User's Manual

pulseEKKO PRO Appendix B: GPR Signal Processing Artifacts
B-3
Dewow Artifacts
1) Pulse Precursor
The result of high pass filtering can induce two types of artifacts into the plotted data sections. The first is
a pre-cursor to the onset of a pulse (Figure B-1 (b) and Figure B-3, left). When the original data are high
pass filtered the wavelet is stretched in time with additional oscillations occurring before and after the
original pulse. This is what gives rise to a pre-cursor in the data before time-zero when looking at the
plotted sections.
2) Clipping
The second artifact is caused by electronic clipping. The basic concept is depicted in Figure B-4a, b and c.
The original signal which is acquired by the antenna and presented to the receiver electronics looks like
that sketched in Figure B-4a. The radar electronics clip any signals above the 50 millivolt level. Figure B-
4b shows the type of result that will be measured if clipping occurs in the signal pre-conditioning circuitry.
When these data are finally high pass filtered, a blank zone in the area where the original signal had a
large wow above the clipping level of the electronics can appear. This results in a blank section on the
record with no reflections visible.
Figure B-4: Inductive WOW clipping.
In the second case where signal saturation clipping occurs, there is no software correction for this. The
user must be aware that this can occur in some geologic settings and the antenna spacing should be
increased to reduce the wow signal amplitude below the clipping level of the receiver electronics. As a
general rule, we recommend using an antenna separation at least equal to the length of the antenna being
employed. For example, 100 MHz antennas should have an antenna spacing of at least 1 metre, a 50
MHz antenna system should use an antenna spacing of 2 metres. Obviously this rule depends on the level
of wow or inductive response present at the particular site. Again note that the processing affects only
displayed data, not recorded data.