ABOUT THE AUTHOR: Terry Banach has had a 29-year career in the field of ultrasonic nondestructive testing, and has worked for the last 20 years with tube and bar inspection systems. In 1978 he joined Magnetic Analysis Corp. as an engineering manager, to oversee the design of ultrasonic equipment used for tube and bar inspection. He has continued to work in various engineering, training, and marketing capacities with ultrasonic products for tube inspection systems, and is currently Ultrasonic Products Manager for the company.

Mr Banach, whose interest in ultrasonic technology began during his service as sonarman in the U.S. Navy, holds an associate's degree in chemical engineering and a bachelor of science degree in electrical engineering. A registered professional engineer in Connecticut and Wisconsin, he also has ASNT Level III certification in ultrasonic testing methodology. He has written and published technical papers on the application of ultrasonic equipment to tube inspection in production environments. In addition to current membership in the International Tube Association he is also a member of ASNT.

Every welded tube manufacturer has as a goal shipment of tubing free from defects. To achieve this goal, the manufacturer must have a weld process which provides the best possible control of weld quality. Although each new year finds welding equipment suppliers offering improvements, the ideal perfect weld production process has not yet appeared. Confronting the reality of an imperfect process eventually leads to one of two situations:

• The tube manufacturer searches for an inspection technique which will insure that defective tubes will not be shipped to the customer; or,

• The customer demands insurance that the tubes received from the manufacturer will not be defective.

A search for available inspection techniques will reveal at least three popular methods: eddy current, flux leakage, and ultrasonic. From the manufacturer's viewpoint, the most perfect and cost-effective method is desirable. From the customer's viewpoint, simply the most perfect method that guarantees zero defects is desirable. Unfortunately a simple truth exists: just as there is no perfect weld process, there is no single perfect inspection technique.

The truth is that not even using two or three techniques synergistically will guarantee a perfect inspection. Synergism simply offers an improved inspection with a higher probability of approaching perfection during the inspection process.

The reason for re-stating the obvious grows out of the typical surprised reactions of both welded tube suppliers and customers when examining results from a less than perfect inspection. This article will describe some of the major truths which must be accepted when the willing or unwilling choice is to use the ultrasonic method to inspect small diameter welded tubing. In addition, we will look at ways to maximize the positive consequences and minimize the negative consequences of the welded tube ultrasonic inspection process.

As a first step, let us define some of the limitations, terms, and test methods that will apply.

Inspection Process Definitions

Small Diameter Welded Tube: This includes welded tubes with diameters less than 1-1/4 in. (32mm). Wall thickness is typically less than 12 per cent of the diameter.

Off-line Inspection: We will discuss only the inspection of cut lengths of tubing. Inspection of cut lengths is most often performed as a final inspection, before shipping. In some cases "off-line inspection" is performed at the customer's plant as an incoming inspection.

Rotational Ultrasonic Inspection of Tubes: We will talk only about an ultrasonic inspection process which provides rotation of the ultrasonic sensor around a passing tube.

Flaw Inspection: Topics in this discussion will be limited to flaw inspection. Thickness measurement of the weld bead will not be included.

Welded Tube Material: The inspection technique discussed here is applicable to any welded metal tube. The tube material may be steel, titanium, brass, etc.

Defect Types: Primarily, we will be applying the ultrasonic inspection method to the detection of defects oriented parallel to the tube length axis, and also defects oriented perpendicular to this same axis. Those parallel to the axis are most commonly called longitudinal defects. Those oriented perpendicular to the axis are most commonly considered transverse defects. Defects oriented in-between these two conditions are considered oblique.

Full-body Inspection: Because the ultrasonic sensor (transducer) is rotating as the tube passes by the sensor, the entire tube body can be inspected. Therefore the weld zone as well as the remainder of the tube wall is inspected.

Overtest: The inspected tube rejection criteria are established so that a small quantity of good tubes will be included along with the rejected bad tubes. The expectation with overtesting is that no bad tubes will be shipped to the customer.

Transducer: A device which transmits sound energy into the tube and also receives any sound energy returning from the tube. Sound energy returns from the tube whenever the sound, transmitted into the tube wall, encounters a condition inside the tube that reflects sound energy. Return sound energy is called an echo signal. Typically echo signals result from defects, although acceptable tube conditions can sometimes also cause echo signals.

Coupling Mechanism: Ultrasound cannot simply be sent from the transducer through the air into the tube. Even a thin film of air essentially blocks the sound from entering the tube wall. Therefore a coupling medium with properties that enhance the transfer of sound energy into the tube must be used. This medium is typically water.

Preliminary Ultrasound Basics Applied to Tube Inspection

Let us first establish some feeling for the action of the ultrasound during the tube inspection process by means of an overly simplified description of the theories. We will limit the theory to inspection that uses a rotating ultrasound transducer.

In most cases, while inspecting small diameter tubing we will be looking for two primary defect orientations. One orientation includes defects with lengths that are essentially parallel to the tube length axis. These are commonly called longitudinally oriented defects. The second orientation includes defects with lengths that are essentially perpendicular to the tube length axis. These are commonly called transversely oriented defects.

To find defects with either of these two orientations, we must control the travel direction of the sound energy within the tube wall. Figure 1 shows the sound propagation, within the wall, required for detection of longitudinally oriented defects.

Figure 2 shows the sound propagation, along the wall, required for the detection of transversely oriented defects. In both figures it can be seen that the sound energy travels along what appears to be a sawtooth path inside the tube wall. Furthermore, both figures show that sound energy is sent in two opposite directions for each defect orientation. This is called bidirectional testing. Bidirectional testing is simply used to improve defect detection probabilities.

Production of these sound energy travel patterns requires that the sound energy enter the tube wall at a specific entrance angle (Figure 3). The diagram shows the angle of the sound beam axis, which exists in the coupling medium (water), with respect to the tube surface. For longitudinally oriented defects, the beam axis angle is measured with respect to an imaginary line coincident with the tube radius. For transversely oriented defects, the beam axis angle is measured with respect to an imaginary line perpendicular to the tube surface.

Sound energy is not supplied continuously during an ultrasonic inspection. Rather it is transmitted as pulses or bursts of sound. Therefore, if 100 per cent inspection of the tube volume is desired, the sound must be supplied frequently enough to ensure adequate coverage of the tube wall. Figure 4a shows this effect. The diagram also demonstrates the effects of transducer rotation on the density of sound beam interrogation of the tube wall. The faster the rotation, the more frequently must the sound energy be transmitted into the tube wall.

The sound energy, when visualized as a beam, has a specific beam width. When a transducer rotates around a moving tube, it paints a virtual inspection stripe around the tube. When a tube moves past a rotating transducer, this inspection stripe becomes a helical stripe along the tube length. Figure 4b shows a diagram of this inspection paint stripe condition. It can be seen that the speed at which you may send a tube past a rotating transducer is limited by how wide the required inspection paint stripe must be to achieve 100 per cent inspection coverage. This speed is commonly called the tube inspection throughput speed. That is, the tube must not pass by the rotating transducer so fast that two helical stripes exist. Under excessive throughput speed, one helical stripe represents an inspected portion of the tube while the other stripe represents an uninspected portion of the tube. (This is sometimes called "barber-poling,Ó from the spiral-striped pole in front of a barber shop.) For 100 per cent volumetric inspection of welded tube, barber-poling is unacceptable.

Calibration Standards

The calibration standard is the key element in the welded tube ultrasonic inspection process. It serves as a benchmark for determining the effectiveness of the ultrasonic system in maintaining a stable acceptance and rejection criteria for inspected tubing. A calibration standard should be made for each welded tube diameter and wall thickness. This may seem an expensive proposition if the plant manufactures a wide variety of diameters and wall thicknesses. But the absence of a correct standard for each welded tube size will, in general, cost much more in falsely rejected material or costly retest time. Worse yet, it could bring about the incalculable harm of producing defective material which has mistakenly passed inspection and been shipped to the customer.

Construction of a Calibration Standard

The calibration standard should be constructed from the same material, diameter, and wall thickness as the welded tubes that you wish to inspect. Prior to having each standard made, select approximately 20 samples from the type and dimension of welded tubes that will be inspected. These samples should possess: the best quality in straightness and roundness; minimal roll forming marks (as clean and smooth a surface as possible) on both OD and ID; and the highest quality welds in the tube lot. A benchmark will be established with final selection from these sample tubes. For a given tube material, diameter and wall, this selection process need only be made once. The standard produced from one of these samples can be used repeatedly for inspecting later lots of welded tubing with the same material, diameter, and wall.

Once these samples have been selected, a cursory ultrasonic inspection can be performed on each of the sample tubes to determine which tube is most ultrasonically quiet. An ultrasonically quiet tube is one that contains no defect signals as well as no other echo signals from reflectors in or on the tube wall. Reflectors result from mechanical conditions in the tube which can reflect sound energy back to the ultrasonic receiver. These same mechanical conditions may not necessarily be considered defects.

To qualify these samples an ultrasonic system can be set up to perform a cursory inspection. Simply adjust each of the bidirectional test transducers to obtain a nominal 45 degree sawtooth sound propagation pattern within the tube wall. This can be done by adjusting each transducer for this angulation according to a recommended index setpoint on each rotary transducer manipulator. Next, adjust each transducer's electronic inspection window to examine a wide sonic view of the tube wall. As the samples are fed past the rotating transducers, the resulting noise condition for each sample tube can be recorded on a paper or electronic strip chart recorder. Figure 5 shows a comparison of two typical recordings: the upper, for a noisy, unacceptable tube sample; the lower, for a quiet tube sample. Select the most quiet tube by comparing the baseline signal conditions captured on the recorded traces from the sample tubes. The choice for the most sonically quiet tube should be the tube that produced a recording with the smoothest as well as lowest amplitude baseline indications. When possible, select a sample tube having a ratio ofbaseline noise height to full-scale recorder grid height that is even better than the signal condition shown in the lower trace of Figure 5.

Comparison of the Baseline Noise from Two Welded
Tubes Using Strip Chart Recording Traces from a Single
Longitudinal Inspection Channel

Upper Trace shows Noicy Welded Tube
Lower Trace Shows Quieter Welded Tube

Calibration Standard Reference Notches

Some very popular specifications which apply to the production of reference notches for small diameter tubing were created by ASTM. Some of these specifications which can apply to welded tubing include ASTM specifications numbers A450, B338, and E213.

Basically, each of these specifications prescribes similar techniques for producing the notches. Each specification prescribes similar dimensions and locations for notches. Though transverse notches are mentioned, the specifications give more attention to the details for longitudinal notch recommendations.

Most often the notches are rectangular and are made by means of an EDM (electrical discharge machining) process, although other machining methods are acceptable. The length of the notches is typically 0.5 in. (12.7mm), but lengths up to 1 in. (25.4mm) are sometimes used. In many cases the notch length must be negotiated with the customer. Generally, a consequence of shallower and shorter agreed-upon notch lengths is slower inspection throughput speeds, required to insure 100 per cent volumetric inspection coverage.

The depths of the notches typically range between 10 per cent to 12.5 per cent of wall thickness. Minimum recommended depth should be greater than 4/1000 in. (0.102mm).

ASTM specification E213 states that the width of the notch should be as small as possible but should not exceed twice the depth. Typically, on small diameter tubing, the notch width is specified as 10/1000 in. (0.254mm) and sometimes as small as 4/1000 in. (0.102mm).

As a typical practice the notches should be located 180 degrees away from the weld. Separation between two adjacent notches should be greater than 10 in. (254mm). This distance should permit easy discrimination of the notch signals on a strip chart recording at most normal inspection throughput speeds. The ID notches should be located approximately 8 in. (200mm) from the tube ends. This proximity to the end ensures that the machining process is able to maintain the precise specifications required for each ID notch.

Each longitudinal notch orientation must be machined on a tube radial, with its length parallel to the tube axis, on both inside and outside of the tube. Both lengthwise sides of any notch must be parallel to each other. Each transverse notch orientation must be machined within a cross-sectional plane perpendicular to the tube axis. No notch must be allowed to be skewed. Figure 6 diagrams these conditions.

When searching for a supplier who will manufacture these notches, it is wise to choose one who can certify the notches as well as provide visual qualification of the notches with molded rubber impressions.

Finding a quality reference notch producer is paramount. While quality at low cost is desirable, low cost without quality will only result in increased inspection costs due to setup time problems, as well as excessive unqualified rejected material.

Mechanical System Alignment

The rotational mechanics must be centered to within 5/1000 in. (0.127mm) of the stationary mechanics to ensure that the sound beam angle is able to properly track the small diameter tubing. Out-of-center conditions beyond this level will prevent inspection repeatability and often cause excessive material rejection, as well.

Rotational transducer systems normally use some form of guide bushing arrangement. These guide bushings are most often manufactured from either high molecular weight plastics or urethane. High-durability urethane material provides superior wear service. Guide bushing pairs are used at the entrance and exit of the rotary. In machining each set of four bushings for a given tube diameter, the relative concentricity of all four bushing outer and inner diameters must be maintained. Clearance for each guide bushing inner diameter must be no more than 10/1000 in. (0.254mm) greater than the inspected tube diameter. This small clearance helps to conform the tube, even bowed tubing, to the test center. Ensuring that the tube center is as close as possible, coincident with the test center, is very important to the inspection of small diameter tubing.

Figure 7 represents the offset distance between the transducer beam axis and the diameter of the tube. This offset distance determines the angle of the sound wave energy sawtoothed pattern within the tube wall. The larger this offset distance, the grater the angle of the sawtooth. Table I lists theamount of offset distance required to produce a 45degree angle in different tube diameters. The larger the diameter, the greater the required offset distance to obtain a desired angle for the sawtoothed sound energy pattern.

Figure 8 demonstrates that if a tube is not straight its cross-section shifts while passing the transducer. The transducer offset was originally set up to produce a 45degree sawtooth sound pattern with a tube which had a center coincident with the test center. When a bowed tube's cross-section shifts the tube center away from the original test center, the offset distance will either increase or decrease from its original value. This results in new values of offset distance as the transducer rotates around the tube. This condition causes a series of changing sawtooth angles throughout each revolution of the transducer. These changing sawtooth patterns frequently cause increased noise conditions from any reflectors within the tube, especially reflectors from a weld bead that is not relatively smooth. Very often these reflectors are acceptable mechanical weld conditions or roll formed conditions on the tube wall. These acceptable conditions may exist on either OD or ID. The result is usually a large amount of rejected tubing for which, very often, no one can find any true cause for rejection. The more subtle and dangerous problem is that defective tubes may have slipped through the inspection process.

Table II shows the effects of these tube-induced variations in offset distance about a central setting of 45degree. As stated above, this central setting would be the transducer offset distance established by the operator during an adjustment for the anticipated inspection tube diameter. If, during inspection, its lack of straightness causes the tube to shift 0.025 in. (0.64 mm) to the left of the rotary test center, the new tube-induced offset will be 0.152 in. (3.9 mm,) resulting in a 37.4degree sawtooth sound pattern within the tube wall for a 1 in. (25.4mm) OD tube. On the other hand, if its lack of straightness causes the tube to shift 0.025 in. (0.64mm) to the right of the rotary test center, the new tube-induced offset will be 0.202 in. (5.1mm) resulting in a 53.8degree sawtooth sound pattern within the tube wall for a 1 in. (25.4mm) OD tube. This represents a variation of the original 45degree sawtooth angle by minus 7.6degree and plus 8.8degree. As the inspected tube diameter increases, Table II shows that the same amount of tube-induced offset variation (0.025 in. / 0.64mm) causes smaller and smaller sawtooth angle variations. This simply means that larger diameter tubes are less sensitive to the effects of a lack of straightness. Unfortunately, the consequences of lack of straightness to the inspection of increasingly smaller diameter welded tubing is more severely pronounced.

The preceding discussion was meant to emphasize the supreme importance of maintaining coincidence of the tube center with the test center while inspecting small diameter tubes. Guide bushings with tight clearances are one of the means that will aid this effort. The disadvantage of tight clearance guide bushings occurs when extreme lack of straightness and/or oversized OD variation is encountered. Obviously, the inevitable tube jam will occur. In most cases straightness and oversized OD conditions can be reasonably controlled during the welding process. They must be, if it is intended to ultrasonically inspect the welded tube product.

Mechanical System Alignment

Not only must the rotary mechanics and the guide bushing system be centered and aligned, but the tube pinch drive system, before and after the rotary, must also be aligned to the test center. This pinch drive system can be either a constant center pinch drive or a V-roll pinch drive. Both types of systems require relatively simple means to check out or adjust the alignment of the rotary mechanism with respect to the pinch drive mechanics. Unfortunately far too often operators, pressed for time to get the product inspected, skip the manufacturer's instructions concerning alignment checkout and adjustment procedures. Immediately after a system is installed there are usually no significant problems. Later, as the drive rolls wear, conditions causing centering problems quickly grow. As a consequence test repeatability becomes an increasing problem.

Test Repeatability

Test repeatability is here defined as the ability to run a tube through the rotary inspection system any number of times and obtain identical test results each time. If a rejected tube is run through the system, it must be rejected each time. If an acceptable tube is run through the system, it must be accepted after each test.

The most frequently used test of repeatability is a series of repeated runs using the calibration standard. During setup with the calibration standard, the operator adjusts the strength of the echo signal received from each notch in the calibration standard to be approximately equal. The operator next adjusts the rejection alarm level to trigger an alarm output when the echo signal is received from each notch. If two notches, an OD and ID longitudinal notch, are present on the standard, the system should produce two audible alarms, one for each notch. The calibration tube should, in addition, be sorted to the reject bin after each test run.

A much more reliable method of evaluating repeatability is to use a strip chart recorder connected to each receiver channel output. A typical bidirectional longitudinal notch inspection uses two transducers, and thus two receiver channels. Each receiver channel output should appear on a separate recording trace. In the case of a bidirectional longitudinal inspection, two strip chart traces will be active. When the calibration standard is run by the two rotating transducer channels, the traces should indicate all the echo signals as spikes with a specific amplitude. In the case of OD and ID notches, each trace should show two spike indications: one for the OD notch and one for the ID notch. When the calibration standard is re-run to evaluate repeatability, the traces should show nearly the same height spike indications each time.

Repeatability is considered good when the run-to-run variation of the spike signal height, obtained from each notch on each trace, is less than several minor divisions on a typical strip chart grid. If the inspection is set up to detect both longitudinal and transverse OD and ID notches, four transducer channels will be used. Therefore four traces on the strip chart recorder will be under monitor. In this case, the spike signal heights from all of the notches should remain essentially constant on each trace. There should be two spike signals on each of the longitudinal channel traces as well as two spike signals for each of the transverse channel traces. One spike will always be from an OD notch; the other spike will be from an ID notch.

One final point should be mentioned regarding the evaluation of repeatability of the rotary inspection system. To ensure that the entire mechanics is centered and that the transducers are set up properly, the notches on the calibration standard should be rotated to different points in the cross-sectional test plane of the rotary inspection unit. Figure 9 shows this concept in diagram form. The notches are shown at four different points in the cross-sectional test plane. These points are shown as 12 o'clock, 3 o'clock, 6 o'clock, and 9 o'clock.

Figure 10 depicts four simulated strip chart recorder traces that might result from the detection of the notches at these same clock orientations. Traces at 3 o'clock and 9 o'clock show excessive amplitude variation of OD and ID spike signals. This could be due to poorly constructed guide bushings, poorly aligned pinch drive mechanics with respect to the rotary mechanics, or poorly centered rotary mechanics with respect to the stator mechanics.

In addition, repeatability problems resulting from excessive tube throughput speed will show up on the strip chart traces. Furthermore, problems caused by not transmitting the sound energy into the tube frequently enough will also appear while evaluating the repeatability traces. Both these problems will be recognized by a wide variation in the notch signal amplitudes among the recorder traces collected for all the repeatability trial runs. With experience, the operator learns to quickly determine the reason for the lack of repeatability and then remedy the cause.

Consequences of the Higher-Sensitivity Inspection Method

For determining whether welded tubing is defective, there must be willingness to bear the consequences. Ultrasonic testing, as the choice of inspection methods for small diameter welded tubing, will drive home the validity of this statement. Ultrasonic inspection can be an excellent means of ensuring that defective welds will be detected. The operator may feel confident that his tube weld quality is mechanically sound and is within mechanical specifications. But the welded tube conditions that yield mechanical acceptability may cause tubing to be ultrasonically unacceptable. That is, these conditions will cause alarms and rejections even though these conditions may be seen as totally acceptable.

Accordingly when the decision in favor of ultrasonic inspection is made, by operator or customer, the operator must be prepared to make a higher quality welded tube that will meet the stringent requirements for a good ultrasonic inspection. This means that he will have to control the quality of the welded tube process so that the primary reflectors within the tube will be only truly defective conditions. Conditions that may previously have been considered as acceptable in terms of surface marks, poor straightness, slight weld upset, slight undercut, slight overcut, and minor lack of penetration may easily cause material rejection under the more sensitive inspection system.

This is especially true if it is decided to use shallow reference notches, approaching 5 per cent of wall thickness. One might think that a standard based on a 5 per cent notch is more than adequate to ensure that no defective tubes will ever be passed on to the customer. It will be a surprise to find that no tubes will pass inspection, and that therefore no tubes will be passed on to the customer. Reinspection of these tubes very likely will show that the setup required to detect 5 per cent notches has resulted in a very poor signal-to-noise ratio test. The inspection criteria have been made so severe that the operator is unable to discriminate actual defect conditions from noise reflector conditions. Unfortunately for some manufacturers of welded tubing, reference notches whose depth is even 10 per cent of the wall thickness get into similar trouble with their normal weld quality.

Once the decision has been taken to overcome these problems through the achievement of higher weld quality, the ultrasonic method is an excellent inspection tool. There must, however, be willingness to keep the ultrasonic system maintained to the highest standards possible. Unattended wear on bushings and centering mechanisms will produce the same problems as occur with lowered weld process quality. The sensitive geometric element of a small diameter welded tube originates from the tube's high rate of surface curvature. This condition makes the tube extremely sensitive to misalignment with respect to the test system center. Practices of reduced mechanical system maintenance and alignment may give rise to a false belief that the plant is saving time and money. The fact is, problems arising from the small diameter tube's sensitivity to misalignment will cost far more in both time and money.

Conclusion

The overriding positive factor is that with experience and with the application of good practices, rotary inspection will permit high speed inspection with reliable results.

A very powerful saying frequently heard in the nondestructive testing community is, "You cannot test quality into your product." This is just as appropriate to welded tube manufacturing. Ultrasonic inspection is a reliable tool to help sort out defective material, but only if it is used with care.

Magnetic Analysis Corporation
535 S. Fourth Avenue, Mount Vernon, NY 10550, USA
Fax: +1 914 699 9837