5/20/04

Article 4
Nonmandatory Appendix N
Time of Flight Diffraction (TOFD) Interpretation
N-410 SCOPE

This appendix is to be used as an aid for the interpretation of Time-Of-Flight

Diffraction (TOFD) ultrasonic images. Diffraction is a common ultrasonic
phenomenon and occurs under much broader conditions than just longitudinal-
longitudinal diffraction as used in typical TOFD examinations. This interpretation
guide is primarily aimed at longitudinal-longitudinal diffraction TOFD setups using
separated transducers on either side of the weld on a plate, pipe or curved
vessel. Other possibilities include:

• Shear-shear diffraction
• Longitudinal-shear diffraction
• Single transducer diffraction (called “back diffraction” or the “tip-
echo method”)
• Twin transducer TOFD with both transducers on the same side of
the flaw/weld.
• Complex inspections, e.g. nozzles.

N-420 GENERAL

N-421 TOFD Images - Data Visualization

(a) TOFD data is routinely displayed as a grayscale image of the digitized

A-scan. Figure N-421 (a) shows the grayscale derivation of an A-scan (or
waveform) signal.

Figure N-421(a) Schematic showing waveform transformation

into grayscale
 (b) TOFD images are generated by the stacking of these grayscale
transformed A-scans as shown in Figure N-421(b). The lateral wave and
backwall signals are visible as continuous multicycle lines. The midwall flaw
shown consists of a visible upper and lower tip signal. These show as
intermediate multicycle signals between the lateral wave and the backwall.

Figure N-421(b) Schematic showing generation of gray scale

B-scan from multiple A-scans

(c) TOFD grayscale images display phase changes, some signals are
white-black-white; others are black-white-black. This permits identification of the
wave source (flaw top or bottom, etc.), as well as being used for flaw sizing.
Depending on the phase of the incident pulse (usually a negative voltage), the
lateral wave would be positive, then the first diffracted (upper tip) signal negative,
the second diffracted (lower tip) signal positive, and the backwall signal negative.
This is shown schematically in Figure N-421(c). This phase information is very
useful for signal interpretation; consequently, RF signals and unrectified signals
are used for TOFD. The phase information is used for correctly identifying signals
(usually the top and bottom of flaws, if they can be differentiated), and for
determining the correct location for depth measurements.
 Transmitter Receiver
Lateral wave

Back-wall reflection

+ +
LW BW

- -
Upper tip Lower tip

Figure N-421(c) Schematic showing standard TOFD setup

and display with waveform and signal phases

(d) An actual TOFD image is shown in Figure N-421(d), with flaws. The
time-base is horizontal and the axis of motion is vertical (the same as the
schematic in Figure N-421(c). The lateral wave is the fairly strong multicycle
pulse at left, and the backwall the strong multicycle pulse at right. The flaws show
as multicycle gray and white reflections between the lateral and backwall signals.
The scan shows several separate flaws (lack of fusion, porosity, incomplete
penetration and slag). The ultrasonic noise usually comes from grain reflections,
which limits the practical frequency that can be used. TOFD scans may only
show the lateral wave (OD) and backwall (ID), with “noise”. There is also
ultrasonic information available past the backwall (typically shear wave
diffractions), but this is generally not used.
Lack of fusion

Porosity

Lack of fusion

Incomplete
penetration

Slag

Figure N-421(d) TOFD display with flaws and displayed

A-scan. Time is horizontal and the axis of
motion is vertical
N-450 PROCEDURE

N-451 Measurement Tools

TOFD variables are: probe spacing, material thickness, sound velocity,

transducer delay, and lateral wave transit and backwall reflection arrival time. Not
all the variables need to be known for flaw sizing. For example, calibration using
just the lateral wave (front wall or OD) and backwall (ID) signals can be
performed without knowing the transducer delay, separation, or velocity. The
arrival time, Figure N-451, of the lateral wave (t1) and the backwall signal (t2) are
entered into the computer software and cursors are then displayed for automated
sizing.

A-

d1
d1
h
t1 t2

Error! Cursors
Build-in L

t1,t2 ⇒ d1, d2 and h are automatically

calculated. P

Figure N-451 Measurement tools for flaw heights

N-452 Flaw Position Errors

Flaws will not always be symmetrically placed between the transmitter and
receiver transducers. Normally, a single pair of transducers is used, centered on
the weld axis. However, multiple TOFD sets can be used, particularly on heavy
wall vessels, and offsets are used to give improved detection. Also, flaws do not
normally occur on the weld centerline. Either way, the flaws will not be positioned
symmetrically, Figure L-452(a) and this will be a source of error in location and
sizing.

S S
Transmitter Receiver

t0 t0

Figure N-452(a) Schematic showing the detection of off-axis

flaws

There will be positional and sizing errors associated with a non-centered flaw, as
shown in Figure N-452(b). However, these errors will be small, and generally are
tolerable since the maximum error due to off-axis position is less than 10% and
the error is actually smaller yet since both the top and bottom of the flaw are
offset by similar amounts. The biggest sizing problems occur with small flaws
near the backwall. Exact error values will depend on the inspection parameters.
 Flaw Position Uncertainty
S S
Transmitter Receiver

t0

t1 t2

In practice, the maximum error on absolute depth position lies below 10%.

The error on height estimation of internal (small) flaws is negligible. Be

careful of small flaws situated at the backwall.

Figure N-452(b): Measurement errors from flaw position

uncertainty
N-453 Measuring Flaw Length

Flaw lengths parallel to the surface can be measured from the TOFD image by
fitting hyperbolic cursors, Figure N-453, to the ends of the flaws.

Figure N-453 TOFD image showing hyperbolic “tails” from

the ends of a flaw image used to measure
flaw length
L-454 Measuring Flaw Depth

Flaw height perpendicular to the surface can be measured from the TOFD image
by fitting cursors on the top and bottom tip signals. The following are two
examples of depth measurements of weld flaws in a 1 in. (25mm) thick plate.
Figure N-454(a) is midwall lack of fusion and Figure N-454(b) is a centerline
crack. Note that TOFD signals are not linear, so midwall flaws show in the upper
third region of the image. It is possible to linearize the TOFD scans by computer
software.

0.43 in.
0.59 in.

Lateral Top Bottom Backwall

wave echo echo echo

Figure N-454(a) TOFD image showing top and bottom

diffracted signals from midwall flaw and
A-scan interpretation
0.62 in.

0.88 in.

Front Top Bottom Backwall

wall signal signal

Figure N-454(b) TOFD image showing top and bottom

diffracted signals from centerline crack and
A-scan interpretation
N-480 EVALUATION

This section shows a variety of TOFD images and the interpretation/explanation.

Unfortunately, there are significant variations amongst flaws and TOFD setups,
and displays, so the following images should be used as a guide only. Evaluator
experience and analysis skills are very important as well.

N-481 Single Flaw Images

(a) Point flaws, Figure N-481(a), like porosity, show up as single multicycle
points between the lateral and backwall signals. Point flaws typically display a
single TOFD signal since flaw heights are smaller than the ring-down of the pulse
(usually a few millimeters, depending on the transducer frequency and damping).
Point flaws usually show parabolic “tails” where the signal drops off towards the
backwall.

A-scan

Indication

Lateral
Backwall
wave

Figure N-481 (a) Schematics of image generation, scan

pattern, waveform, and TOFD display
showing the image of the point flaw
 (b) Inside (ID) far-surface-breaking flaws, Figure N-481(b), shows no
interruption of the lateral wave, a signal near the backwall, and a related
interruption or break of the backwall (depending on flaw size).

Transmitter Receiver
Lateral

Back wall echo

No back wall echo

Error! Ti

Figure N-481(b) Schematics of image generation, flaw

location, and TOFD display showing the
image of the inside (ID) surface-breaking flaw
 (c) Near-surface-breaking flaws, Figure N-481(c), shows perturbations in
the lateral wave. The flaw breaks the lateral wave, so TOFD can be used to
determine if the flaw is surface-breaking or not. The lower signal can then be
used to measure the depth of the flaw. If the flaw is not surface-breaking, i.e. just
subsurface, the lateral wave will not be broken. If the flaw is near-subsurface and
shallow (that is, less than the ringing time of the lateral wave or a few millimeters
deep), then the flaw will probably be invisible to TOFD. The image also displays
a number of signals from point flaws.

Surface-
breaking
flaw

Figure N-481(c) Schematics of image generation, flaw

location, and TOFD display showing the
image of the outside (OD) surface-breaking
flaw
 (d) Midwall flaws, Figure N-481(d), show complete lateral and backwall
signals, plus diffraction signals from the top and bottom of the flaw. The flaw tip
echoes provide a very good profile of the actual flaw. Flaw sizes can be readily
measured from these two signals. Note the phase inversion: top echo is white-
black-white, while the lower echo is black-white-black. Also note the hyperbolic
curve that is easily visible at the left end of the top echo; this is similar to the
effect from a point flaw (see N-481(a)) and permits accurate length measurement
of flaws (see N-450(a)).

1
2

Figure N-481(d) Schematics of flaw location, signals, and

TOFD display showing the image of the
midwall flaw

If a midwall flaw is shallow, i.e. less than the transducer pulse ring-down (a few
millimeters), the top and bottom tip signals cannot be separated. Under these
circumstances, it is not possible to differentiate the top from the bottom of the
flaw, so the evaluator can only say that the flaw is less than the ringdown
distance (which depends on transducer frequency and damping, etc.).
 (e) Lack of root penetration, Figure N-481(e), is similar to an inside (ID)
far-surface-breaking flaw (see N-481 (b)). This flaw gives a strong diffracted
signal (or more correctly, a reflected signal) with a phase inversion from the
backwall signal. Note that whether signals are diffracted or reflected is not
important for TOFD characterization; the analysis and sizing is the same. Also
note even though there is a perturbation of the backwall signal, the backwall is
still visible across the whole flaw. This material also shows small point flaws and
some grain noise, which is quite common. TOFD typically overemphasizes small
point flaws, which are normally undetected by conventional shear wave pulse-
echo techniques.

Error!

2
3

Figure N-481 (e) Flaw location, and TOFD display showing the
image of the lack of root penetration
 (f) Concave root flaws, Figure N-481(f), are similar to lack of root
penetration. The top of the flaw is visible in the TOFD image, as well as the
general shape. The backwall signal shows some perturbation as expected.

2
3

Figure N-481(f) Flaw location and TOFD display showing the image the
concave root flaw.
 (g) Sidewall lack of fusion, Figure N-481(g) is similar to a midwall flaw (see
N-481(d)) with two differences. First, the flaw is angled along the fusion line, so
TOFD is effectively independent of orientation, which is not a problem for TOFD.
Second, the upper flaw signal is partly buried in the lateral wave for this particular
flaw. In this instance, the upper tip signal is detectable since the lateral wave
signal amplitude is noticeably increased. However, if this were not the case, then
the evaluator would be unable to accurately measure the flaw depth.

1
2
3

Figure N-481(g) Flaw location, TOFD display showing the

image of the midwall lack of fusion flaw, and
the A-scan
 (h) Porosity, Figure N-481(h), appears as a series of hyperbolic curves of
varying amplitudes, similar to the point flaw (see N-4811(a)). The TOFD
hyperbolic curves are superimposed since the individual porosity pores are
closely spaced. This does not permit accurate analysis, but the unique nature of
the image permits characterization of the signals as “multiple small point flaws”,
i.e. porosity.

2 1

Figure N-481(h) Flaw location and TOFD display showing the

image of the porosity
 (i) Transverse cracks, Figure N-481(i), are similar to a point flaw (see N-
481(a)). The TOFD scan displays a typical hyperbola. Normally, it would not be
possible to differentiate transverse cracks from near-surface porosity using
TOFD; further inspection would be needed.

1
2

Figure N-481(i) Flaw location and TOFD display showing the

image of the transverse crack
 (j) Interpass lack of fusion, Figure N-481(j), shows as a single, high
amplitude signal in the midwall region. If the signal is long, it is easily
differentiated from porosity or point sources. It is not possible to distinguish the
top and bottom, as these don’t exist as such. Note the expected phase change
from the lateral wave. Interpass lack of fusion signals are generally benign.

Transmitter Receiver
Lateral

Reflected

Back wall

L B

Reflection

1
2

Figure N-481(j) Schematics of image generation, flaw

location and TOFD display showing the
image of the interpass lack of fusion
N-482 Multiple Flaw Images

TOFD images of flawed welds containing four flaws each

(a) Plate 1 (Figure N-482(a)):

• Root crack (right): ~40–64 mm from one end.

• Lack of side wall fusion (mid-left): ~100–125 mm
• Slag: ~163–183 mm
• Lack of root fusion (left): ~237–267 mm

Lack of Slag Lack of

fusion fusion root

Figure N-482(a) Schematic of flaw locations and TOFD image

showing lateral wave, backwall and three of
the four flaws
Figure N-482(a) clearly illustrates the significant advantages of TOFD (midwall
flaw detection, flaw sizing), the limitations due to dead zones, and that:

• The midwall lack of fusion shows up clearly, as does the slag.

• The lack of fusion in the root was not easily detected, though it did disturb
the backwall. This is not surprising in the backwall dead zone. However,
the lack of root fusion shows clearly later on in the time base, due to a
shear-shear diffracted wave. This example illustrates the potential value of
using information later in the time base, but this is outside the scope of
this interpretation manual.
• The root crack is not visible at all due to the backwall dead zone.
 (a) Plate 2 (Figure N-482(b)):

• Lack of root fusion (left): ~15–45 mm from one end

• Toe crack (top left): ~81–101 mm
• Porosity: ~139–159 mm
• Lack of side wall fusion (upper right): ~205–235 mm

Lack of Toe crack Porosity Lack of

root fusion (crown) sidewall fusion

Figure N-482(b) Schematic of flaw locations and TOFD display

showing lateral wave, backwall and four flaws

Figure N-482(b) shows that:

• All four flaws are detectable.

• The lack of root fusion shows up clearly in this scan because it is deeper.
Both the backwall perturbation and the flaw tip signals are clear.
• The crown toe crack is clearly visible, both by complete disruption of the
lateral wave and by the bottom tip signal. Both the lack of root fusion and
crown toe crack are identifiable as surface breaking by the disruption of
the lateral wave and backwall signal respectively.
• The porosity is visible as a series of signals. This cluster of porosity would
be difficult to characterize properly using the TOFD scan alone, since it
could be identified as slag or a planar flaw.
• The lack of sidewall fusion is clearly visible, and could be easily sized
using cursors.
N-483 Typical Problems with TOFD Interpretation

TOFD images can be corrupted by incorrect setups or other problems such as

electrical noise. The following images were all made on the same plate to show
some of the typical problems that can occur. Starting first with an acceptable
scan, and then subsequent scans made to show various corruptions of this
image.

(a) Acceptable scan, Figure N-483(a) . The gain and gate setting are
reasonable, and the electrical noise is minimal.

OD surface-breaking flaw

Lateral wave

Near surface Buried flaw

flaw
Region of
porosity –
often difficult
to detect

Backwall

Figure N-483(a) Acceptable noise levels, flaws, lateral wave

and longitudinal wave backwall
 (b) Incorrect low gain setting, Figure N-483(b). The lateral wave and some
of the diffracted signals are starting to disappear. At yet lower gain levels, some
of the diffracted signals would become undetectable.

Signals
becoming
invisible
in this
area.

Figure N-483(b) TOFD image with gain too low

 (c) Incorrect high gain setting, Figure N-483(c). The noise level increases
to obscure the diffracted signals; this can lead to reduced probability of detection,
and poor sizing. High noise levels can also arise from large grains. In this case,
the solution is to reduce the ultrasonic frequency.

Signals are
becoming
confused
in these
areas.

Figure N-483(c): TOFD image with the gain set too high
 (d) Correct gate settings are critical, because TOFD A-scans are not that
easy to interpret since there are multiple visible signals. As a minimum, the gates
should encompass the lateral wave and longitudinal wave backwall signal; the
gate can extend to the shear wave backwall, if required. Typically, the best signal
to use as a guide is the first (longitudinal wave) backwall, since it is strong and
always present (assuming the transducer separation is reasonably correct). The
following figures show examples of incorrect gate positioning, which will
inherently lead to poor flaw detection.

The first example, Figure N-483(d)(1) shows the gate set too early, the lateral
wave is visible and the backwall is not. Any inside (ID) near-backwall flaws will be
missed.

Lateral wave

Figure N-483(d)(1) TOFD image with the gate set too early
The second example, Figure N-483 (d)(2), shows the gate set too late. The
lateral wave is not visible. The first signal is the backwall, and the second signal
is the shear wave backwall. With this setup, all the outside (OD) near-surface
flaws will be missed.

L-wave
backwall

S-wave
backwall
signal

Figure N-483(d)(2) TOFD image with the gate set too late
The third example, Figure N-483(d)(3), is with the gate set too long. Though this
is not technically incorrect, the image will show the diffracted backwall shear-
shear wave signal. These S-S waves may show additional and confirmatory
information. The diffracted shear waves show the porosity more clearly than the
diffracted longitudinal waves and there is a strong mode-converted signal that
occurs just before the shear wave gate, which could cause interpretation
problems. Normally, the gate is set fairly short to enclose only the lateral wave
and the longitudinal wave backwall to clarify interpretation.

Lateral wave

L-wave
backwall
signal

S-wave
backwall
signal

Figure N-483 (d)(3) TOFD image with the gate set too long
 (e) Incorrect (too far apart) transducer separation, Figure N-483(e),
results in the backwall signal becoming distorted, the lateral wave becomes
weaker, and some of the diffracted signal amplitudes drop.

Distorted
L-wave
backwall

Figure N-483(e) TOFD image with transducers set too far apart
 (f) Incorrect (too close together) transducer separation, Figure N-483(f),
results in the the lateral waves becoming stronger, and the backwall weaker.
Again, the TOFD image of the flaws is poor.

Strong
lateral wave

Weak L-wave
backwall
signal

Figure N-483(f) TOFD image with transducers set too close

together
 (g) If the transducers are not centred on the weld, Figure N-483(g), the
diffracted signal amplitudes will decline to the point where flaw detection is
seriously impaired.

Figure N-483(g) TOFD image with transducers not centered on

the weld axis
 (h) Noise levels, Figure N-483(h) can seriously impair TOFD
interpretation. Noise can come from a number of sources such as electrical,
ultrasonic, grains, and coupling. Typically, ultrasonic and grain noise appears
universally across the TOFD image. Electrical noise appears as an interference
pattern, depending on the noise source. Once the occurrence of the electrical
noise increases beyond a certain point, interpretation becomes essentially
impossible.

Figure N-483(h) TOFD image showing electrical noise

interference