Back to Basics

[ click here for the Back to Basics Archive ]

Limitations on the Detection of Casting Discontinuities Using Ultrasonic and Radiography

by Stuart Kleven* and Malcolm Blair⁺

 
INTRODUCTION

The testing of castings can be a costly process in relation to the cost of the casting itself. The need for testing is dictated by the end use of the casting. Is it a critical application where human life is at risk? Or is it an application where failure would cause expensive downtime or require costly maintenance? Designers of equipment or systems usually specify the amount of testing required. Each industry has specifications and acceptance criteria developed around the type of product manufactured under those specifications or standards. For instance, castings supplied for military applications are normally ordered to AMS-STD-2175. This has four classes of application and four quality grades. Class one has the highest rating and requires 100% radiography as well as a surface examination, such as magnetic particle or liquid penetrant testing. Subsequent classes require sampling and have safety margins that are lower since they may not have a critical application.

The testing of castings is a tremendous undertaking because there are so many variables to consider. Typically, the type of testing may be set by the design activity or customer. Sometimes the tests that are specified may not be the best tests for the component. Consideration as to what types of discontinuities are anticipated should dictate the type of testing. For instance, radiography and ultrasonics are volumetric examinations. This means that they examine the entire thickness of the part. However, they may not be the best examination for surface discontinuities. Magnetic particle and liquid penetrant testing are excellent for surface examinations, but cannot detect discontinuities below the surface. Each method has its advantages and disadvantages. Knowing how and when to balance these out is the key to successfully testing castings. The process starts in the inquiry stage before an order is received. Casting houses should be familiar with the capabilities and detection limits for the various test methods. If a customer specifies a method that will not detect the type or degree of discontinuity indicated, exception should be taken during the time of quotation. Companies producing castings should develop a history that will help predict the type and location of discontinuities. This can be a valuable aid in determining test method options.

---

The testing of castings is a tremendous undertaking because there are so many variables to consider.

---

 

ULTRASONIC TESTING

Uses and Potentials
While radiography is the generally accepted test method for castings, ultrasonics may have some advantages. There may be circumstances where orienting a casting properly for radiography is impossible due to configuration or accessibility. In such cases, ultrasonics may be specified. Ultrasonic testing has gained in popularity because it is environmentally friendly. This means that a particular area does not have to be isolated to protect personnel, as is the case in radiography, where radiation must be contained. Ultrasonic testing relies on the propagation of ultrasonic sound waves through the full thickness of the material. With plain carbon and alloy steel materials (martensitic, pearlitic, ferritic and so on) wave penetration through thicknesses up to 0.3 m (12 in.) is no problem. With austenitic stainlesses, attenuation of the sound beam occurs and penetration of parts with thicknesses of 51 to 76 mm (2 to 3 in.) can be a problem. When grains are large or nonuniform in size and distribution, they can scatter the sound waves and cause attenuation. The noise produced can also mask true indications since the signal produced by the noise often is 40 to 80% of screen height. Other problems result from surfaces that are not parallel to each other. It is difficult to monitor the back reflection for discontinuities that may not necessarily produce a signal with any relative screen height but are severe enough to be considered nonconforming (which produce a loss of back reflection greater than the standard permits).

Scanning in two directions may not produce the desired results either. Contact testing has difficulties with rough surfaces and working in the transducer near zone. Interpretation can be difficult as well. Large castings may need to be turned over to access surfaces or areas that cannot be viewed from the upward surface. This can create handling problems and add to the cost. Parts with multiple thicknesses may produce a variety of indications on the screen that may or may not be relevant to the test. These may be echoes that return later in time due to beam spread and from mode converted signals that result from longitudinal waves being converted to shear waves. Parts that are rough or have surfaces which are as cast or machined surfaces with a rough finish similar to a record album can be difficult to test because proper or consistent contact with the part's surface cannot be attained. On rough surfaces, loss of back reflection is common. The problem lies in determining whether this loss is due to a discontinuity, nonuniform grain structure or simply the rough surface. A rule of thumb for determining surface requirements is as follows: for a flat bottom hole size less than 1.6 mm (0.06 in.) to be detected, the surface finish must be 32 to 63 root mean square; for a flat bottom hole size of 2 to 3.6 mm (0.08 to 0.14 in.) to be detected, the surface finish must be 63 to 125 root mean square; and for a flat bottom hole size greater than 4 mm (0.16 in.) to be detected, the surface finish must be 125 to 250 root mean square.

To properly test a discontinuity, a calibration or reference standard must be made. This is usually made from the same or similar material as the part and with the same heat treat condition. A range of standards must be machined or have electric discharge machined holes or notches placed in them. This is a costly process and should not be considered unimportant. Near surface detection requirements should be accounted for if the part is to be machined with little material removed. Standards material should be tested for acoustic compatibility, thickness and freedom from discontinuities before machining flat bottom holes or other standard test shapes, such as side drilled holes or notches.

 

ASTM A609 Ultrasonic Examination
A thorough search for specifications discussing ultrasonic testing on castings produced one that is frequently used. It is ASTM A609, Standard Practice for Castings, Carbon, Low Alloy, and Martensitic Stainless Steel, Ultrasonic Examination Thereof. As can be noted from the title, austenitic and other nickel based alloys are not included. This is due to the difficulty in penetrating these types of materials with adequate sound waves. This standard does provide a clue that there are difficulties when processing castings. Under the paragraph entitled "Personnel Requirements," it indicates that the technician should be aware of the effects of several things on test results, including transducer material, size, frequency and mode, material structure (grain size, cleanliness and so on), test distance, nonlinearity, thickness and orientation of discontinuities and surface roughness.

Although this provides guidance into the type and effectiveness of the ultrasonic examination, persons specifying and personnel performing the test seldom read this portion of the document ahead of time. In addition, the standard flat bottom hole test blocks specified are at metal paths of 25 mm (1 in.), 51 mm (2 in.), 76 mm (3 in.), 152 mm (6 in.), 254 mm (10 in.) and so on. This does not account for any voids in the near surface that can easily be missed due to surface roughness. The standard does address the use of dual element transducers with flat bottom holes at 3.3 and 10 mm (0.13 and 0.5 in.) or at half thickness and 20 mm (0.75 in.) using a 2.3 mm (0.09 in.) diameter flat bottom hole for near surface. The document also describes a means of transfer mechanism to match attenuation from the test block to the casting as well as adjustment for surface roughness. Two methods are addressed: procedures A and B. Both are fairly detailed and require significant effort to calibrate and adjust for the material attenuation and for sizing of discontinuities based on area.

 

Testing of Ultrasonic Indications
Discontinuity type, shape and orientation should be understood if a successful test is to be conducted. Each type of discontinuity produces a different type of profile. Based on this profile, the discontinuity will either be detectable, undetectable or slightly detectable. Sound waves may need to be manipulated to a specific direction or angle to improve detectability. In castings, the lack of homogeneity in grain structure can limit the use of ultrasonics to compressional wave, straight beam scanning. Shear (transverse) wave or angle beam testing is rarely employed on castings due to this microstructural condition. This would be the very instance where it would be valuable in producing sound waves that are oriented perpendicular to the discontinuity. Cracks can be very difficult to detect using ultrasonics when the orientation is unfavorable. The closer the angle at which the sound waves interact with the plane of a crack or cracklike discontinuity is to being perpendicular, the greater the amplitude of the signal on the screen of the ultrasonic unit. Gross planar discontinuities are readily detectable if the sound strikes the face of the discontinuity, but almost undetectable if it hits on the edge of the discontinuity.

Shrinkage is another problem that can, at times, be difficult to test for and produce a signal on the screen. The irregular shape causes sound waves to reflect in different directions or reflect back at different points in time. This usually produces a wide, broad based indication with numerous small spikes. It can easily be confused with noise or rough surface. Depending on the density of the shrinkage, it may not cause a drastic loss of back reflection. Gas holes are probably the easiest to detect. They normally produce a sharp signal and the signal can be maintained until the sound beam moves completely off the void. Multiple pores in a concentrated area, however, can cause confusion during interpretation, since interference signals can occur due to reinforcement or reduction of the sound wave. In summary, operator skill, a proper surface finish, valid calibration standards and scans from multiple surfaces can improve detectability.

 

RADIOGRAPHIC TESTING

Uses and Potentials
Radiography has long been the method of preference for testing castings for critical service. Radiography may have been selected due to the fact that a hard copy (film) is produced and can be viewed by a number of different inspectors as necessary for interpretation and testing. The image on the film is a result of the interaction of the radiation with the material under examination and the silver halide crystals on the film. Radiography has inherent dangers because the radiation produced can have a detrimental affect on workers if they are overexposed to it. A means of isolating employees and testing personnel from the radiation can be a costly project. The alternative is to have a lot of space and ensure enough distance (2 mR in any one hour boundary) between workers and the radiation. Radiographic equipment costs are high, as well as the need for governmental/state licensing if a radioactive source is needed. Supplies such as chemicals, lead screens and film must also be purchased on a regular basis.

Equivalent steel thicknesses of up to 89 mm (3.5 in.) can be radiographically tested using an X-ray machine. Above this thickness, radioactive isotopes such as Ir-192 and Co-60 are commonly used. Specialty units, such as linear accelerators or betatrons, are used to examine heavier materials. These units are not uncommon in industry use but are expensive to purchase, provide shielding for and maintain operability. Parts handling can also be a problem with heavier castings since they need to be rotated to provide access for different views. Another problem that can be encountered is the fact that most castings are tested in the "as cast" state. Machining and further processing can reduce wall thicknesses and discontinuities may be present afterward. A discontinuity may have been less than 2% of the wall thickness as cast, however it may now be detectable using radiography after machining. Parts requiring clean surfaces (with no discontinuities evident) after machining should be tested after a preliminary machining operation to prevent passing nonconforming parts. Many codes and standards address this very situation and specify the thickness for final testing using radiography when a significant amount of material will be removed.

 

Radiographic Testing of Indications
Similar to ultrasonics, radiography has limitations with regard to discontinuity detection based on the orientation. Cracks and planar discontinuities must be oriented parallel to the radiation beam since they are thin and do not produce enough change in film density to be detectable. This is just the opposite of ultrasonics, which relies on sound waves striking the planar surface to be detectable. Radiography relies on detecting a change in density based on 1% to 2% of the part thickness. This is evident in the radiographic call out in specifications for 1-2T or 2-2T. This means that a 1% or 2% change in thickness can be detected and a 2T hole is discernible on a test plaque called a penetrameter. The T represents the thickness of the penetrameter.

Various factors can affect the detection of discontinuities. Surface roughness or depressions could be interpreted as discontinuities. The part surface should be viewed, if possible, to verify whether or not a surface type indication is present on the film. Undercutting can also lower detectability. Undercutting is caused by excessive radiation reaching the film due to scattered radiation. This can cause the outer edge of the image to be overexposed and not visible on the film. The type of material can also cause radiation to scatter or attenuate. This can appear on the film as an indistinct indication that is difficult to interpret. A classic example called diffraction mottling frequently occurs with austenitic or high nickel base materials. Grain boundaries from large grains are oriented in such a manner as to cause a reflection of the radiation. These appear as light or dark spots similar to shrinkage or inclusions in the material. A second shot at a higher radiation value or with a part shift of about 5 degrees will usually allow the interpreter to differentiate between an actual discontinuity and diffraction. This can be costly because the number of exposures could double.

Shrinkage is often found in castings to varying degrees. It is difficult, even for experienced radiographic interpreters, to be exacting when it comes to interpreting shrinkage. Shrinkage can appear almost anywhere and can be any size and density. Various classes or reference films exist that depict differing degrees of shrinkage for each type, such as dendritic, sponge, filamentary or cavity type. Production radiographs seldom match the reference radiographs in density, area or size. For this reason, the interpreter must try to match the indication to the closest type in appearance. The extent of the indications within the area prescribed by the reference radiograph must be considered as well as the density. Some leeway must be given due to interpretation. A casting may have more of a particular type of shrinkage, but it may be much less dense than the standard. One interpreter may take liberties and adjudge the casting as acceptable, while another may take a more strict interpretation and consider it unacceptable. Effective specifications, where none exist or where there is room for error in judgment, should eliminate most of these situations. If a customer does not have clear cut definitions on discontinuity spacing, size, area of interpretation and number of indications permitted within a specific area, casting houses should develop their own logical definitions and propose meaningful limits at the time of quotation.

 

Other Factors Affecting Radiographic Testing
A good maintenance schedule is required for film processing equipment. Otherwise, interpretation problems could arise. Discontinuities on the surface of the film from processing or on the lead screens in the film cassette can also cause indications that could be misjudged by the film interpreter.

Acceptable lighting, viewing conditions and interpreter vision must also be considered. Radiographs should be viewed in a darkened room with subdued lighting having no more than 21.5 to 32.3 lx (2 or 3 footcandles) of ambient light available. Glare from light striking the film from behind the interpreter should be minimized. Film viewers should be qualified by measuring the amount of illumination available and posting the maximum film density for which the film viewer can be used. Specifications such as ASTM E1742 and MIL-STD-453 contain information on viewer qualification. Film should be exposed to obtain as much information as possible. In other words, darker film is better and more light from the viewer is better. Radiographic technique should also optimize radiographic contrast. Interpreters should have their vision checked annually and should have corrective eyewear if necessary. New tests for contrast sensitivity are now available that can be used to test interpreters. Contrast sensitivity falls off significantly with age and with the onset of certain illnesses or medical conditions.

 

CONCLUSION 
Limits of detectability are of serious concern to both manufacturers and users of castings. Due to the complex nature of castings, it is difficult to set exacting rules. What may work in one case may not work in another, in spite of theory and past experience. Operator training, experience and vision should be considered and practical and meaningful examinations administered for qualification.

The most important characteristics to consider are discontinuity orientation, shape and size, casting surface condition and the amount of material that will be removed during subsequent operations. At times a reasonable balance may need to be achieved. Some discontinuities may be easier to detect with ultrasonics than radiography and vice-versa. Sometimes a combination of both would be best. Radiographic and ultrasonic discontinuity areas may not always exactly equate. For this reason, designers of castings may need to perform some fracture mechanics qualification using actual specimens with discontinuities to determine equivalency prior to accepting a test method and discontinuity size. Surface examinations using magnetic particle or liquid penetrant testing may be necessary to complement volumetric examinations. Communication with the customer during the quotation stage and during production concerning limits of detection should alleviate many of the problems associated with acceptance of nonconforming material. Table 1 provides a quick overview of the advantages and disadvantages of each test method.

 

Table 1 Comparison of ultrasonic and radiography for the testing of castings
	Characteristic	Ultrasonic	Radiography
	Equipment cost	+	-
	Operating training	-	+
	Near surface detection	-	+
	Portability	+	-
	Large part size	=	-
	Rough surface	-	=
	Need for calibration	-	+
	Part attenuation/grains	-	=
	Intricacy of part geometry	-	=
	Limited plant space	+	-
	Discontinuity orientation	=*	=**
	Permanent record	-	+
-	not advantageous
+	advantageous
=	moderate advantageous
*	if oriented parallel to scan surface (planar or cracklike)
**	if oriented perpendicular to part surface/radiation beam (planar or cracklike)

 

REFERENCES 
ASTM International, ASTM A609: Standard Practice for Castings, Carbon, Low Alloy, and Martensitic Stainless Steel, Ultrasonic Examination Thereof, Conshohocken, Pennsylvania, 2002.

ASTM International, ASTM E1742: Standard Practice for Radiographic Examination, Conshohocken, Pennsylvania, 1995.

Society of Automotive Engineers, SAE AMS-STD-2175: Classification and Inspection of Castings, Warrendale, Pennsylvania, 1998.

US Department of Defense, MIL-STD-453: Radiographic Inspection, Washington, DC, 1996.

 

*Alloyweld Inspection Company, Inc., 796 Maple Lane, Bensenville, IL 60106; (630) 595-2145; fax (630) 595-2128; e-mail <alloyweldinsp@aol.com>.

⁺Steel Founders' Society of America, 205 Park Ave., Barrington, IL 60010; (847) 382-8240; fax (847) 382-8287; e-mail <blairm@sfsa.org>.

 

Copyright © 2003 by the American Society for Nondestructive Testing, Inc. All rights reserved.