TOFD

Digitization Principles

2/5/2005 ©Lavender International NDT 1

Consultancy Services Ltd Issue 2.
January 2005
 Advantages of Digital Recording

„ Permanent record of all collected data

„ Various signal processing options
„ Results can be compared easily
„ Re-analysis of raw data, at any time
„ Variety of visual displays available
„ Integrated data acquisition and probe
manipulation control
„ Multiple channel data compared with overlays

2/5/2005 ©Lavender International NDT 2

Consultancy Services Ltd Issue 2.
January 2005
 Digitization of Analogue Waveforms
Analogue waveform
+100%

-100%
Separation set
by digitizer speed Points of sampling analogue data

2/5/2005 ©Lavender International NDT 3

Consultancy Services Ltd Issue 2.
January 2005
 Digital Amplitude
„ Most contemporary TOFD instruments are
fitted with 8 bit digitisers (28)
„ In computer terms this means a digital
range of numbers from 0 to 255
„ Full wave rectified data = 0 to 255 this is
used for phased arrays and pulse-echo
„ RF data = -128 to +127 this is used for
TOFD imaging

2/5/2005 ©Lavender International NDT 4

Consultancy Services Ltd Issue 2.
January 2005
 Points of Amplitude
+127
+100%

Zero

-100%
-128
8 Bit digitizer in RF

2/5/2005 ©Lavender International NDT 5

Consultancy Services Ltd Issue 2.
January 2005
 Logarithmic Amplification
Rectified
„ Amplification is usually logarithmic thus
signal heights can be expressed with the
following equation:
„ 20 log10 H1/H2

„ Rectified = 20 log10 255/1 = 48dB

2/5/2005 ©Lavender International NDT 6

Consultancy Services Ltd Issue 2.
January 2005
 Logarithmic Amplification
RF Unrectified
„ Unrectified data uses half of the digital
numbers available to produce the screen
display
„ For a range of 0 to 255 this represents
approximately +128 and – 128 digital
numbers
„ RF = 20 log10 128/1 = 42dB

2/5/2005 ©Lavender International NDT 7

Consultancy Services Ltd Issue 2.
January 2005
 Sampling Rate

„ Working from first principles

„ 1µs = 1.10-6s
„ 25MHz = 25 x106Hz
„ Thus 1µs = 25 samples
„ So 1 sample = 1/25 = 0.04µs

2/5/2005 ©Lavender International NDT 8

Consultancy Services Ltd Issue 2.
January 2005
 Digitization at 25MHz

+100%

-100% 0.04µs

Separation of 0.04µs between sample points

2/5/2005 ©Lavender International NDT 9

Consultancy Services Ltd Issue 2.
January 2005
 Time Across One Wavelength

„ Time taken to complete 1λ is

termed a period
„ If probe frequency = 5MHz
(5x106Hz)
„ Then 1µs = 5 cycles
„ Thus 1 period = 1/5 = 0.2µs

2/5/2005 ©Lavender International NDT 10

Consultancy Services Ltd Issue 2.
January 2005
 One Time Period

+100%

Time across one wave cycle is called a “Period”

Distance across one wave cycle is wavelength

-100%

2/5/2005 ©Lavender International NDT 11

Consultancy Services Ltd Issue 2.
January 2005
 Samples Taken In A Period

„ Using the previous data

„ 1 period = 0.2µs
„ Digitization rate = 0.04µs per
sample
„ Which equates 0.2µs/0.04µs = 5
„ Therefore 5 samples are taken in
each period (full cycle)

2/5/2005 ©Lavender International NDT 12

Consultancy Services Ltd Issue 2.
January 2005
 Absolute Minimum Sampling

„ In order to digitally recreate a sinusoidal

wave from sampling an analogue signal,
at least 2 samples per period must be
taken ( Nyquist limit )

2/5/2005 ©Lavender International NDT 13

Consultancy Services Ltd Issue 2.
January 2005
 Potential Error Of Only
2 Samples Per Period

Sample points on analogue

waveform

No digitally recreated signal

2/5/2005 ©Lavender International NDT 14

Consultancy Services Ltd Issue 2.
January 2005
 Ideal Minimum Sampling

5 sample points on
analogue waveform

Digitally recreated
waveform within 10%
of analogue signal

2/5/2005 ©Lavender International NDT 15

Consultancy Services Ltd Issue 2.
January 2005
 Pulse Length Of Excitation Voltage

Rise of excitation Return to zero of

voltage excitation voltage
+ -

Waveform from rise voltage Waveform from return voltage

2/5/2005 ©Lavender International NDT 16

Consultancy Services Ltd Issue 2.
January 2005
 Effect Of Changing Excitation
Pulse Width

Half a One
period period

2/5/2005 ©Lavender International NDT 17

Consultancy Services Ltd Issue 2.
January 2005
 Pulse Voltage Time
„ Setting the driver voltage to half a period
of the array dominant frequency:
„ 1µs/Dominent frequency in Hz
„ 1µs/5MHz = 0.2µs = 200ns
„ 0.5 of 200ns = 100ns
„ Most TOFD systems have software that
incorporate calculation function to
accomplish this

2/5/2005 ©Lavender International NDT 18

Consultancy Services Ltd Issue 2.
January 2005
 Signal Averaging

Sample 1

Sample 2

Sample 3

Sample 4

Averaged
signal

2/5/2005 ©Lavender International NDT 19

Consultancy Services Ltd Issue 2.
January 2005
 Features Of Signal Averaging

„ Random electronic noise is

averaged out reducing influence
„ Grain interference not reduced
„ Standing echoes not reduced
„ True indications not affect

2/5/2005 ©Lavender International NDT 20

Consultancy Services Ltd Issue 2.
January 2005
 Use Of Averaging
„ Using an averaging of 16 requires
each degree increment or ‘A’ scan to
be pulsed 16 times in order to
produce 1 ‘A’ scan image
„ Use of high averaging values will
restrict practical scan speed possible

2/5/2005 ©Lavender International NDT 21

Consultancy Services Ltd Issue 2.
January 2005
 Pulse Repetition Frequency PRF

„ Rate of voltage pulses

transmitted from pulser to
transducer
„ Not to be confused with probe
frequency
„ Selectable PRF values typically
range from ~ 60Hz to ~
15000Hz

2/5/2005 ©Lavender International NDT 22

Consultancy Services Ltd Issue 2.
January 2005
 Effect Of Erroneously
Low Pulse Rate

„ Can lead to intermittent data

imaging
„ Drop out of lines of data
„ Serious risk of missing defects

2/5/2005 ©Lavender International NDT 23

Consultancy Services Ltd Issue 2.
January 2005
 Lines of Missed Data

Causes: Low PRF, High travel speed,

High resolution, High averaging, Low Com speed,
Small increments of data collection

2/5/2005 ©Lavender International NDT 24

Consultancy Services Ltd Issue 2.
January 2005
 Effect Of Excessively High PRF

„ Low velocity materials or thick

cross-sections may not return
pulses before subsequent pulsed
waves are transmitted
„ Potential to cause standing
echoes from colliding wavefronts
(ghosts)

2/5/2005 ©Lavender International NDT 25

Consultancy Services Ltd Issue 2.
January 2005
 Probe center frequency
( Dominant frequency )

Peak
value

0 1 2 3 4 5 6 7 8 9 10 11 12
MHz

2/5/2005 ©Lavender International NDT 26

Consultancy Services Ltd Issue 2.
January 2005
 Bandwidth
100%

90%

80% Bandwidth measured

at .707 of peak value
70%

60%

50% Bandwidth ~
3.5MHz
40%

30%

20%

10%

0%
0 1 2 3 4 5 6 7 8 9 10 11 12
MHz

2/5/2005 ©Lavender International NDT 27

Consultancy Services Ltd Issue 2.
January 2005
 Probe Frequency Spectrum
100
%

80%

60%
Low frequency High frequency
long Short wavelength
40% wavelength

20%

0%

2/5/2005 ©Lavender International NDT 28

Consultancy Services Ltd Issue 2.
January 2005
 Wavelengths Associated With
A Broad Band Probe

2/5/2005 ©Lavender International NDT 29

Consultancy Services Ltd Issue 2.
January 2005
 Band Pass Filters

High pass filter Low pass filter

set at 4MHz set at 7MHz

General guide is to set the low pass at 2x centre frequency

And set the high pass to 0.5 of the centre frequency

2/5/2005 ©Lavender International NDT 30

Consultancy Services Ltd Issue 2.
January 2005
 Caution !
„ If the high pass filter is set incorrectly you will
filter out relevant indications
„ Flaw responses typically are received back at
slightly lower frequencies than originally
transmitted
„ Therefore individual circumstances require careful
adjustment and may warrant the high pass filter
switching off

2/5/2005 ©Lavender International NDT 31

Consultancy Services Ltd Issue 2.
January 2005
 Grey Scale Palette
+127
+100%

Zero

-100%
-128

2/5/2005 ©Lavender International NDT 32

Consultancy Services Ltd Issue 2.
January 2005
 “A” Scan as Greyscale Image

Typically used for TOFD

2/5/2005 ©Lavender International NDT 33

Consultancy Services Ltd Issue 2.
January 2005
 Typical TOFD Image
Using RF Waveform

2/5/2005 ©Lavender International NDT 34

Consultancy Services Ltd Issue 2.
January 2005
 Acknowledgements
„ This presentation is one of a series written by
Tim Armitt of Lavender International NDT Ltd.
„ Lavender International have been involved in
TOFD for many years and have become
recognised as one of the principle training
organisations involved in TOFD internationally.
„ Full courses are available from Lavender
International with the option of level 1, 2 & level
3 certification in TOFD
„ For additional information email us using the
address: [email protected] or visit our
website: www.lavender-ndt.com

2/5/2005 ©Lavender International NDT 35

Consultancy Services Ltd Issue 2.
January 2005
 Any Questions?

2/5/2005 ©Lavender International NDT 36

Consultancy Services Ltd Issue 2.
January 2005