questioning to narrow down the root causes and design
of experiments to identify and control the variables are The International Atlas of Casting Defects handbook
explored. An actual case history in solving shrinkage divides the defects into seven major categories. The
porosity is described to illustrate these techniques in entire defect codes and their names are given in the
practical use. appendix. New names were added shown in italics. In
addition to the codes listed in the handbook, AFS
INTRODUCTION committee 4E has added codes at the end of the current
codes to further identify the material of the casting such
A tremendous amount of productivity is lost through as gray iron, aluminum, etc. and to identify the process
defective castings. By employing a disciplined approach by which the castings were made. Use of these codes,
to understand the nature of defects and the mechanism in addition to in-house names or codes, will enable
of defect formation and controlling the key process foundrymen to use data from one foundry to another
variables we can significantly reduce the scrap both in- without any confusion.
house and at the customer. Misidentification of defects
leads to costly scrap, lost time and customer MAJOR DEFECT CODES
dissatisfaction. It is not uncommon to have different A – Metallic projections
names for the same defect, making it difficult to compare B – Cavities
the successes and failures of other foundries in solving C – Discontinuities
the same problems. The AFS has published a book D – Defective surface
standardizing the names of almost all the defects seen E – Incomplete casting
in the industry, but the foundries have not yet adopted F – Incorrect shape and dimension
the standardized codes and names. Even if in-house G – Inclusions and structure
names unique to the foundry are used they can be
cross-referenced with the published codes to help ALLOY CATEGORIES
disseminate the knowledge for solving casting defect GI – Gray iron
problems. DI – Ductile iron
CS – Carbon steel
Shrinkage porosity revealed during machining plagues SS – Stainless steel
the customers as well as foundries. Foundries are now OF – Other ferrous
increasingly required to pay for the machining expenses AL – Aluminum
for scrapped castings. Shrinkage porosity occurs MG – Magnesium
sporadically making it difficult to determine the root CU – Copper
cause. It is very difficult to verify in the foundries by ZN – Zinc
using NDT techniques due to complex shapes, and the OA – Other non ferrous
cost or time required to test the castings.
PROCESS
BACKGROUND GS – Green sand
BK – Baked sand
Ever since the start of production of ductile iron castings, NB – No bake
foundries have been working on understanding the CB – Cold box
effect of process variables on the shrinkage formation HB – Hot box/shell
and controlling these variables to minimize defects. PM – Permanent mold
DC – Die casting
2003 Keith Millis Symposium on Ductile Cast Iron

LF – Lost foam also have a lot of information collected through the years
IV – Investment casting in helping solve foundry problems. It is important to have
VP – ‘V’ Process good knowledge of the mechanics of casting defect
CN – Centrifugal formation, probable locations and the nature of the
IG – Ingot defects. In some difficult cases, analytical tools such as
CC – Continuous cast the SEMs are used to pinpoint the nature and cause of
OM – Other methods the defects. One should use any and all tools available
to make sure the defect is identified correctly.
A fully coded defect will appear as B214-DI-GS, which
will be read as shrink at the riser contact in ductile iron GATHERING DATA – INFORMATION
castings made in green sand molding process. This
information is available at the website given below. In solving casting quality problems, we must start with
http://www.castingsolutions.com reliable data for key process variables that affect the
quality of the castings. Data gathered in the foundries
As we have come up with new processes and new should be reliable. Gage R and R for the measuring
materials, we also have created new types of defects equipment should be in the acceptable range. For
with new names not covered in the present atlas. Some evaluating casting defects, instead of having two
of the new defect names that are considered, for the categories (acceptable castings and rejects), a range of
newer materials and processes and to classify them in quality standards should be established with a
more detail are listed here. reproducible measuring system, e.g. 1 through 5, 1
zVermicular graphite denoting no defects and 5 representing worst case
zChunk graphite defect. Once a valid measurement system is established
zMesh graphite then any experimentation can be undertaken to study
zExploded graphite the effect of variables on the casting quality.
zInverse chill carbides
zGrain boundary carbides Routine documentation of all process data should be
zGas voids with graphite layer easily obtainable for analysis of conditions causing the
zGas void with oxide layer defects. More frequent data gathering may be necessary
zBlister to zero in on the causes of defects. Any data that are
zDoughnut (fisheye) used for analysis should be dependable.
zLeaker
zRiser break-in DEFECT MAPPING
zShrink at riser contact
It is very important to analyze the defects by carefully
cataloging details of the defects. It will be very
IDENTIFICATION OF DEFECTS informative when the defects are plotted on a pattern
layout diagram. The details will indicate how many
Most of the defects seen in the foundries are easy to defects of one kind occur at a particular location of a
recognize and identify by the correct defect name and cavity in the casting, cope or drag and relative location
code. There are a few defects that may appear to be with respect to a gating schematic. Defects
similar but may be entirely different types. If we do not documentation should also contain the time the castings
correctly diagnose these defects we may get into more were made, especially for persistent types of defect.
problems trying to correct something or change From this kind of detailed mapping it is easy to see
something that may not be necessary. It may be better where the defects are mainly located, what type of
to keep the defect as neutral prior to classifying into a defects and which cavities are prone to the defects
more specific defect. If distinction could not be made under consideration. For shrink type of defects,
between gas defect and shrinkage defect, then it should information about the risers (how well it is functioning-
be kept as porosity until it is clearly determined to be piping) should also be documented.
either shrink or gas defect. Ambiguous defects include
inclusions (sand or slag). There are several places From the defect mapping information and process data
information is available to help determine the true defect during high scrap times and low scrap times, useful
name and code. Every foundry should have known contrast information can be derived. By asking and
defect standards that have been identified previously for answering questions listed below, one can remove
reference and training new personnel. Defect processes not contributing to the defect. By eliminating a
handbooks, in-house manuals, and the internet are a broad process, quite a few variables need not be
few of the sources that are readily accessible. Suppliers considered thus simplifying the work ahead. One should
2003 Keith Millis Symposium on Ductile Cast Iron

be careful not to eliminate variables that may be riser, at the location shown, Figure 2. shows two
influencing the defect formation. It should also be noted different locations where a riser is attached for different
that there may be interaction between variables as well cavities, due to layout of patterns for maximizing number
as design features. of cavities in the mold.
What else is known:
¾ Porosity shows up after machining the casting.
Processes not
When the defect is: ¾ Porosity is three times more when the riser is at
contributing to the
location ‘B’ than when the riser is at location ‘A’.
defect
¾ Occassionally in some heat codes there is a spike
In only one of several Melt, molding, pouring in the defect severity.
cavities core, ¾ Overall, the incidence of porosity is low.
Shrink at riser connection Sand, molding, core
Shrink at hot spot Sand, molding, core
Tear-up, stickers Core, melt, pouring
Crush at certain times Core, tooling, melt, pouring
Misrun at start of a shift Tooling, sand, molding,
core
Sand at the bottom Melt, pouring
Table 1. Examples of elimination of processes not
contributing to the defect

Prior to investigating casting defects, one should gain

knowledge about the nature of defects, the appearance
and possible location of the defects and the mechanism
of defect formation. During the investigative phase, by
asking and answering pertinent questions, more insight
into the defects can be gained
.
Some of the questions for root cause analysis:
¾ Where are the defects?
o Drag, side walls, above or below core Porosity
o Closer to gates or farther away from gates
o At the surface, interior or sub surface
¾ Are the defects found at the same location?
¾ Are the defects occuring all the time?
¾ Are the defects people related? Figure 1. Schematic view of cut view of casting
¾ Are the defects time related?
¾ Are the defects cavity related?
¾ Do these defects occur in other similar jobs? Type
Rise Type 2,
(tw
locatio
locati cavities
Further questions as to related defects can shed light as cavities
to the causes of defects.
Are there related defects in the same castings?
o Misrun, short pour, run-outs, laps
o Rat tails, buckles, scabs
o Rough surface, burn-in, penetration
o Swells, erosion, stickers
o Shrinkage, gas porosity, leakers Location B Location A
When related defects occur in the same castings, it may
indicate a system related problem.

We are going to look at a case history of a porosity

defect and the procedure that was followed to reduce
Area of
the defect occurrence. The schematic view of a
sectioned casting is shown in Figure 1. Porosity occurs Figure 2. Alternate locations for riser
sporadically in a heavier section not fed directly by a
2003 Keith Millis Symposium on Ductile Cast Iron

ANALYSIS OF THE DEFECT design and riser design is marginal, then shrinkage
defects may exhibit different forms.
By the appearance (dendritic) and the location (below
the core) and in the center of the isolated heavy section,
it was concluded this was a shrinkage defect. This
defect is remote area (centerline) porosity. The
international code for this defect is determined to be
B222 – Centerline or axial shrinkage porosity. As the
section where the defect occurs is isolated from the riser
by a thinner section of the casting, the riser is not able to
compensate for the shrinkage that occurs at the end of
solidification of the isolated section. Solidification models
predict there will be shrinkage at this location. Even
then, over 98% of the time the castings are sound. Most
of the time the expansion due to graphite precipitation
compensates for the shrinkage occuring at this late
stage of the solidification process.

Figure 4. Riser piping behavior

Figure 3. Example of micro shrinkage in the center Exudation

of unfed area of a casting

When there is microshrinkage, there has been

insufficient graphite expansion at the end of solidification
to compensate for shrinkage due to austenite formation.

EXPANSION AND SHRINKAGE

Expansion and shrinkage occcurs simultaneously during

solidification of a casting. In many instances expansion
due to graphite precipitation pushes liquid metal into the Figure 5. Exudation of metal is clearly seen into the
voids or into the free surface of a riser. This can readily riser piping cavity.
be seen by observing the piping behavior in the risers as
shown in figures 4 and 5 There is absence of Evidence of exudation is clear in cases where the metal
continuous piping in most cases, which indicates at was semisolid and solidified very quickly maintaining its
different times expansion is greater than the shrinkage, shape. There are some rare instances where the
pushing liquid metal into the riser. Process variations exudation is into a primary liquid shrinkage cavity inside
within mold and from mold to mold affect the rate and a casting. One such instance is shown in figure 6.
relative values of shrinkage and expansion. If the casting
2003 Keith Millis Symposium on Ductile Cast Iron

¾ Silicon level
¾ Precipitation rate and timing of graphite
nodules

Factors influencing the volume of liquid metal:

¾ Pouring temperature
¾ Green sand heat conductivity-density
¾ Core sand heat conductivity
¾ Mold quality – mold wall movement
¾ Pouring rate

Design of casting
¾ Core length-affecting mass at hot spot

The top ranked factors listed in bold letters were thought

to influence the shrink most, and warranted further
Figure 6. Exudation of semisolid metal into a void in study. The graphite precipitation rate and timing is an
the center of a casting important variable but is not directly controlled like the
other variables. Graphite nodule precipitation is
In the case study under discussion here, the factors influenced by several other variables listed and/or by
inherent in the design of both the casting and gating inoculation effectiveness. For instance, over-inoculation
makes the casting susceptible to shrinkage porosity. or very high carbon equivalent can result in primary
Since a very small percentage of castings have defects graphite precipitation, which results in duplex nodule
and they occur in spikes, it suggests that there are some size distribution. There is some evidence this type of
significant variables or combination of variables that graphite nodule distribution leads to increased
influence the shrinkage porosity. shrinkage. A DIS research project at CANMET showed
that excess inoculation increases shrinkage.
The fact that riser location ‘A’ figure 2. results in much
less frequency of shrinkage defects than riser location Also from this study and other experiences, we know
‘B’ suggests that casting design has a significant effect. that an increased magnesium residual level for the
By looking at the results of casting solidification section thickness and final sulfur increases shrinkage.
simulation, we can detect that the feeding path is open Increased level of silicon contributes to increased
for a slightly longer duration for the ‘A’ location than for shrinkage. Silicon level is also controlled to minimize
the ‘B’ location. This means that when the feeding path carbide formation. There may be less freedom in some
is cut off, the remaining volume of liquid in the isolated shops to lower the silicon level, due to common iron
area is slightly smaller with ‘A‘ riser than with ‘B’ riser. being poured for many jobs, some of which require
This fact then leads us to conclude that the variables higher silicon to control carbides and/or pearlite levels in
which influence the amount of liquid remaining in the the casting. In some cases a minimum silicon may be
isolated area have a direct effect on the shrinkage specified by the customer for operational benefits.
defects. The other set of variables that will affect the
shrinkage are metallurgical,variables which affect DESIGN OF EXPERIMENTS
graphite precipitation especially at the end of
solidification. Considering all the above factors it was decided to
conduct a design of experiment to determine the
VARIABLES variable factors that influence shrinkage the most. The
first set of variables considered for the study are:
Prior to running design of experiments, a matrix study 1. Magnesium residual level
was done to determine the key variables that in this case 2. Core design
will affect the shrinkage porosity. Listed below are the 3. Pouring temperature
variables considered to have effect on the shrinkage. From the feedback from the customer and analyzing the
internal process data, a strong correlation was found
Metallurgical factors that affect the shrinkage: that pointed to the magnesium residual level contributing
¾ Carbon equivalent to an increase in shrinkage defects on some days.
¾ Cumulative level of forward segregating elements
¾ Magnesium and cerium residuals
¾ Level of base iron nucleation
2003 Keith Millis Symposium on Ductile Cast Iron

Core design was chosen as a factor that affects the

volume of liquid metal in the hot spot area. By
lengthening the core, liquid mass could be reduced. It
was expected that the longer core will result in reduced Table 3. Shrinkage severity response
shrinkage.
Short cores Long cores
Pouring temperature was selected as another variable
Low High Low High
due to its influence on the volume of liquid metal
temp temp temp temp
remaining when the feed path is cut off. Generally in
larger casting sections lower temperature seemed to Low
0.7 0 0 0
help minimize shrinkage. Mg
110 35 101 31
Hi Mg
A full factorial design was used to determine if there
were any interactions between the three variables
selected. A set of 8 heats were poured as shown in At lower magnesium levels there was no shrinkage
Table 2, Numbers 1 through 8 show the sequence when seen. At higher magnesium levels, temperature has an
the heats were poured. effect, with higher temperature reducing the amount of
shrinkage. Core design did not affect the shrink. The
Table 2. Three factor full factorial experiments major effect is Mg, followed by temperature. Interaction
between magnesium and temperature can be seen.
Short cores Long cores
Low High Low High After the study the process variables for magnesium
temp temp temp temp residual and pouring temperature were adjusted to a
more favorable range. This change resulted in a
Low significant improvement in the level of defects found
1 5 3 7
Mg after machining. Even with the reduced defect severity,
2 6 4 8 porosity still occurs at the same pattern/location as
Hi Mg
before. As we know, the silicon level also has an
Castings from these heats were radiographed and rated influence on shrinkage, so the silicon level is also
by two people as to the severity of shrinkage. Shrinkage adjusted lower, while monitoring for an increase in
was rated from 0 to 5, 0 being no shrinkage and 5 being carbides.
the worst case observed. The points for each condition
were totalled and used as a response from the We continue to audit the castings with radiography and
are not seeing any defects in the radiographs. We are
experiments for analysis. See Table 3. and Figure 7.
still investigating other variables such as green sand
density and core sand density that affect the volume of
liquid when the feed path is cut off. .
250
SUMMARY
200
A systematic approach to defect identification and
150 L.Mg analytical techniques to find root causes is explained in
detail. One example of a defect case history is shown to
100 H.Mg explain the procedure that was followed to identify and
resolve the problem.
50
ACKNOWLEDGMENTS
0 The author is grateful to Citation Corporation for
L.temp H.temp permission to present this paper. Thanks also to Mike
Barstow, Gene Muratore and Rosa Armstrong for help in
editing the manuscript.
Figure 7 : Total shrinkage porosity at low and high
levels of Mg and temperature
2003 Keith Millis Symposium on Ductile Cast Iron

REFERENCES B 222 Centerline or Axial Shrinkage Porosity

1. International Atlas of casting defects. AFS B 223 Center line porosity at isolated hot spots*
publication B 300 Porous structures caused by many small cavities
2. Analysis of casting defects. AFS publication
B 311Macro,Micro,shrinkage porosity-leakers
3. Citation Corporation in-house publications
4. Sillen, Rudolf., Shrinkages in Iron Castings,
Communication to Ductile Iron Society, 2002 C - Discontinuities
5. Sparkman, David., “Offsetting Macro-Shrinkage in C 100 Discontinuities caused by Mechanical effects
Ductile Iron” DIS News, Issue 3, 2001 C 111 Breakage (Cold)
C 121 Hot Cracking
C 211 Cold Tearing
Appendix 1. C 221 Hot Tearing

Defect Codes From International Atlas of Casting C 300 Discontinuities caused by lack of fusion.
Defects Handbook C 311 Cold Shut or Cold Lap
C 321 Interrupted Pour
A - Metallic Projections C 331 Cold Shut (At Chill or Insert ) Unfused Chaplet
A 100 Metallic projections C 400 Discontinuities caused by
in the form of fins or flash metallurgical defects
A 111 Joint flash or Finning C 411 Conchoidal or Rock-Candy Fracture
A 112 Veining or Finning C 412 Intergranular Corrosion
A 113 Heat Checked Mold or Die
A 114 Fillet Vein D. Defective Surface
A 121 Cope Raise, Raised Mold D 100 Casting surface Folds, Gas Runs
A 122 Sag or Strain D 112 Cope Defect, Elephant Skin, Laps
A 123 Cracked or Broken Mold D 113 Seams or Scars
A 200 Massive projections D 114 Flow Marks
A 211 Swells D 121 Rough Casting Surface
A 212 Erosion, Cut, or Wash D 122 Severe Roughness, High-Pressure Molding
A 213 Crush D 131Buckle
A 221 Mold Drop or Sticker D 132 Rat Tail
A 222 Raised Core or Mold Element, Cutoff D 133 Flow Marks, Crow Feet
A 223 Raised Sand D 134 Metal-Mold Reaction, Orange Peel
A 224 Mold Drop D 135 Soldering, Die Erosion
A 225 Corner Scab D 141Sink Marks, Draw, Suck-In
A 226 Broken or Crushed Core D 142 Slag Inclusions
D 200 Serious surface defects
B - Cavities D 211 Push-Up, Clamp Off
B 100 Cavities with generally rounded D 221 Burn On
smooth walls detectable to the naked eye D 222 Burn In
B 111 Blowholes, Pinholes D 223 Metal Penetration
B 112 Blowholes near inserts, chills, chaplets D 224 Dip Coat Spall, Scab
B 113 Slag Blowholes D 231 Scab, Expansion Scab
B 121 Surface or Subsurface Blowholes D 232 Cope Spall, Boil Scab, Erosion Scab
B 122 Corner Blowholes, Draws D 233 Blacking Scab
B 123 Surface Pinholes D 241 Oxide Scale
B 124 Dispersed Shrinkage D 242 Adherent Packing Material
B 200 Cavities with rough walls, shrinkage D 243 Scaling
B 211 Open or External Shrinkage
B 213 Core Shrinkage
B 214 Shrink at riser contact*
B 221 Internal or Blind Shrinkage
2003 Keith Millis Symposium on Ductile Cast Iron

G 214 Ferritic Skin*

E. Incomplete Casting G 221 Primary Graphite White Iron
E 100 Missing portion of casting (no fracture) G 222 Excessive Pearlite Layer
E 111 Misrun G 223 Localized Hard Spots, Inclusions
E 112 Defective Coating (Tear-Dropping) or Poor Mold Repair G 224 Flake Graphite*
E 121 Misrun (a) G 225 Chunk graphite*
E 122 Poured Short G 226 Exploded Graphite*
E 123 Runout G 262 Kish Graphite Inclusions
E 124 Excessive Cleaning G 263 Carbon Floatation
E 125 Fusion or Melting during Heat Treatment G 264 Faceted (Dendritic) Fracture
E 200 Missing portion of casting (with fracture) * additional codes
E 211 Fractured Casting
E 221 Broken Casting (At Gate, Riser or Vent)
E 231 Early Shakeout

F. Incorrect Dimensions or Shape

F 100 Incorrect dimensions, correct shape
F 111 Improper Shrinkage Allowance
F 121 Hindered Contraction
F 122 Irregular Contraction
F 123 Excessive Rapping of Pattern
F 124 Mold Expansion during Baking
F 125 Mold wall Movement, Mold Cavity Enlargement
F 126 Distorted Casting
F 200 Casting shape incorrect or in certain locations
F 211 Pattern Error
F 212 Pattern Mounting Error
F 222 Shifted Core
F 223 Ramoff, Ramaway
F 231 Deformed Pattern
F 232 Mold Creep, Deformed Mold, Springback
F 233 Casting Distortion
F 234 Warped Casting

G. Inclusions
G 100 Inclusions or Structural Anomalies
G 111 Metallic Inclusions
G 112 Cold Shot
G 113 Internal Sweating, Phoshphide Sweat
G 121 Inclusions of Slag, dross or Flux: Ceroxide
G 122 Slag-Blowhole Defect
G 131 Sand Inclusions
G 132 Blacking or Refractory Coating inclusions
G 141 Black Spots
G 142 Oxide Inclusions or Sinks, Seams
G 143 Lustrous Carbon Films, Kish Tracks
G 144 Hard Spots
G 200 Structural anomalies-macroscopic
G 211 Primary Chill, Chilled Spots or Edges
G 212 Unmottled Chill, Clear Chill
G 213 Inverse Chill*